The impact of domestic borders on acquisition
performance: Evidence from China
Joint Thesis MSc BA SIM & MSc IB&M
4
thof April 2016
Supervisor (SIM): K. J. McCarthy
2
ndSupervisor (IB&M): S. R. Gubbi
Eefje van Stralen Schuitemakersstraat 2-44
9711 HW Groningen
e.j.stralen@gmail.com
s1877267 Word count: 12.396
MSc BA Strategic Innovation Management, & MSc International Business & Management
University of Groningen, Faculty of Economics and Business Duisenberg Building, Nettelbosje 2
2
The impact of domestic borders on acquisition
performance: Evidence from China
Abstract
This study examines whether domestic borders matter in explaining acquisition performance. Classical international business studies argue that national borders create a ‘’liability of foreignness”. However, findings from international economic geography suggest that a simple host-home country dichotomy is insufficient to explain international business activities in its full extent. Given the presence of intra-country differences, particularly in developing countries, this paper suggests that the costs and benefits associated with national borders, are also present when firms cross subnational borders. Using a sample of 910 domestic acquisitions in China, announced in the period Jan 1990 to June 2015, this study investigates the impact of administrative and cultural borders within China on acquisition performance. The results suggest that domestic borders matter, but not in the way we expected. We found that both types of borders have a positive impact on acquisition performance. This study contributes to extant international business research by providing evidence for; (1) the relevance of domestic borders in explaining acquisition performance, and; (2) that crossing these borders can be an asset, rather than a liability.
Keywords: M&A, China, subnational borders, liability of foreignness, institutional & cultural differences.
I. INTRODUCTION
The idea that operating across borders creates a ‘liability of foreignness’ sets international management apart from management in general (Zaheer, Schomaker, & Nachum, 2012). International expansion through acquisitions1 offers significant value-creation opportunities for firms, but it presents significant challenges as well. National borders imply differences in culture, customer preferences, business practices, and institutional forces that cause difficulties for firms to realize their strategic objectives (Aybar & Ficici, 2009).
While classical international business theories use a country as the primary geographic unit of analysis to explain border effects on M&A behavior and performance, findings from international economic geography suggest that this simple host-home country dichotomy is insufficient to explain international business activities in its full extent (Fujita, Krugman, & Venables, 1999; Lammarino & McCann, 2013). Intra-national variations can often be as significant as cross-national differences (Boschma et al., 2015; Tung, 2008). This perspective, suggesting that spatial variation is also present within countries, contributed to the emergence
1
3
of research that is using finer geographic units as a proxy to study the impact of geographic location on business activities (Chan, Makino, & Isobe, 2010).
Following this line of research, this study examines the impact of subnational borders in China on M&A performance. Because, notwithstanding a small number of studies, the fact that most international business studies have ignored the importance of intra-national differences is particularly limiting our understanding of acquisitions in developing countries. Developing countries consist of a large number of subnational regions, such as; Brazil, China, and India, and the regional development of these regions varies in respect to each other (Meyer & Nguyen, 2005). Further, cultural heterogeneity, and linguistic and institutional differences are common features in these countries (Dikova, Sahib, & van Witteloostuijn, 2010; Erel, Liao, & Weisbach, 2012; Lebedev, Peng, Xie, & Stevens, 2014). For this reason, China’s acquisition market is chosen to examine the impact of domestic borders. Although being the fourth largest economy in the world with a strong general economic performance, China’s economic development among its regions has diverged (Xiaolu, 2007). China is characterized by a high level of regional disparity and cultural differences among its regions, and is therefore, an interesting and suitable economy to explore the consequences of crossing subnational borders on acquisition behavior and performance. Further, the accelerated growth of foreign direct investment in China has increased the interest in China’s cultural and institutional characteristics and their effect on organizations (Ahlstrom, Chen, & Yeh, 2010).
This paper suggests that, given the presence of intra-country differences, particularly in developing countries, it is likely that domestic borders are associated with costs and benefits, and have the potential to create a liability as well. Therefore, the research question addressed in this paper is: What is the impact of subnational borders on M&A performance in
China? To answer this question, a sample of 910 domestic acquisitions in China, announced
4
instead of developed economies, and; (2) firms crossing domestic borders instead of national borders.
The results of this study indicate that domestic borders matter in explaining acquisition performance, but not in the way as we expected. We find that, first, crossing institutional borders enhances the acquisition performance by 1.3%, even controlling for institutional distance and cultural differences, there exists a positive border-effect between the provincial-level regions of China on acquisition performance. And second, the results suggest that crossing cultural borders will improve the acquisition performance with 1%. Overall, this study found proof for the relevance of intra-national differences in explaining the heterogeneity in acquisition performance. However, this study fails to support the suggestion that domestic borders create additional costs for acquiring firms, but shows that ‘’foreignness’’ can be an asset, instead of a liability.
Besides using finer geographic units as a proxy to study the impact of geographic location on business activities, this paper offers a theoretical contribution to the M&A literature. By examining the impact of domestic borders on M&As, it enriched our understanding of the factors that drive M&A strategic decisions and performance. Since the mid-1980s, acquisitions have been a popular way of doing business through buying, selling, dividing, and combining different companies and similar entities. Acquisitions can help an organization to grow rapidly in its current industry or country of origin, or a new industry or new location, without the creation of a subsidiary or the use of a joint venture. However, despite the commonly cited benefits and the increasing number of acquisitions, many acquisitions are described as failures. With a failure rate of acquisitions estimated by scholars in a range of 60%-80%, the empirical reality of M&A performance is inconclusive (Moeller, Schlingemann, 2005b). To address the inconsistency between the theoretical advantages and the practical failures this paper explores contextual conditions that impact acquisition performance in a large emerging economy, and thereby this paper contributes to the understanding of the heterogeneity in M&A performance.
5
2014 is accounted for by domestic acquisitions (Brown & Chan, 2015). Despite the fast growing merger market of China, research on Chinese M&As lags behind because empirical research on the Chinese merger markets started only from the late 1990s (Peng, Luo, & Sun, 1999), and most of this studies used case studies and descriptive statistics (Lin, Peng, Yang, & Sun, 2009). Yang, Sun, Lin, & Peng (2011) observed that only six previous papers deal with M&As in China, whereof three use quantitative methods. Overall, there is a lack of research on regional differences within China, or in large and diverse countries in general for that matter. The impact of the distinctive subcultures in China is rarely considered by researchers, which limits our understanding on a potentially crucial driver on organizational behavior and performance (Schlevogt, 2001).
In the next section, extant literature will be reviewed, to provide a more in depth elaboration regarding the existence of border effects, and the intra-country differences that are present in China. Thereafter, the methods section offers an overview of the conducted research, by which the sample, variables and statistical methods will be elaborated. Finally, the results will be presented followed by the discussion.
II. THEORY
Borders matter
6
1995). In those cases, the gains from doing business abroad do not out weight the notable costs that firms face when they cross borders.
Historically, transactions cost economics (TCE), first proposed by Williamson (1979), have provided the theoretical foundation on which cross-border M&A research is based. The focus in research following TCE has been the uncertainty and risk associated with geographical distance and different institutional and cultural environments (Shimizu, Hitt, Vaidyanath, & Pisano, 2004). Firstly, the TCE explains how geographical distance increases the transaction costs. Doing business in a non-proximate location causes a ‘’liability of distance’’ (Boeh & Beamish, 2012). This liability is a consequence of costly frictions that impede the ability to control a non-proximate business entity, consisting of transportation costs, monitoring costs (Chapman, 2003), and information asymmetries (Coval & Moskowiz, 1999). The impact of institutional differences can also be explained by the TCE, as differences between the political, economic, and regulatory environments of the acquirer and target increase the costs of cooperation (Davidson & McFetridge, 1985).
7
business network. However, the process of becoming an ''insider'' is complicated by being a ''foreigner'' (Johanson & Vahlne, 2009). .
Do domestic borders matter?
8
The EG perspective, acknowledging that spatial variation is also present within countries, contributed to the emergence of two streams of international business research that focuses on different levels of geographic units to study locational effects. At one hand, some researchers started to zoom out and use supra-national regions instead of individual countries to define the geographical scope of business. For instance, Qian, Li, & Rugman (2013) distinguish four global market areas: the triad regions (i.e., Asia, Europe, and North America) and the rest of the world, to study the ‘liability of regional foreignness’ and its effect on geographic diversification and firm performance. On the other hand, researchers started to zoom in, and use finer geographic units as a proxy to study the impact of geographic location. For example, Chan, Makino, & Isobe (2010) examine whether subnational regions matter in explaining foreign affiliate performance in two host-country settings, the United States and China. They found that the variation in foreign affiliate performance across subnational regions is larger in emerging economies than it is in advanced economies, as the former tends to have more disconnected economies due to a lack of national markets, efficient market intermediaries, and infrastructure.
This study will follow the second stream of research, by acknowledging the existence of intra-country differences that cannot be ignored. Therefore, this study examines the impact of subnational borders in China on acquisition performance, and whether or not these borders create the same liabilities as national borders, suggested by many international business scholars. The impact of two types of domestic borders within China will be investigated. The first logical step in zooming in, is to use a finer administrative division as a proxy to study the impact of borders on acquisition performance. Whereas traditional international business studies take country-level administrative divisions as a proxy – the impact of going from country A to country B, this study will use provincial-level administrative divisions – the impact of going from province A to province B.
9
examining the role of organizational cultural differences (e.g., Cartwright & Cooper, 1993; Marks, 1982; Weber, 1996). Since culture is not discontinuous in space, which means that societal cultures do not change abruptly at administrative borders, this provincial-level division is unable to capture the border effect caused by cultural differences. Therefore, the cultural division of China will be used as a proxy to examine the influence of crossing cultural borders.
Domestic borders in China
Provincial-level administrative borders
With an average rate close to 10 percent, China has experienced an exceptional growth during the past decades (Xiaolu, 2007). However, alongside this fast economic growth has come a challenge: increasing regional disparity, as China has been struggling to follow an equitable growth path with new pressures and challenges (Bin, 2015; Fujita & Hu, 2001). Since the 1970s, there has been a transition process in China from a command and closed economy to a market-oriented and open economy, which has been conceptualized as a triple process of decentralization, marketization, and globalization (Wei, 2001). This gradual, partial, and spatially uneven process has had profound impacts on the institutional and economic development of China’s provinces (He, Wei, & Xie, 2008).
The institutional framework of a country consists of both formal (e.g., regulations and law) and informal (e.g., codes of conduct and norms) dimensions (North, 1990). Countries vary in their level of institutional development, but the development of institutions also varies within them (Chan, Isobe, & Makino, 2008). In China, the transition process from a planned to market economy required the establishment of an almost entirely new set of formal institutions, including political and judicial regulations, economic rules, and third-party enforcement. In general, the institutional development towards a market-oriented society is slow, whereby inefficiencies, corruption and low transparency still mark China’s governmental sector. Additionally, due to the gradual development of legal institutions and the decentralization of political institutions, the degree of institutional development differs among China’s administrative regions (Hasan, Wachtel, & Zhou, 2009).
10
central government exhibits a step-by-step shift of attention from the coastal areas to the inland and rural regions of China. This growth pattern prioritized the development of the eastern provinces, whereby the inland regions had to catch up later (Bin, 2015). The gap between the east coast provinces and the less developed inland provinces, especially the western provinces, is growing fast2. To give an example: the per capita income of Shanghai, which is the most developed provincial city, is 10 times higher as compared to Guizhou, the less developed province of China (NBS, 2006). The outflow of capital and human capital from the inland regions to the eastern provinces contributed to the regional divergence (Xiaolu & Gang, 2004). Second, Demurger et al. (2002) found that the combination of geographic location, economic policy, and other factors such as the quality of the infrastructure resulted in regional disparities.
In sum, the degree of economic and institutional development in China differs among its provinces. Therefore, we suggest that the provincial-level administrative borders capture those institutional and economic differences. Figure 1 presents a map of China’s provincial-level administrative divisions.
Subcultural borders
Beside the substantial differences in economic and institutional development of China’s regions, the degree of cultural diversity in China is high. Comparable to the global cultural differences between the ‘’East” and the “West”, China’s subcultural variation is considerable and may lead to significant differences in organizational behavior (Ahlstrom, 2010). Today, China’s government recognized 56 ethnic groups, whereof the Han Chinese is the largest ethnic group, constituting approximately 92% of the population of mainland China. The Han Chinese inhabits mainly the eastern part of China, while Central and Western China is the residence of most ethnic minorities. Alongside with ethnic diversity, there is a high degree of linguistic or dialect diversification. This diversity in dialect is something that has a long tradition in China, and despite the emergence of Mandarin as the national language, cultural differences cannot be separated from the diversity in languages that it entails (Erbauch, 1995).
2
11 Figure 1 - Map of China's provincial-level administrative regions
12
Dialects carry clear subcultural attributes, as they play an important role in forming individuals’ identification and influencing their perception of organizations (Gong, Chow & Ahlstrom, 2010).
Overall, China can be divided in eight cultural regions, in which the inhabitants share a relatively homogeneous cultural trait, language and human activity (Wu, 1996). In the methods section, the cultural traits belonging to these cultural regions are elaborated. We suggest that the borders of those cultural regions capture the most important cultural differences that exist within China. Figure 2 presents a map of China’s cultural regions.
III. HYPOTHESES
Given the diversity of China’s administrative and cultural regions, this study examines whether the theories that explain the cross-border effect on the country level can also be applied to firms that cross subnational borders instead of national borders. First, the relationship between China’s institutional borders and acquisition performance will be elaborated. Second, we will discuss the impact of cultural differences on acquisition performance.
Institutional borders and acquisition performance
Institutions are defined as ‘’the rules of the game in society or, more formally, are the human devised constraints that shape human interaction’’ (North, 1990, ch. 3-5). Institutional theory argues that the beliefs, goals, and actions of individuals and organizations are strongly influenced by environmental institutions (Scott, 1987). The theory emphasizes the critical interaction between institutions and organizations, and suggests that they co-evolve in close interaction (North, 1990). Accordingly, the success of firms over time is fundamentally influenced by the institutional environment, as informal and formal institutional constraints reduce the uncertainty in human and organizational interaction (Dikova et al., 2010). Since institutions affect the performance of the economy by their effect on the costs of exchange and production (North, 1990), the institution-based view could be regarded as a complement of the TCE.
13
inconsistencies are common in emerging economies, due to the decentralization of authority and the delegation of responsibility and autonomy for regional development to regional governments, which increases the difficulty in achieving legitimacy in the host environment (Montinola, Qian, & Weingast, 1995). For example, Dikova et al. (2010) found that institutional differences have a negative effect on acquisition deal-completion likelihood and would extend deal-completing duration, due to the hurdles caused by unforeseen institutional differences.
Further, environmental institutions necessary to accomplish economic exchange vary in their complexity, since the formal structure of rights in a specific exchange determine how costly it is to make the exchange. Especially in the case of acquisitions, the complexity of the exchange can be high; as such exchanges are subject to regulatory constraints caused by bureaucratic self-interest, political extraction and the protection of local firms (Bittlingmayer & Hazlett, 2000). Second, rapidly changing institutions in transition economies generate inconsistencies between the requirements of different institutions, as well as uncertainty over institutional changes in the future. To overcome these uncertainties, firms have to rely on network-based coordination mechanisms that cause transaction costs (Meyer, 2001). Overall, differences between the political, economic, and regulatory environments of the acquirer and target increase the costs of cooperation (Davidson & McFetridge, 1985), which in turn decreases the potential for M&A success in the case of cross-border deals.
Research employing the institutional theory in China has tended to highlight the constraining nature of institutions (e.g. Peng & Heath, 1996), while institutions not only specify limits, but enable certain actions as well. For example, Luo (2001) found that relations between firms and governments are mainly cooperative, and have a positive influence on sales and financial performance. Additionaly, Peng (2000) found that co-operation with local authorities can help the firm to obtain various approvals from central authorities. However, firms that are located in another institutional environment find difficulties in creating such relations, as it requires a deep understanding of, and integration with the local institutional environment (London & Hart, 2004). Therefore, local firms are at an advantage in creating a competitive advantage through relations with authorities.
14
administrative borders implies that the acquiring firms face institutional differences that increase the complexity and costs of obtaining legitimacy in the host environment, which in turn decreases the potential for M&A success in the case of cross-border deals. This leads to the following hypothesis:
Hypothesis 1: Acquisitions in the same provincial-level administrative region will perform better than cross-province deals.
Cultural borders and acquisition performance
15
language, educational systems and many other intangible factors (Calori, Lubatkin, Very, & Veiga, 1997). This unique administrative style is often more clearly apparent to foreigners than to the nationals themselves (Hofstede, 1980).
In sum, extant literature suggests that organizational cultures arise to a large extent from societal culture. The organizational culture, in terms of organizational and administrative practices, management styles, and employee behavior, is either directly affected by the societal culture, or is indirectly affected by the value and beliefs of its organizational members, which are in their turn, deeply shaped by historical experiences and institutions. Since the organizational culture is embedded in, and influenced by the macro societal culture where the organization is located in, this study will research the impact of societal culture, rather than the organizational culture, on acquisition performance.However, the difficulties, costs, and risks associated with cross-cultural interaction that will likely affect acquisition performance arise at the organizational level. Previous studies indicate that the complications experienced by the acquiring firm in managing the acquisition stem from cultural differences between the target and acquiring firm (e.g., Chatterjee, Lubatkin, Schweiger, & Weber, 1992).
Several theories have been developed to explain the role of culture in acquisitions, which can be grouped into three categories: the cultural fit perspective, the acculturation perspective, and the social constructivist perspective on culture (Dauber, 2012). The cultural
fit perspective suggests that the degree of compatibility between the firms involved in an
16
specific, such as marketing and labor relations (Geringer, Beamish, & DaCosta, 1989). Finally, cultural distance decreases the likelihood that local market knowledge is readily available to the acquirer, which increases the coordination and monitoring costs, and subsequently, decreasing the firm’s market value at the time of the acquisition (Doukas & Travlos, 1988). The second perspective focuses on the ‘’acculturation’’ process, instead of static cultural differences between the acquirer and target (Shahl & Voight, 2005). In the context of acquisitions, acculturation is defined as ‘’the outcome of a cooperative process whereby the beliefs, assumptions, and values of two previously independent work forces form a jointly determined culture’’ (Larsson & Lubatkin, 2001). This shared culture consists of a common organizational language, mutual understanding and trust, and values that promote shared interests, and is a prerequisite for M&A performance. The course and outcome of the acculturation process depend on several factors, including ‘managerial’ aspects and ‘human-sociological’ aspects. For example, Cartwright and Cooper (1992) found that the more autonomy is restricted, the higher the target’s perception of being dominated and inferior to, the acquiring firm, which will cause acculturative problems, such as conflicts and tension. The human-sociological aspect points to cultural differences, and Weber et al. (2006) found that a high degree of cultural differences negatively affects the acculturation process, as it results in post-merger stress, negative attitudes about the acquisition, and a decrease in actual cooperation. The social constructivist perspective suggests that culture is based on shared patterns of interpretation, which are continually produced, reproduced and changed by individuals identifying with them (Vaara, 2002). The theory emphasizes symbolization and communication, and culture is seen as an interpretative and evolving process. In the context of M&A, social constructivists suggest that difficulties in the post-acquisition integration process may be best understood in terms of in-group out-group bias and the pursuit for social identity. Often observed in acquisitions is an exaggerated focus on the differences and a lack of attention to the similarities, which causes a sense-making mechanism, focused on ‘’our’’ uniqueness and ‘’their’’ otherness (Kleppestø, 1998).
17
Olie, 1990). The purpose of this paper is to explore if theories used to explain border effects at the country level also can be applied to the subnational level in China. Therefore, based on theoretical and empirical evidence about the impact of national culture differences on acquisition performance, we expect the same impact of subcultural differences on acquisition performance, given the heterogeneity of China’s subcultures. This leads to the following hypothesis:
Hypothesis 2: Acquisitions in the same cultural region will perform better than cross-cultural region deals.
IV. METHODS
Research settings
The data used is this study is obtained from multiple databases. The database used to attract information about the M&A deals in China for the chosen time period is obtained from the
Thomson Reuters SDC Platinum database. Thomson Reuters SDC Platinum compiles and
analyzes information on a wide range of financial topics, such as; private equity, M&A deals, project finance, and other types of financial transactions (Thomson Reuters, 2016). Since the 1970s up till today, Thomson Reuters SDC Platinum collected information about approximately 900.000 M&A deals that have been conducted over the whole world.
The stock price information concerning the acquiring firms and the information on the stock exchanges these firms are listed on is gathered from the database Datastream. Datastream is a database that provides financial and economic data on firms, like for example; share prices, indices, equities and options (Thomson Reuters, 2016).
Data & sample
The acquisitions researched in this study are deals between acquiring and target firms both located in China. By defining China, we include: China, Hong Kong, Macau and Taiwan. Hong Kong and Macau are special administrative regions under the jurisdiction of the ‘’People’s Republic of China’’. Taiwan, officially the ‘’Republic of China’’, is formally a sovereign state. In practice, however, only 22 countries recognize the sovereignty of Taiwan, and Taiwan is treated as a province of China by the rest of the world (PRC, 2005).
18
methodology. Subsequently, the M&A data is refined to include all the deals: (1) announced between Jan, 1990 and June, 2015; (2) with a transaction value above US$ 10 million; (3) for 100% purchase of the target, and; (4) deals that do not involve a repurchase of own shares, recapitalization, or a spin-off to existing shareholders. This refinement led to a sample of 1,805 acquisitions. The geographic location of the firms involved in the M&A is the focus of interest in this study, for a number of deals it was not possible to identify the location of the acquiring and/or target firms’ headquarters3, reducing the sample size by 174 deals.
Once the 1631 acquisitions were identified, the SEDOL numbers of the acquirer firms, also provided by the SDC database, were used to obtain the share prices of these firms from Datastream. For 464 acquisitions, the SEDOL- or any other identification numbers were unable to be identified, reducing the sample size to 1167. Secondly, for nine listed companies, Datastream was unable to find the return index, reducing the sample size to 1158 deals. Another 163 deals were removed from the sample due to acquiring firms involved in multiple acquisitions within a period of three months, which will result in ambiguity about the attribution of a particular deal on the cumulative abnormal returns after the announcement. Thereby, a correct estimation of the normal returns cannot be guaranteed when the effects of a previous deal may be incorporated in the estimation period (T-30, T-60), which will bias the estimated normal returns. Further, if the acquiring firms did not have a sufficient amount of stock price information available the days surrounding the acquisition announcement, we were compelled to remove those deals from the sample. Overall, this refinement led to a sample of 910 deals.
Returns on the relevant indices the acquiring firms are listed on were also derived from Datastream. The firms are listed on: the Shanghai Stock Exchange (n=208); Shenzhen Stock Exchange (n=382); DAX 30 Stock Exhange (n=2); FTSE World Stock Exhange (n=3), Hang Seng Stock Exhange (n=181), Straits Times Stock Exhange (n=7); S&P Composite Stock Exhange (n=47); Taiwan Stock Exhange (n=79), and; TOPIX Stock Exhange (n=1).
Measures
Dependent variable
The performance of the acquiring firms after conducting an acquisition are measured based on stock price performance (Duso, Gugler, & Yurtoglu, 2010). By making use of event study
19
methodology, we follow the majority of acquisition performance studies (Zollo & Mierer, 2008). Fama, Fischer, Jensen and Roll (1969) were the first to point out the relevance of event study methodology in measuring the adjustment of stock prices to new information. An event study analyzes the impact of a specific event on the value of the firm (MacKinlay, 1997), as it attempts to reflect on the informational value of an event by calculating abnormal returns (Shah & Arora, 2014). In this study, event study methodology is used to analyze any fluctuations in the stock prices in the time around the acquisition announcement.
1. Time sequence of the event study.
To depict the performance of an acquisition, an event window and an estimation window were established. The event window is defined as the time interval chosen for the study where τ = 0 is the announcement date of the acquisition. Usually for the study of M&A announcements the event window chosen is larger than solely the announcement day itself. In practice, the period of interests is often expanded to multiple days, before and after the announcement (Shah & Arora, 2014; MacKinlay, 1997). The pre-event period is considered to incorporate any leakages of information, such as rumors. The post-event period is considered to estimate the delay in the reach of the information being dispersed (Peterson, 1989). For the acquisitions in the sample announced on a non-trading day, the announcement date is converted into the first upcoming trading day.
The calculation of an event’s impact requires a measure of the abnormal return. The abnormal return is the actual return of the security in respect to the normal return of the firm over the event window (MacKinlay, 1997). A market model is used to estimate the normal returns, by making use of benchmarks. The market indices on which the acquiring firms were registered represent the benchmarks in this study, including: the Shanghai Stock Exchange; Shenzhen Stock Exchange; DAX 30 Stock Exhange; FTSE World Stock Exhange, Hang Seng Stock Exhange, Straits Times Stock Exhange; S&P Composite Stock Exhange; Taiwan Stock Exhange, and; TOPIX Stock Exhange. The market indices provide information about ‘normal’ movements in the market, by which fluctuations in firm stock prices can be assigned to an entire market anomaly or by a certain, firm specific, event.
Figure 3 exhibits the timeline for the event study. The M&A announcement is equal to
20
& Wolf, 2002), the event window consists of a total number of three days (T-1, T1), measuring the abnormal return a day before, after, and at the announcement day itself.
To calculate the normal returns, an estimation window of 30 days is established. It is important that the estimators for the parameters of the normal return model are not influenced by the returns around the event. Consequently, it is necessary for the event window and the estimation window not to overlap (MacKinlay, 1997), therefore the estimation window is the time between T-30 and T-60.
2. Measurement of the normal and abnormal returns
The abnormal return is calculated by identifying the difference between the actual returns and normal returns of the acquiring firms. Abnormal return of a security i;
𝐴𝑅𝑖,𝑡 = 𝑅𝑖,𝑡− 𝐸[𝑅𝑖,𝑡] (1)
Where Ri,t is the actual return and E[Ri,t] is the estimated or normal return (Duso, Gugler, & Yurtoglu, 2010). In equation (2) is displayed how the normal return is calculated;
𝐸[𝑅𝑖,𝑡] = 𝛼𝑖 + 𝛽𝑖 𝑅𝑚,𝑡+ 𝜀𝑖,𝑡 (2)
Where αi is the intercept coefficient, βi is the slope, Rm,t the formation of the used benchmarks, and εi,t is expected to be equal to 0 (MacKinlay, 1997). If the abnormal returns in the event window are added the cumulative abnormal returns (CAR) become apparent;
𝐶𝐴𝑅𝑖[𝑡1, 𝑡2] = ∑𝑡2 𝐴𝑅𝑖,𝑡
𝑡=𝑡1 (3)
To make assumptions about the data, the CAR has to be calculated by equation (3) (MacKinlay, 1997).
21 Independent variables
1. Institutional border effect
China consists of 34 provincial-level administrative divisions, including: 23 provinces, five autonomous regions, four centrally administered municipalities, and two special administrative regions. The autonomous regions: Guangxi, Inner Mongolia, Ningxia, Tibet and Xinjiang, are mainly inhabited by ethnic minorities and have their own local governments. The centrally administered municipalities, Shanghai, Beijing, Tianjin and Chongqing, are a higher level of city, with status equal to that of provinces.
To measure the effect of institutional borders, a dummy variable is used to identify deals by which the acquiring and target firm are located in the same administrative region (‘0’) and cross-administrative region deals (‘1’), by which the acquiring and target firm are located in different administrative regions.
2. Cultural border effect
Subcultural boundaries are used as the proxy of China's regional cultural differences. The assumption is that within one cultural region, individuals share a relatively homogeneous cultural trait or human activity. To identify subcultures, several factors have to be taken into account, such as: climate, topography, geography, and political and economic ideologies (Lenartowicz & Roth, 1999). The division of China's subcultural regions is primarily influenced by three factors: (1) functional variables, like typography and climate; (2) historical development, like linguistic systems and the time that a region is ruled by core empires in the past; (3) locational conditions, which represent the dynamic integration of functional variables and historical development.
By taking these three factors into account, Wu (1996) established a partition of China's subcultural regions, which has been widely accepted by researchers. He identified eight subcultural regions, including: (1) Inner Mongolia Culture Region (R1) - dominated by Mongal ethnics that live on stock farming and have their own language. The climate in this region is arid; (2) Xinjiang Culture Region (R2) - dominated by Uyghur ethnics that live on stock farming and have their own language. The climate in this region is arid; (3) Tibetan
Culture Region (R3) - dominated by Tibetan ethnics that live on stock farming and have their
own language. The climate in this region is the plateau climate; (4) Central Plains Culture
Region (R4) - this region is the origin of Han ethnics that show the traits of typical
22
monsoon climate of medium latitudes; (5) Northeast Culture Region (R5) - this region is close related to the central plains culture region (R4), due to the migration of many Han ethnics in the 19th century; (6) Yangtze Culture region (R6) - this region has the atmosphere of business as the Yangtze is providing developed water transportation. The climate is the subtropical monsoon climate, and rice is the staple food of this region; (7) Southwest Culture Region (R7) - Han ethnics and other minorities lived in this region. Therefore, the culture in this region shows mixes traits; (8) Southeast Culture Region (R8) - this region is far away from the core circle of Han ethnics, though people living here are mainly migrated from the Central Plains Culture Region (R4) and Northeast Culture Region (R5) more than 1000 years ago. This region shows traits of business culture as the coastline provides the opportunity to communicate and drive business with the world.
To measure the effect of cultural borders, a dummy variable is used to identify deals by which the acquiring and target firm are located in the same subcultural region (‘0’) and cross-cultural deals (‘1’), by which the acquiring and target firm are located in different subcultural regions.
Controls
Several factors have proven to have an impact on post-acquisition performance (King, Dalton, Daily, & Covin, 2004). Since the purpose of this paper is to investigate the impact of subnational factors on acquisition performance and all the deals occurred in China’s merger market, international factors are excluded and subnational factors included. Additionally, there is controlled for deal- and organizational specific factors. Thus, in total there are three types of controls added to the model.
First, there is controlled for locational factors, including: (1) The Geographic distance between the acquiring and target firm – calculated by the GPS addresses of the two firms4 – as geographic distance creates a ‘’liability of distance’’ (Boeh & Beamish, 2012); (2) The level of Institutional distance5 between the home and host regions – using the standard IPD6
4
First we identified the GPS coordinates of the acquiring and target firm using GPS Visualizer (http://gpsvisualizer.com/). After this identification, we used the Haversine formula to calculate the absolute distance between the GPS addresses of the firms.
5 As a proxy for institutional distance, the measure introduced by Kogut and Singh (1988) in their paper on the relation between cultural distance and entry mode is used.
23
institutional measures for the institutional distance between China, Hong Kong and Taiwan, and the NERI7 institutional measures for the institutional distance between China’s provinces.
Secondly, we control for firm specific factors that are likely to impact post-acquisition performance. We control for: (4) Prior performance – measured by the acquirer’s return on assets (ROA), at the end of the year ahead of the year of the acquisition – as according to Bloom (1999), a firm’s prior performance has the potential to influence the firms’ next year performance; (5) Market-to-book ratio of the acquiring firm in the year of the acquisition - as high market-to-book ratio, so-called ‘value’ firms, perform better than low market-to-book ‘glamour’ firms (Rau & Vermaelen, 1998); (6) Relative size of the acquiring firm – measured by the number of employees a year prior to the acquisition – as comparatively, larger acquirers are known to underperform (e.g., Moeller, Schlingemann, & Stulz, 2004); (7)
Public/private indicator – by which a dummy variable is used to identify public (‘1’) and
private (‘0’) targets – as the returns of the acquiring firms in publicly listed targets differs significantly from private targets (Officer, 2007). The information used to measure prior performance, market-to-book ratio and relative size is obtained from Datastream, information on the target’s status is obtained from Thomson Reuters SDC Database.
Finally, we control for deal specific factors that are likely to impact the post-acquisition performance, including: (8) Relatedness indicator – by which dummy variables are set ('1') if the SIC codes of the target and the acquiring firm share the same first-two digit, and are therefore in the same industry – as several studies (e.g., Hitt, Ireland, Camp, & Sexton, 2001) suggest that deals between related firms increase deal performance as industry familiarity diminish the need for acquiring firm managers to learn the business of the target firm; (9) Deal attitude indicator – by which a dummy variable is used to identify hostile (‘1’) and unsolicited (‘0’) deals – because it is suggested that hostility impacts performance (Eckbo, 2009); (10) Method of payment – which is the ratio of cash to equity when the deal is financed – because it is suggested that cash-financed deals outperform stock-financed deals
stakeholders, strategic vision and innovation; (6) security of transactions and contracts; (7) market regulations, social dialogue; (8) openness, and; (9) social cohesion and social mobility. The data is publically available at: http://www.cepii.fr/institutions/EN/ipd.asp.
24
(Heron & Lie, 2002). Information on the deal related controls is obtained from Thomson Reuters SDC Database.
To remove extreme outliers, all the control variables are winsorized between 0.5% and 99.5%. Additionally, for the variables that failed the Shapiro-Wilk test for normality, the logs were employed (Institutional distance, Prior performance, Market-to-book ratio, Relative
size, Method of payment).
Estimation
To analyze the impact of administrative and cultural borders on the CAR after the acquisition announcement, an Ordinary Least Squares (OLS) regression technique is used to estimate the model. OLS regression techniques are in this case appropriate since the dependent variable, the CAR, is a non-count, normally distributed variable. 278 observations were lost in the regression analysis due to missing values on one of the control variables, such that 632 were used in the regression analysis.
V. RESULTS
Table 1 reports the descriptive and correlation statistics for all the variables. The two
independent variables, the cross-province and cross-culture indicators are highly correlated (ρ = 0.76), which is as expected, since the administrative and cultural regions for the most part overlap each other. Except this, other correlations above the 0.7 are not observed, which is the first sign of that there is no multicollinearity among the key variables.
M&A activity in China
The degree of M&A activity in China has increased substantially in the time period between 1990 and 2015. Figure 4 makes use of the full sample of global acquisitions and Chinese acquisitions in the time period between 1990 and 2015, and illustrates the share of Chinese acquisitions in global M&A activity. The graph shows that the share of Chinese acquisitions has grown rapidly, even more extensively compared to the overall increase of acquisitions around the world; from 0.5% in 1990 to almost 10% today.
25
Mean S.D. Min Max 1 2 3 4 5 6 7 8 9 10 11 12 13
1 Acquisition performance (CAR) 0.03 0.08 -0.53 0.98 1.00
2 Cross-province indicator 0.41 0.49 0 1 0.11 1.00 0.00
3 Cross-cultural region indicator 0.31 0.46 0 1 0.08 0.76 1.00 -0.01 0.00
4 Geographic distance (km) 725.59 1496.46 0.00 5987.52 0.03 0.32 0.34 1.00 -0.35 0.00 0.00
5 Institutional distance (IPD) (ln) 0.31 1.25 0.00 5.42 -0.12 0.30 0.20 0.10 1.00 0.00 0.00 0.00 0.00
6 Institutional distance (NERI) (ln) 0.16 0.70 0.00 11.23 0.03 0.28 0.22 0.08 -0.06 1.00 -0.35 0.00 0.00 -0.02 -0.08 7 Prior performance (ln) 1.59 1.01 -3.51 5.45 -0.11 -0.03 -0.09 -0.15 0.00 -0.01 1.00 0.00 -0.36 -0.02 0.00 -0.97 -0.82 8 Market-to-book ratio (ln) 0.83 0.97 -4.61 5.47 0.14 0.25 0.19 0.11 -0.06 0.09 0.17 1.00 0.00 0.00 0.00 0.00 -0.10 -0.01 0.00 9 Relative size (ln) 7.11 1.59 0.00 12.82 -0.03 -0.06 -0.04 0.01 -0.04 -0.03 -0.03 -0.03 1.00 -0.47 -0.08 -0.27 -0.73 -0.30 -0.44 -0.41 -0.38
10 Target status indicator 0.18 0.39 0 1 -0.16 -0.17 -0.16 -0.07 0.04 -0.09 -0.03 -0.28 0.17 1.00 0.00 0.00 0.00 -0.04 -0.22 -0.01 -0.38 0.00 0.00
11 Relatedness 2-digit 0.38 0.49 0 1 -0.02 -0.01 -0.07 -0.03 0.00 0.05 0.15 -0.04 0.13 0.06 1.00 -0.54 -0.86 -0.05 -0.38 -0.95 -0.17 0.00 -0.29 0.00 -0.06
12 Deal attitude indicator 0.97 0.18 0 1 0.04 0.02 -0.02 0.02 0.02 -0.01 -0.06 0.03 -0.03 0.04 -0.01 1.00 -0.27 -0.46 -0.65 -0.63 -0.59 -0.79 -0.09 -0.42 -0.45 -0.27 -0.74
13 Method of payment (ln) 22.30 35.82 0.00 100.00 -0.13 0.00 -0.01 -0.05 0.11 -0.03 0.05 -0.03 0.01 -0.02 0.06 -0.06 1.00 0.00 -0.91 -0.70 -0.11 0.00 -0.43 -0.19 -0.44 -0.80 -0.59 -0.08 -0.08
26 0 1.000 2.000 3.000 4.000 5.000 6.000 7.000 1990 1995 2000 2005 2010 2015 N um ber of M& A s
Year Outward M&A
Domestic M&A
Figure 5 illustrates the amount of China’s domestic and outward acquisitions, and shows that
the share of domestic acquisitions increased to more than 90% of total M&A activity in China.
Figure 4 - Share of Chinese M&As in global M&A activity
0 2 4 6 8 10 12 14 1990 1995 2000 2005 2010 2015 Shar e of g lobal M% A ac ti v it y (% ) Year
27
Of the domestic deals, there is a notable difference in the distribution of acquisitions among the macro-regions of China, as 85.4% of the acquisitions are conducted within the eastern coastal region. The share of M&A activity in the rural western region is only 4.5%, and 10.1% in the central region. Further, we find that of the domestic deals, acquiring firms tend to favor an acquisition within their own administrative and/or cultural region, as 58.8% of the acquisitions are within the same administrative region, and even 69.3% of the acquirers stay within the same cultural region. Acquisitions within the same cultural region are mostly conducted in the Southeast culture region (52.9%) and the Yangtze culture region (23.1%). Not surprisingly, these two cultural regions show traits of a business culture.
Against the suggestion of scholars, who estimated the failure rate of acquisitions in a range of 60%-80% (e.g., Moeller & Schlingemann, 2005b), we find that 65.9% (n=600) of the Chinese acquisitions create positive abnormal returns for the acquiring firms, and only 34.1% (n=310) of the acquisitions failed to create value. On average, acquisitions in China create positive abnormal returns of 3%, appendix 1 reports the distribution of the CAR. Even more surprisingly, cross-border acquisitions performed, on average, better than acquisitions of which the acquiring and target firm were both located in the same administrative and/or cultural region. Table 2 presents a summary of the CARS for within-region deals and cross-border deals.
CAR(T-1,T1)
Cross-province acquisition Mean S.D. Freq.
Yes 0.0437 0.0635 375
No 0.0254 0.0886 535
Cross-culture acquisition
Yes 0.0429 0.064 279
No 0.0285 0.0854 631
Table 2 - Acquisition performance and cross-border behavior
The impact of domestic borders on M&A performance
28
We conducted a number of checks before considering the results. First, a Breusch Pagan test is used to test for heteroscedasticity in the models8. The test suggests that heteroscedasticity is in all models not a concern, with a chi-square of 1.95 And p-value of 0.1622 in the base model. Second, as a high multicollinearity among the independent variables reduces the reliability of the relationship between the dependent variable and the independent variables, a variance inflation factor (VIF) test is conducted to check for multicollinearity within the models. In the base model, the highest VIF founded is 1.21, with a mean of 1.09 which is well below the maximum threshold for the level of multicollinearity of 10 and even 5 (Kutner, Nachtsheim & Neter, 2004). This suggests that multicollinearity is unlikely to bias the regression coefficients.
Model 1 reports on the set of control variables. The adjusted R-squared of the base
model is 10.9%, which is notably high compared to other event studies. For example, Moeller and Schlingemann (2005b) in their study on acquiring firm returns in the recent merger wave, conducted an event study, and OLS regression, and reported adjusted R-squares ranging from 2.4-5.6%. Further, a study of the same authors comparing cross-border and domestic acquisitions, using event study methodology and regression estimations, reported adjusted R-squares ranging from 3.6-4.7% (Moeller & Schlingemann, 2005a).
In model 2, the first independent variable, the administrative border variable, is added to the set of control variables. We excluded the institutional distance variable, as the dependent variable is assumed to capture this distance. Contrary to the expected effect of institutional borders, the result shows a positive and significant coefficient (p < 0.01), which indicates that crossing an administrative border results in a higher CAR of 1% as a result of the acquisition.
In Model 3, we added the institutional distance control to model 2. Surprisingly, the significant coefficient (p < 0.01) of the administrative border variable is higher compared to model 2. This means that, controlling for institutional distance, crossing an administrative border results in a higher CAR of 1.3% as a result of the acquisition. Overall, the results from model 2 and 3 suggest that administrative borders matter, but not in the way we expected. Instead of resulting in higher costs and lower acquisition performance as a consequence of crossing a border, administrative borders create value. Therefore, this surprisingly result fails to support hypothesis 1. Additionally, it is not the institutional distance between two
29
Table 3 - OLS regression
(1) (2) (3) (4) (5)
VARIABLES CAR(T-1,T1) CAR(T-1,T1) CAR(T-1,T1) CAR(T-1,T1) CAR(T-1,T1)
Cross-province indicator 0.01*** 0.01*** 0.01** (3.012) (3.579) (2.478) Cross-culture indicator 0.01** 0.00 (2.264) (0.132) Geographic distance -0.04 -0.13 -0.12 -0.12 -0.13 (-0.448) (-1.448) (-1.108) (-1.177) (-1.160)
Institutional distance (IPD) (ln) -0.51*** -0.66*** -0.58*** -0.66***
(-3.114) (-4.160) (-4.424) (-4.013)
Institutional distance (NERI) (ln) 0.28 -0.29 0.12 -0.28
(0.302) (-0.299) (0.116) (-0.311) Prior performance (ln) -0.44*** -0.46*** -0.43*** -0.42*** -0.43*** (-2.911) (-2.920) (-3.130) (-2.804) (-3.033) Market-to-book ratio (ln) 0.34 0.32 0.26 0.30 0.26 (1.506) (1.514) (1.170) (1.400) (1.168) Relative size (ln) -0.30** -0.28* -0.25* -0.29** -0.25* (-2.152) (-1.963) (-1.819) (-1.987) (-1.916)
Target status indicator -0.98* -1.03* -0.97 -0.94 -0.97
(-1.651) (-1.687) (-1.628) (-1.534) (-1.613)
Relatedness 2-digit -0.35 -0.32 -0.41 -0.31 -0.40
(-0.840) (-0.808) (-1.008) (-0.776) (-0.967)
Deal attitude indicator -0.14 -0.15 -0.13 -0.12 -0.13
30
administrative regions that causes the border effect, since coefficient of the administrative border variable increases when we control for institutional distance.
Model 4 presents the second independent variable, the cross-culture variable, added to
the set of control variables. Again, the result fails to support the suggestion that borders create costs. The significant coefficient (p < 0.05) is positively related to the CAR, which suggests that the acquisition of a target firm located in another cultural region results in a 1% higher performance, compared to acquisitions within the same cultural region. Given this result, it supports the suggestion that subnational borders matter, but not as expected, and fails to support hypotheses 2.
Model 5 includes both independent variables and all the control variables. Whereas
cultural borders in model 3 have a significant positive effect on acquisition performance, the results in model 5 show an insignificant impact of these borders. The fact that the results in model 4 change materially is what we expected, given that the province and cross-culture indicators are highly correlated (ρ = 0.76). However, this result suggests that, controlling for cultural differences, there is still a statistically significant impact (p < 0.05) of administrative borders on performance.
Robustness
To determine the robustness of the results additional analyses have been conducted.
Additional event study
First, an additional event study, with a different event window, is conducted to determine the robustness of the results. The methodological process of the event study described in the methods section is repeated for the additional event study. Instead of an event window of three days (T-1,T1), an event window of five days (T-2,T2) is established. The CAR resulting from this event window are used to re-estimate the models 1, 3, 4 & 5. All the re-tested models do not report significant differences, the positive and significant results on both hypotheses are again confirmed. Appendix 2 reports the results on model 6-9, which are the re-tested OLS estimations.
Endogeneity
31
identify superior strategies. However, this approach does not account for the fact that strategy choice is endogenous and self-selected, as strategy performance is estimated in an setting where there is assumed that firms randomly choose their strategy (Shaver, 1998). Firms choose the strategy that is optimal given their attributes and those of their industry, and the often used empirical models do not account for this choice process (Masten, 1993).
A potential endogeneity problem is this research is, for example: the question whether crossing a border improves performance, or that only high performing firms make the strategic choice to cross a border. If the latter is the case, it means that the choice of crossing a border is self-selected, and that there are omitted variables influencing a firm’s cross-border behavior. Subsequently, the conclusions drawn from the regression models may be incorrect, since they solely have proven that cross-border acquisitions will enhance the performance of already good performing firms, and do not provide evidence that crossing borders is a good strategy for all Chinese firms. If firms choose to cross borders, a proper evaluation of the impact of crossing borders would be incomplete without the consideration of the underlying characteristics that influence the decision to cross a border. Dastidar (2009) addressed the endogeneity problem by examining whether the self-selection concern plays a role in international diversification, and found evidence of endogeneity in the international diversification decision.
To check for endogeneity in our models, a Heckman two-stage model is used (Shaver, 1998). This model is able to address the concern whether strategy choice is endogenous and self-selected. The first stage of the Heckman approach is a probit regression, used to obtain estimates of lambda (λ), which is the correction of endogeneity. This probit model tests whether a firm’s decision to cross a border is associated with firm-level characteristics. Following partly the Heckman procedure conducted by Dastidar (2009), the independent variables; the cross-border indicators, are regressed on the acquirer-specific characteristics incorporated in our base model, including: (1) Prior performance, since there is evidence that successful firms are more likely to diversify by crossing borders (Dastidar, 2009); (2)
Market-to-book ratio, as firms that have a relatively high market-Market-to-book value tend to be purchasers
(Erel, Liao, and Weisbach, 2012), and; (3) Relative size, since cross-border acquirers have to be large to overcome the advantages of domestic competitors (Caves, 1971).
In the first probit model, the independent variable (cross-administrative region
indicator) is regressed on the acquirer-specific control variables. The second probit model
32 acquirer-specific controls. For both independent variables, we calculated the Inverse Mills Ratio (IMR), or lambda (λ), which represents the estimates of the self-selection correction.
In the second stage, the OLS regression models are re-estimated with the IMRs as additional controls. Appendix 3 presents the results. Model 10 reports the re-estimation of hypothesis 1, by adding the first calculated IMR to model 3. Second, model 11 re-estimated hypothesis 2, this model consists of model 4 and the second calculated IMR. Finally, model
12 re-estimates model 5, which includes all the variables, and the two IMRs. In all the re-tested models, the IMR is not statistically significant, which indicates that endogeneity is not driving the results. More specifically, the impact of crossing borders on acquisition performance is not dependent on acquirer-specific characteristics.
Re-estimation using ‘’pure’’ Chinese firms
Although our definition of China corresponds with the way China is defined by the most part of the world, we want to be sure that the results are not driven by this definition. Therefore, we re-estimated the basic models using only the 513 purely Chinese firms in our sample, by excluding the special administrative regions; Hong Kong and Macau, and Taiwan; which is formally a sovereign state. Appendix 4 presents the results. The re-estimated models demonstrate the same results: the significant positive effect of administrative and cultural borders on acquisition performance is re-confirmed. This suggests that the potential bias caused by our definition of China is not a concern.
VI. DISCUSSION & CONCLUSION
The main findings of this study suggest that domestic borders matter in explaining acquisition performance. However, the impact of China’s domestic borders on acquisition performance is not in the way as we expected.
33
firms rapidly to acquire a portfolio of locational assets (UNCTAD, 2000). Despite these benefits, empirical findings suggest that cross-border acquisitions, on average, do not create value for acquiring firms (Datta & Puia, 1995). This is because, on the other hand, there is argued that; foreignness is a liability. This liability consists of the additional costs that arise from cultural, political and economic differences, and the unfamiliarity with the host environment (Zaheer, 1995). Further, the liability of foreignness is associated with the lack of knowledge, the challenge to obtain legitimacy in the host environment, and discriminatory attitudes toward foreign firms by local stakeholders, such as employees and customers that prefer to interact with local firms (Kostova & Zaheer, 1999). The international business literature argues, that at the bottom line, the benefits of crossing a border, do not out weight the additional costs of doing business abroad (Hymer, 1976). However, we find that, first, crossing administrative borders enhances the acquisition performance by 1.3%, even controlling for institutional distance and cultural differences, there exists a positive border-effect between the provincial-level regions of China and acquisition performance. Thus, it is not so much the institutional distance between the location of the acquiring and target firm causing the administrative border effect, as we expected. But, the results demonstrate that operating in a other administrative region, irrespectively the institutional distance, improve the acquisition performance. Second, the results suggest that crossing cultural borders will improve the acquisition performance with 1%. Following the suggestion of the economic geography literature, this study found proof for the relevance of intra-national differences in explaining the heterogeneity in acquisition performance. However, this study fails to support the suggestion that domestic borders create additional costs for acquiring firms.
34
complicate the acquisition can be an advantage as well. Different cultures come along with different types of knowledge, and are likely to be associated with higher levels of capability complementarity and greater learning opportunities (Stahl & Voigt, 2005). Accordingly, Ghoshal (1987) found that a high degree of similarity between two firms impedes the potential to learn from each other. Beside the benefits, on the other hand, it could be that the costs associated with crossing domestic borders, rather than national borders, are lower. As such, that at the bottom line, the benefits of crossing a border turn out to be greater.
Another explanation for the results could be that the established international business literature on cross-border acquisitions does not apply to Chinese firms. Using traditional international business theories, we have tested whether theories explaining the border effect are applicable in; (1) the case of a large emerging economy, and; (2) firms crossing domestic instead of national borders. However, Chinese firms are in various aspects different than Western firms, upon which the traditional international business theories are based. For example, China’s business environment is characterized by a high degree of; family ownership (Luo, Huang, & Wang, 2011), political involvement (Liu, Wang, & Zhang, 2013), state-owned enterprises (SOEs), and conglomerate management structures. And therefore, for example, the disadvantages that arise from institutional differences and the lack of efficient formal institutions within China are less detrimental for Chinese firms, since these special ownership structures can act as alternative governance structures (Xu & Wang, 1990). Therefore, the risks and costs associated with crossing institutional borders are potentially less present in the context of China.
Clearly, further research is needed to understand the cost and benefits of domestic borders in China, as well as domestic borders within other countries, and the relationship of these borders on firm performance. However, despite the remaining question marks, the empirical findings of this study prove the relevance of using finer geographic units as a proxy to examine the impact of geography on business activities. This is in line with the suggestion of the economic geography literature that intra-national variations can often be as relevant as cross-national differences (Fujita et al., 1999).
Limitations
This study is subject to several limitations that provide pathways for future research.
35
matter in explaining acquisition performance, we do not know if domestic borders matter in explaining acquisition performance in other contexts as well. Additionally, in this study we assumed that Chinese firms behave and respond in the same way as Western firms do, on which the main part of the international business literature is based. Accordingly, our hypotheses and empirical models are built upon the suggestions of the international business literature. Looking at our set of control variables, we find that some of them are in the opposite direction related to acquisition performance, than is suggested by the established acquisition literature. For example, the coefficients of prior performance and the method of
payment on acquisition performance are both negative and significant, while there is
demonstrated in prior studies that performance and cash-financed deals are positively related to acquisition performance (Bloom, 1999; Heron & Lie, 2002). Although the notable high adjusted R-squared of the base model, this observation raises the question whether our base model, which is based on findings of studies on western firms, is applicable to Chinese firms. Due to the lack of knowledge about the firm-level characteristics that predict acquisition performance of Chinese firms in particular, our model may have missed importance variables necessary to fully isolate the effect of domestic borders. Therefore, we call for future researchers, to continue to investigate Chinese acquisitions and firm-level characteristics that explain the heterogeneity in acquisition performance. Second, to address the question whether domestic borders matter in other contexts, future research using finer geographic units as a proxy is needed to investigate if different types of domestic borders impact firm performance in other developing, and developed, countries as well.
