The Impact of Country Borders on Mergers and Acquisitions:
A Firm-Level and Country-Level Analysis
Martijn Hendrik Boeve
2
The Impact of Country Borders on Mergers and Acquisitions:
A Firm-Level and Country-Level Analysis
Martijn Hendrik Boeve
University of Groningen, Faculty of Economics and Business
__________________________________________________________________________________
Abstract:
This study examines to what extent country borders matter for cross-border mergers and acquisitions (M&As). Throughout this thesis I make a clear distinction between two levels of analysis: firm-level and country-level. Country borders influence both the firm performance and the volume of M&As between a country-pair. The results of this study show a significant difference between the stock market reaction of domestic deals compared to cross-border deals. On average, the following main indicators are present when the cross-border M&A activity is likely to be high: relatively high economic performance and strong investor protection of the target country, relatively open economy of the acquirer country, similar legal and cultural environment, and geographical proximity. A country-pair with a high degree of M&A activity has less uncertainty and consequently this will positively influence the stock market reaction of acquisitions between these countries.
Keywords: Cross-Border Mergers and Acquisitions, Abnormal Returns, Country-Specific
Characteristics
JEL codes: G34
__________________________________________________________________________________
MSc Thesis Finance Martijn Boeve, s1879650
3
Table of contents
1. Introduction ... 4
2. Literature review ... 6
2.1 Wealth creation of cross-border mergers & acquisitions ... 6
2.2 The influence of country-specific differences ... 7
3. Methodology ... 9
3.1 Firm-level methodology: Event study ... 9
3.2 Country-level methodology: Cross-sectional ordinary least-squares regression analysis 11 4. Data ... 13
4.1 Data collection ... 13
4.2 Hypotheses ... 15
5. Results ... 16
5.1 Testing for differences between domestic and cross-border M&As ... 16
5.2 The influence of the country-specific variables ... 18
6. Conclusion ... 22
References ... 24
Appendix A: Number of cross-border acquirers, targets, and domestic deals ... 26
Appendix B: CAARs for each subsample, t-statistics, and p-values ... 30
Appendix C: Comparison of CAARs for symmetrical event windows ... 32
4
1. Introduction
This study examines the importance of country borders for mergers and acquisitions (M&As).1 Although already a lot of research has been done about domestic mergers and acquisitions, less is explored about cross-border M&As. Country borders influence both the firm performance as well as the volume of M&As between a country-pair. One of the main aims of this thesis is to complement the existing literature on M&As by clearly highlighting the role of country borders in the valuation differential between domestic and cross-border M&As. The announcement of a domestic M&A deal is valued substantially higher compared to cross-border M&As. This differential is highly relevant for the managers decision-making process. Furthermore, I link a twofold levels of analyses to enhance our understanding of the impact of country border on cross-border M&As. This thesis bridges the gap between former firm-level and country-level research. Finally, by using a larger country sample and a different, more recent time frame, the research has a much more global and topical coverage compared to other authors. The total sample contains 73,230 cross-border M&As, of which 15,602 belong to publicly traded firms. This larger sample size is more representative of the population and limits the influence of outliers.
In this thesis I continue to build on empirical literature which state that country borders matter for M&As. The research objective of the thesis will be to answer the following research question: To what
extent do country borders matter for mergers and acquisitions on firm-level and country-level?
A particular country is considered to be attractive for acquisitions when it has a relatively high volume of M&As. Subsequent to the assumption that managers focus on maximizing shareholders’ value, I assume that when the degree of cross-border M&A activity between a certain country-pair is high (high attractiveness), the stock market reaction to such an acquisition will also be higher compared to countries with less cross-border M&A activity.
Throughout this thesis I make a clear distinction between two levels of analysis: firm-level and country-level. In the context of cross-border M&As these different levels are strongly interrelated. On a firm-level, I indicate the valuation difference between domestic and cross-border M&As. On a country-level, the country-specific differences and similarities between the acquirer’s country and the target country are analyzed in an attempt to explain this valuation differential. In order to make the research question more strictly defined and quantifiable, it is divided into the following subquestions:
1. On a firm-level, what is the influence of country borders on the stock market reaction of a cross-border merger or acquisition?
2. On a country-level, to what extent do country-specific differences explain the degree of cross-border M&A activity between a certain country-pair?
1 The terms mergers, acquisitions, M&As, and deals will be used interchangeably throughout this thesis but all
5 Stock markets react differently to domestic M&As compared to acquisitions across borders. Country borders affect the risk-return characteristics of these deals because of various external factors. From the perspective of cross-border M&As, country-specific differences that exist because of country borders can be measured in multiple ways. Obviously, there exists a certain spatial distance between the country of the acquirer and the country of the target. Besides this physical distance, there exist inter alia economic, cultural, and legal differences between one country and another, e.g. religion, language, accounting standards, et cetera. These differences might influence the volume of a cross-border acquisitions between a certain country-pair.
The firm-level subquestion will be answered by performing an event study, the country-level subquestion by a cross-sectional ordinary least-squares regression analysis to measure the influence of several country-specific variables on the degree of cross-border M&A activity.
6
2. Literature review
Mergers and acquisitions refer to the consolidation of two firms, in which a new company is formed or in which one company is purchased by another. There are multiple reasons for M&As to occur. Among others, M&As take place to create synergies, to generate economies of scale, to enlarge the product scope, to diversify risk, to obtain market (or monopolistic) power, to change management after poor performance of the target management, or to maximize managers utility and minimize agent conflict. Generally, mergers arise when an acquirer perceives that the value of the combined firm is greater than the sum of the values of the separate firms (Ruback & Jensen, 1983). However, especially for larger firms, agency problems play a role in acquisitions because of empire building and hubris of the acquirer. Hubris of the acquirer implies manager’s overconfidence about the benefits of an acquisition. It can explain why firms place a bid above the market price and therefore pay too much for a target (Roll, 1986).
2.1 Wealth creation of cross-border mergers & acquisitions
Although already a lot of research has been done about domestic mergers and acquisitions, less is explored about cross-border M&As. However, due to the increasing globalization of financial markets, expansion of economies, greater capital mobility, and technological improvements cross-border acquisitions become more and more common (Hur et al., 2011).
The eclectic paradigm is a theoretical framework developed by Dunning (1980) to analyze the motives for undertaking foreign direct investments, including cross-border M&As. According to this theory cross-border acquisitions create wealth by internalizing the firm's ownership-specific advantages in foreign locations. These are financial, technological, informational, or organizational advantages. Financial variables of targets that have a significant effect on the firm value of an acquirer include (Gonzalez et al., 1997): price-to-earnings ratio, leverage, dividend payout ratio, capitalization or investment in fixed assets, current ratio, and firm size. These financial variables, however, are beyond the scope of this research.
The shareholders’ wealth creation by the internalization of the ownership-specific advantages can be different for cross-border M&As compared to domestic M&As. Uysal et al. (2008) find evidence that acquirers earn significantly higher returns in local M&A transactions relative to non-local (not necessarily cross-border) transactions. Here local transactions imply that the acquirer and target firm are located within 100 km of each other. These findings indicate that distance between two locations can play a role in the wealth creation. According to Uysal et al. (2008), information advantages arising from the geographical proximity seem to facilitate this higher stock market reaction.
7 difference in wealth creation between a domestic and a cross-border acquisition will be referred to as the cross-border effect. The factors influencing this firm-level effect will be estimated by several country-level factors. Bilateral differences and similarities between the acquirer’s and target’s countries will be discussed in the following section.
2.2 The influence of country-specific differences
When a firm seeks to acquire another firm across the borders of its country, it is subject to different external factors compared to its country of origin. It faces inter alia economic, legal, and cultural differences which can make a cross-border M&A either more, but likely less attractive.
On the one hand, some country-level characteristics (for example, a currency appreciation or above-average macro-economic performance) could make cross-border M&As more attractive than domestic ones. On the other hand, differences between two countries might increase the risk-return characteristics associated with cross-border M&As.
Several authors have empirically tested the effects of various economic factors. They often make a distinction between developed and emerging markets for this matter. Zhu & Jog (2012) find that when the target firm is from an emerging market, a cross-border acquisition is value enhancing in both the short-term, as well as the long-term stock performance. An explanation for this is that the acquisition lowers the risk of the target due to changes in the international shareholder base and the strength of the investor right protection of the acquirer. Furthermore, firms that acquire foreign firms in developing countries perform significantly better in the period before the acquisition than those that acquire foreign firms in developed countries (Liu & Qiu, 2013). Also, the greater the difference in stock market performance between two countries, the more likely that firms in the superior-performing country purchase firms in the worse-performing country (Erel et al., 2012). So based to the findings of these authors, I expect that the acquirer is located in the economically better performing country compared to the target firm’s country.
8 of targets from such countries pay a price for their country's institutional framework that makes accounting information less value relevant.
An illustrative example of the relevance of institutional changes for cross-border M&As is the European integration (Coeurdacier et al., 2009). In the euro area financial liberalization policies, government policies and regional agreements led to a sharp increase in the cross-border M&A activity in the past two-and-a-half decade. This applies to both intra-euro area M&As as well as cross-border M&As from non-euro area to euro area countries. This European integration has a positive influence on the worldwide cross-border M&A activity. A reason for this is that the institutional changes, such as the single market of the EU and the establishment of the EMU, triggered a reallocation of capital. So, harmonization of the institutional frameworks and legal systems is likely to increase the degree of M&A activity between a pair of countries.
The same harmonization principle applies for cultures and cross-border M&A activity. Cultural differences between countries usually increase the costs of integration. Ahern et al. (2012) find evidence that cultural differences negatively influence the volume of cross-border M&As and the gains from synergies. Also Steigner & Sutton (2011) agree that, on the one hand, cultural differences are likely to increase the transaction costs. However, on the other hand, greater cultural differences might also have a positive effect on the long-term performance of the acquirer with high intangibles. The reason for this is that significant internalization benefits can be realized from technological know-how. Although in the short-term this effect can be neglected.
9
3. Methodology
The methodology section is separated into two parts: firm-level and country-level. Both levels of analysis require a different approach. For the former an event study is performed to approximate the stock market reaction of both cross-border and domestic deals. For the latter a cross-sectional ordinary least-squares (OLS) regression analysis will take into account the differences and similarities between the acquirer’s and target’s country on their effect on the degree of cross-border M&A activity.
3.1 Firm-level methodology: Event study
There are multiple ways to measure the performance of M&As. Zollo & Meier (2008) categorize studies on M&A by their performance metric. The majority of the literature measures M&A performance by accounting performance (21%), long-term financial performance (15%), or short-term financial performance (30%). This research uses the last approach by performing an event study.
Using an event study I measure the cumulative abnormal returns (𝐶𝐶𝑅) of the acquirer of each deal. This performance measure indicates the shareholders’ valuation of each M&A deal. I examine the stock market reaction of the acquiring firm following the methodology of Brown & Warner (1980, 1985) and MacKinlay (1997). To determine the stock market reaction I compare realized returns with a benchmark return. This is a conventional method for determining the effect on firm value.
The return R of a stock with price S for an acquirer i on event day t is defined as Equation 1. 𝑅𝑖,𝑡= 𝑆𝑡𝑆− 𝑆𝑡−1
𝑡−1 (1)
Abnormal returns (𝐶𝑅𝑖,𝑡) are defined as the difference between the actual return on a certain date t and its expected return (Brown & Warner, 1980). The expected return for the event window is derived from the estimation window (see Figure 1).
Figure 1: Timeline for an event study
Note: The expected returns from the event window are derived from the estimation
window (MacKinlay, 1997).
The estimation window [T0, T1] is 200 up to and including 60 days prior to the announcement day.
The event window [T2, T3] is 25 days prior to and 5 days after the announcement day. According to
Duso et al. (2010), a long event window of 25 or 50 days prior to the announcement day correctly anticipates the stock market reaction of a M&A deal. They suggest an event windows of 25,5] (or [-50,5]). When the announcement day is in the weekend, I adjust 𝑡 = 0 to the next available trading day.
10 Market returns (𝑅𝑚,𝑡) adjusted for risk determine the AR𝑖,𝑡: 2
AR𝑖,𝑡 = 𝑅𝑖,𝑡− 𝛼�𝑖− 𝛽̂𝑖𝑅𝑚,𝑡 (2)
The cumulative abnormal returns (CAR) are the sum of the AR within a certain window [t, t] for a single stock:
CAR𝑖,𝑡 = CAR𝑖,𝑡−1+ AR𝑖,𝑡 (3)
To examine the effect of the acquisition on the firm value, the null hypothesis that an acquisition does not affect the stock market reaction can be translated into: 𝐸(CAR𝑖,𝑡) = 0. Based on Brown & Warner (1985) the t-statistic for testing this null hypothesis is calculated as follows:
𝑡 = � AAR 𝑇2 𝑡=𝑇1+1 � � 𝑆̂2 𝑇2 𝑡=𝑇1+1 (AAR)� 1/2 � (4)
where AAR is the average abnormal return3 and 𝑆̂(AAR) is the standard deviation of the AAR.4 The
t-test assumes that the sample is normally distributed.
The cumulative average abnormal return (CAAR) are calculated as the sum of the AARs within the event window [T2, T3]. To compare the CAARs of the domestic deals subsample with the cross-border
deals subsample, I use the two sample t test for comparing two means: 𝑡 =CAAR𝐷− CAAR𝐶𝐶 �𝑠𝐷2 𝑛𝐷+ 𝑠 𝐶𝐶2 𝑛𝐶𝐶 (5)
where the subscript D represents the domestic subsample and CB the cross-borders subsample, 𝑠𝐷 and 𝑠𝐶𝐶 are the standard deviation of the two samples, and 𝑛𝐷 and 𝑛𝐶𝐶 represent the sample sizes. The degrees of freedom are the smaller of 𝑛𝐷− 1 and 𝑛𝐶𝐶− 1.
2 where 𝛼�
𝑖 is the constant term and 𝛽̂𝑖 is the slope, determined by the ordinary least squares obtained by the estimation window.
3 The average abnormal return is calculated as:
AAR = 𝑁 � AR1 𝑖,𝑡 𝑁𝑡
𝑖=1
(6)
4 The standard deviation of the average abnormal returns, 𝑆̂(AAR), is calculated as:
11
3.2 Country-level methodology: Cross-sectional ordinary least-squares regression analysis
Besides the event study, a cross-sectional ordinary least-squares (OLS) regression analysis will be performed to estimate the relationship between the degree of cross-border M&A activity between a certain pair of countries and several not mutually exclusive, relative country-specific variables of acquiring country a compared to target country b. From the total sample of acquisitions I construct a matrix with the number of acquisitions between each country-pair. Subsequently, following Erel et al. (2012) I estimate the degree of cross-border M&A activity between a certain country-pair (Cross-border M&A activity𝑎𝑎) as the fraction of the number of cross-border M&As between both countries divided by the sum of the domestic deals of the acquiring country and the numerator. In this case the value will have an upper limit of one and a lower limit of zero. This fraction thus corrects for the domestic M&A volume to measure the degree of cross-border M&A activity.
Based on the literature review the regression will include the following variables. Cross-border M&A activity𝑎,𝑎
= 𝛽0+ 𝛽1relative GDP per capita𝑎,𝑎+ 𝛽2relative trade flows𝑎,𝑎 + 𝛽3relative strenght of investor protection𝑎,𝑎
+ 𝛽4dummy same legal origin𝑎,𝑎+ 𝛽5dummy same language𝑎𝑎 + 𝛽6dummy same religion𝑎,𝑎+ 𝛽7distance𝑎,𝑎+ 𝜀𝑎𝑎
(9)
First, the relative GDP per capita will capture the differences in the average state of the economies between the acquirer and target country within the time frame. Second, because the unavailability of bilateral trade flows between each country-pair, I calculate the relative trade flows of the acquirer’s and target’s country. They are as the average of the sum of the imports and exports as a percentage of GDP within the time frame. The trade flows are included to measure the influence of the degree of openness of the involved countries. Third, the average relative rate that measures the strength of investor protection is included to capture corporate governance differences.
Compared to the methodology of Erel et al., instead of using absolute differences I use relative variables to indicate the differences between two countries. In this way the variables better cover how firms experience the difference compared to absolute values. To illustrate how these relative variables are formed I use relative average GDP per capita in the period 1997 to 2013 as an example. The relative GDP per capita of the bidder’s country a and target’s country b is calculated as:
relative GDP per capita𝑎𝑎= �1𝑛 � GDP per capita𝑎,𝑡 𝑡=2013 𝑡=1997 1 𝑛 � GDP per capita𝑏,𝑡 𝑡=2013 𝑡=1997 � � − 1 (10)
12 Fourth, a dummy variable is added to account for the same legal system, i.e., common law or civil law. The latter is subdivided into distinct groups: French, German, or Scandinavian civil law. Fifth, dummy variables for same language and for same religion between acquiring and target country are added to capture the effects of cultural differences. And finally, I put a geographic proximity for the distance between the two countries in the regression. I construct a matrix of the great-circle distances between each country-pair. A general way to calculate the great-circle distance𝑎,𝑎 in kilometers between the capital cities of acquirer’s country a and target’s country b is:
distance𝑎,𝑎 = 𝑟∆𝜎𝑎,𝑎 (11)
where 𝑟 is the radius of the earth, which is on average 6,371 kilometers, and ∆𝜎𝑎,𝑎 is the central angle between both cities given in radians. The central angle is calculated as follows:
∆𝜎𝑎,𝑎 = cos−1[sin(𝜙𝑎) sin(𝜙𝑎) + cos(𝜙𝑎) cos(𝜙𝑎) cos(∆𝜆)] (12) where 𝜙𝑎 and 𝜙𝑎 are respectively the geographical latitudes of the capital cities of the country of acquirer 𝑎 and target 𝑏. ∆𝜆 is the absolute difference between geographical longitudes of both capital cities.
13
4. Data
4.1 Data collection
In order to answer the research question on firm-level and country-level I draw two samples. The first sample includes deals with both listed as well as unlisted acquirers (referred to as the total
sample). Because firm-level data is only available for listed acquirers, our second sample contains
only these deals (listed acquirers sample). In other words, the listed acquirers sample is a subsample of the total sample.
I collect comprehensive, global M&A data from Bureau van Dijk’s Zephyr database with the announcement date between the year 1997 up to and including 2013. This data includes from both the acquirer and the target: International Securities Identification Number (ISIN), country (code), four-digit Standard Industrial Classification (SIC) codes, company name, and business description. Also the deal value, deal type, offer price, bid premium, announcement date, and completed date are collected.
I select only pure (100 percent) mergers and acquisitions. This means that I omit IPOs, institutional buy-outs, joint-ventures, management buy-ins, management buy-outs, demergers, minority stakes, and share buy backs. Furthermore, only completed deals that are confirmed retain in the samples. Another criteria is that both the country of the acquirer as well as the country of the target is known. In order to correct for overlapping firm value effects I omit the deals that are closed by the acquirer in the same year. This means that only the first deal of an acquirer in a particular year is preserved. In line with most literature about cross-border M&As (for example: Erel et al., 2012; Liu & Qiu, 2013) acquirers who are financial institutions are taken out of the sample.5
Table 1 gives the number of observations for both the total sample as well as the subsample with only the listed acquirers. The total sample contains 290,003 M&As of which 73,230 are cross-border, whereas the listed acquirers sample contains 44,317 M&As of which 15,602 are cross-border. The number of cross-border acquirers, targets, and domestic deals for each country of the total sample is listed in Appendix A.
The daily stock returns of the acquirers as well as the daily local MSCI market returns are gathered from Thomson Reuters Datastream. The ISINs are used as the link between the multiple databases for deal and company information.
5
These are the acquirers with SIC codes between 6000 and 6799. This is the ‘Finance, Insurance, And Real
Estate’ division. This division includes: Depository Institutions; Non-depository Credit Institutions; Security
14
Table 1: Number of observations
Note: This table shows the number of observations (M&A deals) for the total sample and
the listed acquirers sample for each year in the time frame, divided into cross-border or domestic.
Total sample
Listed acquirers sample
Year Cross-border Domestic Cross-border Domestic 1997 1,569 1,908 332 176 1998 3,340 4,123 672 417 1999 3,950 5,276 749 534 2000 4,403 9,847 894 1,486 2001 3,679 9,521 743 1,372 2002 3,228 9,280 659 1,230 2003 3,588 11,157 716 1,347 2004 4,428 13,691 892 1,861 2005 5,037 15,163 1,070 2,263 2006 5,245 15,490 1,158 2,282 2007 5,725 17,578 1,316 2,687 2008 5,461 16,879 1,205 2,334 2009 3,980 16,058 758 1,812 2010 4,534 17,197 1,125 2,450 2011 5,080 17,145 1,223 2,336 2012 4,987 17,909 1,134 2,192 2013 4,996 18,551 956 1,936 Total 73,230 216,773 15,602 28,715
For the country-level analysis, I collect the following country-specific information. From the World Development Indicators of The World Bank I obtain annual economic information of the acquirer and target countries. These include inter alia GDP per capita, and exports and imports of goods and services. Also from The World Bank I collect for all available countries ratings on the strength of the minority investor protection in order to measure legal differences. This rating is an average index of multiple scores on disclosure, director liability, ease of shareholder suits, conflict of interest regulation, shareholder rights, the strength of the governance structure, corporate transparency, and shareholder governance (Djankov et al., 2008). The cultural variables are obtained from Stulz & Williamson (2003). From this paper I collect the primary language, primary religion, and the legal origin of each available country. Finally, to estimate the distances of each country-pair I collect the latitudes and longitudes of the capital cities of each country in the sample6.
15
4.2 Hypotheses
The research is split into a two-fold analyses. To recall, the research question is divided into the following subquestions: What is the influence of country borders on the stock market reaction of a cross-border merger or acquisition? And to what extent do country-specific differences explain the degree of cross-border M&A activity between a certain country-pair?
The underlying assumption is that when the degree of cross-border M&A activity between a certain country-pair is high (high attractiveness), the stock market reaction to such an acquisition will be higher compared to countries with less M&A activity. A country-pair with a high degree of M&A activity experiences less uncertainty and consequently this will positively influence the stock market reaction of acquisitions between these countries.
The first subquestions lead to the following null hypothesis:
H01: Country borders do not affect the stock market reaction of a cross-border merger or acquisition.
H01 makes no distinction between domestic and cross-border M&As with respect to the stock market reaction. In other words, there is no cross-border effect on a firm-level. I will compare the stock market reaction of domestic M&As with cross-border M&As to estimate this cross-border effect.
If there exists a significant cross-border effect on a firm-level, then which country factors influence the attractiveness of a cross-border M&A? This attractiveness will be tested on a country-level in terms of M&A volume. The second null hypothesis (H02) concerns with the country-specific characteristics that exist because of country borders:
H02: Economic, cultural, legal and spatial differences do not affect the degree of cross-border M&A activity between a certain country-pair.
16
5. Results
5.1 Testing for differences between domestic and cross-border M&As
To test the first null hypothesis (H01) that country borders do not affect the stock market reaction of a cross-border M&A I calculate the cumulative average abnormal returns (CAARs) from the listed acquirers sample. Figure 2 illustrates the CAARs for each day in the event window [-25, T2] for both
the cross-border and the domestic deals in the sample7. For the event window [-25, 5] the CAARs of both the cross-border deals subsample (t-statistic: 4.52; p-value: 0.000) as well as the domestic deals subsample (t-statistic: 13.12; p-value: 0.000) are significantly different from zero. This means that, whether it be domestic or cross-border, both types of M&As have a substantial impact on the shareholders’ valuation of the acquirer.
Figure 2: Cumulative Average Abnormal Returns
Note: This figure shows the cumulative average abnormal returns (CAARs) for each day
in the event windows ([-25, -25] up to and including [-25, 5]) from the announcement date (t = 0) for both the cross-border deals (the solid line) and the domestic deals (the striped line) in the sample. Also, the difference between the CAARs of the domestic deals sample and the cross-border deal sample is shown by the dashed line. For the event window [-25, 5] the CAARs of both the cross-border deals subsample (t-statistic: 4.52; p-value: 0.000) as well as the domestic deals subsample (t-statistic: 13.12; p-value: 0.000) are significantly different from zero. Furthermore, there exists a significant positive difference (t-statistic: 6.84; p-value: 0.000) between the CAAR of the domestic deals and the cross-border deals. The results underlying this figure can be found in Appendix B.
7 with T
17 However, I find a significant positive difference (t-statistic: 6.84; p-value: 0.000) between the CAAR of the domestic deals and the cross-border deals in the event window [-25, 5]. From the shareholder’s value maximization perspective this results implies that, on average, cross-border M&As are valued less than domestic ones. Therefore, H01 is rejected: country borders do affect the stock market reaction of a cross-border M&A substantial. Because the market reaction is much higher for domestic deals compared to cross-border deals, it can be concluded that, on average, the additional wealth creation of an acquisition of a firm in a foreign location does not outweigh the additional uncertainty compared to an acquisition of a similar firm in a domestic location. To answer the first subquestion, on a firm-level, the influence of country borders of a cross-border merger or acquisition is that it substantially lowers the stock market reaction compared to domestic deals.
Also when using other, symmetrical event windows ([-2, 2], [-5, 5], [-10, 10], and [-25, 25]) these results prove their robustness (see Appendix C). Only for the longest event window of [-25, 25] the stock market reaction of cross-border M&As is not significantly different from zero. An explanation for this finding is that the gains with respect to the synergies of the acquisition in the longer term are lower than initially anticipated due to higher uncertainty or more information asymmetry.
Figure 3: Annual Cumulative Average Abnormal Return
Note: This figures shows the cumulative average abnormal return for each year in the
sample for the event window [-25, 5]. The dashed Difference-line illustrates the consistency of the cross-border effect, since it is positive over the entire time frame. The correlation between the solid Cross-border-line and striped Domestic-line equals 0.926. The results underlying this figure can be found in Appendix D.
cross-18 border and domestic deals over time are highly correlated (R2 = 0.926), I will not concentrate on these annual differences, since they move in a similar pattern. The difference between the CAARs of the domestic and the cross-border deals is positive for the entire time frame but is only significant for the years 1997, 1999, 2001-2003, 2006, and 2009 on (at least) a 5% significance-level (see Appendix D).
5.2 The influence of the country-specific variables
From the total sample that consists of both listed and unlisted acquirers, an estimation of the degree of cross-border M&A activity between a certain country-pair is calculated using the country-matrix with the number of acquisitions between each country-pair. With these data I am able to calculate a proxy for the cross-border M&A volume in a similar way as Erel et al. (2012) (see section 3.2). I end up with a country-matrix of 160×159 with countries who have at least one acquirer. After omitting the country-pairs that do not have any mergers or acquisitions within the time frame, 3,960 observations remain. In other words, about 15.6% of all possible country-pairs have at least one cross-border merger or acquisition. This study focuses only on this fraction where cross-border M&A activity actually exists within the time frame.
In the previous section I have identified a cross-border effect on a firm-level. There are many possible underlying factors that influence this average valuation effect. However, I cannot test for country-specific variables on a firm-level, although the differences and similarities between countries are likely to be relevant. Differences between country-specific characteristics might affect the uncertainty regarding the deal and therefore the bilateral attractiveness of cross-border M&As. For this purpose, I will estimate the underlying factors from a broader perspective of which country-specific differences and similarities make a country-pair either more or less attractive for cross-border M&As, measured by the degree of cross-border M&A activity. I test for economic, cultural, legal, and spatial differences between the acquirer and target country.
Table 2 provides the output of the cross-sectional OLS-regression analysis on the relationship between the degree of cross-border M&A activity and the (relative) country-specific variables of acquiring country a and target country b. To recall, for the relative variables, when the coefficients are positive, the acquirer country has a higher score on the particular variable compared to the target country, and vice versa.
19 (1951) test is close to 2 (DW-statistic = 1.746). In other words, the correlation of the disturbance terms is close to zero.
To test the multiple explanatory variables simultaneously that all of the slope parameters are jointly zero, H02, the F-test of the regression gives a p-value of 0.000 (F-statistic: 44.129). The second null hypothesis that the degree of cross-border M&A activity between a certain country-pair is the same regardless of the economic, cultural, legal and spatial differences is therefore rejected.
Table 2: Cross-sectional Ordinary Least Squares Regression Analysis
Note: The dependent variable of the cross-sectional OLS regression analysis is the fraction
of the number of cross-border M&As betweenacquirer’s country a and target’s country b to the sum of the number of domestic M&As of country a and the numerator. The explanatory variables can be divided into four categories: economic, legal, cultural, and distance. Since the errors of the regression analysis do not have a constant variance, they are modified to heteroskedasticity-consistent standard error estimates (White, 1980). The number of asteriks, *, **, or ***, imply that the t-statistic is significant on respectively 10%, 5%, or 1%-level. The p-values are given between parentheses.
Variable Category Coefficient t-statistic
(p-value)
Relative GDP per capita Economic -0.000 -7.56***
(0.000)
Relative trade flows Economic 0.003 5.34***
(0.000) Relative investor protection rate Legal -0.005 -4.39*** (0.000)
Same legal system dummy Legal 0.003 2.25**
(0.024)
Same language dummy Cultural 0.023 4.25***
(0.000)
Same religion dummy Cultural -0.002 -0.05
(0.959)
Geographic proximity Distance -0.000 -8.59***
(0.000) Constant 0.019 11.33*** (0.000) Observations 1,216 R2 0.204 Adjusted R2 0.199
20 5%-level; the religion dummy is not significant. The directions of the relationship are largely consistent with the expectations, although there are some exceptions.
First I discuss the economic variables, which consist of the GDP per capita and the trade flows of the involved countries in a cross-border M&A deal. The relative GDP per capita variable shows that cross-border M&A activity is significantly higher when the target country has a higher GDP per capita relative to the acquirer country. This is a conflicting result. From the literature review we learn that it is more likely that a firm from an economically superior-performing (developed) country will acquire a firm in worse-performing (developing) country (Zhu & Jog, 2012; Liu & Qiu, 2013). Developed countries are typically characterized by better corporate governance practices which the target firm eventually will adopt. In both the short-term and the long-term, these practices turns out to be value enhancing in the stock performance and firm performance (among others productivity and profitability). An explanation for this conflicting result is that this thesis does not solely focus only on extremes in terms of state of economies as in the distinction of developed and developing countries. Developing countries are generally characterized by high economic growth rates, rather than high GDP. When the target firm is located in a developing country , the acquirer’s wealth creation is rooted in the growth potential. Nevertheless, in terms of M&A activity, when the target country has a relatively higher GDP per capita, the acquirer might benefit from this economic differential as well. This can make a target country more attractive. Furthermore, when a country has a relatively open economy, it is more likely that it has a higher degree of M&A activity across its borders. The openness of the economy in this regression analysis is measured by the total of exports and imports as a percentage of GDP.
Second, the relevance of legal differences is analyzed by the strength of the investor protection and whether both involved countries have a similar legal system. The regression shows that relatively strong investor protection of the target country increases the probability that a cross-border acquisition takes place. This contradicts the expectations based on the empirical literature (Bris et al., 2008; Rossi & Volpin, 2004; Bris & Cabolis, 2008; Black et al., 2007). These authors find a positive relation between the quality of the shareholder protection and accounting standards in the country of the
acquirer and the volume of cross-border M&A activity. In this thesis the strength of the investor
protection is measured by an average index of multiple scores on disclosure, director liability, ease of shareholder suits, conflict of interest regulation, shareholder rights, the strength of the governance structure, corporate transparency, and shareholder governance (Djankov et al., 2008). It is self-evident though that strong investor protection of target countries can only be beneficial for an acquirer. The relevance of the legal system is emphasized by the ‘same legal system’ dummy. Having a similar legal system (civil or common) has a positive influence on the degree of cross-border M&A activity.
21 results of Ahern et al. (2012). As already mentioned above, I find no significant result for the influence of the same primary religion on M&A activity.
22
6. Conclusion
In the previous section I analyzed the impact of country borders for mergers and acquisitions on both firm- and country-level. On a firm-level, a cross-border effect is identified as the differential between the stock market reaction of domestic and cross-border M&As. In line with Dunning (1980), I find empirical evidence that cross-border deals create wealth by internalizing the firm's ownership-specific advantages in foreign locations. However, this wealth creation is significantly lower compared with the internalization of these advantages in domestic locations due to more uncertainty and information disadvantages. Differences between the characteristics of the target’s and acquirer’s country are likely to be relevant for the explanation of this valuation gap, the cross-border effect. On a country-level, these relevant country-specific characteristics are estimated by the relative attractiveness of a target country by analyzing the degree of cross-border M&A activity between a certain country-pair. The results suggest that economic, legal and cultural differences as well as physical distance have a significant impact on the degree of cross-border M&A activity between a country-pair.
Returning to the research question, country-borders matter to a large extent for mergers and acquisitions on two levels. Country borders influence both the acquirer’s performance and the volume of M&As between country-pairs. The results suggest that, on average, cross-border M&A activity is likely to be high when the following main indicators are present: relatively high economic performance and strong investor protection of the target country, relatively open economy of the acquirer country, similar legal and cultural environment, and geographical proximity. A country with a high degree of M&A activity is considered as attractive and reduces the uncertainty and information disadvantages with respect to the synergies, which eventually will positively influence the wealth creation of an acquisition.
This thesis makes three main contributions. First, it clearly highlights the role of country borders in the valuation differential between domestic and cross-border M&As. Second, this thesis bridges the gap between two levels of analyses. On a country-level, the attractiveness of a particular country for cross-border M&As is measured and linked to acquirer’s performance through reduced uncertainty and information advantages. Third, it uses a larger country sample and a different, more recent time frame, the research has a much more global and topical coverage compared to other authors. This larger sample size is more representative of the population and limits the influence of outliers.
23 In fact, information advantages also arise from domestic M&As within the United States of firms that are located close to each other (Uysal et al., 2008). Therefore, this is not solely a cross-border characteristic. Finally, because I measure the degree of cross-border M&A activity by the total volume of deals within the entire time frame the cross-sectional regression model uses average values for GDP per capita, trade flows, and strength of investor production rather than annual values. Unfortunately, a regression model for each year independently would have too few observations. This limits the thesis in the sense that it cannot correct for annual developments of countries.
24
References
Aharon, D. Y., Gavious, I., & Yosef, R., 2010. Stock Market Bubble Effects on Mergers and Acquisitions. The Quarterly Review of Economics and Finance 50, 456-470.
Ahern, K., Daminelli, D., & Fracassi, C., 2012. Lost in Translation? The Effect of Cultural Values on Mergers Around the World. Journal of Financial Economics, forthcoming.
Black, E. L., Carnes, T. A., Jandik, T., & Henderson, B. C., 2007. The Relevance of Target Accounting Quality to the Long-Term Succes of Cross-Border Mergers. Journal of Business
Finance and Accounting 34, 139-168.
Bris, A., & Cabolis, C., 2008. The Value of Investor Protection: Evidence from Cross-Border Mergers.
Review of Financial Studies 21, 605-648.
Bris, A., Brisley, N., & Cabolis, C., 2008. Adopting Better Corporate Governance: Evidence from Cross-Border Mergers. Journal of Corporate Finance 14, 224-240.
Brown, S. J., & Warner, J. B., 1980. Measuring Security Price Performance. Journal of Financial
Economics 3, 205-258.
Brown, S.J., & Warner, J.B. (1985), Using Daily Stock Returns: The Case of Event Studies, Journal of
Financial Economics, 1, 3-31.
Coeurdacier, N., De Santis, R., & Aviat, A., 2009. Cross-border Mergers and Acquisitions and European Integration. Economic Policy 24, 55-106.
Djankov, S., La Porta, R., Lopez-de-Silanes, F., & Shleifer, A., 2008. The Law and Economics of Self-Dealing. Journal of Financial Economics 88, 430-465.
Dunning, J. H., 1980. Towards an Ecletic Theory of International Production: Some Empirical Tests.
Journal of International Business Studies 11, 9-31.
Durbin, J., & Watson, G. S., 1951. Testing for Serial Correlation in Least Squares Regression.
Biometrika 38, 159-171.
Duso, T., Gugler, K., & Yurtoglu, B., 2010. Is the Event Study Methodology Useful for Merger Analysis? A Comparison of Stock Market and Accounting Data. International Review of Law and
Economics 30, 186-192.
Erel, I., Liao, R., & Weisbach, M., 2012. Determinants of Cross-border Mergers and Acquisitions. The
Journal of Finance 67, 1045-1082.
Georgopoulos, G. J., 2008. Cross-border Mergers and Acquisitions: Does the Exchange Rate Matter?
Canadian Journal of Economics 41, 450-474.
Gonzalez, P., Vasconcellos, G. M., Kish, R. J., & Kramer, J. K., 1997. Cross-border Mergers and Acquisitions: Maximizing the Value of the Firm. Applied Financial Economics 7, 295-305. Hur, J., Parinduri, R. A., & Riyanto, Y. E., 2011. Cross-Border M&A Inflows and Quality of Country
25 Liu, Q., & Qiu, L. D., 2013. Characteristics of Acquirers and Targets in Domestic and Cross-border
Mergers and Acquisitions. Review of Development Economics 17, 474-493.
MacKinlay, A. C., 1997. Event Studies in Economics and Finance. Journal of Economic Literature 1, 13-39.
Roll, R., 1986. The Hubris Hypothesis of Corporate Takeovers. Journal of Business 59, 197-216. Rose, A. K., 2000. One Money, One Market: Estimating the Effect of Common Currencies on Trade.
Economic Policy 30, 7-45.
Rossi, S., & Volpin, P. F., 2004. Cross-Country Determinants of Mergers and Acquisitions. Journal of
Financial Economics 74, 277-304.
Ruback, R. S., & Jensen, M. C., 1983. The Market for Corporate Control: The Scientific Evidence.
Journal of Financial Economics, 11, 5-50.
Steigner, T., & Sutton, N. K., 2011. How Does National Culture Impact Internalization Benefits in Cross-Border Mergers and Acquisitions? Financial Review 46, 103-125.
Stiebale, J., & Trax, M., 2011. The Effects of Cross-Border M&As on the Acquirers' Domestic Performance: Firm-Level Evidence. Canadian Journal of Economics 44, 957-990.
Stulz, R.M., & Williamson, R., 2003. Culture, Openness, and Finance. Journal of Financial
Economics 70, 323-324.
Uysal, V.B., Kedia, S., & Panchapagesan, V., 2008. Geography and Acquirer Returns. Journal of
Financial Intermediation 17, 256-275.
White, H., 1980. A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity. Econometrica 48, 817-838.
Zhu, P., & Jog, V., 2012. Impact on Target Firm Risk-Return Characteristics of Domestic and Cross-Border Mergers and Acquisitions in Emerging Markets. Emerging Markets Finance & Trade 48, 79-101.
26
Appendix A:
Number of cross-border acquirers, targets, and domestic deals
Table 3: This table shows the number of cross-border acquirers, targets, and domestic deals per country in the full sample.Country code Country Cross-border acquirers Cross-border targets Domestic deals AD Andorra 10 2 1
AE United Arab Emirates 220 163 116
AG Antigua And Barbuda 3 9 3
AL Albania 2 24 6 AM Armenia 3 54 22 AO Angola 9 6 3 AR Argentina 72 522 359 AT Austria 1,207 685 883 AU Australia 1,333 1,681 5,324 AW Aruba 1 0 0 AZ Azerbaijan 10 22 27
BA Bosnia And Herzegovina 3 111 181
BB Barbados 17 30 6 BD Bangladesh 0 18 15 BE Belgium 1,517 1,480 1,927 BF Burkina Faso 0 6 0 BG Bulgaria 52 712 1,845 BH Bahrain 34 14 18 BI Burundi 0 5 0 BJ Benin 0 3 0 BM Bermuda 736 456 70 BN Brunei Darussalam 6 12 4 BO Bolivia 2 35 22 BR Brazil 214 1,198 1,596 BS Bahamas 36 32 5 BW Botswana 5 21 7 BY Belarus 4 100 54 BZ Belize 57 12 3 CA Canada 3,662 2,924 7,294
CD Congo Democratic Republic Of 1 23 3
CF Central African Republic 0 1 0
CG Congo 1 8 0 CH Switzerland 2,205 1,411 2,107 CI Cote D' Ivoire 2 28 11 CL Chile 119 368 461 CM Cameroon 2 11 0 CN China 591 3,131 8,246 CO Colombia 96 242 239 CR Costa Rica 17 68 32
CS Serbia And Montenegro 26 57 19
27 (continued) Country code Country Cross-border acquirers Cross-border targets Domestic deals ER Eritrea 0 1 0 ES Spain 1,415 1,783 7,583 ET Ethiopia 0 4 5 FI Finland 1,278 842 9,839 FJ Fiji 0 10 6 FK Falkland Islands 1 0 0 FO Faroe Islands 2 1 1 FR France 4,291 3,722 8,647 GA Gabon 0 11 0 GB United Kingdom 8,091 7,237 26,892 GE Georgia 10 75 63 GF French Guiana 0 3 0 GG Guernsey 0 1 0 GH Ghana 2 39 11 GI Gibraltar 27 28 5 GL Greenland 1 6 1 GM Gambia 0 3 0 GN Guinea 0 11 1 GQ Equatorial Guinea 0 2 1 GR Greece 352 185 898 GT Guatemala 10 34 15 GU Guam 1 1 1 GW Guinea-bissau 0 1 0 GY Guyana 0 12 2 HK Hong Kong 1,455 1,471 469 HN Honduras 1 26 8 HR Croatia 76 195 249 HT Haiti 0 2 1 HU Hungary 154 535 769 ID Indonesia 61 382 741 IE Ireland 1,013 788 618 II Supranational 9 0 0 IL Israel 412 360 455 IN India 840 1,047 2,361 IQ Iraq 0 3 3
IR Iran (Islamic Republic Of) 1 3 0
IS Iceland 202 40 158 IT Italy 1,415 1,813 4,265 JM Jamaica 14 25 20 JO Jordan 19 43 37 JP Japan 1,086 490 4,483 KE Kenya 17 58 37 KG Kyrgyzstan 2 18 47 KH Cambodia 0 31 4 KI Kiribati 0 1 0 KM Comoros 0 1 0
KN Saint Kitts And Nevis 11 2 0
KP Korea Democratic People's Republic Of 0 3 0 KR Korea Republic Of 278 343 1,191
KV Kosovo 0 11 0
KW Kuwait 81 17 38
KY Cayman Islands 894 703 112
KZ Kazakhstan 44 135 142
LA Lao People's Democratic Republic 0 5 0
LB Lebanon 32 13 14
LC Saint Lucia 3 5 1
LI Liechtenstein 40 19 3
28 (continued) Country code Country Cross-border acquirers Cross-border targets Domestic deals LR Liberia 5 8 0 LS Lesotho 0 2 0 LT Lithuania 102 320 407 LU Luxembourg 676 290 100 LV Latvia 64 304 267
LY Libyan Arab Jamahiriya 12 4 0
MA Morocco 18 68 64 MC Monaco 8 33 4 MD Moldova Republic Of 4 60 30 ME Montenegro* 4 23 7 MG Madagascar 0 14 0 MH Marshall Islands 26 8 2 MK Macedonia (Fyrom) 5 77 24 ML Mali 1 8 0 MM Myanmar 0 1 0 MN Mongolia 0 11 0 MO Macau 7 31 6 MR Mauritania 1 8 0 MT Malta 38 37 37 MU Mauritius 114 66 14 MV Maldives 0 3 0 MW Malawi 5 3 2 MX Mexico 223 648 350 MY Malaysia 622 502 4,477 MZ Mozambique 0 29 7 NA Namibia 4 34 18 NC New Caledonia 2 4 0 NE Niger 0 1 0 NG Nigeria 27 59 141 NI Nicaragua 1 27 5 NL Netherlands 3,729 2,456 7,558 NO Norway 973 1,014 2,861 NP Nepal 1 2 2 NR Nauru 2 0 0 NZ New Zealand 262 476 672 OM Oman 15 17 20 PA Panama 84 100 34 PE Peru 33 217 118 PF French Polynesia 2 1 2
PG Papua New Guinea 7 21 13
29 (continued) Country code Country Cross-border acquirers Cross-border targets Domestic deals SI Slovenia 121 101 211 SK Slovakia 60 248 158 SL Sierra Leone 1 8 3 SM San Marino 1 1 0 SN Senegal 0 5 0 SR Suriname 0 5 1
ST Sao Tome And Principe 0 1 0
SV El Salvador 4 35 9
SY Syrian Arab Republic 1 1 0
SZ Swaziland 0 1 5
TC Turks And Caicos Islands 2 4 1
TD Chad 0 3 1 TG Togo 7 4 0 TH Thailand 80 247 493 TJ Tajikistan 1 6 1 TM Turkmenistan 0 1 0 TN Tunisia 10 33 16 TO Tonga 0 1 1 TR Turkey 135 386 464
TT Trinidad And Tobago 15 17 9
TV Tuvalu 0 2 0
TW Taiwan 795 196 674
TZ Tanzania United Republic Of 2 37 2
UA Ukraine 62 870 1,624
UG Uganda 2 41 9
US United States Of America 13,940 9,059 59,217
UY Uruguay 7 84 24
UZ Uzbekistan 3 53 18
VC Saint Vincent And The Grenadines 6 1 0
VE Venezuela 27 40 53
VG Virgin Islands (British) 1,784 1,198 419
30
Appendix B:
CAARs for each subsample, t-statistics, and p-values
Table 4: This table underlies the values of Figure 2. It gives the cumulative average abnormal returns for each event window
[-25, T2] with its t-statistic. The number of asteriks *, **, *** imply that the t-statistic is significant on respectively 10%, 5%,
or 1%-level. The p-values are given between parentheses.
(a) listed acquirers sample (b) cross-border (c) domestic (d) difference (c)-(b) (e) listed acquirers sample
32
Appendix C:
Comparison of CAARs for symmetrical event windows
Cumulative average abnormal returns for symmetrical event window [-2,2], [-5,5], [-10,10], and [-25,25] with its t-statistic.
The cumulative average abnormal return (CAAR) are calculated as the sum of the AARs within the event window [T2, T3]
(see Figure 4). The number of asteriks *, **, *** imply that the t-statistic is significant on respectively 10%, 5%, or 1%-level. The p-values are given between parentheses.
(a) listed acquirers sample (b) cross-border (c) domestic (d) difference (c)-(b) (e) listed acquirers sample
(f) cross-border (g) domestic (h) difference (g)-(f) Event window CAAR t-statistic (p-value) [-2,2] 0.015 0.011 0.017 0.006 19.79*** 11.56*** 17.72*** 5.08*** (0.000) (0.000) (0.000) (0.000) [-5,5] 0.019 0.012 0.024 0.012 17.41*** 8.49*** 16.51*** 6.48*** (0.000) (0.000) (0.000) (0.000) [-10,10] 0.019 0.010 0.024 0.014 12.14*** 5.10*** 11.95*** 5.48*** (0.000) (0.000) (0.000) (0.000) [-25,25] 0.017 0.003 0.025 0.022 7.01*** 0.90 8.02*** 5.60*** (0.000) (0.369) (0.000) (0.000)
Figure 4: Daily Average Abnormal Return
Note: The daily average abnormal return (AAR) from 25 days prior to and 5 days after the
33
Appendix D:
Annual CAARs, t-statistics, and p-values
Table 5: Cumulative average abnormal returns for event window [-25, 5] for each year in the (sub-)samples with its t-statistic, The number of asteriks *, **, *** imply that the t-statistic is
significant on respectively 10%, 5%, or 1%-level, The p-values are given between parentheses.
Year Number of observations Standard deviation CAAR [-25,5] t-statistic
34
(continued)
Year Number of observations Standard deviation CAAR [-25,5] t-statistic
