The long-term performance of Chinese bidders after cross-border M&A

University of Amsterdam

Amsterdam Business School

MSc Finance (Quantitative Finance)

Master Thesis

The Long-term Performance of Chinese Bidders

after Cross-border M&A

Xu, Yumin 11830611

Email: yumin.xu@student.uva.nl Thesis supervisor: Vladimir Vladimirov

Statement of Originality

This document is written by Student Yumin Xu who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents

Acknowledgements

Thanks to my supervisor, professor Vladimir Vladimirov, gave me advice on the research direction. Thanks to my parents for supporting my decision on studying abroad. Thanks to all of my professors who enlighten my thinking about finance. Thanks to my friends who help me on data collection.

Abstract

There are no consistent conclusions on post-merger performance of acquirers in the literature, because of the failure to account for motivations. Started with the motivations of Chinese firms engaging in cross-border M&A, i.e. seeking high efficiency and developing R&D, this paper aims to explore the post-merger performance of Chinese acquirers in the long time. Difference-in-difference method and propensity score matching method are used in the paper to investigate the causality. Instead of using abnormal return, return on asset is used in the paper to measure the post-merger performance of acquirers. This paper finds that return on asset of Chinese acquirers has significantly decreased in the third year and it is robust in different industries and technology intensity sectors. Although, there is a positive effect on the improvement of efficiency and R&D investment of acquirers in the third year on average, the influence is heterogeneous among industries and technology intensity sectors.

1. Introduction ... 6

2. Literature Review ... 8

2.1 Determinants of cross-border M&A ... 8

2.2 Post-merger performance of acquiring firms ... 11

3. Methodology ... 14

3.1 General difference-in-difference methodology ... 14

3.2 Propensity score matching method ... 16

4. Data and sample construction ... 17

4.1 Data selection ... 17

4.2 Variable construction ... 18

4.3 Sample construction... 20

5. Difference-in-difference results ... 26

6. Robustness checks ... 30

6.1 Technology intensity sectors ... 30

6.2 Sample excluding domestic M&A ... 32

7. Conclusion ... 33

Reference list ... 36

Appendix ... 38

1. Variable definition ... 38

1. Introduction

Since Chinese government announced policy “go abroad” in 2001, an increasing number of Chinese firms have engaged in outward foreign direct investment (OFDI). In fact, China has become the second largest investing country in the world (UNCTD,2017). Cross-border M&A, as a mode of OFDI, is frequently adopted by Chinese firms. McKinsey company reports “Chinese firms spent $227 billion on acquiring foreign firms, six times what foreign firms acquired Chinese firms, and involved in ten of the largest deals around world in 2016.” Such huge volume of Chinese firms engaging in international transaction do make us curious about the motivations and the post-merger performance of these firms. Do these cross-border mergers really succeed? In fact, Child (2001) has pointed out that both target firms and acquirers face a great challenge on the post-merger stage and the probability of successful post-merger management is very low (KPMG, 2004). It is interesting to investigate the truth of these transaction in China. This paper will focus on what post-merger performance of Chinese bidders is.

In the paper, cross-border M&A refers to transactions that headquarters of acquirer and target firms are not in the same countries. Attentively, economic policies in Hong Kong, Macao and Taiwan are different from those in mainland in China, so deals conducted in these areas are also viewed as cross-border M&A.

Extensive research on cross-border M&A is conducted in the OFDI context (e.g. Andersen, 1997; Brouthers 2000; Harzing, 2002). Recently, scholars tend to carry on the research from the value-creating perspective where post-merger performance is measured relatively in longer time. In value-creating context, views on post-performance of firms are controversial. Morck (1991) believes that cross-border M&A helps both acquirers and target firms diversify risk, obtain synergy and benefit from internalization. However, other scholars e.g. Kaplan (1992), tend to believe only target firms can take benefits from such transactions, while Markides and Ittner (1994) hold a contrary view that bidders could create value for themselves. Shimizu et al (2004) thinks that conflicting results in the previous research might be the failure to account

for the different motives of each acquisition. Therefore, this paper will explore the post-merger performance of firms based on motivations of cross-border M&A.

Most empirical evidences are draw from developed countries, because cross-border M&A frequently takes place in these countries and substantial firm data are accessible. Recently, firms whose headquarters established in emerging markets increasingly participate in cross-border M&A, e.g. China, so findings drew from developed countries may not be applicable for these firms. Especially, firms in developing and developed countries have different motives to take part in cross-border M&A. With a great deal of cross-border M&A carrying on in emerging market, firm data are more accessible compared to former time, and it is realistic to conduct these research in emerging market now. Also, extensive research focuses on the post-merger performance of target firms rather than acquiring firms (Girma et al, 2004; Benfratello et al 2006; Bandick et al, 2010) and only few research focuses on post-merger performance of acquiring firms. To enrich the literature, this paper intends to figure out the effect of cross-border M&A on acquiring firms rather than target firms.

When scholars test whether the value of firms has increased after cross-border M&A, abnormal return or accounting indicators are frequently used to measure performance. Since the stock market is immature in China, stock price cannot correctly reflect the fundamental performance of a firm as developed market can. Also, sustainable value creating comes from improvement of firm fundamental performance. Therefore, true effect of cross-border merger on acquiring firms in China is more reasonably measured by accounting indicators in the long time. Enlighted by Stiebale and Trax (2011), who has conduct an empirical research to explore the acquirers’ domestic performance on firm level in France and United Kingdom, I will use difference-in-difference method and propensity score matching method to complete empirical investigation in this paper. The method is also used by Liu Yali (2016), Jiang Guanhong (2017) and Epstein (2007). Moreover, I will take a close look at the effect of cross-border M&A on post-merger performance of firms in different industries.

advantages to fill the gap between them and firms in developed countries. Therefore, they prefer to seek strategic assets (McKinsey, 2017), e.g management experience, brand, technology, innovations, etc. to compensate their disadvantages. Based on reverse spillover effect (Driffield and Love, 2003), bidders motivated by these reasons can result in the improvement of innovations and efficiency after cross-border M&A. Reverse spillover effect is usually used in the research on OFDI without distinguishing entry modes. However, there are a lot of differences between green field investment and cross-border M&A in terms of investment methods, theory and ownership (Bucklyey et al., 2008), therefore the effect of reverse spillover may be different. This paper will use return on asset, R&D investment and total factor productivity to explore whether Chinese bidders have significantly increased firm value by achieving the aims of efficiency and innovation improvement. There are three main contributions of this paper. Firstly, this paper adds a new evidence to literatures about the effect of the cross-border M&A on acquiring firms in emerging market. Secondly, this paper will try to answer the debate under value-creating context whether the wealth of acquiring firms is created after the transaction by inducing cross-border merger motivations. Thirdly, the paper provides an empirical example on cross-border M&A research in China, where many studies on it mainly focusing on individual case. The paper is divided into six parts next: part two is literature review of determinants of cross-border M&A and post-merger performance of bidders; part three is methodology, and difference-in-difference and propensity score matching method will be discussed carefully in this part. Part four and part five are data selection and results. Part six is robustness checks and part seven is the conclusion of this paper.

2. Literature Review

2.1 Determinants of cross-border M&A

Basically, there are four modes that a firm can choose when it decides to conduct OFDI, i.e. greenfield investments, cross-border M&A, exports, alliances and joint ventures (Dunning, 1981). Dunning had concluded that firms investing abroad aimed to acquire

natural resource, to gain access to new markets, to improve efficiency, or to obtain strategic assets. Derived from the theory, Galan (2007) distinguishes economic development in host country from that in the home country and makes the conclusion followed: (1) firms are encouraged by resource seeking when they acquire foreign firms incorporated in relatively high developed countries; (2) firms are motivated by society factors when they acquire firms whose headquarters are located in relatively less developed countries. Goh and Wong (2011) point that there is a positive relationship between the volume of OFDI and market capacity, capital liquidity in host countries. Hymer (1960) stresses out the motives of OFDI through monopolistic advantage theory. He believes the determinants of OFDI are the exclusive advantages of intangible assets (technology and management experience) and the scale of economics owed by one party. Similarly, Helpman and Yealpe (2004) also agree that firms with high productivity have very high possibility of engaging in OFDI and it is obvious when heterogeneity in firms is taken into consideration. Blomstrom and Kokko (1998) think that technology spillover can boost technology and efficiency of firms in host country. Specific to Chinese market, Laforet et al (2012) find that, in terms of ownership of firms, SOEs prefer to acquire firms located in countries where natural resource is abundant and policy risk is high, while private owned firms are more likely to gain access to new markets. Jiang (2012) uses gravity mode of trade to investigate how acquirers in China choose investing countries. Slightly different from findings of Galan, the scholar finds that Chinese firms acquire firms for natural resource and new market in developing countries and for strategic assets in developed countries. Except these motivations, many scholars believe that policy factor plays an important role in the international transactions for Chinese firms. Since Chinese government uses OFDI as a channel of diplomatic strategy across countries, the volume of OFDI has shapely grew (Song and Zhang, 2011). Kafouros and Wright (2012) find empirical evidences that the investment scale and the choose of host counties are influenced when Chinese government engaging in OFDI.

The motivations of cross-border M&A are similar to those of OFDI, since it is one of OFDI methods. However, the importance of the motivations is different. Compared

transaction form frequently used in OFDI: (1) cross-border M&A provide bidders higher control right over assets relative to alliance; (2) relative to greenfield investment (building own establishments), cross-border M&A ensures acquirers to enter foreign market with less transaction cost (Newburry et al. 1997). These advantages of cross-border M&A encourage firms to use this mode in OFDI.

Wemerfelt (1984) puts forward that a firm is featured by various endowment of complementary intangible assets and “capability”, which can be viewed as strategic assets nowadays. These strategic assets are the main motivations that firms are willing to participate in cross-border M&A. Nocke and Yeaple (2006) argue that firms taking part in cross-border M&A are driven by gaining international mobile and non-mobile assets. They indicate that bidders acquiring foreign firms want to obtain mobile assets, i.e. advanced R&D, management experience and skilled employees, while Delio (2002) believes that firms are attracted by non-mobile assets, which are country-specific resource, i.e. marketing network, brand, customs and mature technology. Especially when developing new technology is difficult for acquirers at home country, these firms prefer to acquire foreign firms in industries where R&D is high (Delio et al. 1999). In fact, technology is the main reason driving firms engaging in cross-border M&A (Frey et al. 2006), compared firms which take domestic M&A. Moreover, emerging-market firms as latecomers tend to acquire abroad, mainly encouraged by the pursuit of strategic resource (Cogman et al. 2015; Erel et al. 2012), and China is no exception. Unlike the argument Helpmen (2004) has proposed, Chung (2002) thinks that firms in lower-tech industries are more likely to invest in firms which own advanced tech. And this activity can bring reverse spillovers to acquirers (Driffield and Love, 2003). It is much reasonable for firms in emerging market, because these firms can quickly compensate the competitive weakness and fill capacity gaps, compared with developed-market firms (Luo and Tung, 2007). To achieve the aim, Chinese firms would like to establish R&D research centers in developed countries which own advance technology, knowledge-intensive products to improve their innovation ability and boost firms’ efficiency (Bonaglia et al.,2007). Rugman (2007) also points out that Chinese firms would like to acquire firms with country-specific advantages, i.e. brand, mature technology, management, which are shortcomings in Chinese firms. In all, Chinese

firms are mainly motivated by obtaining strategy asset, namely R&D expansion and high efficiency, to initiate a cross-border M&A, aiming to benefit from reverse spillover and to improve their weaknesses.

If Chinese firms acquire firms in developing countries for non-mobile assets, they benefit from reverse spillover in two aspects. First, bidders can get close to local research centers and learn customers’ preference of products, which is known as proximity-concentration trade-off (Horstmann et al., 1992; Markusen et al., 2000; Helpman et al., 2004). The information is transferred to parent firms and ultimately promotes synergy. Second, acquiring firms can achieve economies of scales and greatly reduce product cost, resulting in improvement of efficiency (Kleinert et al., 2007). Also, Chinese firms can acquire firms to get mobile assets in countries with high technology intensity. According to the life cycle of products, products experience new, maturing, standardized and declined stage and the first three stages usually take place in these countries. Unlike the former acquiring situation, firms can obtain innovations, product trend and the newest customer demand under this acquisition, and it is beneficial for firms’ R&D investment. Additionally, firms can rapidly improve innovations by establishing R&D centers with acquired firms in foreign countries or using the advanced technology of targets (Pradhan et al., 2009; Buckley et al., 2008). By analyzing the literature of determinants of cross-border M&A, I propose the hypothesis in the paper:

Hypothesis 1: The cross-border M&A has a positive effect on efficiency of acquiring

firms in the long time in China.

Hypothesis 2: The cross-border M&A has a positive effect on R&D expenditure of

acquiring firms in the long time in China

2.2 Post-merger performance of acquiring firms

Cross-border M&A is significantly different from domestic M&A. Except risks involved in domestic M&A, firms engaging in cross-border M&A face more risks, e.g. economic risk, regulatory risk, cultural structures risk (Hofstede, 1980; House et al., 2002), liability of foreignness(Zaheer,1995), and double-layered acculturation (Barkema et al.,1996). Therefore, the post-merger performance of acquirers in

cross-border M&A is more uncertain than in domestic M&A. Actually, there are no unified answers to whether firms engaging in cross-border merger can increase their wealth. In the context of OFDI, scholars tend to study the effect from the change of employment, wage, product efficiency and output after firms invest in foreign countries. Barba and Castellani (2004), using Italian data, conclude that parent firms can increase their outputs, employment and wealth after OFDI in 1997. Stiebale (2011) finds that cross-border M&A boosts on average acquirers' domestic sales, investment and capital in bidders. Similarly, Debaera (2010) also finds that employment of acquiring firms goes up when Korean firms acquire firms located in less developed countries. Chen (2012) takes a close investigation on the investment of emerging-market firms in developed market, confirming the existence of reserve spillover between the two markets. In the context of M&A, empirical research focuses on the change of stock price. As suggested by value-creating theory, these research cares about whether shareholders earn positive abnormal return after completion of cross-border M&A. Scholars (Buckely, 1976; Morck, 1992; Wilson, 1980) believe that firms can earn positive return after completion of cross-border merger, if firms participate in such a transaction for seeking market advantages. Also, similar results are found in US when Markides (1994) investigates the post-merger performance of acquirers. Kang (1993) provides the empirical evidence that both bidders and target firms gain significantly positive abnormal return when Japanese firms acquire US firms.

However, some scholars show us opposite evidences. Starting with the performance of domestic M&A, extensive empirical research suggests that domestic M&A has negative effect on post-merger performance of firms. Coontz (2004) says companies fail to perform well after M&A in all parameters understudy. In the long time, acquirers significantly underperform over the three-year post-event period and cross-border deals perform poorly in the long run (André et al., 2004). Taking transaction cost into consideration, Datta (1995) thinks that acquiring firms cannot gain positive abnormal return from the completed transactions. Similarly, Li and Guisinger (1991) point out that cross-border M&A is likely to fail if firms cannot deal with inherent risks of the transaction and cannot make good integration after completion of cross-border M&A.

Especially, when scholars use event study to test whether bidders earn abnormal return after cross-border M&A, many empirical evidences indicate that acquiring firms cannot make it (Jarrell and Gregg, 1987; Frank et al., 1989). Moreover, Ravenscraf (1989) believes that cross-border M&A indeed cause underperformance of acquirers.

Literatures about post-merger performance of acquiring firms are different. It may be caused by the neglect of different motivations of cross-border M&A (Seth et al.,2002). Instead of concentrating on the final performance of acquiring firms, it is more reasonable for us to investigate how different motives of cross-border M&A can affect the post-performance of acquirers. Combined with literatures of the influence of OFDI on investing firms, Bertrand and Zuniga (2006) carry on research about the influence of M&A, including cross-border and domestic deals, on R&D expenditure, here are conclusions: (1) there is little effect on R&D expenditure; (2) the effect of cross-border M&A and domestic M&A are different. Stiebale (2013) finds that the R&D investment has been improved in Germany.

Admittedly, firms located in developed countries are often motived by increasing market shares in host countries and raise profits of firms. Nevertheless, as suggested in first part of literature review, the motivation of Chinese bidders engaging in cross-border M&A is not to increase market shares but to obtain strategic asset for the purpose of innovations and efficiency, so cross-border M&A will have little influence on market shares of bidders. On the other words, cross-border M&A may have little influence on the firm profits. Instead of using abnormal return of stock price, which only accounts for the interest of shareholders, this paper uses the return on asset to measure firm profit. Since Chinese firms aim to improve R&D investment after the completion of cross-border M&A, these firm will increase the expenditure on R&D and transaction cost will increase, because they need to apply the technology of target. Based the analysis above, the third hypothesis is:

Hypothesis 3: The cross-border M&A has a negative effect on return on asset (ROA)

3. Methodology

Difference-in-difference methodology and propensity score matching method will be used in this paper. As the requirement of application of difference-in-difference method, firm-year data will be used in this paper to explore the causality relationship between the cross-border M&A and long-term performance of Chinese bidders.

3.1 General difference-in-difference methodology

Difference-in-difference (DID) method is universally used in social area to test the effect of a policy. In 1978, Ashenfelter introduced DID to estimate how training program affected the earnings of employees, since then this method has been widely used in economic area. Bertrand (2004) showed us that there have been 92 papers published in six economic journals from 1990 to 2000. However, the first application of DID in China happened in 2005, which was used by Zhou Lian and Chen Ye. The basic intuition of DID method is to evaluate the effect of treatment on experimental subjects. Namely, how the performance of Chines firm i has changed s years later it engaging in cross-border M&A at time t:

𝐷𝐷_",$%&=E(𝑦_",$%& 𝐷 = 1 − 𝐸(𝑦_",$%& 𝐷 = 0 (1)

For the equation (1) above, 𝑦_",$%& represents the performance of firm i after s years of

the completion of cross-border M&A. D stands for the truth whether firm i has engaged in cross-border M&A or not, and it takes value one if it is treated otherwise takes value

zero. 𝐷𝐷_",$%& represents the change of performance of Chinese firm i after taking

cross-border merger s years later, compared to the performance of itself if it had not engaged in the activity. However, it is impossible to observe the performance of a firm not involved in cross-border merger once it has taken the activity. Previously, “before-after” comparison is adopted to measure the effect. Namely, the difference between the performance of a firm before and after taking cross-border M&A is used as an estimator of the effect of treatment:

𝐷𝐷_",$%&=E(𝑦_",$%& 𝐷 = 1 − 𝐸(𝑦_",$01 𝐷 = 1 (2)

The obviously advantage of “before and after” method is that it is easy for scholars to do research. However, during the experiment period, some factors which are not caused by treatment can influence firm performance. If it happens, the estimator of result will

be biased. To solve the problem, some scholars propose another method which is to find a replacement of the counterfactual situation, that is to find a comparable firm whose performance can replace the unobservable performance of the firm engaging in cross-border merger if it was not. The difference between performance of the firm involved in border merger and that of the similar firm not involved in cross-border merger is regarded as the estimator of effect of cross-cross-border M&A on firm performance:

𝐷𝐷_",$%&=E(𝑦_",$%&$234$35 _{𝐷 = 1 − 𝐸(𝑦}

",$%&67$234$35 𝐷 = 1 (3) However, there is an important assumption that the two firms are indifferent before one of the them takes cross-border merger. If the assumption is hold, equation (3) can translate into equation (1).

DID method solves the problem by evaluating the change of the treated subjects under the counterfactual context. Firms are naturally allocated into two groups: firms acquiring foreign firms are allocated into treatment group while firms not acquiring foreign firms are assigned to control group. The difference between the treatment group and control group “before and after” difference is the unbiased estimator of effect of cross-border M&A on Chinese bidders:

𝐷𝐷_",$%&=E(𝑦_",$%&$234$35_{− 𝑦}

",$01$234$35 𝐷 = 1 − 𝐸(𝑦",$%&67$234$35 − 𝑦",$0167$234$35 𝐷 = 0 (4)

Specifically, to estimate the 𝐷𝐷_",$%&，regression below is used:

𝑦_"$ = 𝛽₉+ 𝛽₁𝐺_"𝑇_$+ 𝜆_"+ 𝜈_$+ 𝑋 + 𝜖_"$ (5)

In equation (5), 𝑦_"$ is the dependent variable that measures the performance of

Chinese bidders. In this paper, long-term performance of bidders after the completion

of cross-border M&A will be measured by ROA. 𝐺_" is a dummy variable which takes

value one if firm i initiates a cross-border M&A and takes value zero if it is not take

the transaction. 𝑇$ is also a dummy variable which takes value one if year t is the year

after firms taking cross-border M&A and takes value zero if year t is year that fims have not taken cross-border M&A yet. Many variables will be used as control variables. 𝜆_" and 𝜈_$ is the industry and time fixed effect, respectively. They are used to control

unobservable factors which bias estimator 𝛽₁. X is a vector to control firm characters.

In equation (5), cross term 𝐺_"𝑇_$ stands for firm i at year t after taking cross-border

negative, it means that cross-border M&A decrease the value of acquirer based on other condition are unchanged. This is exactly what expect results based on the hypothesis 3. The result will be biased if the two groups are significantly different before treatment and there is great possibility exist in natural trial. Only if firms are randomly allocated into two groups, the decision of a firm makes to engage in cross-border merger is exogenous. Thereafter, the causality is held. To avoid the selection bias, propensity score matching method (PSM) can be use. PSM method ensures that firms in the two

group are similar before they take cross-border M&A. Only then can 𝛽₁ be viewed as

the true effect of cross-border M&A on bidders.

3.2 Propensity score matching method

PSM method is proposed firstly by Paul Rosenbaum and Donald Rubin in 1983. As mentioned above, PSM method solves the problem under counterfactual situation that counterfactual outcome is not observable. It is able to match firms in treatment group with firms in control group. On other words, firms in the both groups are similar to each other respect to features of firms, e.g. firm size, industry, employees, etc. The performance of firms in control group can be used as proper benchmark to evaluate the effect of treatment, i.e. PSM method ensures that the effect of treatment is concluded based on the comparable group. There are two assumptions which ensure the successful matching: (1) conditional independence. Whether a firm takes cross-border M&A or not is randomly, given the covariates X controlling firm features, i.e. the selection into treatment group is not influenced by unobservable factors except firms feature, the same with outcomes. Therefore, the probability of a firm taking cross-border merger is conditional independent on X:

P(x) = Pr [D=1|X=x] (6)

P(x) ⊥ D | X (7)

(2) common support condition. For each variable x in vector X, the probability of x existing in both treatment group and control group is positive:

0< Pr[D=1|X=x]<1 (8) If these two assumptions are held, then the matching is successful. In equation (6) and (7), P(x) is the propensity score. Firms with the same P(x) have the similar

characteristics. Also, once assumption (1) has held, the potential outcome is independent under the propensity score.

To obtain the comparable treatment group and control group, first, I use logit regression to obtain propensity score of a firm taking cross-border M&A:

𝑡𝑟𝑒𝑎𝑡_"$ = 𝛼_"$X_"$ + 𝛾_J+ 𝛿_" + 𝜀_"$ (9) In equation (9), treat represents the result if a firm completed at least one cross-border deal over 2007-2014. It takes value one if it did complete one transaction otherwise it

takes value zero. X_"$ is a vector of variables which capture characteristics of firms.

𝛾_$ 𝑎𝑛𝑑 𝛿_" is the year and industry fixed effect to control the unobservable factors.

Second, matching the propensity score. I choose one treated firm matched by three untreated firms because sample is screwed toward to control group. Third, balancing test. The purpose of this step is to ensure the firms after matching in both groups are significantly indifferent before they are selected into groups. Mean covariates t-test is used to detect the difference.

4. Data and sample construction

4.1 Data selection

The research subjects are firms that have completed cross-border M&A between 2007 and 2014. On the one hand, accounting standards used in China have changed in 2007. Selecting firm data after 2006 ensures accounting data consistent through the period and avoids substantially missing data, e.g. total wage and intangible asset were unavailable in financial reports before 2007. On the other hand, cross-border M&A are not active before 2007. Dunning (1981) points out that firms could achieve the maximum profit three years later of investment. Therefore, to explore the long-term performance, this paper will choose the performance of bidders in China in the third year after firms have completed cross-border M&A.

Data used in the paper mainly comes from two databases: Thomson One and China Stock Market and Accounting Research Database (CSMAR). Thomson One database provides cross-border M&A information of firms. In Thomson One, merge information,

i.e. acquirer nation, target nation, effective date, firm status, type of firms, can be acquired. Subjects of this research are Chinese bidders, so deals that acquirer nation is China while target nation not are selected in the research. As the definition in Thomson One, “China” only refers to mainland China while Hong Kong, Macao, Taiwan are excluded. Whether firms engage in cross-border M&A is suggested by the “Cross-Border” item in the database. According to the item, I can identify public firms that have engaged in cross-border M&A. The completed cross-border M&A are suggested by the effective date. From CSMAR, balance sheets and income statements of public firms from 2007 to 2017 can be acquired. Recent years, a number of Chinese firms are listed abroad, for these firms, balance sheets and income statements are reported by foreign currency. Before combining financial data of these firms with firms listed in China mainland, these financial statements should be reported in Chinese currency by translation. Since the established date of most firms is missing in both databases, to keep comprehensive data, I get the established date by hand collection. After all these processes have been done, I combine Thomson One and CSMAR together. I also refine the sample by excluding: (i) financial firms, utilities and International Organization whose two-digit China National Industrial Classification of 66-69,76-79, 91-97; (ii) firms with abundantly missing data on sales, PP&E, total wages and employees;(iii) firms are defined as government, joint venture, subsidies in Thomson One. Also, if a firm has completed more than one cross-border merge deals over this period, only the latest deal is included in sample.

Imposing all the filter above, 124 public firms which completed cross-border M&A from 2007 to 2014 are included in sample. Together with firms have not engaged in cross-border M&A among the period, there are 24,434 firm-year observations in the original sample.

4.2 Variable construction

This paper uses data in financial reports to describe the long-term performance of a firm and many variables are used to measure features of a firm. When scholars need to measure the value of a firm, ROA is a usually better measurement of fundamental performance of a firm than ROE (Hagel et al, 2010); innovation is usually measured by

R&D expenditure and a proxy for R&D expenditure is intangible asset ratio. Total factor productivity (TFP) can measure the efficiency of a firm, which is better than labor efficiency. What we need to pay attention to is the calculation of TFP. The traditional method of TFP calculation is Solow residual method. However, this method neglects endogeneity: when inputs affect the level of the outputs, the output also influences inputs in a firm. In this thesis, I adopt the LP method (Levinsohn and Petrin, 2003), which uses the material as semi-parameter to avoid the simultaneity problem. The first step of PSM method is logit regression. In this regression, variables capture characteristics of firms are used to predict the propensity of a firm engaging in cross-border M&A. I include the TPF to control the productivity level of firms to avoid that only high efficiency firms are included in treatment group, which will ultimately upward bias. At the same time, it can test whether a firm taking cross-border M&A is motivated by efficiency seeking. Since few firms reported research and development (R&D) expenditure in China, to avoid the loss of massive observations, I use intangible asset ratio, measured by the ratio of intangible asset to total asset, as a proxy of R&D expenditure to control the firm growth. Similar to TFP, the coefficient of intangible asset ratio indicates whether a firm involving cross-border M&A is encouraged by improving R&D expenditure. The growth of sales is used to control the domestic growth. The log of PP&E is use to capture firm size. The log of average wage is used to measure the various skills of labors. The log of capital stock is used to measure the production process and control for the influence of multinational firms with higher capital intensity. The log of age reflects the growth potential and experience. I use the working capital ratio to control the ability of a firm completing a cross-border M&A without concern of financial liquidity. Table 1 describes these variables.

Dummy variables are included to capture firm other characteristics. SOE is used to capture the status of firms whether they are state owned enterprises or not. It takes value one if acquirers are SOEs and takes value zero if they are not. Past M&A, including both cross-border and domestic deals, experience of firms is captured by experience dummy, which can influence the possibility and efficiency of firms taking cross-border M&A (Jamie et al, 2009). It takes value one if firms have at least taken a merger and

otherwise take value zero. Industry fix effect is used to control the different level of technology and industrial features. Year fixed effect is used to control unobserved shocks, e.g. policy changing, to the performance of firms along the years.

TABLE1.

Summary statistic

This table is descriptive statistics of key variables used in propensity score matching method. Variable “treat” comes from Thomson One and other variables come from CSMAR. Column (2) is the mean of each variable in the sample. Column (3) is the standard deviation of the variable, which shows the level of discretion of observations. Column (4) and column (5) is the minimum and maximum value of each variable in the sample, respectively. The time span in sample is from 2007 to 2017and the whole firm-year observation is 24,434.

Variables Mean Standard

deviation Minimum maximum (1) (2) (3) （4） treat 0.2527 0.1569 0 1 TFP 14.5149 1.5881 1.0692 19.5191 sales 20.4967 1.8729 6.6846 27.6998 PP&E 20.7816 1.422 11.7207 27.5985

intangible asset ratio -3.2273 1.3945 -13.999 2.5253

capital 19.8439 1.0299 17.0736 25.9329

wage 10.89 1.4585 -2.6615 18.3915

working capital ratio 0.1806 0.4313 -13.60755 0.9717

age 2.4626 0.5042 0 3.7842

SOE 0.4169 0.4931 0 1

4.3 Sample construction

PSM method, which excludes the selection bias before firms selected into groups, is used here to construct firm sample. Results are showed in TABLE 2. From the table, TFP and R&D expenditure have significantly negative impact on the propensity of a firm engaging in cross-border M&A. It suggests that a firm with lower TFP is more

likely to initiate a cross-border M&A relative firms not engaging in the transaction in the same year and in the same industry, given other conditions are unchanged. The same with R&D expenditure. Firms with lower R&D expenditure are also more likely to acquire foreign firms under the same situation. This result confirms the conclusion of Chung (2002) that firms with low tech are more likely to engage in cross-border M&A and proves the analysis of the determinants of Chinese firms engaging cross-border M&A in literature review. Suggested by SOE, stated owned enterprises plays less important role in border merger transaction. In fact, less SOE initiating cross-border M&A during this period compared to non-stated owned enterprises, given other condition constant. Growth of firms themselves have positive influence on the decision that firms take part in cross-border M&A. The evidence can be found in the growth of sales, growth of PP&E, working capital ratio and growth of the wage. Also, prior domestic and international merger experience increases the likelihood of completing merger in foreign countries (Jamie et al, 2009).

After the first step of PSM method, propensity score is used to match firms and the matched firms should have the similar features before any firm takes a cross-border merger. TABLE 3 gives the results of the balancing test. Proved by p-value, original treatment group and control group are significantly different before matching. However, this bias is substantially reduced and both groups are similar after matching. GRAPH 1 shows it directly.

TABLE 2.

Logistic regression.

This table shows the results of logistic regression, which is the first step of propensity score matching method. In this regression, dependent variable is Treat, which takes value one if a firm has completed at least one cross-border merge over the period. TFP is a measurement of firm efficiency and it is calculated by L&P method. Intangible asset ratio is firm intangible asset to total asset. Working capital ratio is measured by net working capital to total asset. Sales, PP&E, capital, wage and age are all in log form. SOE, exporter and experience are dummy variables. In the regression, industry and year fixed effect are used. The standard errors in the regression are robust and showed in the parentheses. ***, **and * denotes significance at 1%,

Independent Variable Treat TFP -0.319*** (0.115) Sales 0.330*** (0.108) PP&E 0.480*** (0.0972)

Intangible asset ratio -0.0948**

(0.0432)

Capital -0.0748

(0.0996)

Wage 0.157***

(0.0420)

Working capital ratio 1.648***

(0.262) Age 1.132*** (0.136) SOE -1.319*** (0.131) Exporter 0.439*** (0.170) Experience 0.220* (0.126) Industry Yes Year Yes TABLE 3. Balancing test

This table is the result of balancing test of propensity score matching. Column (1) is type of sample. U means the sample is unmatched, while M means the sample has been matched. Column (2) and (3) is the treatment group and control group, respectively. Column (4) and (6) is bias reduced. Column (6) and (7) is the result of mean t-test, represented by t-value and p-value, respectively. From column (7), p-value of matching sample is all bigger than 0.3, which

means there is no significant difference between firms in treatment group and control group before they are selected into two groups.

Sample Treated Control % bias Bias

reduced T-test P-value (1) (2) (3) (4) (5) (6) (7) TFP U 15.01 14.666 26.1 5.13 0.000 M 14.972 15.004 -2.4 90.70 -0.42 0.677 Sales U 21.198 20.616 34.3 7 0.000 M 21.116 21.155 -2.3 93.30 -0.37 0.713 PP&E U 20.945 20.09 22.6 5.13 0.000 M 20.87 20.914 -2.9 87.00 -0.44 0.664

Intangible asset ratio U -3.4233 -3.2734 -11 -2.29 0.000

M -3.4322 -3.3672 -4.8 56.60 -0.75 0.451

Capital U 19.971 19.713 23.5 5.5 0.000

M 19.909 19.917 -0.7 97.00 -0.1 0.916

Wage U 11.151 10.902 15.6 3.58 0.000

M 11.15 11.173 -1.5 90.60 -0.22 0.823

Working capital ratio U 0.25418 0.18413 17.5 3.1 0.002

M 0.25597 0.23906 4.2 75.90 0.95 0.334 Age U 2.6092 2.4763 29.4 5.81 0.000 M 2.6069 2.6025 1 96.70 0.17 0.864 SOE U 0.27754 0.42011 -30.2 -6.18 0.000 M 0.27922 0.2684 2.3 92.40 0.37 0.713 Exporter U 0.11653 0.05512 22 5.36 0.000 M 0.10173 0.09307 3.1 85.90 0.44 0.658 Experience U 0.22034 0.1682 13.2 2.96 0.003 M 0.21645 0.22294 -1.6 87.50 -0.24 0.812 Pscore U 0.10624 0.03518 82.7 30.68 0.000 M 0.09553 0.0956 -0.1 99.90 -0.01 0.990

GRAPH 1.

Propensity score comparison before and after matching

These two graphs are the comparison of probability of firms in treatment and control group taking cross-border merger between matched sample and unmatched sample. Graph above is the probability distribution of firms in two group taking cross-border merger in unmatched sample. Graph below is the probability distribution of firms in two group taking cross-border merger in matched sample.

0 10 20 30 kd e n si ty _ p sco re 0 .2 .4 .6 .8 pscore treat control

before matching

After matching firms in treatment group with firms in control group, 117 firms which engaged in cross-border M&A among the period matched successfully. I use 1:3 to match firms, resulting in 468 firms and 4106 firm-year observation after dropping missing data are included in sample in the end.

TABLE 4 is the description of matching sample. From the industry distribution of bidders, more than 70% firms initiating cross-border M&A belong to manufacture industry. This propensity suggests that bidders tend to seek for strategic asset, e.g. technology, efficiency and capital intensity.

TABLE 4.

Industry distribution of bidders.

This table shows industry distribution of bidders in China from 2007 to 2014. Column (1) is the absolute number of firm at least have completed a cross-border merger over the period and column (2) is the percentage of industry distribution of bidders. In all, 117 firms have completed cross-border M&A. 0 2 4 6 8 kd e n si ty _ p sco re 0 .2 .4 .6 .8 pscore treat control

after matching

Industry Absolute number % number (1) (2) Agriculture/Fishing/Forestry 2 1.71% Mining 4 3.42% Manufacture 88 75.21% Construction 2 1.71% Transportation 3 2.56% Telecom 2 1.71% Restaurant 11 9.40%

Real Estate Development 3 2.56%

Entertainment/Leisure 2 1.71%

Total 117 100%

5. Difference-in-difference results

After successfully matching firms which have completed cross-border M&A with firms not engaging in international mergers, these firms selected into two groups are similar. Therefore, the matching procedure can reduce a substantial amount of biases which result from differences in observed covariates. DID can be directly used with fixed effects to explore the effect of cross-border M&A on acquirers, since firms are already similar after PSM. This paper focuses on the long-term performance of Chinese bidder, which is measured in the third year. It also explores the persistence of effect, so performance in the first and second year are also considered. TABLE 5 is the main results of the regression.

As showed in TABLE 5, if firms are motivated by strategic assets to initiate an international M&A, this transaction dose increase TFP of the firm. The results confirm hypothesis 1. It is also similar to the findings of Stiebale (2013). Taking a look at the second row of TABLE 5, the effect is significant at 5% level for the first two years and less significant at 10% in the third year. Compared to firms not engaging in cross-border merger, TFP of bidders goes up about 20% in the third year when control the industry and year. The decreasing level of significance can be explained from two aspects: (1)

marginal decreasing effect of acquired knowledge. In the initial stage after cross-border M&A, acquirers can learn the whole new knowledge from target firms. However, the amount of learnable knowledge decreases with acquirer gradually absorbing knowledge from targets, i.e. once a firm acquires advanced technology and management experience, marginal effect of the knowledge that can be acquired reduces. (2) marginal decreasing effect of brand. At the beginning of completion of cross-border M&A, bidders can make use of brand to increase TFP, but this effect also gradually reduces with the exploration of the target brand. Since the marginal diminishing effect on TFP, firms have to invest more on R&D to sustain the situation or to exaggerate the margin of knowledge of target firms.

From the third row of TABLE 5, the growth of R&D investment significantly increases after firms have completed cross-border M&A, compared to firms not in the same year and in the same country. This result supports hypothesis 2. In the third year, the growth of R&D goes up about 20% for firms taking international merger relative to firms not taking the transaction in the same industry and in the same year. The results in the first two year are more significant than that in the third year and the growth rate roughly decreases in the next two years. R&D investment growth is consistence with the growth of TFP. Jovanovic (2008) finds hat R&D investment is encouraged by efficiency increase. Also, if firms intend to boost TFP in the long time, they also need to raise the expenditure of R&D.

ROA is used to measure the fundamental performance of a firm. Hypothesis 3 cannot be rejected in this paper, which means that acquirers experience the loss of value after the completion of cross-border M&A. Based on the first row of TABLE 5, in the third year, there is a significantly negative effect of cross-border M&A on the profitability of a firm. Specifically, average ROA decreases by about 1.4 percentage point for firms taking international merger relative to firms that are in the same industry and in the same year but not taking cross-border merger. Looking backward, the results suggest that the profitability of the firm becomes worse and worse with the end of transaction. This finding is consistent with Meeks (1977) and Ghosh (2001), whose findings show that post-merger ROA continually declines. The reason is that Chinese firms engaging

ability. Therefore, these firms spend more earnings on R&D expenditure to achieve their aims, which cause the decrease on ROA when firms complete transaction. Moreover, as Morgan Stanley reported in 2017 that increasing number of Chinese firms have been engaging in international merger in recent ten years, but post-merger performance is bad because of poor integration. Cultural distance, different management style and inefficient communication cause poor post-merger integration. Increasing management cost and transaction cost also cause the decrease of ROA for acquirers in the long time.

Considering the different motivations of initiating cross-border M&A in China, these changes may be not the same for different industries. As showed in TABLE 4, the majority of firms initiating cross-border M&A are manufactural firms. Only a small part of samples coming from service industry. Since manufacture and service industries have different production, operation and profit method, the influence of cross-border M&A on them may be different. Therefore, according to the classification of industry form by National Bureau of Statistics of China, the sample is divided into two parts: manufacture industry and service industry. The results are reported in TABLE 6. In both industries, ROA is significantly negative affected in the third year, and the negative influence is worse in the service industry. As for TFP, cross-border M&A have significantly positive effect for firms in manufactural industry while it is not significant for service industry. On the contrary, cross-border transactions have a significantly positive effect on R&D expenditure in the service industry while it has negative but not significant influence in the manufacture industry. It makes sense. In the manufacture industry, e.g. automobile industry, acquiring firms can use the advanced technology and production line with high efficiency in target firms to improve themselves efficiency when transactions are completed. For these firms, they prefer to use the mature knowledge and production lines in target firms to improve their efficiency instead of investing more on R&D to develop innovations. And the effect persists in the long time as suggested in TABLE 6. In the service industry, e.g. software development sector, innovations are very important to improve profitability. Therefore, they would like to spend more on R&D investment. It is supported by the findings in TABLE 6. Compared firms in the service industry without taking cross-border M&A,

firms completed the transaction have significantly increased R&D expenditure by 75 percentage point in the third year.

TABLE 5

Average effect of cross-border M&A on performance of Chinese bidders.

This table shows how cross-border M&A influences the performance of acquirers. Performance is measured by ROA, R&D expenditure and TFP based on profitability, research ability and

efficiency, respectively. Column (1) is the average treated effect. 𝑦$%& stands for performance

of a firm s years after completing cross-border merger, while 𝑦_$01 stands for performance of

a firm one year before it completes cross-border M&A. Column (2) is the performance of bidders in the first year after it completed cross-border merger. Column (3) and (4) stands for post-performance of bidders in the second year and in the third year, respectively. The standard errors are showed in the parentheses. Industry and year fixed effect are used in this regression. ***, **and * denotes significance at 1%, 5% and 10%, respectively.

𝑦$%O− 𝑦$01 First year Second year Third year

(1) (2) （3） （4） ROA -0.009** -0.013*** -0.014*** (0.004) (0.004) (0.005) TFP 0.201** 0.210** 0.202* (0.078) (0.092) (0.116) Ln(IA) 0.213*** 0.195*** 0.197** (0.0632) (0.0701) (0.0817)

Industry yes yes yes

Year yes yes yes

N on support 3958 3476 2997

TABLE 6

Heterogeneous effect of cross-border M&A on manufacture and service industry.

This table shows the average effect of cross-border M&A is heterogeneous in manufacture and service industry. Column (1) – (3) is the average effect of cross-border M&A on ROA, TFP and R&D expenditure of manufactural firms in the first year, the second year and the third year,

respectively. Column (4) – (6) is the average effect of cross-border M&A on ROA, TFP, R&D expenditure of firms in service industry in the first year, the second year and the third year, respectively. The standard errors in the regression are showed in the parentheses. Industry and year fixed effect are used in this regression. ***, **and * denotes significance at 1%, 5% and 10%, respectively.

Manufacture Industry Service Industry

First year Second Year Third Year First year Second Year Third Year 𝑦_$%& − 𝑦_$01 (1) (2) (3) (4) （5） （6） ROA -0.00371 -0.00783* -0.0108** -0.0224*** -0.0272*** -0.0323*** (0.0040) (0.0045) (0.0054) (0.0079) (0.0091) (0.0117) TFP 0.178** 0.203** 0.223* 0.252 0.191 0.15 (0.0859) (0.1030) (0.1320) (0.1910) (0.2270) (0.2980) Ln(IA) 0.0592 -0.0143 -0.069 0.831*** 0.839*** 0.754*** (0.06160) (0.06940) (0.08270) (0.19900) (0.22600) (0.26400)

Industry yes yes yes yes yes yes

Year yes yes yes yes yes yes

Observation 3,179 2,839 2,462 746 637 528

6. Robustness checks

6.1 Technology intensity sectors

The results above indicate that firms with lower efficiency and R&D expenditure are more likely to take part in cross-border M&A and it has positive effect on R&D expenditure and TFP overall. But when test the effect among manufacture and service industry, the effect is different according to the characteristic of industries. It is also necessary to check if the effect on R&D expenditure and TFP holds in different technology intensity sectors. According to the classification of high technology set by

China National Bureau of Statistics, the sample is divided into two subgroups only including firms in manufacture and service industries. TABLE 7 is the results of the average treated effect of cross-border M&A on ROA, TFP and R&D expenditures in the high technology intensity sector and in the low technology intensity sector. In the high technology intensity sector, in the third year, cross-border M&A has significant effect on the performance of bidders. Specifically, Chinese bidders experience a significant decrease on ROA after completion of the international M&A in the third year relative to firms not engaging in the transaction in the same industry. As for the R&D expenditure and TFP, the effect is opposite: acquiring firms significantly increase TFP and R&D expenditure by 31.6% and 21.5%, respectively, compared to similar firms in the third year and in the same industry. Therefore, the conclusion is held in high technology intensity sectors. In low technology intensity sector, whose results are represented in column (4) - (6) in TABLE 7, these conclusions are held in the third year except that R&D investment loses significance. The reason why the increase of R&D expenditure is not significant is that firms with low demand of technology can benefit most from technology of target firms when cross-border M&A have completed. Therefore, these firms do not need to spend much on R&D expenditure to explore complicated and advanced innovations as high technology intensity sector dose.

TABLE 7

Heterogeneous effect of cross-border M&A on different technology-intensity sectors.

This table shows the various effect of cross-border M&A on the high technology intensity sector and low technology intensity sectors. Column (1) – (3) is average effect of cross-border M&A on ROA, TFP and R&D expenditure of high technology intensity sector in the first year, the second year and the third year, respectively. Column (4) – (6) is average effect of cross-border M&A on ROA, TFP, R&D expenditure of firms in low technology intensity sector in the first year, the second year and the third year, respectively. The standard errors in the regression are showed in the parentheses. Industry and year fixed effect are used in this regression. ***, **and * denotes significance at 1%, 5% and 10%, respectively.

High Technology Intensity Low Technology Intensity First year Second Year Third Year First year Second Year Third Year 𝑦_$%&− 𝑦_$01 (1) (2) (3) (4) (5) (6) ROA -0.01 -0.0120* -0.0164* -0.01 -0.0122** -0.0136** (0.0057) (0.0067) (0.0085) (0.0050) (0.0055) (0.0062) TFP 0.0850 0.1240 0.316* 0.312*** 0.313** 0.284* (0.1140) (0.1400) (0.1840) (0.1170) (0.1380) (0.1710) Ln(IA) 0.233*** 0.218*** 0.215** 0.171 0.17 0.167 (0.0755) (0.0828) (0.0996) (0.1060) (0.1170) (0.1330)

Industry yes yes yes yes yes yes

Year yes yes yes yes yes yes

Observations 1,664 1,458 1,252 1,737 1,529 1,323

6.2 Sample excluding domestic M&A

Some people concern that effect of cross-border M&A on bidders is biased when firms engaging in domestic M&A are included in control groups. To test whether the results are different in the long time, firms engaging in domestic M&A among this period are excluded before PSM. By reconstructing samples, only firms completing cross-border M&A and firms neither taking cross-border M&A nor involving domestic M&A are included in the sample. The results are reported in TABLE 8. In third year, relative to firms taking no M&A, ROA of firms participating in cross-border M&A significantly decrease and R&D expenditure significantly increases in the same year and in the same industry. We can find that the degree of decrease on ROA of bidders are stable at the same magnitude in both samples, while the increase of R&D expenditure in TABLE 8 is almost twice the increase of R&D expenditure in TABLE 5. We conclude that firms engaging in cross-border M&A do significantly boost R&D expenditure. However, TFP of bidders decreases but not significantly in the third year. This can be explained when transaction cost is taken into consideration. One the one hand, if firms not take M&A, they have no such transaction cost. On the other hand, integration after cross-border M&A is difficult, which causes the increase of transaction cost, e.g. agency cost.

When firms take domestic M&A, they suffer the loss of it only but cannot benefit from “proximity concentration”. Therefore, when domestic M&A included in samples, TFP relatively increases.

TABLE 8

Average effect of cross-border M&A on performance of Chinese bidders.

This table shows how cross-border M&A influence the performance of firms compared to firms have not engaged in both cross-border and domestic M&A. Column (1) – (3) is average effect of cross-border M&A on ROA, TFP and R&D expenditure of firms in the first year, the second year and the third year, respectively. The standard errors are showed in the parentheses. Industry and year fixed effect are used in this regression. ***, **and * denotes significance at 1%, 5% and 10%, respectively.

First year Second Year Third Year

𝑦_$%&− 𝑦_$01 （1） （2） （3） ROA -0.0117** -0.0167*** -0.0213*** (0.0055) (0.0064) (0.0080) TFP -0.197** -0.232** -0.237 (0.0976) (0.1170) (0.1490) Ln(IA) 0.447*** 0.418*** 0.350*** (0.08610) (0.09750) (0.11600)

Industry yes yes yes

Year yes yes yes

Observations 1,999 1,743 1,487

7. Conclusion

In all, this paper uses a combined dataset of Thomson One and CSMAR to explore the long-term performance of acquirers, which are motivated by seeking strategic asset, after cross-border M&A. This paper first reviews the literature about determinants of border M&A and indicates nowadays Chinese firms are more likely to take cross-border M&A for the purpose of efficiency improvement and R&D expenditure, aiming

the analysis of literature on post-performance of acquirers, the paper proposes the opinion that the value of acquirers after cross-border mergers may decrease if these firms taking the transaction pursue the improvement of efficiency and innovations. Therefore, difference-in-difference methodology is adopted in the paper to test the causal effect between cross-border M&A and post-performance of acquirers in the long time, from motivations aspect. Although this method can avoid endogeneity to some extent, it will fail if firms are significantly different from each other before they are selected into groups. To avoid selection bias, propensity score matching method is used to construct sample before using difference-in-difference method.

From the regression results showed in tables, when Chinese bidders acquire foreign firms to seek high efficiency and innovations, these firms go through a significant decrease on ROA three years after completion of cross-border M&A. However, the effect of cross-border M&A on improvement of TFP and R&D expenditure is heterogeneous among industries and different technology intensity sectors. Specifically, firms in manufactory industry or in low technology intensity sector have significantly boosted TFP, while firms in service industry or in high technology intensity sector significantly have increased R&D investment in the third year after completion of cross-border M&A. For the firms which can benefit most from efficiency improvement have relative low requirement of innovations, therefore they are more likely encouraged by gaining access to knowledge and technology of target firms to participate in cross-border M&A. As a result, efficiency of these firms can significantly improve after completion of cross-border M&A rather than R&D expenditure. Firms in manufacture industries or in low technology intensity sector show this feature. On the contrary, firms in service industry or in high technology sector have high demand of innovations, so they prefer to develop new products and conduct research by using resource of target firms, resulting in significant increase of R&D expenditure after transaction.

Previous research uses abnormal return, measuring the value-creating of shareholders, to figure out the effect of cross-border M&A on post-performance of acquirers, and few research uses ROA. Therefore, this paper uses ROA to measure the value-creating for the whole firm, which is more reasonable. This paper also explicitly explores the different performance of acquirers after cross-border M&A by motives, providing a

specific empirical evidence on the research of post-merger performance. Additionally, since it is difficult to obtain transaction data in China, previous research usually used case study to explain the effect, which is not universal. This paper provides an example of overall research on long-term performance of Chinese bidders.

Admittedly, sample used for empirical investigation in this paper is relative smaller, compared research conducted in developed countries. On the one hand, there are extensive missing data of key variables, e.g. wages, employees, sales, before 2007. On the other hand, cross-border deals are not as many as in developed countries, and extensive Chinese firms engaged in cross-border M&A in 2016 which is out of our sample. Therefore, to get more reliable results, it is better in the future to test the result again with more firm samples and extend the long-term performance of acquirers to five years later or more.

Reference list

André P, Kooli M and Jean-François L'Her.(2004). The effects of cross-border M&A on the acquirers' domestic performance: firm-level evidence. (2004). Financial Management.

Chung, W. and Alcácer, J. (2002). Knowledge Seeking and Location Choice of Foreign Direct Investment in the United States. Management Science, 48(12), pp.1534-1554. Delios, A. and Beamish, P. (2002). Ownership Strategy of Japanese Firms: Transactional, Institutional and Experience Influences. SSRN Electronic Journal. Deng, P. and Yang, M. (2015). Cross-border mergers and acquisitions by emerging market firms: A comparative investigation. International Business Review, 24(1), pp.157-172.

EREL, I., LIAO, R. and WEISBACH, M. (2012). Determinants of Cross-Border Mergers and Acquisitions. The Journal of Finance, 67(3), pp.1045-1082.

Franco, C., Rentocchini, F. and Vittucci Marzetti, G. (2008). Why Do Firms Invest Abroad? An Analysis of the Motives Underlying Foreign Direct Investments. SSRN Electronic Journal.

Hennart, J. and Park, Y. (1993). Greenfield vs. Acquisition: The Strategy of Japanese Investors in the United States. Management Science, 39(9), pp.1054-1070.

Huang, Y. and Wang, B. (2011). Chinese Outward Direct Investment: Is There a China Model? China & World Economy, 19(4), pp.1-21.

Helpman, E., Melitz, M. and Yeaple, S. (2004). Export Versus FDI with Heterogeneous Firms. American Economic Review, 94(1), pp.300-316.

Kogut, B. and Singh, H. (1988). The Effect of National Culture on the Choice of Entry Mode. Journal of International Business Studies, 19(3), pp.411-432.

Kaplan, S. and Weisbach, M. (1992). The Success of Acquisitions: Evidence from Divestitures. The Journal of Finance, 47(1), p.107.

Li, J., Li, P. and Wang, B. (2016). Do cross-border acquisitions create value? Evidence from overseas acquisitions by Chinese firms. International Business Review, 25(2), pp.471-483.

Morck, R. and Yeung, B. (1991). Why Investors Value Multinationality. The Journal of Business, 64(2), p.165.

Markides, C. and Ittner, C. (1994). Shareholder Benefits from Corporate International Diversification: Evidence from U.S. International Acquisitions. Journal of International Business Studies, 25(2), pp.343-366.

McKinsey & Company. (2018). Why emerging-market companies acquire abroad. [online]

Newburry, W. and Zeira, Y. (1997). Generic differences between equity international joint ventures (EIJVs), international acquisitions (IAs) and international greenfield investments (IGIs): Implications for parent companies. Journal of World Business, 32(2), pp.87-102.

Nocke, V. and Yeaple, S. (2007). Cross-border mergers and acquisitions vs. greenfield foreign direct investment: The role of firm heterogeneity. Journal of International Economics, 72(2), pp.336-365.

Porta, R., Lopez-De-Silanes, F., Shleifer, A. and Vishny, R. (1997). Legal Determinants of External Finance. The Journal of Finance, 52(3)

Rossi, S. and Volpin, P. (2003). Cross-Country Determinants of Mergers and Acquisitions. SSRN Electronic Journal.

Rosenbaum, Paul R.; Rubin, Donald B. The Central Role of the Propensity Score in Observational Studies for Causal Effects. Biometrika. 1983, 70 (1): 41–55.

Shimizu, K., Hitt, M., Vaidyanath, D. and Pisano, V. (2004). Theoretical foundations of cross-border mergers and acquisitions: A review of current research and recommendations for the future. Journal of International Management, 10(3), pp.307-353.

Stiebale, J. and Trax, M. (2010). The Effects of Cross-Border M&A on the Acquirers’ Domestic Performance – Firm-Level Evidence. SSRN Electronic Journal.

Seru, A. (2014). Firm boundaries matter: Evidence from conglomerates and R&D activity. Journal of Financial Economics, 111(2), pp.381-405.

Walker, M. (2000). Corporate Takeovers, Strategic Objectives, and Acquiring-Firm Shareholder Wealth. Financial Management, 29(1), p.53.

Xu, E. (2017). Cross-border merger waves. Journal of Corporate Finance, 46, pp.207-231.

Appendix

1. Variable definition

Sales Log of sale.

Intangible asset ratio Intangible asset/total asset

Working capital ratio Net working capital to total asset

Wage Log of total wages to employees

Age The period from the established year to

the year firms engaging in cross-border M&A.

PP&E Log of PP&E

TFP Based on the method which is proposed

by Levinsohn and Petrin (2003). Levinsohn and Petrin used intermediate input as the proxy variable of unobserved shock to productivity instead of investment as used by Olley and Pakes

Capital Log of capital stock

SOE Dummy variable which takes value one

if acquirers are SOE, otherwise it takes value zero.

Exporter Dummy variable which takes value one

if it has export products, otherwise it takes value zero.

Experience Dummy variable which takes value one

if acquirers have engaged in cross-border or domestic M&A before, otherwise it takes value zero.

2. Sample used in robustness check

The sample used in robustness check exclude domestic merger deals. PSM is also used to construct this sample, ensuring that firms are similar before selected into groups. GRAPH 2 showed the propensity probability of firms may taking cross-border merger in both groups. According to the graph, the result showed in robustness check part is reliable.

GRAPH 2.

Propensity score comparison before and after matching in sample excluding domestic merger deals.

Graph above is the probability distribution of a firm may take cross-border merger in two groups before using PSM method, while graph below shows possibility of firms taking cross-border merger in two groups after using PSM method. Obviously, the possibility of a firm engaging cross-border merger is similar in matching groups.

0 5 10 15 20 kd e n si ty _ p sco re 0 .2 .4 .6 .8 1 pscore treat control

before matching

0 1 2 3 kd e n si ty _ p sco re 0 .2 .4 .6 .8 1 pscore treat control

after matching