1
Cross-border M&A and a new measure of cultural distance
*
Bram J. Raatjes
aba
University of Groningen, Faculty of Economics and Business, Groningen, The Netherlands b
University of Uppsala, Department of Economics, Uppsala, Sweden
Abstract The culture-performance relationship in cross-border mergers and acquisitions has been extensively studied with, sometimes, conflicting results. To measure cultural distance most research draws upon the outdated proxy, which was developed by Kogut and Singh in 1988. Based on their measure I create an updated measure of cultural distance and test the culture-performance relationship using 36-month buy-and-hold abnormal returns as the dependent variable on a sample of 598 cross-border acquisitions in the period 2000-2009. I compare my results to the original Kogut and Singh (1988) index measure and another alternative. I find significant results for the cultural distance measure in all three variations. As I hypothesize the evidence is the strongest for the updated measure. My results are robust to various alternative model specifications.
JEL Classification: G15; G34
2
1. Introduction
In the past thirty years cross-border mergers and acquisitions (M&A) have become common among the different forms of internationalization for companies seeking a global presence. Without a doubt they have become one of the most popular growth and entry mode strategies, thus substantially shaping our international business environment. The resources dedicated to M&A activity are tremendous. After the technology bubble burst M&A activity started to pick-up again in 2003 and continued to grow until in 2007 growth came to a halt and the financial crisis unfolded. In 2007 worldwide M&A activity peaked with more than $3.3 trillion spent on deals globally1.
The significance of cross-border M&A within the field of International Business has invoked a substantial amount of research on this topic. Since the beginning of the twenty-first century many scholars (see e.g. Cartwright and Schoenberg, 2006; Stahl and Voigt, 2008) have reviewed the field of M&A and in particular the relationship between culture and cross-border M&A. One of the usual suspects blamed for poor cross-cross-border M&A performance is cultural disparity between merging firms. Although scholars agree that culture affects M&A performance, findings on the direction of the relationship often contradict each other. Cultural distance is frequently associated with negative results, referring to negative abnormal returns (see e.g. Datta and Puia, 1995; Krug and Nigh, 1998), however, it is also associated with positive abnormal returns (see e.g. Chakrabarti, Jayaraman, and Mukherjee, 2003; Chakrabarti, Gupta-Mukherjee, and Jayaraman, 2008; Morosini, Shane, and Singh, 1998; Weber, Shenkar, and Raveh, 1996) and insignificant results, thus neither a positive nor negative relation (see e.g. Markides and Ittner, 1994; Markides and Oyon, 1998). The construct of culture and cultural distance is extensively reviewed by scholars (see e.g. Shenkar, 2001; Taras, Kirkman and Steel, 2010; Kirkman, Lowe and Gibson, 2006) and within the field of M&A various scholars pointed out directions for future research based on comprehensive reviews (Cartwright and Schoenberg, 2006; Stahl and Voigt, 2008; Shimizu, Hitt, Vaidyanath and Pisano, 2004) and meta-analyses (Stahl and Voigt, 2005; Stahl and Voigt, 2004; King, Dalton, Daily and Covin, 2004). Although these reviews and meta-analyses contain valuable suggestions, scholars so far have mostly failed to draw on the findings of these studies. Scholars have not updated their proxy of the cultural distance construct despite the suggestions and new evidence from practitioners in the field (see e.g. Shenkar, 2001).
The aim of this study is twofold. First, I update the cultural distance construct and use it in a regression model to test the culture-performance relationship in cross-border M&A. The most significant advancement is the use of a more recent and complete cultural distance construct. I update the Kogut and Singh (1988) index measure, which according to the best of my knowledge has not been done before. Second, I clarify our understanding of the culture-performance relationship in cross-border M&A. Hereto I use a model adopted from Chakrabarti et al. (2008) and study the effect of cultural distance on post-merger performance of acquiring firms, while using an event study methodology with a long-term oriented performance measure.
1 Global M&A monthly. A middle-market perspective on U.S., Europe, and Asia mergers & acquisitions.
3
I test the culture-performance relationship using three different specifications of the Kogut and Singh (1988) index measure. Kogut and Singh (1988) created a cultural distance measure based on the Hofstede dimensions (Hofstede, 1980). Hofstede identified four universal cultural value dimensions, on which Kogut and Singh (1988) based their measure. This measure is still used today even though a fifth and sixth dimension have been developed and were validated in later years. The measure is thus outdated (Shenkar, 2001). In the first specification of the Kogut and Singh (1988) index measure, I use only the original four Hofstede dimensions. In the second specification I include the fifth dimension, and in the third specification I add the sixth dimension to the measure.
I find significant results for the cultural distance measure in all three specifications. This means that the greater the cultural disparity between the home countries of two merging firms, the greater the potential positive abnormal returns. The results show that the evidence is the strongest for the specification using all six dimensions and thereby validates the inclusion of both the fifth and sixth dimension. My results are robust and remain significant in various robustness checks.
In the literature section the Hofstede dimensions will be discussed at length. In addition I control for various factors such as firm-level, country-level and deal-specific factors. Herein, I adopt the definition of Shimizu et al. (2004) who define cross-border M&A as “those involving an acquirer firm and a target firm whose headquarters are located in different home countries”. In doing so, I answer to the call of Cartwright and Schoenberg (2006) who identify “substantial methodological bridges to cross” in order to create a closer link between models of acquisition performance and insights from organizational and behavioural studies. Furthermore many authors invited scholars to increase the empirical examination of the culture-performance relationship in future research on cross-border M&A (see e.g. Shimizu et al., 2004). In the following section I draw upon previous research and argue the justification for the improved measure of cultural distance. In the section thereafter I elaborate on the calculation of the improved measure of cultural distance and the model specifications used to test the culture-performance relationship. In the fourth section I present my results and some robustness checks. In section 5 I conclude my study and discuss implications for future research.
2. Literature
In this section I first review and analyse the extent literature concerned with the culture-performance relationship in cross-border M&A. Thereafter I discuss the developments within the field of cross-cultural management and its implications for the model design. Lastly, I review the performance construct and indicate its consequences for the model specifications.
2.1. The culture-performance relationship in M&A
4
For example, Markides and Ittner (1994) studied the valuation consequences for a sample of international acquisitions by U.S. acquirers between 1975 and 1988. They found international acquisitions to create value for the acquirers. They controlled for the relationship between the target‟s country and the acquirer‟s country using among other variables, the cultural distance as measured by the four dimensions developed by Hofstede (1980). None of the four dimensions however proved to be significant. Acquisitions in English speaking countries were also found to be insignificant even though many have claimed language barriers to be a strong impediment to successful cross-cultural management2. For a sample of 236 acquisitions in a similar study, Markides and Oyon (1998) found U.S. international acquisitions in continental Europe to create value, whereas acquisitions in Canada and Britain did not. To control for the socio-cultural distance between the acquirer‟s country and target‟s country they used the Masculinity dimension (Hofstede, 1980) which proved to be insignificant.
Studies that find a negative relationship between culture and performance in cross-border M&A are somewhat less abundant in the field. Krug and Nigh (1998) studied manufacturing firms in the U.S. who were acquired by a foreign firm. They find cultural distance between the home country of the acquirer and the U.S. to be positively related to the top management departures in the six-year period after the merger. To measure cultural distance Krug and Nigh (1998) apply Kogut and Singh‟s (1988) index measure (on which I elaborate in paragraph 2.2.2.) based on the first four cultural dimensions developed by Hofstede (1980). In studying the influence of relatedness and cultural fit on shareholder value creation, Datta and Puia (1995) reach a similar conclusion. They study 112 cross-border acquisitions of U.S. firms between 1978 and 1990 and used the same index measure for cultural distance, but they split their sample in two, based on low/high cultural distances and perform a univariate analysis of the variable, whereas Krug and Nigh (1998) use a regression analysis. Datta and Puia (1995) find that for four out of five event windows included in their model, the „high cultural distance‟ subsample shows significantly lower cumulative excess returns than the „low cultural distance‟ subsample.
In strong contradiction to the findings of the research mentioned in the two previous paragraphs, many scholars find evidence for a positive relation between culture and performance in cross-border M&A. Based on correlation matrices of their samples Weber et al. (1996) find organizational cultural differences to have a negative effect on top management commitment in domestic M&A. For cross-border M&A they find an inverse relationship between organizational culture and top management commitment. Furthermore they find the dimension Power Distance to be significantly correlated with integration and the dimension Individualism with employees‟ attitudes towards cooperation with other top management team in cross-border M&A, suggesting smoother integration as cultural distance increases. On a different note Morosini et al. (1998) hypothesize that national cultural distance improves cross-border acquisition performance by providing the merging firms access to each other‟s diverse set of routines embedded in national culture. For a sample of 52 acquisitions in Italy they find cultural distance as measured by the Kogut and Singh (1988) measure as well as the dimension Uncertainty Avoidance (Hofstede, 1980) both to be
2
5
significant. In a different study on the longevity of foreign entries cultural distance had a positive sign for a subsample of acquisitions in where culture was measured using both Kogut and Singh‟s (1988) measure and an alternative measure developed by Ronen and Shenkar in 1985 (Barkema, Bell and Pennings, 1996). Chakrabarti et al. (2008) study 800 cross-border acquisitions and find that cross-border acquisitions perform better in the long run if the acquirer and the target come from countries that are culturally more disparate. They do this using an event-study methodology and Kogut and Singh‟s index measure for cultural distance, while controlling for various deal-specific and country-specific factors.
2.2. Cultural distance
„Collective programming of the mind‟ is the term coined by Hofstede (1980) to describe culture and values in societies, organizations and at the individual level. Systems of values which constitute culture are part of the collective programming of our minds. In line with Hofstede (1980) I define cultural distance as: the difference between value systems in societies. Thus, in comparing culture and measuring cultural distance I compare the different „value systems‟ across countries. There are various theoretical perspectives on the role of culture in M&A which can be grouped in to the following three categories (Stahl and Voigt, 2008):
the cultural fit perspective;
the acculturation perspective;
and the social constructivist perspective.
The cultural fit perspective is based on the notion that the cultural fit between the merging firms is a critical determinant of the post-merger integration process. Cultural fit models seek to explain the relationship between pre-merger cultural differences and post-merger performance. A different perspective to the role of culture in M&A is the acculturation perspective. This perspective is based on the premise that during the acculturation process changes are induced in the two cultural systems of the merging firms. The outcome is a jointly developed culture with diffused cultural elements of both pre-merger cultures. The higher the degree of post-merger integration is desired from a strategic point of view, the more important the acculturation process becomes. Lastly, in contradiction to the more functional and objective understanding of culture in acculturation and cultural fit models, the social constructivists view culture as a more dynamic construct that constantly changes due to the production and reproduction of shared patterns of interpretation. This model rather contradicts the view of culture as a stable system of norms and values. In this study I adopt the cultural fit perspective in explaining the culture-performance relationship in cross-border M&A.
6
made by researchers contribute further to the incomparability. Cultural differences are often not clearly defined and there is a great variety in measures and approaches used (Teerikangas and Very, 2006). To avoid further confusion surrounding the level of analysis, the construct used in this study is national culture.
2.2.1. Hofstede’s cultural dimensions
Using factor analysis based on survey data3, Hofstede (1980) constructed four main dimensions that addressed basic problems all societies have to deal with. In 1991 these dimensions were formulated as follows: Power Distance, Uncertainty Avoidance, Individualism, and Masculinity (Hofstede). And this period is described as the „four-dimension period‟ (Minkov and Hofstede, 2011). In his monograph Cultures and
organizations: Software of the mind, Hofstede (1991) introduced a fifth dimension in
collaboration with Michael Bond and they labelled it: Long-Term Orientation. This dimension addressed the focus of people‟s efforts on the past, present and future, and was significantly correlated with the years before and after the CVS study4, proving its validity (Minkov and Hofstede, 2011). Upon years of analysing the data from the World Values Survey5 and other databases Michael Minkov found a sixth universal dimension which he labelled Indulgence versus Restraint and was adopted by Hofstede and added to his model (Hofstede, Hofstede and Minkov, 2010).
Hofstede is an influential and highly cited researcher in the field of cross-cultural management. According to the Google Scholar Citation Index, Geert Hofstede has an h-index of 72 and an i10-index of 169. His 1980‟s Culture’s Consequences: International Differences
in Work-Related Values and consecutive work has made a significant impact on academic
field and inspired thousands of empirical studies (Sivakumar and Nakata, 2001). For example, Hofstede´s work is cited more than 44,000 times in the last five years (Google Scholar Citation Index)6.
In addition to research on the culture-performance relationship in M&A Hofstede‟s dimensions have been used in various fields. La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1988) used them to study the institutional protection of shareholders and creditors. Tung (1982) used the dimensions to study effectiveness of training, Govindarajan, Gupta and Wang (2001) to study efficiency of global value chains and Newman and Nollen (1996) used the dimensions to study the fit between national culture and management practices. The number of applications is endless.
Advocates of Hostede‟s metrics of cultural distance often cite his book of 1980 and refer to its academic impact. Critics on the other hand commonly mention the possible bias due to the fact that the findings of his work are based on a survey on employees of only one large multinational business organization (Hofstede, 1980). These statements suffer some severe
3 Hofstede collected data within subsidiaries of IBM in 40 countries. The survey data was collected in 1968
and for a second time in 1972, producing in total over 116,000 questionnaires. In addition data was collected among managers participating in international management development courses which were unrelated to IBM.
4
The Chinese Value Study (CVS), developed by Bond and colleagues, was designed specifically for people having Eastern life values. It focused on four dimensions and showed considerable difference to responses on Western value questionnaires (Matthews, 2000).
5 www.worldvaluesurvey.org 6
7
limitations. They largely neglect consecutive work by Hofstede and others, ignore more recent evidence supporting claims of the use of Hofstede‟s dimensions, and they fail to follow the suggestions for improving the specification of the culture-performance relationship in cross-border M&A models.
The work of Hofstede from 1980 is criticized (see e.g. Shenkar, 2001). The critiques can be approximately aligned along the following categories: (1) non-representative; (2) too simplistic; and (3) lacking updates and outdated. The claim that the Hofstede dimensions are not representative is invalid. The dimensions are validated by other researchers based on multiple large value studies such as the CVS study (Minkov and Hofstede, 2011). Although the Hofstede dimensions have some limitations and suffer legitimate scrutiny, they have been proven to be very helpful in explaining the direction of the relation between culture and M&A outcomes, notwithstanding the immense strengthening of our understanding of the culture construct. And even though Hofstede‟s dimensions can be seen as a quantitative measure of a „soft‟ concept used to measure or explain „hard‟ data, the concept is highly accessible and logically convincing (Drogendijk and Zander, 2010). Additionally, by updating the measure based on more recent findings and including more dimensions I reduce the simplicity of the metrics.
2.2.2. The Kogut and Singh index measure
In his review of the cultural distance construct, Shenkar (2001) refers to the Kogut and Singh measure as “a rather simplistic aggregate of Hofstede‟s (1980) dimensions” and therefore considers the measure to be liable to the same critiques as Hofstede is. The index has not been updated since its inception, even though Hofstede and others have continued to produce consecutive work providing new and improved insights (Shenkar, 2001). Furthermore, Shenkar (2001) argues that in constructing the Kogut and Singh measure the authors assume equivalence which is invalid. The index measures and weighs all dimensions equally, suggesting that their explanatory powers are equal. This is intuitively unlikely and multiple scholars have focused on individual dimensions. In acknowledging the possibility for some dimensions to have a stronger or weaker influence on the dependent variables I account for this by adding the different dimensions separately as independent variables to one of my model specifications which is common in the field (see e.g. Chakrabarti et al., 2008; Kogut and Singh, 1988; Morosini et al., 1998). The reason why no scholars have updated the Kogut and Singh measure to date remains somewhat of a mystery. Researchers seem to stick to the status quo and do not deviate from the existing paths. I follow Shenkar (2001) in his suggestion and provide the field with an updated index measure.
2.3. M&A performance
8
investors in the acquirer is more questionable. In their analysis of the acquisition performance construct Zollo and Meier (2008) argue that short-term window event-studies measure something different than actual acquisition performance, but rather they measure the collective cognitive heuristic, i.e. the overall market sentiment. Apparently there exists a certain degree of information asymmetry or lack of foresight as to predict the outcome of an acquisition based on common knowledge and public information available at the announcement date (Zollo and Meier, 2008). This stands to reason that the efficient market hypothesis on which the justification for the use of announcement date returns are based cannot hold in practice. Therefore, in this study I use buy-and-hold abnormal returns (BHAR) for a 36-month window which is considered to be the appropriate long-term performance measure as it said to “precisely measures investor experience” (Barber and Lyon, 1997; Lyon, Barber and Tsai, 1999). The downside of the BHAR methodology is that possible biases can arise from new listings, rebalancing of benchmark portfolios and skewness due to long-horizon returns (Ikenberry, Lakonishok and Vermaelen, 1995). However, these biases can be mitigated using large sample sizes (Mitchell and Stafford, 2000).
2.4. Hypotheses
Based on the above literature review, I hypothesize that cultural distance is positively related to the long-term performance of cross-border mergers and acquisitions. Thus, greater cultural disparity between the home countries of merging partners will generate positive abnormal returns for acquirer´s shareholders.
Furthermore it has become clear that the fifth and sixth Hofstede dimensions, together with the original four dimensions most accurately represent the cultural value systems of different countries. Therefore, I expect the updated Kogut and Singh (1988) index measure using all six Hofstede dimensions to provide stronger evidence for the influence of culture on cross-border M&A performance than the original Kogut and Singh index.
By testing this hypothesis I finally provide the academic field with an updated measure of cultural distance and I contribute to our standing of the culture-performance relationship in cross-border M&A. In the next section I elaborate on the calculation of the long-term performance measure7 and the research approach I take in this study.
3. Methodology
King et al. (2004) advised future researchers to build on past research models instead of creating new models. Ignoring or overlooking of control variables for which theoretical justification has been proven leads to a decreased fit of the model and can potentially cause non-normality in model specifications (Brooks, 2008). Chakrabarti et al. (2008) studied the culture-performance relationship in M&A through a well specified model including all important and regularly used control variables. Furthermore, the model developed by Chakrabarti et al. (2008) is one of the most recent and complete models developed in the field, if not the most recent. In studying the culture-performance relationship in my sample of
7 See for an extensive narrative on the pros and cons of the BHAR and its possible biases Mitchell and
9
cross-border M&A deals I use an improved measure of the cultural distance construct and draw on a model used by Chakrabarti et al. (2008).
The analysis and empirical test are based on a sample of deals collected through FactSet over a 10-year period [2000-2009] and observations are limited to cross-border acquisitions. I follow Chakrabarti et al. (2008) in applying the following criteria to the deals in choosing the sample. I include transactions:
1. that are completed;
2. that are over $100 million in value;
3. where the acquirer and target are from different countries;
4. where the acquirer owns 100% target shares after the transaction8;
5. where the acquirer is publicly traded; and where both the target‟s and the acquirer‟s nation are known.
The sample is limited to transactions of over $100 million in value for two reasons. First, economic gains from M&A are most likely to be detected when the target firm is large in relation to the acquirer and, second, although less in number of transactions, the total dollar value of the sample accounts for a significant portion of the dollar value of all cross-border M&A activity in the period9. Additionally the threshold of $100 million increases the comparability of this study. In constructing the sample I use the announcement date of the acquisition. Thus completed deals announced on and after January 1st, 2000 and before or on December 31st, 2009 are included. Stock market data for acquirers and total market index returns for the acquirer‟s country are obtained through FactSet. To be able to compare and have uniformity across countries I use stock market indices from FactSet.
To avoid contamination of the returns in the event window I exclude acquirers who conduct deals in the three-year horizon following a previous deal. This results in the exclusion of many transactions, due to the fact that many firms conduct large deals every one to two years, sometimes even twice a year. Finally, as is a common practice in research concerning cross-border M&A and national culture I exclude observations of which either the acquirer or target firm is from a tax haven10 to avoid “shell” operations. Furthermore, reverse mergers were excluded from the sample to avoid distortions in the calculations of the variable size, which is measured as the deal-value divided by the market-value of the acquirer. Additionally
8
This also means that deals with multiple buyers (e.g. club deals) are excluded as the percentage stake in the target by each acquirer is less than 100%.
9 Similar arguments where provided by Cornett and Tehranian (1992), who studied post-acquisition
performance of bank mergers. Since Cornett and Tehranian (1992), the threshold of $100 million has become common practice in M&A research (see e.g. Cybo-Ottone and Murgia, 2000; Chakrabarti et al., 2003; Ismail and Davidson, 2005; Chakrabarti et al., 2008).
10 Tax havens excluded are: Andorra, Anguilla, Antigua and Barbuda, Aruba, Commonwealth of the
10
I exclude related-party11 transactions as these transactions also contaminate the variable size and have been shown to create significant negative excess returns at both the initial announcement and the 12-month period following the announcement (Cheung, Rau and Stouraitis, 2006). Eliminating the reverse mergers and related-party transactions from the sample only decreased the sample size by six observations.
The sample characteristics by country are presented in table 1. By far the United States is most often the country of origin for both acquiring and target firms, and constitutes roughly one-third of the total sample. For both acquiring and target firms the United Kingdom follows in a second place and Canada in the third place. What is immediately apparent is that most cross-border M&A seems to be limited to developed countries. The sample lacks acquirer and target country observations from the African continent. Furthermore observations from Latin America and Asia are limited. Observations from Latin America are limited to Argentina, Chile, Brazil and Mexico, while observations in Asia consist for the most part of deals involving firms from China, Japan, Hong Kong and India. The sample includes one acquiring firm and one target firm from the United Arab Emirates, making the UAE the only Middle-Eastern country in the sample. However, unfortunately Hofstede‟s (1980) work did not cover the United Arab Emirates and therefore these observations are not included in models that include the cultural distance variable.
When compared to the sample of Chakrabarti et al. (2008) which was collected in the period beginning in 1991 till 2004 there are some notable differences. First of all they have numerous deals involving target firms from New Zealand and deals involving acquiring firms from South Africa and Singapore, both which are absent in my sample. Furthermore, their sample includes many more deals involving acquiring firms and/or target firms from Scandinavian countries such as Norway, Sweden and Finland, whereas my sample only contains one acquisition of an Norwegian firm.
11 Related-party transactions include deals between a listed firm and either its large shareholders or entities
11
Table 1 Breakdown of the sample on the most common home countries of acquirers and targets.
Acquirer home countries No. of acquisitions Target home countries No. of acquisitions
United States 217 United States 211
United Kingdom 74 United Kingdom 120
Canada 60 Canada 68 France 38 Germany 55 Netherlands 33 France 36 Japan 32 Netherlands 26 Germany 31 Australia 20 Australia 27 Switzerland 17 Switzerland 27 Italy 11 Spain 17 China 9 Italy 16 Brazil 9
India 15 Hong Kong 8
Hong Kong 13 Spain 7
Mexico 8 Portugal 6 Ireland 8 Japan 6 China 5 Mexico 6 Austria 4 Ireland 5 Malaysia 4 Russia 5 Others 7 Others 11 N = 636 N = 636
In table 2 I report the most common pairs by home country of the merging firms and the number of unique pairs12. The twelve most common pairs constitute little more than half the sample. The highest numbers of transactions are closed between partners in the US and the UK. Together, partners within these two countries closed 104 transactions. Furthermore, a large part of cross-border M&A is conducted in North America alone. One hundred two deals out of the 636 deals are solely conducted between partners in the US and Canada.
Table 2 Common pairs and number of unique pairs.
Acquiring countries Target countries Unique pair No. of acquisitions
United States United Kingdom 1 65
United States Canada 1 56
Canada United States 1 46
United Kingdom United States 1 39
United States Germany 1 24
Japan Unites States 1 19
France United States 1 17
Netherlands United States 1 17
United States France 1 15
Australia United States 1 14
Germany United States 1 12
India United States 1 10
Other pairs 122 302
Total number of unique pairs and number of acquisitions 134 636
12
12
The characteristics of the sample are shown in table 3. The portion of deals that are neutral/hostile is only 3%. And in 14% of the deals a tender offer was made. These percentages are similar to the sample used by Chakrabarti et al. (2008). About 50% of the transactions are paid for with 100% cash, whereas Chakrabarti et al. (2008) had 76% of transactions paid for with cash. Most deals are unrelated and only involve one bidder.
Table 3 Sample description of cross-border deals in 2000-2009.
Number Percent
Total number of acquisitions 636 100%
Cash 320 50% Non-cash 316 50% Friendly 614 97% Hostile/neutral 22 3% Tender offer 86 14% No tender offer 550 86% Competing bidders 14 2% One bidder 622 98% Unrelated 345 54% Related 291 46%
Cash vs Non-cash, friendly vs hostile/neutral, tender offer vs non-tender offer, competing bidders vs one bidder, unrelated vs related (at either the four or three-digit SIC level) are the transaction characteristics used to describe the sample.
3.1. Long-term performance
The buy-and-hold returns on an investment in a security (e.g. firm‟s stock), minus the return on a buy-and-hold investment in a portfolio with an appropriate return (e.g. the return on the market), is the buy-and-hold abnormal return. It is computed as follows:
∏[ ]
∏[ ( )]
(1)
where is the number of periods used to compute the returns, is the return on the individual security for time , and ( ) is the return on the market. In my case I use the return of the primary index in the acquirer‟s country of domicile.
3.2. Cultural distance
I measure cross-cultural differences with the Hofstede dimensions which are adopted from Hofstede (1980, 2001, 2005) and (Hofstede et al., 2010). The distances are calculated using the methodology first developed by Kogut and Singh in 1988. Kogut and Singh (1988) formed a composite index based along the four dimensions identified by Hofstede in 1980. Kandogan (2012) algebraically simplified the measure and constructed it as follows:
∑
( )
13
where is the cultural distance between countries and , is one country‟s score on dimension , is the other country‟s score on dimension , is the variance of the index for the dimension, and is the number of dimensions. However, research of the past two decades calls for inclusion of the other two dimensions into our model, i.e. Long-Term Orientation and Indulgence versus Restraint. From the above algebraic simplification it becomes obvious that Kogut and Singh‟s (1988) measure can be easily mathematically extended to all six dimensions. Upon doing so the measure is computed as follows:
∑
( )
(3)
where is the cultural distance between the target‟s and the acquirer‟s country and is calculated from the numerical values of the six dimensions: Power Distance (PDI), Uncertainty Avoidance (UAI), Individualism (IDV), Masculinity (MAS), Long-Term Orientation (LTO), and Indulgence versus Restraint (IVR).
3.3. Deal-specific variables
Prior research has indicated and identified multiple deal-specific variables that affect the outcome of M&A. Based on the characteristics of those variables I construct dummy variables to control for these effects. First, independent of the type of takeover bid, i.e. friendly or hostile, the method of payment is found to be significantly related to abnormal returns for stockholders. Bidding firms using pure stock exchange as the method of payment experience negative abnormal returns, whereas bidding firms using pure cash as the method of payment show neither negative nor positive abnormal returns (Travlos, 1987). The method of payment is commonly studied in M&A research (King et al., 2004) and the cash_dummy variable takes the value 1 if the acquirer paid with 100% cash for the shares of the target and 0 otherwise.
Second, the attitude of the acquirer has been cited to influence the post-merger performance and is often controlled for (see e.g. Healy, Palepu and Ruback, 1992; Finkelstein and Haleblian, 2002). Acquisitions can be considered hostile or friendly. According to Mallette and Fowler (1992) for example hostile acquisitions are less likely to lead to success. The control variable friendly_dummy takes a value of 1 when an acquisition is friendly and 0 when an acquisition is hostile/neutral13.
Third, Bruner (2002) has found indications that acquisitions made through a tender offer enjoy high abnormal returns (16%)14. The argument here is that it can pay to directly appeal to target firms‟ shareholders and sidestep the target firms‟ management. An earlier study found statistically significant results of abnormal returns of 9% for acquisitions made through tender offers, even after adjusting for book-to-market ratios (Raghavendra Rau and
13 For more on the role of hostile attitudes in takeovers see e.g. Martin and McConnell (1991), or Frank and
Mayer (1996).
14 See Bruner (2004), p. 10 for references to numerous other studies concerning tender offers and acquirer
14
Vermaelen, 1998)15. To control for tender offers the tender_dummy variable takes a value of 1 when an acquisition was made through a tender offer and 0 otherwise.
Fourth, multiple bidders have been shown to influence the performance in mergers and acquisitions (see e.g. Servaes, 1991) and therefore the variable no_of_bidders is added which represents the number of firms bidding for a target.
Finally, I control for firm size. The firm size effect has been widely acknowledged by leading scholars in the field for many years (see e.g. Agrawal, Jaffe and Mandelker, 1992; Comment and Jarrell, 1995). The average market value of firms in the sample is $10,967 million, while the median market value is only $2,604 million. To control for this variability in size, I measure firm size as the natural logarithm of the acquirer‟s market value in the month prior to the acquisition and refer to the variable as log(mv).
3.4. Country-level variables
Other than cultural differences various country-specific factors are likely to influence the outcome of cross-border M&A. Economic differences such as differences in income, bilateral trade, international trade and exchange rates can play their part, but also the difference in corporate governance between countries can influence M&A outcome.
To reflect the difference in wages the per capita income differences between two countries is considered a suitable proxy. PCI_diff is used to measure this economic disparity and is calculated as follows:
(4)
I use openness_target to represent the target country‟s economy‟s interconnectedness to the global economy. A target country‟s openness can influence the ease of how factors of production, personnel and financial means can be transported across borders. It is computed as follows:
(5)
The bilateral trade between the two countries is simply the logarithm of the target country‟s exports to and the import from the acquirer‟s country summed, taken in the year prior to the effective year of acquisition.
( ) (
) (6)
The fourth economic variable is exchange rate volatility which has been related to foreign direct investment (see e.g. Chakrabarti and Scholnick, 2002) and therefore is likely to have an influence on M&A decisions. I incorporate the variable forex_volatility into my model specifications using the standard deviation of the exchange rate volatility measured over the
15 Five years previously, Fama and French (1993) argued that abnormal positive returns on tender offers
15
period –36 months to –1 month. To make sure exchange rate volatility can be compared across country pairs I first compute percentage changes for all 36 months, thus resulting in 35 data points per country pair. All data regarding per capita GDP, GDP, and import and export figures are obtained from the World Bank. Bilateral trade data was provided by the OECD and exchange rates come from the FactSet Database.
Lastly, it is often argued that the target firm incorporates the corporate governance system from the acquirer‟s home country. The difference between corporate governance in the two countries says something about the difference in investor protection between the countries. To measure corporate governance I use the Antidirector Rights Index (ADRI) of each country. The ADRI was pioneered by La Porta et al. (1998) and in 2005 the ADRI received an update by Spamann which was necessary as the institutional environment changes with time. Spamann published his Revised ADRI in The “Antidirector Rights Index” Revisited, 2010. I use the Revised ADRI as it better represent the investor protection in a time that coincides with the period over which the sample was collected. The corp_gov_diff variable is computed fairly simple:
(7)
3.5. Socio-cultural variables
In addition to the Kogut and Singh index measure I also use alternative measures of culture. Stulz and Williamson (2003) have used religion and language to proxy cultural differences in a study examining the reason for differences in investor protection across countries. Data on these two variables is obtained from the CIA World Factbook 2012 and then matched with all the countries in the sample. The religion_dummy takes a value of 1 when both countries share the same primary religion and 0 otherwise. The language_dummy takes a value of 1 when the primary language spoken in the two countries is the same and 0 otherwise. Another socio-cultural factor often used in finance is the origin of the legal system. La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1997) have used it in studying the relationship between investor protection and the size of capital markets, but also in a study on the protection of corporate shareholders and creditors. The socio-cultural control variables are used in separate regression due to high correlations between them, for example the language dummy variable and law dummy variable show a positive correlation coefficient of 0.73. The natural logarithm of the Kogut and Singh index allows capturing of a non-linear relationship (Chakrabarti et al., 2008).
3.6. Regression
16
errors within the acquirer country16. By adopting this method the model estimates robust standard errors that are robust to heteroskedasticity and serial correlation of arbitrary forms for fixed T and large N (Arellano, 2003, p.18). And potential cross-section correlation does not affect cross-section inferences in event-studies (Prabhala, 1997).
Acquisitions occur fairly random throughout the year, i.e. some days see multiple deals closed and on some days no deals are closed. Due to infrequent distribution of observations in the time variable the “period fixed effects option” cannot be applied. Therefore I manually add dummy variables for all years. I do the same to apply target country fixed effects as I am bound to set the acquirer country variable as the cross-section identifier. The way the data is structured groups are constructed as directional pairs of countries. Thus within group estimation for Dutch target-UK acquirer pairs allows for different standard errors compared to the estimation of UK target-Dutch acquirer pairs.
4. Results
The results of six different model specifications are presented in table 4. The regressions are run in line with research conducted by Chakrabarti et al. (2008). All models presented in table 4 are regressed on the 36-month return dependent variable. The explanatory variables used are the various firm-level, country-level and deal-specific variables. Macro-economic trends are controlled for using year fixed effects. The residuals in all models are normally distributed.
In model 1, only deal-specific variables are used to explain the dependent variable. None of the explanatory variables are significant. The deal attitude variable which has been found to be significant by Chakrabarti et al. (2008) can easily deviate as only 22 deals have a hostile attitude in my sample. The previous mentioned authors also found the method of payment to be significant. The sample characteristics regarding this variable, however, differ a lot. I only find 50% of deals to be paid for with 100% cash, while Chakrabarti et al. (2008) have 76% of deals paid for with 100% cash in their sample. The coefficient on No_of_bidders remains negative throughout all model specifications. The coefficient on the intercept is negative in the first model specification, but switches signs in the second model and remains positive for all other model specifications. A similar paradox is observed on the friendly_dummy coefficient which changes signs to negative and remains negative in all other model specifications. This suggests that acquirers are to gain more from hostile takeovers than from friendly takeovers, however the percentage of deals with a hostile attitude is only a mere 3% in my sample.
In the second model country-variables are added. All country-level variables and their coefficients hold the same sign throughout all model specifications. Both openness_target and
log(bilateral trade) are significant at respectively the 10% and 5% level. This is in line with
the findings of Chakrabarti et al. (2008).
In model 3 a measure of corporate governance is added which proxies the difference between investor protection in the country of the acquirer and the country of the target. Contrary to the findings of Chakrabarti et al. (2008) corp_gov_diff is insignificant; however,
16 Covariance of standard errors in panel data at section level is sometimes referred to as “White
17
in model 5 corp_gov_diff is significant. Perhaps this can be attributed to the smaller sample size in this study.
In model 4 I finally add the construct of cultural distance as measured by the logarithm of the Kogut and Singh index measure. Model 4 is presented three times, in slightly different specifications. In model 4a, the cultural distance measure is calculated using the original four dimensions, then in model 4b using the five dimensions and in model 4c using all six dimensions. Openness_target and PCI_diff are both significant at the 10% level. In accordance with popular believe the coefficient of the cultural distance construct is positive at the 10% level. Theory posits that cultural differences may enhance the M&A performance through the transfer of capabilities, create access to partners‟ unique set of routines and mutual learning (Morosini et al., 1998; Chakrabarti et al., 2008). The cultural distance variable‟s significance (p-value = 0.0539) corresponds with Chakrabarti et al. (2008) and other studies mentioned previously. The evidence for using only the first five dimensions is less strong (p-value = 0.0959) than using just the original four dimensions (p-value = 0.0890). Evidently the results, thus, suggests that culture does play a significant role in the culture-performance relationship in cross-border M&A. In the model where the cultural distance regressor is computed using all six dimensions, R-squared is 9.81%. In model 4c R-squared remains equal at the three-digit level compared to the traditional approach of using just 4 dimensions. Model 4b has a slightly lower R-squared. Even more interesting therefore is the failure of scholars to incorporate the dimensions Long-Term Orientation and Indulgence vs Restraint.
Table 4 Regressions of buy-and-hold returns of acquirers for the three-year horizon following the acquisition.
Independent variable
Model 1 Model 2 Model 3 Model 4a Model 4b Model 4c
Coeff. Prob. Coeff. Prob. Coeff. Prob. Coeff. Prob. Coeff. Prob. Coeff. Prob.
Intercept -0.052 0.8001 -0.220 0.5536 -0.141 0.7406 -0.351 0.4259 -0.320 0.4703 -0.339 0.4566 Friendly_dummy 0.013 0.9298 -0.023 0.8796 -0.028 0.8571 -0.027 0.8631 -0.026 0.8692 -0.028 0.8559 Tender_dummy 0.121 0.2514 0.114 0.2713 0.114 0.2729 0.113 0.2888 0.112 0.2874 0.113 0.2858 Cash_dummy 0.015 0.7862 0.022 0.7109 0.022 0.7066 0.020 0.7259 0.021 0.7237 0.017 0.7700 No_of_bidders -0.147 0.2368 -0.143 0.2768 -0.149 0.2574 -0.138 0.2897 -0.137 0.2925 -0.137 0.2963 Log(mv) 0.013 0.6902 0.012 0.7158 0.010 0.7627 0.011 0.7321 0.010 0.7643 0.010 0.7760 Openness_target 1.035* 0.0622 1.016* 0.0695 1.054* 0.0620 1.061* 0.0607 1.079* 0.0564 PCI_diff 0.199 0.1076 0.192 0.1241 0.237* 0.0623 0.218* 0.0871 0.215* 0.0898 Forex_Volatility -0.271 0.8340 -0.303 0.8139 -0.279 0.8236 -0.304 0.8106 -0.304 0.8108 Log(Bilateral Trade) -0.115** 0.0290 -0.121** 0.0276 -0.097* 0.0817 -0.100* 0.0673 -0.096* 0.0927 Corp_Gov_Diff -0.018 0.5504 -0.014 0.6474 -0.014 0.6408 -0.013 0.6726 Log(Hofstede dist) 0.078* 0.0890 0.068* 0.0959 0.082* 0.0539 R²(%) 8.3 9.5 9.5 9.8 9.7 9.8 Durban-Watson stat. 2.08 2.08 2.08 2.09 2.09 2.09 N = 600 598 598 598 598 598 The dependent variable in all these models is the 36-month BHAR. All regressions are estimated using target-country fixed effects and year fixed effects. In model 4a, log(Hostede dist) is calculated using the original four dimensions. In model 4b
log(Hofstede dist) is calculated using the five dimensions and in model 4c it is calculated with all six dimensions.
18
In models 5 through 7 I control for various socio-cultural variables, beginning with religion in model 5, predominant language in model 6 and lastly controlling for legal origin in model seven17. The socio-cultural control variables; religion (-0.41), language (-0.84) and legal origin (-0.67), each show a strong correlation with the cultural distance18. Therefore the socio-cultural variables substitute the cultural distance measure in each model. The
religion_dummy takes the value of 1 when the two countries share the same primary religion.
The variable has a positive sign, but is insignificant. In this model specification the variable
corp_gov_diff, however, is significant at the 10% level.
In model 6 the language_dummy takes a negative sign and is also insignificant. The
corp_gov_diff variable loses its significance. In the last model specification the legal_dummy
is negative, but not significant. In comparison with the results found by Chakrabarti et al. (2008), the language_dummy lacks significance, which in their case proved to be highly significant (1% level).
Not all dimensions are assumed to have an equally large influence on performance, a critique on the use of the Kogut and Singh index measure. As indicated in section 2 I would include a regression with the individual dimensions added as separate regressors. The distances on each dimension are computed as an absolute value. The Kogut and Singh index also creates an absolute distance measure. Chakrabrti et al. (2008) also include the individual dimensions, however, contrary to their findings none of the individual dimensions turns out to be significant19. Thus, the measure is perhaps more coherent than previously assumed as none of the dimensions can individually explain the variation in the buy-and-hold abnormal return of the acquirers.
4.1. Robustness checks
I conduct several robustness checks to how the regression coefficients estimates behave when the model specifications are altered by removing or adding regressors. To confirm the structural validity of my model I, first, regress my model of interest on a 24-month window BHAR. Other robustness checks I conduct include acquisition relatedness, an alternative measure of firm size and separate regressions of subsamples wherein the sample is divided into one containing small and one containing large firms. The residuals in all these models are normally distributed. The regression results for the robustness checks can be found in table 5.
4.1.1. Results using 24-month BHAR
In a regression run using the 24-month BHAR as my dependent variable the results of my model remain stable. I use the 24-month window BHAR in model 4c as it was presented previously in table 4. Although the model fit decreased somewhat (R-squared = 8.9%) the
log(Hofstede dist) remains significant at the 10% level (p-value = 0.0821). log(bilateral trade)
remains significant at the 5% level (p-value = 0.0148), but openness_target and PCI_diff lose their significance. In contrast the intercept becomes highly significant (p-value = 0.0072) and
no_of_bidders becomes significant at the 10% level (p-value = 0.0698). These robust results
17
For results of model 6, 7 and 8; see Appendix B.
18 Correlation coefficients are calculated based on the socio-cultural dummy variables and the natural
logarithm of the Kogut and Singh index measure using all six dimensions. Using only the four dimensions yields similar results, i.e. correlation coefficients of -0.44, -0.86 and -0.64 respectively.
19
19
are in line with the findings of Chakrabarti et al. (2008) who also found significant results for their proxy of cultural distance regressed on a 24-month window.
4.1.2. Results controlling for relatedness
When I control for relatedness of acquisitions the results remain stable again. Related acquisitions are characterized as acquisitions between partners that have equal SIC codes at either the three-digit or four-digit level. I thus add relatedness to the specification of model 4c and I find relatedness to be insignificant. Log(hofstede dist) is significant at the 10% level value = 0.0573). The same goes for log(bilateral trade) value = 0.0818) and PCI_diff (p-value = 0.0890). And openness_target has turned significant at the 5% level (p-(p-value = 0.0489). The model fit increased as compared to the original model specification (R-squared = 10.1%). Again the robustness of the results are conclusive with the findings of Chakrabarti et al. (2008) who also find no evidence for relatedness to influence the estimates of the explanatory variables.
Table 5 Results for robustness checks
Dependent variable 24-month BHAR 36-month BHAR
Independent variable
Model R1 Model R2 Model R3
Coeff. Prob. Coeff. Prob. Coeff. Prob.
Intercept 1.142 0.0072 -0.328 0.4711 -0.311 0.5127 Friendly_dummy 0.108 0.4021 -0.034 0.8287 -0.028 0.8562 Tender_dummy 0.147 0.1224 0.107 0.3149 0.115 0.2735 Cash_dummy 0.051 0.3177 0.019 0.7470 0.019 0.7513 No_of_bidders -0.215 0.0698 -0.140 0.2818 -0.139 0.2915 Log(mv) -0.003 0.9345 0.009 0.7863 Size -0.002 0.9606 Openness_target 0.272 0.3431 1.124** 0.0489 1.073* 0.0584 PCI_diff 0.053 0.7142 0.217* 0.0890 0.217* 0.0825 Forex_Volatility 0.467 0.6773 -0.332 0.7917 -0.312 0.8055 Log(Bilateral Trade) -0.145 0.0148 -0.100* 0.0818 -0.096* 0.0921 Corp_Gov_Diff -0.043 0.1877 -0.011 0.7070 -0.014 0.6511 Log(Hofstede dist) 0.069* 0.0821 0.080* 0.0573 0.082* 0.0520 Relatedness 0.073 0.1835 R²(%) 8.9 10.1 9.8 Durban-Watson stat. 2.13 2.08 2.09 N = 626 598 598
All regressions are estimated using target-country fixed effects and year fixed effects20. ***Significant at the 1% level; **significant at the 5% level; *significant at the 10% level.
4.1.3. Alternative measure of size
Thirdly, I control for an alternative measure of size. Moeller, Schlingemann and Stulz (2004) argue that the dollar-value of acquisitions is relatively larger for small firms than for large firms. To control for this they propose the following measure:
20 For space considerations results of the dummy variables for target-country fixed effects and year fixed
20
(8)
which I substitute for the natural logarithm of the market value of the acquirer in model 4c. The result is a small decrease of the R-squared; to 9.7%. Opennes_target, PCI_diff, log_trade and log(Hofstede dist) all remain significant at the 10% level. And the respective p-values are: 0.0584; 0.0825; 0.0921; and 0.0520. The variable size is insignificant (p-value = 0.9606).
4.1.4. Large firms and small firms
Finally, the last robustness test I conduct is two-fold. I divide the sample into two subsamples and run separate regressions on both of them. The results for this robustness test can be found in Appendix C. Moeller et al., (2004) find evidence for a size effect in acquisitions. They argue that small firms tend to enjoy higher returns from acquisitions and test this by running separate regressions on subsamples of small and large firms. The sample is divided into two subsamples based on the Tobin‟s q value of the acquirer. The median Tobin‟s q is the threshold. Unfortunately, I do not find evidence that supports their claim and neither does it strengthen my results. Obviously the sample size of both subsamples is now halved, leading to a loss of the significance of most variables. And most importantly none of the measures of firm size are significant. Neither log(mv) nor size. The smaller sample size and the small number of observations in the subsample that are either hostile or have multiple bidders decreases the reliability of the results.
5. Conclusion
The body of literature on the culture-performance relationship in cross-border M&A is inconclusive. I argue that this is due to a varying methodical quality, the inadequate control for various variables and the often different units of analysis. This study does not aim to resolve the inconclusiveness, rather it tries „to open a door‟, and provides the field with an updated cultural distance measure. One for which it had been so eager. The Hofstede‟s dimensions of cultural values are legitimately scrutinized, but the academic field itself is to blame. For a long time scholars have ignored consecutive work by Hofstede and others.
By combining best practices and following up on the work of other researchers I lay the path for a new stream of updated research. With more empirical research – built on the updated measure of cultural distance – in the future the field may resolve the inconclusive results that sprawl the field to this day. Based on a regression of 598 cross-border M&A deals I find evidence for the inclusion of the fifth and sixth dimension into the Kogut and Singh index measure of cultural distance. I compare three different specifications of the Kogut and Singh (1988) index measure. The results show that the measure that includes all six dimensions is also the one to be the most significant.
21
5.1. Limitations
Adopting a long-term performance measure has its benefits, but also comes with some limitations. From the behavioural learning perspective Finkelstein and Haleblian (1999) looked at the relationship between organizational acquisition experience and post-merger performance. They find evidence for both positive and negative effects of acquisition experience. However, by avoiding contamination of the event window, firms with more acquisition experience or at least more recent experience in large acquisitions are per definition excluded from our analysis. Unfortunately this makes it impossible to control for acquisition experience without creating bias even though multiple scholars have identified it as an important factor influencing acquisition performance21.
My findings are consistent with the view that cultural disparity has a positive effect on acquisition performance, these findings cannot be generalised to all cross-border M&A activity. Studies on samples of small firms, i.e. firms with market values of less than $100 million, must prove if these findings can be generalised to the whole population. Furthermore, the sample does not include many observations from the African continent nor from the Middle-East, and only limited observations from Asia and Latin America. The findings therefore cannot be generalised to the former two regions. Implications for cross-border M&A activity involving countries from Asia or Latin America should be interpreted cautiously.
5.2. Future research considerations
First of all do I hope that in future research scholars will include the updated measure of cultural distance based on all six dimensions as I propose in this paper. More empirical justification for this „update‟ is desirable and therefore I ask scholars to pursue the same venue as I have taken. Additionally scholars may look for a way to incorporate „acquisition experience‟ into models studying long-term acquisition performance. Lastly, although I find clear evidence for the updated measure of cultural distance there are many other possible factors influencing the culture-performance relationship in cross-border M&A. Therefore scholars are encouraged to continue to improve the existing models used to test the culture-performance relationship.
References
Agrawal, A., Jaffe, J., Mandelker, G. 1992. The post‐merger performance of acquiring firms: a re‐examination of an anomaly. The Journal of Finance 47, 1605-1621.
Alexandridis, G., Mavrovitis, C., Travlos, N. 2012. How have M&As changed? Evidence from the sixth merger wave. The European Journal of Finance 18, 663-688.
Arellano, M. 2003. Panel Data Econometrics. Oxford University Press, New York.
Aybar, B., Ficici, A. 2009. Cross-border acquisitions and firm value: an analysis of emerging-market multinationals. Journal of International Business Studies 40, 1317-1338.
Barber, B., Lyon, J. 1997. Detecting long-run abnormal stock returns: the empirical power and specification of test statistics. Journal of Financial Economics 43, 341-372.
21 For more on acquisition experience and acquisition performance see Evans, Pucik and Barsoux (2002,
22
Barkema, H., Bell, J., Pennings, J. 1996. Foreign entry, cultural barriers and learning. Strategic Management Journal 151-166.
Berg, R. 2012. The anonymity factor in making multicultural teams work: virtual and real teams. Business Communication Quarterly 75, 404-424.
Brooks, C. 2008. Introductory Econometrics for Finance. Cambridge University Press, New York.
Bruner, R. 2002. Does M&A pay? A survey of evidence for the decision-maker. Journal of Applied Finance 12, 48-68.
Cartwright, S., Schoenberg, R. 2006. Thirty years of mergers and acquisitions research: recent advances and future opportunities. British Journal of Management 17, S1-S5.
Chakrabarti, R., Gupta-Mukherjee, S., Jayaraman, N. 2008. Mars–Venus marriages: culture and cross-border M&A. Journal of International Business Studies 40, 216-236.
Chakrabarti, R., Jayaraman, N., Mukherjee, S. 2003. Cross-border M&A: does culture matter?. Unpublished working paper. Georgia Institute of Technology, Atlanta.
Chakrabarti, R., Scholnick, B. 2002. Exchange rate expectations and foreign direct investment flows. Weltwirtschaftliches Archiv 138, 1-21.
Cheung, Y., Rau, P., Stouraitis, A. 2006. Tunneling, propping, and expropriation: evidence from connected party transactions in Hong Kong. Journal of Financial Economics 82, 343-386
Comment, R., Jarrell, G. 1995. Corporate focus and stock returns. Journal of Financial Economics 37, 67-87. Cornett, M., Tehranian, H. 1992. Changes in corporate performance associated with bank acquisitions. Journal of Financial Economics 31, 211-234.
Cybo-Ottone, A., Murgia, M. 2000. Mergers and shareholder wealth in European banking. Journal of Banking & Finance 24, 831-859
Datta, D., Puia, G. 1995. Cross-border acquisitions: an examination of the influence of relatedness and cultural fit on shareholder value creation in US acquiring firms. Management International Review 337-359.
Drogendijk, R., Zander, L. 2010. Walking the cultural distance: in search of direction beyond Friction. Advances in International Management 23, 189-212.
Evans, P., Pucik, V., Barsoux, J. 2002. The Global Challenge: Frameworks for International Human Resource Management. McGraw-Hill, New York.
Fama, E., French, K. 1993. Common risk factors in the returns on stocks and bonds. Journal of Financial Economics 33, 3-56.
Finkelstein, S., Haleblian, J. 2002. Understanding acquisition performance: the role of transfer effects. Organization Science 13, 36-47.
Franks, J., Mayer, C. 1996. Hostile takeovers and the correction of managerial failure. Journal of Financial Economics 40, 163-181.
23
Haleblian, J., Finkelstein, S. 1999. The influence of organizational acquisition experience on acquisition performance: a behavioral learning perspective. Administrative Science Quarterly 44, 29-56.
Healy, P., Palepu, K., Ruback, R. 1992. Does corporate performance improve after mergers? Journal of Financial Economics 31, 135-175.
Hitt, M., Harrison, J., Ireland, R., Best, A. 1998. Attributes of successful and unsuccessful acquisitions of US firms. British Journal of Management 9, 91-114.
Hofstede, G., Hofstede, G., Minkov, M. 2010. Cultures and Organizations: Software of the Mind, third edition. McGraw-Hill, New York.
Hofstede, G. 1980. Culture's consequences: International Differences in Work-Related Values. Sage Publications, Thousand Oaks.
Ikenberry, D., Lakonishok, J., Vermaelen, T. 1995. Market underreaction to open market share repurchases. Journal of Financial Economics 39, 181-208.
Ismail, A., Davidson, I. 2005. Further analysis of mergers and shareholder wealth effects in European banking. Applied Financial Economics 15, 13-30
Kandogan, Y. 2012. An improvement to Kogut and Singh Measure of cultural distance considering the
relationship among different dimensions of culture. Research in International Business and Finance 26, 196-203. King, D., Dalton, D., Daily, C., Covin, J. 2004. Meta‐analyses of post‐acquisition performance: indications of unidentified moderators. Strategic Management Journal 25, 187-200.
Kirkman, B., Lowe, K., Gibson, C. 2006. A quarter century of Culture's Consequences: a review of empirical research incorporating Hofstede's cultural values framework. Journal of International Business Studies 37, 285-320.
Kogut, B., Singh, H. 1988. The effect of national culture on the choice of entry mode. Journal of International Business Studies 411-432.
Krug, J., Nigh, D. 1998. Top management departures in cross-border acquisitions: Governance issues in an international context. Journal of International Management 4, 267-287.
Lopez-de-Silanes, F., La Porta, R., Shleifer, A., Vishny, R. 1998. Law and finance. Journal of Political Economy 106, 1113-1155.
Lyon, J., Barber, B., Tsai, C. 1999. Improved methods for tests of long‐run abnormal stock returns. The Journal of Finance 54, 165-201.
MacKinlay, C. 1997. Event studies in economics and finance. Journal of Economic Literature 35, 13-39. Mallette, P., Fowler, K. 1992. Effects of board composition and stock ownership on the adoption of “poison pills”. Academy of Management Journal 35, 1010-1035.
Markides, C., Ittner, C. 1994. Shareholder benefits from corporate international diversification: evidence from US international acquisitions. Journal of International Business Studies 343-366.
Markides, C., Oyon, D. 1998. International acquisitions: do they create value for shareholders? European Management Journal 16, 125-135.
