Government intervention and corporate M&A transactions: Evidence

Transcription

1 Government intervention and corporate M&A transactions: Evidence from China Qigui Liu, Tianpei Luo, Gary Gang Tian 1 School of Accounting, Economics and Finance, University of Wollongong, Australia Department of Finance, Deakin University, Australia Abstract This study examines the impact of government intervention on mergers and acquisitions (M&A) decisions and post-m&a performance of Chinese listed firms. We documents that government intervention through majority ownership distorts M&A decisions in SOEs. In order to fulfill its social responsibilities, SOEs conduct more socially desired M&As, i.e., they are more likely to acquire local and politically connected targets and targets controlled state. These M&As reduce firm value which are indicated by the worse market reactions around the announcement of M&As. Our results further indicate that government strengthens its control power over SOEs by appointing politically affiliated managers in SOEs. The impact of government intervention is enhanced when local governments have a stronger intervening motivation. Overall, our results illustrate that the grabbing hand of government reduces firm value in listed SOEs in China. The M&As is one of the possible channels through which government extracts resources from listed SOEs to accomplish social and political objectives. Keywords: Mergers and acquisitions, Political connections, Vertical mergers, Government intervention, Corruption 1 Gary Tian is corresponding author, his is gary.tian@deakin.edu.au. Qigui Liu s address is qigui@uow.edu.au. Tianpei Luo s address is tl979@uowmail.edu.au. 1

2 1. Introduction M&A decisions are among the most important forms of corporate investment because these investments tend to significantly affect future corporate performance and shareholder wealth (Gaspar et al., 2005). Whether M&A performance create value to shareholders and what drives the M&A performance has long been interest in the corporate finance. The empirical results generally show that mergers and acquisitions do not necessarily bring value to acquiring firms, but produce mixed results 2. Another strand of literature focuses on how M&A performance is influenced by factors such as the acquiring and target firms characteristics, merger type, payment method and managers personal characteristics, such as stress or overconfidence (Moeller et al., 2005; Moeller, 2005; Fan and Goyal, 2006; Dong et al., 2006; Rhodes-Kropf et al., 2005; Masulis et al., 2007; Malmendier and Tate, 2008; Cai and Sevilir, 2012; Ishii and Xuan, 2014). What has been less explored in the literature is how firms M&A decision and performance are influenced by the government intervention. Heavy government intervention is a common business feature in most countries. Previous studies argue that the major objectives of interventionist government are rent seeking, extraction of private benefits and protection of local industries from competition 3. As the majority shareholder in state-owned enterprises (SOEs), governments exert intervention by using their voting rights to influence firm decisions (Beuselinck et al., 2015), but such influences may be detrimental to firm value. By appointing a connected manager, interventionist government extracts resources from listed firms under their control to accomplish the social objectives rather than to maximize firm value (Fan et al., 2007; Wu et al., 2010; Chen et al., 2011b). Empirical evidences reveal that government 2 See Dodd (1980), Franks and Harris (1989), Bradley et al. (1983), Land et al., 1991, and Masulis et al. (2007), Firth (1980), Asquith (1983), Malatesta (1983), Franks et al (1991) and Agrawal et al. (1992). 3 See Stigler (1971), Peltzman (1979), McChesney (1987), De Soto (1991), Shleifer and Vishny (1998) and Fan et al. (2007). 2

3 intervention leads to a low government quality (Ginka et al., 2012), investment inefficiency (Chen et al., 2011) and worse post-ipo stock return (Fan et al., 2007), which results less economic efficiency (Qian and Roland, 1998 and Wang et al., 2008). In the current study, we examine the impact of government intervention on M&A decisions and post-m&a performance in SOEs. Previous studies has confirmed that government intervention in SOEs hinders corporate performance, however, there still not enough evidence to indicate whether and how government intervention in SOEs constitutes a friction when they make M&A decisions, which lead to worse post-m&a performance in China. To understand the impact of government intervention on firms M&A decisions and performance, this study aims to provide empirical evidence to answer the following questions: Does government intervention affect their M&A decisions and post-m&a performance in SOEs? How does government intervention influence M&A decisions and performance in SOEs? Does the impact of government intervention on M&A performance in SOEs differ under different level of institutional environment and different level of local SOEs political and social burden from the local government? Given the significantly less developed law and institutional environment in China, including investor protection system, quality of government and corporate governance and underdeveloped capital markets (Allen et al., 2005), the theoretical argument predicts that the government intervention will negatively affect firm performance. This negative relationship is more pronounced in SOEs due to ambiguous clarification of the ultimate property rights in SOEs. This paper examines the role of government intervention in SOEs M&A decisions and post-m&a performance in China. We hypothesize that the government intervention alters SOEs M&A decisions to help local government pursue social desirable and political objectives which destroy firm value. 3

4 In this study, we firstly measure the government intervention through the government ownership in listed SOEs. We test whether listed SOEs will be more likely to conduct M&As for social welfare maximization rather than for shareholders interests. In China, the governments have a conflicted dual role in SOEs. As the majority shareholders and owner of SOEs, they are supposed to maximize firm performance and shareholder wealth. However, the non-transferability of state owned shares and assets create thorny incentive problems among both government and firm managers (Fan et al., 2007 and Shleifer and Vishny, 1986). On the other hand, as an administrator of social welfare, government officials are motivated to influence firm s decisions and undertake projects which pursue social desirable objectives rather than firm value maximization. Moreover, the tournament-style promotion system based on regional economic performance and social stability creates a stronger incentives for local government leaders to exert their influence over SOEs for their promotion potential (Cao et al., 2015). Therefore, strong government intervention usually leads to poor firm performance (Chen et al., 2011). In this study, we conjecture that government intervention distorts M&A decisions in SOEs through the majority ownership of government, which have worse post- M&A performance than non-soes. The politically connected managers in SOEs are argued that to be the channel for government intervention over listed SOEs in China (Fan et al., 2007) and are expected to be better accomplishing social and political goals in contrast to those unconnected managers. Firstly, the government still retained the decision right on the appointment of CEOs in listed SOEs and preferred to appoint affiliated managers on listed SOEs to implement its interventions. These connected managers therefore, have a stronger incentive to pursue political objectives which may run counter to corporate productivity. In addition, the poor corporate governance arising from the multilayered principal-agent framework in SOEs, managers of SOEs are less effectively monitored by the government, which makes the first type of agency issue severe in 4

5 SOEs. In other words, managers of SOEs have a strong incentive to pursue their private benefit through empire-building or even bribe-taking, which further harms the interest of shareholders. The incentive to pursue private benefit is stronger when managers have political connections because these help them to retain their positions when the firm suffers a bad performance (You and Du, 2012). Therefore, we further test the impact of government intervention by examining the influence of political connections on M&A transactions in SOEs. Following Fan et al. (2007) we classify the connected SOEs are those which their chairman or CEO is a current or former government officials. We conjecture that in SOEs, politically connections will negatively affect post-m&a performance in contrast to non-connected SOEs. This study mainly focus on the impact of government intervention on M&A transactions in SOEs in China but also uses non-soes are a control sample to reveal the variations in this relationship. In contrast, non-soes have a simpler goal structure of value maximization, we expect that political connections in non-soes will bring benefits to connected firms because, rather than being nominated by the government, connected managers of non-soes are generally sought and employed by the board, or even the controlling shareholders, to maximize shareholder interests. So the main purpose for employing those connected managers is to overcome financial barriers and obtain external financial resources (Faccio, 2006). We predict that connected managers in non-soes will pursue more value-adding M&As because they are able to obtain government support to acquire quality target firms. As the largest emerging market, the capital market in China provides an ideal institutional environment to conduct our analysis. First, as a controlling shareholder in SOEs, Chinese government plays a crucial role in its operation activities. During the Chinese decentralization in 1980s, the local governments in China at all levels have obtained authority and responsibility for their own local economies and be entailed devolution of the supervision power of SOEs from the central governments (Qian and Roland, 1998). This encourages fiscal competitions 5

6 among local governments and motivates them to intervene and prey local SOEs for local economic growth and social objectives. The prohibition of selling government owned shares in SOEs and the social welfare responsibility of SOEs make it difficult to imagine the government would either grant these firms discretion over staffing levels or subject them to truly enterprise-threatening competition in the market (Megginson and Netter, 2001). Moreover, it has been observed increasing government policies favoring the state sector in recent years. Therefore, the government ownership and policies are likely to continue to influence Chinese SOEs corporate decisions. Secondly, in contrast to the western market, a distinguishing feature of the M&A market in China is government intervention at the various levels of this process (Liu et al., 2013b) especially for SOEs. Although central government has largely granted operation decisions rights to SOE managers during the corporatization process, the government retained the ultimate decision rights conceding the M&As in SOEs (Qian, 1995). Therefore, M&As in SOEs could be a possible channel through which government exert its controlling power over SOEs. Like the IPO process, M&A deals in SOEs can only be conducted if approved by the Expert Advisory Committee for the Merger, Acquisition and Restructuring Listed Companies, which is an affiliated unit of China Securities Regulatory Commission, an authority of the Chinese government. Thirdly, in China, the government maintains its control over listed SOEs by appointing as the top managers of SOEs mainly current or former government officials or bureaucrats. This further enhances the power of government intervention in the decision making process of SOEs. Therefore, M&A activities in SOEs are mainly the result of government intervention, which is facilitated and strengthened by connected managers. Overall, the significant government intervention in SOEs in China together with the vital role of politically connected managers provides us an opportunity to conduct our study. 6

7 Using a sample of 514 acquisition announcements in Chinese publicly listed firms from 2005 to 2011, we conduct a series of empirical analyses to provide evidence for our hypotheses. Our findings confirm that the government intervention distorts M&As behaviors in SOEs and reduce firm value which is indicated by a worse M&A performance in SOEs as compared to non-soes. Moreover, political connections further reduce the post-m&a performance in SOEs which is which is measured by the cumulative abnormal return around the M&A announcement. This result consists with the argument that politically connected managers are the channel to exert government intervention over SOEs. We future test whether SOEs are more likely to conduct social desired M&As through examining the type of targets which acquired by SOEs. We find that in contrast to non-soes, SOEs are more likely to acquire local targets, firms controlled by local governments and politically connected targets. They pay significantly high M&A premium in these deals. These M&As consequently receive significantly worse market reaction and run counter to the shareholders interest. The negative effects of government intervention on M&A performance are more severe in politically connected SOEs in contrast to non-connected SOEs. However, we find that politically affiliated non-soes receive better market reaction in M&A transactions than unaffiliated counterparts, which supports the Helping hand effect of government on firm performance in private firms (Shleifer and Vishny, 1998). In order to address the potential endogeneity issue, we first examine whether the influence of political connections on the M&A decisions of SOEs differs between regions with different levels of government intervention, and our results confirm that SOEs with connected managers pursue more value-destroying M&As in regions with more government interventions. In addition, we provide evidence that SOEs with connected managers conduct more valuedestroying M&A when the local unemployment rate is higher, suggesting that social welfare becomes an important concern for politically connected SOEs when local government is facing 7

8 more pressures for social stability. Taken together, our findings suggest that, through majority ownership in SOEs and the appointment of politically connected managers, government intervention is distort M&A decisions for social and political objectives in SOEs and destroy firm value. This intervention is more severe in region with high level of government intervention and large social pressures. This study contributes to the literature in the following ways. First, our evidence enrich the extent of literature in M&As. Previous studies in this area primarily focus on the impact that different factors, such as the size of the acquiring firm, payment methods, corporate governance and social ties (Moeller et al., 2004; Faccio and Masulis, 2005; Dong et al., 2006; Rhodes- Kropf et al., 2005; Masulis et al., 2007; Cai and Sevilir, 2012; Ishii and Xuan, 2014), have on post-m&a performance in developed market. Our findings reveal that in emerging markets such as China, listed firms M&A decisions and performance are seriously affected by government intervention. Secondly, our findings suggests that SOE s M&A decisions can act as a channel through which government ownership and political connections affect the value of listed firms in China. Moreover, we find that local governments are more likely to grabbing listed firms resources when they have stronger intervening motivations. Of course, M&A transaction is only one of many channels which interventionist government may use to transfer resources from listed firms for their political goals. The government can expropriate listed firms by forcing them to undertake value destroyed related party transactions, accept affiliated appointment and conduct inefficient investments (Chen et al., 2011; Cheung et al., 2010; Fan et al., 2007). Finally, our analysis may also help to enhancing the understanding the different value of political connections between SOEs and non-soes in China. In particular, previous studies mainly focus on the effect of political capital on firm value and access to the financial market (Fisman, 2001; Johnson and Mitton, 2003; Khwaja and Mian, 2005; Faccio, 2006; Fan et al., 8

9 2007; Chen et al., 2011) and their results are mixed. Our results reveal that the different roles by political connected managers depend on the different controlling shareholders in China. The remainder of the paper is organized as follows. Section 2 develops our hypotheses. Section 3 introduces our data, sample, variables and the empirical model employed. Section 4 presents the empirical results and interpretations. Section 5 summarizes and concludes this paper. 2. Institutional environment and hypothesis development 2.1. Institutional environment: mergers and acquisitions in China Prior to the economic reform in China, the Chinese firms were fully controlled by government and the M&A decisions were made based on the macro-economic situation and government needs. After the economic reform, especially after the establishing of the Shanghai and Shenzhen stock exchanges, M&A deals increased significantly and became more marketoriented, but the central and local governments still retained the ultimate right of decision about mergers and acquisition in SOEs (Fan et al., 2007). Moreover, the main purpose of acquisitions in SOEs was still largely to rescue and support poorly performing local SOEs to fulfill social and political objectives, such as regional development, social stability and personal promotion. This value-destroying government intervention is enhanced by the appointment of current or former government bureaucrats as firms executives (Chen et al, 2011b). On the other hand, the M&A deals conducted by non-soes are more market-oriented, without being much influenced by governmental objectives. Given the close ties between the wealth of controlling shareholders and firm performance, non-soes tend to maximize firm value through seeking profitable investment projects, which include quality target firms. However, valuable investment opportunities are scarce in China. This motivates non-soes seeking and establishing political connections to obtain quality M&A target firms. 9

10 Under this unique institutional environment, firms merger and acquisition decisions are very likely to be influenced by managers political connections, both in SOEs and non-soes. Anecdotal evidence from recent media reports has shown that politically connected managers in SOEs (non-soes) are more likely to conduct value-destroying (value-adding) M&As. For instance, a recent ten billion RMB acquisition deal conducted by China Resources, which was the 18 th largest SOE by sales in 2012, has drawn much attention in China 4. This abnormal acquisition deal was initiated and supported by the Chairman of China Resources, Mr. Song Lin, who held an equivalent rank as vice-ministerial level government official and the secretary of the People s Political Consultative Conference. In February 2010, China Resources and its affiliates agreed to acquire the Jinye Coking Group, which was facing significant financial distress and significant decrease in profits. In this acquisition contract, China Resources and its affiliates agreed to pay 10.3 billion RMB for 80% of equity of the Jinye Group, which mainly included three coal mines and seven related companies, and the acquisition price was actually much higher than the assessed fair value of the assets (5.2 billion RMB). Moreover, the exploration licenses of three coal mines, which were the most valuable assets in this acquisition, expired before this deal. The other seven related companies neither generated profits nor started operation after this deal. Under this situation, however, Mr. Song still forced China Resources and its affiliates to conduct the acquisition. Overall, China Resources lost billions of RMB in this acquisition and Mr. Song and other senior executives were dismissed in Hypothesis development Government intervention and acquisition performance 4 The detailed information can be viewed at 10

11 As discussed in above section, government may seek rents from listed SOEs by utilizing their political powers over these firms, especially in countries where institutional constraints are weak (Sheifer and Vishny, 1994; Sheifer and Vishny, 1998; Fan et al., 2007; Cheung et al., 2010; Chen et al., 2011). In China, the government has the conflicting dual roles in SOEs, the administrator of social welfare and ultimate controlling shareholder. This motives the interventionist government to accomplish social and political objectives such as social stability, regional economy development and regional employment. Thus the government intervention will change SOE s objective function to that is preferred by government, which consequently reduce firm productivity and value (Lin et al., 1998). The prohibition of selling government owned shares in SOEs enhance the incentive of the government leaders and firms managers to expropriate listed SOEs for social benefits by exercising their intervention (Megginson and Netter, 2001). Since the government still retains the ultimate decision rights of M&A in SOEs, M&A could be the channel through which government exerts it intervention. It is therefore reasonable to expect that the M&As announced by SOEs would receive a worse market reactions than non-soes. Thus, we hypothesize that: H1a: In contrast to non-soes, the post-m&a performance is worse in SOEs due to the government intervention. Although we argue that government intervention has important impact on SOE s M&A decisions, the establishing the system of state-owned business groups which is so called Multilayer Legal Person system makes it more difficulty for government to intervene listed SOE s decision making process (Fan et al., 2013). However, the appointments of current and former government officials as top managers are expected to enhance the government intervention in listed SOEs. Therefore, the politically connected managers are viewed as a channel though which government leaders grab resources from SOEs (Fan et al., 2007). Moreover, since the social benefits based promotion system, the politically connected managers are also more 11

12 willing to perform government agendas and political objectives for their future promotion. This phenomenon is strengthened by low pay-performance sensitivity in SOEs (Firth et al., 2006). Thus, we expected that in contrast to unconnected SOEs, the politically connected SOEs are able to better fulfill government intervention thought conducting M&A and destroy firm value. We hypothesize that: H1b: The post-m&a performance is worse in SOEs with politically connected managers than unconnected SOEs Government intervention, M&A decisions and post-m&a performance If, as expected, the M&A is a channel through which government exerts its intervention over SOEs for the purpose of maximizing social, political and personal objectives rather than maximizing firm value, we should further expect that a positive relationship between government intervention and value-destroying M&As in SOEs. In this section, we investigate how government intervention affects M&A decisions in SOEs. One direct channel to investigate how government intervention affects post-m&a performance is through examining the M&A premium paid by SOEs. Due to the strong government intervention, the government may expropriate SOEs by requiring a high M&A premium in selling these firms under their control. Moreover, the self-interest government officials and managers may even extract private benefits, such as stolen cash received in bribes, by employing their political power over SOEs to force them pay high M&A premium for personal benefits. This situation becomes worse if the SOEs have politically connected managers. The China Resource case in section 2.1 provides the anecdotal evidence on this argument. In addition, due to the lower pay-performance sensitivity, the managers in SOEs are less likely make efforts on the analysis of the true value of target firms. This further increases the 12

13 likelihood of overpaying in M&As for SOEs. Therefore, we expect that due to the government intervention, SOEs will pay higher M&A premium than non-soes. Thus, we hypothesize that: H2a: SOEs will pay a higher M&A premium in contrast non-soes due to the strong government intervention. The local SOEs in China are usually accused of operating inefficiently because they have to take more responsibility for local social obligations, such as tax delivery, employment, and GDP growth. Therefore, when local firms suffer from financial distress which may risk the local social stability and economy growth, local government are more likely to exercise its controlling power over SOEs, especially for these politically connected SOEs and force them to rescue these distressed firms from going bankruptcy. Connected managers in SOEs also have incentive to acquire other local firms not only because of concerns over future promotion but also to make private benefits, such as kickbacks and bribery, due to the poor corporate governance and lack monitoring by the large shareholders. Consequently, these social and personal benefits distort the M&A decision making process and harm firm performance. Thus, we expected that government intervention increase the incentive of conducting local M&As in SOEs, but receive a worse market reaction around the announcement. We therefore hypothesize that: H2b: In SOEs, government intervention increases the likelihood of conducting local M&As, which reduce the post-m&a performance in those firms in contrast to non-soes. Based on the foregoing arguments, we suggest that government officials may expropriates listed SOEs through M&As for social and political objectives. This government intervention in M&As could be enhanced and more easily applied if the target firm is also controlled by state. As the controlling shareholder of acquiring and target firms, bureaucrats utilize their control right over both firms to fulfill social and event personal objectives. The government 13

14 can force listed SOEs to acquire a poorly performed SOE or sell a SOE to the listed firm at a price higher than the market value. However, such M&A transactions are more difficult to be completed between SOEs and non-soes. Especially after the enactment of 2007 Property Law, private firms are given more protections against the potential expropriation of their assets (Berkowitz et al., 2015). Cheung et al. (2010) find that the related party transactions between local SOEs reduce firm value which in the line with the grabbing hand of government. Thus, we expect that government intervention increase the incentive of listed SOEs to acquire a SOE target, but such M&As have poor post-m&a performance. We therefore hypothesize that: H2c: In SOEs, government intervention increases the likelihood of acquiring a SOE target, which has a negative impact on the post-m&a performance in those firms as compared with non-soes. Politically connected managers are suggested as a channel though which government exerts its intervention in SOEs (Fan et al., 2007). As discussed above, these politically connected managers are expected to be better fulfilled social objectives such as regional economy development, employment and social stability, either because of their political duty or personal benefits. We therefore, expect that the degree of government intervention will be higher if the acquiring and target firms both have top managers who are current or former government officials. This enhanced government intervention further distorts listed SOEs M&A behavior and reduce the post-m&a performance. We hypothesize that: H2d: In SOEs, government intervention increases the likelihood of SOEs to acquire a politically connected target firm, but reduces those firms post-m&a performance in contrast with non-soes. 3. Methodology and measurement of variables 14

15 3.1. Sample The sample used in this paper consists of M&A deals conducted by publicly listed firms on the Shanghai and Shenzhen stock exchanges from 2005 to We use the CSMAR China Listed Firm s M&A Database to obtain announcement dates, information on acquiring and target firms and M&A financial information for completed deals in our sample period. We also collect other information from a series of datasets from the CSMAR database. These include the China Stock Market Financial Statement Database from 2005 to 2011; the China Listed Firm s Corporate Governance Research Database from 2005 to 2011; the China Stock Market Trading Database from 2004 to The CSMAR database is one of the most important and widely used databases in research on the Chinese capital market. Following previous studies in acquisitions, we require M&A deals to meet the following criteria. We require that the acquiring firm obtains at least 51% of the target shares and omit M&A deals in which the acquiring firm already holds at least 51% of the target before the deal (Malmendier and Tate, 2008). Moreover, we exclude small transactions in which the deal value is less than 1% of the acquirer s market capitalization (Cai and Sevilir, 2012 and Masulis et al., 2007). We exclude the announcement if the acquiring firm announces two or more M&A deals within three months. We require the acquirer to make annual financial statement information available (three years prior to acquisition announcements and three years post these announcements) and stock return data (250 trading days prior to M&A announcements) from CSMAR databases. Finally, we exclude deals in which acquiring or target firm information, announcement date and financial data are missing. After meeting these criteria, our final sample yields 514 M&A cases among a total of 10,586 firm year observations. Tables 1 and 2 provide the distribution of our 514 M&A deals by year and industry, respectively. Panel A of table 1 demonstrates that the M&A deals significantly increase during our sample period, especially after This result is consistent with the view that the Chinese 15

16 non-tradable share reform facilitates firms to conduct mergers and acquisitions. Our results indicate that almost 47 percent of M&A are conducted by SOEs. The munber of M&As conducted by non-soes increase significantly after In panel B, we find that around 39 (63) percent of M&As in SOEs (non-soes) are conducted by connected (non-connected) managers in both SOE and non-soe subsamples. While the percentage of M&A conducted by connected managers increases in non-soes from 18.2 percent in 2005 to 39.4 percent in 2011, but there is no great variation in the SOE subsample. <Table 1> Table 2 presents the distribution of M&A by industry. Almost 50% of M&A deals are conducted in the manufacturing industry. There is a cross-industry variation in the likelihood of having an acquisition conducted by politically connected managers. Acquisitions in the infrastructure and public utility sectors, such as construction, real estate, electricity, mining and gas and hot water services, are more likely to be conducted by politically connected managers. <Table 2> 3.2. Measurement of variables Government intervention The reliability of the measurement of government intervention is critical in this study. We test our research questions by measuring the government intervention at two different levels. Firstly, we measure the government intervention by examining whether a firm is controlled by the government. We define a listed firm is a SOEs which the dummy variable SOE equals 1 if the firm s ultimate controlling shareholder is central or local governments, any government departments or a SOEs. Secondly, we further measure the government intervention through the firm s political affiliation. Following Fan et al. (2007), we define a firm as having political connections if either the CEO or chairman of the board satisfies any one of following three 16

17 criteria. The CEO or chairman of the board is: (1) a current or former government official; (2) a current or former member of the People s Political Consultative Conference; (3) a current or former member of the People s Congress. The corporate political connections data is collected manually from the profile of the CEO and chairman of board. The CSMAR corporate government database provides detailed biographical information about top managers. For those who have not been recorded, we collected this information from the firm s annual report. We employ the dummy variable (PC) to measure the acquiring firm s political connections, which equals 1 if the CEO or the chairman of director satisfies the three criteria and 0 otherwise Other variables A series of variables are constructed to measure firms M&A decisions and post-m&a performance. We first define a dummy variable M&A, which is equal to 1 if the firm conducts an M&A in a given year, in order to measure the likelihood of conducting an M&A. Furthermore, we employ several stock and accounting-based measures to evaluate the post- M&A performance of the Chinese listed firms in our sample. The stock performance measures are the three-, five- and 11-days post-acquisition cumulative abnormal market-adjusted stock returns (CARs). We use the Capital Asset Pricing Model (CAPM) to find the expected stock returns during event windows for adjustment in all our stock performance analyses. The market value-weighted market index of both the Shanghai and Shenzhen stock exchanges are employed as market return in this model. The estimate window is 250 trading days, which start from 280 trading days prior to the announcement. Following Huang et al. (2014), total M&A premium is defined as the different between M&A price and the fair value of the target firm. To make our results more robust, we measure the M&A premium using both the relative value of the premium (PREMIUM 1), which is the total premium relative to fair value of the target 17

18 firm, and the absolute premium (PREMIUM 2), which is the natural logarithm of the total premium. We also use three long-term performance measures, which are the change in Tobin s q (Growth in Q), the change in ROA (Growth in ROA) and the growth in earnings (Growth in Earning). The Tobin s q is calculated as the market value divided by replacement value. We calculate ROA as net income divided by total assets. Consistent with previous literature (Fan et al., 2007), we use the pre-m&a accounting figures as a benchmark to evaluate change of accounting performance in the post-m&a period. We calculate the Growth in Q and Growth in ROA by subtracting the average Tobin s q and ROA in the three years prior to the M&A announcement from the three years of annual Tobin s q and ROA after the M&A announcement. The Growth in Earning is the percentage change of the average of annual earnings from three years before the M&A to three years after. In this study, we manually collect target firm information from the M&A announcement. We collect location data for the target firm and include a dummy variable (LOCAL) which equals 1 if the acquiring and target firms are located in the same province and 0 otherwise We measure the target ownership structure by the dummy variable TARGET SOE which equals 1 if the target firm is controlled by state. We also collect target s political capital data as TARGET PC, which equals 1 if the target firm s managers satisfy the previously mentioned three political connection criteria and 0 otherwise. We also collect information about the industry in which target firm operated. If the acquiring firm conducts a vertical merger, we include a dummy variable (VERTICAL) that equals 1, if the industry sector of the target firm is upstream or downstream of the acquiring firm s industry sector. To conduct our regression analysis, we also include various control variables in our regression models to control for factors which may affect M&A performance. The definitions of these variables are reported in detail in Appendix A. 18

19 3.3. Regression model To examine the effect of political connections on the likelihood of a firm conducting M&A and on post-m&a performance, we employ the following equation as the baseline regression model. M&A Dummy/Post M&A Performance/M&A characteristics = α 0 + β 1 Government Intervention i,t + β 2 Control Variables i,t + β 3 Year and Industry Dummy + ε 1 In this equation, the key dependent variables are M&A dummy, stock and accounting post- M&A performance, and M&A characteristics. When the dummy variable is used as a dependent variable, the model becomes a logistic model. Post-M&A performance is measured by CARs, as discussed above. M&A characteristics are variables to proxy the following characteristics of acquisitions: (1) whether the target firm is a SOE; (2) whether the target firm is a local firm; (3) whether the target firm has a politically connected manager; (4) whether the M&A is a vertical merger;. The key independent variable is government intervention, which is measured by the firm s government ownership and political connections. The year and industry dummies are also included in our regression models to control for the effect of year and industry. The key independent variables may interact with other variables when necessary. All variables are defined in Appendix A. 4. Empirical results 4.1. Summary statistics and univariate tests Table 3 presents the summary statistics for our main variables. The results show that the proportion of connected firms accounts for 25% of our firm year observations, while 53% of our firm year observations are SOEs which are controlled by the government or a government agency. For our M&A deals, our results show that the acquiring firm s shareholders earn 19

20 slightly positive returns from conducting M&As. The one-, two- and five-day CARs are around 2% in our sample. The acquiring firms are willing to pay up to 81% more than the fair value of target firms to gain controlling rights, which is shown by the positive value (0.81) of PREMIUM1. Compared with acquiring firms, the size of target firms is relatively small, at about 22% of the acquirer s market value. In our study, 18% of target firms have political connections (TARGET PC) and vertical mergers and acquisitions (VERTICAL MERGE) are about 29% of the 514 M&A deals. <Table 3> In table 4, we present the univariate test for stock performance of M&A conducted by SOEs and non-soes. The results indicate that the overall market performance which is measured by the three-, five- and 11-days post-m&a CARs, is worse for SOEs than for non-soes. Moreover, we find that the SOEs with connected managers underperform firms without political connections, in terms of post-m&a market performance. These differences are statistically significant. These results support our arguments that government intervention distorts M&A performance in SOEs through its controlling rights in listed SOEs and politically connected managers. In contrasted to the results for SOEs, connected non-soes perform significantly better than firms without political connections which support the argument that in the counties with weak legal protection and less developed capital market, politically connected managers provide various benefits to these connected private. Moreover, our results on the M&A premium reveal that politically connected non-soes pay significantly lower premium than their counterparts. The results from other long-term performance measures in Table 5 confirm the above arguments. We find the post-m&a Tobin s q, ROA and earnings are worse for connected SOEs, but improve in connected non-soes. These differences are also statistically and economically significant. Thus, these results confirm our hypotheses H1a and H1b. 20

21 <Table 4> <Table 5> 4.2 Political connections and post-m&a performance Before we examine the implications of government intervention on M&A performance, we first estimate the effect of government intervention on the likelihood of a firm conducting an M&A in listed SOEs. The results presented in Table 6 suggest that SOEs are inactively in the M&A market. This result suggests that managers in listed SOEs have less incentive to bear risks from M&A in order to avoid losses. However, connected SOEs are more likely to conduct M&A. The coefficient of PC is positive and statistically significant at the 1% level of significance in column 2. It indicates that since the government has more direct intervention power over connected SOEs, these firms are forced to undertake more M&As which is for accomplishing social or political objectives, in contrast non-connected SOEs. <Table 6> Table 7 presents the regression results for the first hypothesis, which predict that due to government intervention, listed SOEs have a worse market reaction as compared to listed non- SOEs and the politically connected SOEs are underperformed than unconnected SOEs around the announcements of M&As. From columns 1 to 3, we use the full sample to test whether government intervention which is measured by the government ownership leads to a worse post-m&a performance in listed SOEs in contrast to non-soes. We find the coefficients on SOE significantly negatively relates to the three-day, five-day, and 11-day cumulative abnormal return indicating a worse market reaction when SOEs make M&A announcements. These results are consistent with our argument that government utilizes their controlling power over listed SOEs to intervene their M&A behaviors for social and political goals. These M&A announcements usually receive negative market reactions and reduce firm value. In columns 4 21

22 to 5, we further test the impact of political connection on the post-m&a performance in listed SOEs. The significantly negative coefficients of PC suggest that the politically connected SOEs receive significantly worse market reactions when they conduct M&As as compared to nonconnected listed SOEs. Although politically connected SOEs may receive preferential treatments in M&As, but our results reveals that the negative effect of government intervention offsets these benefits. Thus, the findings confirm our expectation that the interventionist government appoints current or former bureaucrats as the top managers to ensure the control over SOEs and strength their intervention. Overall, the findings in table 7 support our hypothesis H1 and suggest that government intervention reduces the M&A performance in SOEs by utilizing their control rights over SOEs and appointment politically connected managers. For other control variables, consistent with previous literature, we find that relative size (RELATIVE SIZE) is positive and significant, which indicates that acquiring a larger target generates higher returns. Larger firms receive better post-m&a performance which is indicated by the positive coefficient on SIZE. However, cash-financed M&A perform worse than stock or mixed-finance deals, which is consistent with the results of Moeller et al. (2004) and Travlos (1987). <Table 7> In table 8, we further provide the regression results regarding the impact of political connections on the post-ma long-term accounting performance of SOEs. This result confirms the above market performance results. Overall, the results from tables 7 and 8 indicate that the government intervention in acquiring firms are associated with lower market and long-term accounting performance in SOEs, which confirms our hypotheses H1a and H1b. We argue that the possible explanations on this negative impact of government intervention on SOEs M&A 22

23 performance is the heavy social responsibility of SOEs. The following sections will provide more evidence to support our arguments. <Table 8> 4.3 How government intervention decreases post-m&a performance in SOEs? We have provided evidence that government intervention results in worse post-m&a performance in SOEs. In this section, we aim to provide the answer for the question how government intervention affects M&A decision making in SOEs Government intervention and M&A premium In this section, we investigate the impact of government intervention on M&A premium in SOEs. Table 9 presents the regression results. As expected, the coefficient on SOE is significantly positive to the M&A premium. This result confirms our expectation that local government may expropriate listed SOEs by asking a higher price than the fair value of the target firms. However, we find that the political connections in SOEs do not have significant impact on M&A premium. We note that the sample size is quite small due to the missing data in the fair value and trading value of the target firms in the CSMAR database. In contrast, we find the politically connected managers reduce the M&A premium in non-soes. The detail analyses for this finding will be reported in the following section in our study. Overall, we find that SOEs pay higher in M&As than non-soes in China. <Table 9> Government intervention, local acquisitions and post-m&a performance As results in previous sections have shown, we find SOEs perform poorly compared to private firms in M&As. Over the year of decentralization in China, local governments have to take the responsibility of local economy and compete with other provinces for the limited resources. Therefore, they have strong incentives to seek helps and expropriate listed SOEs under their 23

24 jurisdiction to support the local economy and social benefits. We argue that listed SOEs may conduct more M&As to support local economy or buyout financial distressed local firms, no matter whether those mergers create value for shareholders or not. In order to provide empirical evidence for this argument, in this section we examine whether SOEs with political connections are more likely to acquire local firms and whether such acquisitions harm firm value. Table 10 presents the logistic regression results regarding the impact of government intervention on the likelihood of local M&A deals. In column 1, the coefficient on SOE is statistically positive to the likelihood of acquire a local firm. Moreover, we find the politically connected SOEs conduct more local M&As than its non-connected counterparts. These findings support our above arguments that the government officials exert intervention to encourage SOEs conduct M&As under their jurisdiction in order to support local economy. More importantly, in panel 1 of table 11, we find the coefficient on the interaction between SOE and LOCAL is statistically significantly negative to cumulative abnormal return over the event windows, which indicate that the SOEs post-m&a performance are significantly lower when they acquire a local target. This finding is confirmed by the regression results for the impact of political connection in SOEs and local M&As on the post-m&a performance in panel B of table 11. We find that in contrast to non-connected SOEs, the connected SOEs receive worse market reaction when they conduct a local M&A. Our findings support our hypothesis H2c and suggest that interventionist governments are motived to expropriate local SOEs by forcing them undertake local M&As and this government intervention are more pronounced if the firms have connected managers. This result is consistent with the argument that local governments have stronger incentives to intervene SOEs for social and political goals which harms investment efficiency (Chen et al., 2011). <Table 10> <Table 11> 24

25 Government intervention, target firm ownership structure and post-m&a performance In the previous section, we find that politically connected SOEs are more likely to acquire a local target in order to support local economy. In this section, we further investigate whether SOEs are prefer to acquire the targets which also controlled by the state. These results are tabulated in table 12. Consistent with our expectation, we find that the SOEs are more likely to acquire the SOEs target as compared to non-soes which is shown by the significantly positive coefficient on the SOE in column 1 of table 12. Moreover, our results indicate that politically connected SOEs are bale to acquire more SOEs target than non-connected SOEs. One possible explanation for this result is that SOEs can easily obtain valuable sate owned assets compared to private firm. If this argument is correct, consequently, this M&A will receive positive market reactions and increase firm value. On the other hand, due to the government intervention, the M&A decision may be distorted in SOEs and they have to acquire certain poorly performed SOEs for social benefits which harms firm value. In table 13, we present the regression results for the impact of government intervention and target ownership structure on the post-m&a performance. We find the coefficient of the interaction term SOE*TARGET SOE and PC*TARGET SOE are statistically negatively related to the post-m&a performance. These results suggest that acquiring a SOEs target reduce firm value in SOEs which support the government intervention argument. <Table 12> <Table 13> The effect of target firm political connections on post-m&a performance Asa shown in previous sections, our results reveal that politically connected managers are the channel through which government exert intervention over SOEs. We therefore, expect that 25