Cross-Border Acquisitions by Chinese Enterprises: The Benefits and Disadvantages of Political Connections

This paper explores whether and how political connections affect the likelihood of completing a cross-border M&A deal for Chinese publicly listed, but privately-owned enterprises (POEs) and the resulting firm performance. In line with our proposed political connection trade-off theory, we find that POEs with politically connected top managers are more likely to complete a cross-border M&A deal than POEs with no such connections, but that this comes at the cost of negative announcement returns and subsequent lower accounting performance. These findings support the idea that politically connected top managers engage in “political empire building” behavior at the cost of shareholders’ wealth.

China become the world's third largest foreign investor after the U.S. and Japan. Moreover, China is the only developing economy among the top ten foreign investors. In contrast to other major developing economies whose outbound FDIs typically take the form of reinvested earnings, China's FDI mainly consists of new equity investments (UNCTAD, 2016).
In 2006, the Chinese government also incorporated privately owned enterprises (POEs) in its go global strategy by starting to offer tax rebates and access to long-term financing at favorable terms 6 football clubs. They rush into these deals not because they are particularly good investments, but because President Xi Jinping has expressed hopes that China will become a soccer powerhouse (see South China Morning Post, 2017).
A further possible explanation for the failure of outbound investments is that after deal completion, Chinese POEs often find that competition in the host country is much tougher and that some business practices commonly accepted in China, such as relaxing health and safety standards, cannot be mirrored abroad (see . A similar argument was also invoked by Antkiewicz and Whalley (2006) in discussing why most of the cross-border M&A transactions attempted by Chinese SOEs are unsuccessful in Organization of Economic Cooperation and Development (OECD) countries.
We use the term political connection trade-off theory to refer to the oppositional situation whereby politically connected POEs are better positioned (than their unconnected counterparts) to manage the necessary logistics of a cross-border M&A, but the deals often come at the cost of poor financial performance. If productivity and profitability were frequently to matter less than political goals, politically connected POEs would subject themselves to moral hazard and create a "principal-principal" conflict between the state and the firm's shareholders (see Young et al., 2008) Specifically, the "political empire building" behavior of politically connected top managers would have a negative effect on shareholders' wealth.
To test the political connection trade-off theory, i.e. that politically connected top managers of POEs are more likely to complete a cross-border M&A transaction than their unconnected 7 counterparts, but at the cost of poorer performance, we conduct several analyses. First, based on a sample of 1,782 Chinese POEs listed on either the Shanghai or Shenzhen Stock Exchanges, we analyze the POEs' likelihood of completing a cross-border M&A deal. Consistent with our argumentation, we find that politically connected POEs have a greater likelihood of successfully completing a cross-border merger or acquisition than their unconnected counterparts. Our results remain robust after invoking a variety of robustness checks.
In a second set of analyses, we examine stock price returns and the return on equity after the announcement of a cross-border M&A to test for the market reaction and the impact on firm performance. We expect both the short-and long-term post-M&A performance to be lower for politically connected POEs than for non-politically connected POEs. We show that this is indeed the case; POEs with a politically connected chairman or CEO show significantly lower announcement returns (to the tune of about 1.5 to 2 percent) and are less profitable than their nonconnected counterparts within the first three years after deal completion.
Overall, our study supports the political connection trade-off theory and makes the following contributions to the existing literature. First, to the best of our knowledge, this study is the first to deliver a theoretical framework and empirical analysis of how political connections influence a Chinese POE's decision to engage in cross-border M&A activities and what the related costs of these connections are. Second, our study contributes to the literature on how governmental influence in emerging markets can affect the decisions of domestic firms to expand internationally through cross-border M&A deals (see Xiao and Sun, 2005;Rui and Yip, 2008;Peng, Wang, and Jiang, 2008;Luo, Xue, and Han, 2010;and Du and Boateng, 2015). Finally, our study contributes to the research methodology typically used in studies in this area. 8 The remainder of this study proceeds as follows. Section 2 develops our hypotheses, while Section 3 describes the data collection process. Our research methodology is presented in Section 4 and Section 5 provides our empirical results. Section 6 concludes.

Political Connections and Cross-border M&A Deals
Political connections can be valuable to firms' financing activities in both developed and developing countries, as many empirical studies have shown. 3 However, the benefits are generally more pronounced in emerging markets because of their relatively inferior institutional environments, more concentrated ownership structures, and less efficient legal systems (La Porta et al., 1998, 2000. In the case of China, we argue that the political connections of top management team members are more beneficial for POEs than SOEs simply because Chinese POEs face a different institutional environment. SOEs in China are the pillars of the national economy, while POEs must seek ways to overcome the discrimination they face in the capital market. One method is to build political ties with the government by hiring top managers with specific political backgrounds (Chen et al., 2011). Positive influences of political connections on various economic activities of Chinese POEs are documented in many empirical studies. Li et al. (2008) find that POE founders are more likely to obtain financing from state-controlled institutions if they have political party membership. Politically connected Chinese firms are also more likely to obtain loans with longer terms and lower interest rates when borrowing from state-owned banks (see Luo and Zhen, 2008;Yu and Pan, 2008;and Yuan, Jing, and Liao, 2010). Luo and Liu (2009) note that it is 9 easier for politically connected POEs in China to enter industries with high entry barriers, such as banking and telecommunications. Similarly, Li and Zhou (2015) find that politically connected POEs are more likely to get IPO requests approved and that such POEs are less likely to be subjected to on-site auditing from regulatory authorities.
Based on the arguments above, we expect that POEs whose top managers have political ties to the Chinese government are both more willing and more able to complete cross-border M&A transactions. Thus, we postulate Hypothesis 1 as follows: Hypothesis 1: Politically connected POEs are more likely to complete cross-border M&A deals than unconnected POEs.

Corporate Governance and Cross-border M&A Deals
In the previous subsection, our argument for the value of political connections is based on the institutional environment of a POE's home country. Nevertheless, when POEs enter the global market, they are also affected by the institutional environment of the host countries (see Kostova, 1999;Lu et al., 2014;and Regner and Edman, 2014). Meyer and Rowan (1977) argue that when companies enter a foreign market, they are likely to adapt to the prevalent organizational practices and structures in the host country with the goal of enhancing their overall sense of legitimacy.
This issue is more prominent when companies from emerging economies, with relatively poorer institutional environments, enter more advanced economies that typically feature higher-level institutional environments (as is mostly the case in the present study). Therefore, we expect that some Chinese POEs will endeavor to ameliorate their corporate governance to ensure that they meet local governance standards before attempting to conduct cross-border deals. This would give them a greater chance of being successful. Therefore, in Hypothesis 2, we posit the following relationship between corporate governance and POEs' cross-border deals: Hypothesis 2: POEs with better corporate governance are more likely to complete a crossborder M&A deal.

The Performance of Acquiring POEs
It is commonly known that the Chinese government intervenes with SOEs' business activities by appointing managers that have strong political ties. These politically connected managers can assist the government in achieving political and social objectives, which may be prioritized over commercial goals (see Wu, Wu and Rui, 2012). By following the government's recommendations, the managers can increase their political capital, which is vital to their political career. However, decision making based on a manager's political agenda may come at the expense of shareholders' wealth, creating a principal-principal conflict between the intervening government and non-state shareholders (see Young et al., 2008). This conjecture is supported by empirical analysis. For example, Wu, Wu, and Rui (2012) show that SOEs with politically connected top managers have lower accounting performance (measured by ROA) and fewer growth options (measured by Tobin's q). Similarly, Fan et al. (2007) examine the performance of Chinese IPO firms using a sample of 790 partially privatized SOEs. Their empirical analysis shows that IPO firms whose CEO is politically connected to the Chinese government have lower initial returns and lower accounting performance in the three years after going public. They argue that the political rent seeking behavior of politically connected CEOs expropriates the wealth of minority shareholders, which in turn harms firm performance. Evidence of the principal-principal conflict is also found by Sun, Vinig, and Hosman (2017)  firm-year observations.

Identification of Cross-border M&A Transactions by Chinese POEs
We define a POE as acquisitive if a cross-border M&A deal was completed during the observation period. We obtain the cross-border deals of Chinese POEs from CSMAR's China Listed Firms' Merger & Acquisition, Asset Restructuring Research Database. We exclude any cases where the cross-border M&A occurred in tax havens or offshore financial centers because firms acquired in this way are not "real" or "producing" foreign companies, but rather Chinese "shell companies." 4 We find that 290 Chinese POEs completed 385 cross-border M&A deals between 2007 to 2016.
We exclude two POEs that engaged in cross-border M&A activities before that time period, because these acquisitions might follow a different rationale. We consider the remaining 288 firms completing 385 cross-border M&A transactions as acquiring POEs (see Table 1). CSMAR's China Listed Firms' Merger & Acquisition, Asset Restructuring Research Database also provides the 13 country of origin of each overseas target that is acquired. In total, the cross-border deals completed by Chinese POEs span forty countries (see Table 2 for an overview).

Identification of Political Connections
We proxy for political connections by following the recent literature by Faccio (2006), Fan, Wong, and Zhang (2007), Li and Zhou (2015), and Schweizer, Walker, and Zhang (2016). The present study only considers the political background of the Chinese POE's board chairman and CEO. We hand-collect the information for each company in our sample from Stockstar, which provides detailed past and current work experience for the top management of each listed company.
We define a POE's chairman or CEO as politically connected if he or she is or was a representative in the People's Congress (PC), the Chinese People's Political Consultative Conference (CPPCC), an officer in local or central government, or an officer in the military. We code the political connection dummy variable (Connection) as 1 for each year since the chairman or CEO is politically connected, and 0 otherwise (see Li and Zhou, 2015;Schweizer, Walker, and Zhang, 2016). In addition, we measure the strength of the political connections of each firm's chairman or CEO by creating a political connection index (PC Index). The value of this index ranges from 1 to 3 depending on the strength of the political ties (where 3 represents the strongest political connection). 5 5 According to a research report by Harvard University's Kennedy School, the PC, in conjunction with the CPPCC, act as the legislative arm of the government and thus as the highest political entity in China (see Saich, 2015). The PC's functions include: overseeing the work of government departments and electing major officials; amending the constitution; supervising the enforcement of constitutional and legal enactments; and examining and approving the state budget and the economic plan. Members serving on the standing committee of the PC and CPPCC command particular power, as they work actively on law-making. In addition, Saich (2015) notes that the PC and CPPCC play more than a ceremonial role in China. Therefore, we assign the highest PC Index value of 3 to POEs in which the top managers are (or were) members of the standing committee of the PC and CPPCC, as well as to POEs whose top manager is the head of the central or provincial government. If the top managers of a POE are ordinary members of the PC or CPPCC, we assign a value of 2 to the PC Index. Finally, if the chairman or CEO is only an officer of a specific governmental department, or was an officer in the army, his or her political connections are considered more limited; hence, we assign a value of 1 to the PC Index.

Measuring Corporate Governance
To examine the impact of corporate governance on Chinese POEs becoming acquisitive overseas, we construct a comprehensive index that measures the overall corporate governance level of Chinese POEs. Our index aims to reflect a company's overall governance quality more accurately than single governance factors. It also eliminates multicollinearity that may arise in multivariate regressions when using single governance factors (Brown, Beekes, and Verhoeven, 2011). The advantages of a corporate governance index have been elaborated upon quite extensively in the extant literature (see Gompers, Ishii, and Metrick, 2003;Brown and Caylor, 2006;Dutordoir, Strong, and Ziegan, 2014;and Shan, 2015).
We follow Shan (2015) and construct an equally weighted corporate governance index for Chinese listed firms according to China's two-tier board system, but modify where necessary to account for the fact that our sample includes only POEs and no SOEs. Thus, we exclude the factor differentiating SOEs and POEs. We also exclude a factor for cross-listings. 6

Control Variables
We include an array of control variables that could potentially affect the likelihood of a Chinese POE carrying out a cross-border deal: profitability (ROA), leverage (Leverage), firm size (Firm Size), growth opportunities (Tobin's q), and a tangible asset ratio (Tangibility). We also use those variables to conduct a propensity score matching (PSM) technique. Detailed information for the control variables is provided in Panel A of Table A1 in the appendix.
To study the market reaction to the cross-border announcement, we include the following deal characteristics: the cultural difference between China and the country in which the target firm is

Political Connections and Cross-border M&A
To examine how political connections can affect the likelihood of becoming an acquisitive POE, we carry out the following panel logit regressions which take account of the fact that some acquisitive POEs complete more than one cross-border M&A deal: (1/0) , = + 1 • , + 2 • , + • , , +  (1), but the dependent variable is now a count variable that measures the number of overseas targets acquired by a POE over our sample period:

Endogeneity Concerns
We note that the potentially endogenous nature of political connections may impede the robustness of the proposed causal relationship between political connections and the likelihood of acquiring overseas targets for Chinese POEs. We conduct a quasi-natural experiment to cope with this problem. Specifically, we examine whether chairman/CEO turnovers that result in an increase in the PC Index (i.e. that cause firms to be more politically connected) increase the likelihood of POEs engaging in cross-border M&A transactions. Our first step is to apply a similar PSM routine to that used in our previous analysis. We include the PC Index as an additional matching variable and match in the year before the POE completed its first cross-border M&A deal (288 firm-year observations) with POEs that did not acquire any overseas companies during the observation period.
This ensures that acquiring and non-acquiring POEs have "identical" company characteristics just before their first cross-border M&A. If political connections facilitate cross-border deal completions, we expect that companies replacing their top management with more politically connected successors will be more likely to engage in cross-border M&A activities.
To measure this effect, we create a dummy variable (Political Turnover) that equals 1 if the CEO or chairman is replaced in the five years before the firm completed the first cross-border M&A deal with a CEO/chairman with a higher PC Index (stronger political ties), and 0 if there is no turnover or a turnover that does not result in a higher PC Index. 9 Our model reads as follows: The variable of interest in equation (3) is the coefficient on Political Turnover ( 1 ). If political connections indeed increase the likelihood of a Chinese POE going global, we expect 1 to be positive. We also perform a robustness check in which we replace Political Turnover with a variable that indicates the change in the political connections of the blockholders within the five years before the firm's first cross-border M&A (PBH Turnover).

The Financial Performance of Chinese POEs after Cross-border M&A Announcements
We begin our analysis of how the market reacts to cross-border M&A announcements by Chinese POEs by using a standard event study approach. Following Du and Boateng (2015), we use an event window of (-1, 1) and an estimation period of (-240, -21) relative to the first announcement date of an acquisition ( = 0). The cumulative average abnormal returns (CAARs) are calculated using a one-factor market model (employing the value-weighted Shanghai and Shenzhen Composite Index as a market index). For robustness, we also consider the event windows (0, 1), (-2, 2), and (-3, 3).
To examine the link between political connections and the market reaction to cross-border deal announcements, we estimate the following multivariate regression: where is the cumulative abnormal return of the acquiring POE during the period starting one day before and ending one day after the cross-border deal announcement. The independent variables are the same as in equation (1)  ( 1 ) to be negative.
To complete the picture, we also examine the accounting performance of POEs, measured by ROE, after completing a cross-border M&A transaction. However, the decisions to hire politically connected top executives and to become active in acquiring foreign companies are likely to be made simultaneously in an equilibrium setting. This raises a potential endogeneity concern, which ideally would be overcome by finding a suitable instrument. Unfortunately, we were not successful in finding or constructing a convincing instrument. Consequently, the coefficients can be interpreted as indicating correlation only.

21
We compare the financial performance of (politically connected) acquiring companies during the three-year period after completion of the cross-border M&A deal with the performance of nonacquiring POEs. The model is specified as follows: where the dependent variable is the of firm and is a dummy variable that equals 1 if POE completed a cross-border M&A between years and + 3 , and 0 otherwise.
Multi Acquirer is a dummy variable that equals 1 if firm acquires more than one overseas target starting in year when the second acquisition is completed until the end of the observation period, and 0 otherwise. All other variables are as defined in equation (1). Our main coefficient of interest is that for the interaction term, 3 . If politically connected POEs tend to incur moral hazard problems by engaging in political empire building, we expect 3 to be negative and statistically significant. In other words, we expect the effect of political connections on firm performance to decrease for POEs that have completed a cross-border M&A transaction.

Descriptive Statistics
Panel A of   POEs. However, we find no univariate evidence that the corporate governance (Gov Index) of acquisitive POEs is higher than that of non-acquisitive POEs (Hypothesis 2). We will explore this factor further in our multivariate analyses.
For the control variables, we find that acquisitive POEs are on average more profitable (ROA), larger (Size), and have lower growth opportunities (Tobin's q) than non-acquisitive POEs. In addition, we find that non-acquisitive firms have stronger ownership-based political connections, measured by the variables LBH Connection and Multi BH Connections. A lower level of state ownership may help acquisitive POEs in the sense that they could be perceived as less government connected. The correlation matrices in Table 4 show that the pairwise correlations are not greater than 0.5. To further unveil any potential multicollinearity issues, we also calculate the Variance Inflation Factors (VIF) in our multivariate regressions. In line with our bivariate correlation analysis, multicollinearity does not appear to pose any problems in a multivariate context.

Political Connections and M&A Engagement by POEs
To investigate the link between political connections and the probability of becoming acquisitive in international markets, we show the results of a fixed-effects panel logit regression in Table 5.
Our baseline results in column 1 indicate that politically connected POEs are more likely to acquire overseas companies. The coefficient of Connection is 1.474 and statistically significant at the 1% level. Column 2 shows the results when measuring political connections via the PC Index, which likewise supports Hypothesis 1, i.e. that the likelihood of completing a cross-border M&A transaction increases with the strength of political connections. However, we do not find any statistical support for Hypothesis 2, namely that a firm with sounder corporate governance is more likely to complete a cross-border M&A deal.
-Please insert Table 5 about here-Next, to have a one-to-one comparison, we perform cross-sectional logistic regressions based on a balanced matched sample of acquisitive and non-acquisitive POEs using the PSM method.
The results in Table 6 (Panels A and B) indicate that after matching, the firm characteristics of nonacquisitive POEs are not statistically different from those of acquisitive POEs. Thus, the sample is well-balanced.
Using this balanced sample, the baseline results in column 1 ( transactions if they have political connections. These POEs are also more likely to be able to manage the logistics of these transactions and to get preferential treatment by the government after completing a cross-border M&A transaction. This is supported by our sample firms receiving on average an 83% percent higher loan volume in the two years after completing a cross-border M&A deals than in the two years before (see Figure 1). Figure 1 about here-

Corporate Governance and M&A Engagement by POEs
To investigate the link between corporate governance and a POE's likelihood of acquiring an overseas target, we focus on the coefficient of the governance index (Gov Index). When performing post-matching cross-sectional analyses (see Table 7), this coefficient is positive and statistically significant, at least at the 10%-level, indicating that POEs with higher corporate governance standards are more likely to acquire companies outside China (in line with Hypothesis 2). However, the Gov Index was not significant in the panel logistic regression setting in Table 5. Thus, we do not find robust empirical support for Hypothesis 2.
To address a potential endogeneity issue associated with the decision to become an acquisitive POE, we conduct a quasi-experiment in which we focus on the replacement of a CEO or chairman by a successor with stronger political ties than his or her predecessor (i.e. a higher PC Index). We characterize these turnovers using the variable Political Turnover. If political connections result in a higher probability of acquiring a company outside China, we expect to find a higher likelihood of POEs entering the global markets after a political turnover. To ensure a balanced sample of acquisitive and non-acquisitive POEs, we run a similar PSM routine to that used previously, but also require the 288 firm-year observations (corresponding to the POEs' first cross-border deals) to have the same PC Index as those in the control group. The diagnostic tests from Table 8 show that the PSM successfully balances the sample.
We again run a logit regression with the dependent variable of becoming an overseas acquisitive POE and a set of explanatory variables that include the Political Turnover dummy. The results (see Table 9) show that the coefficient of Political Turnover is positive and statistically significant, indicating that the likelihood of a POE proceeding with a cross-border M&A significantly increases after a political turnover. This finding provides strong support for Hypothesis 1 and for a causal relationship between political connections and cross-border M&A activities by Chinese POEs.
-Please insert Tables 8 and 9 about here-

The Financial Performance of Multinational POEs after Cross-border M&As
To explore how POEs fare after completing a cross-border acquisition, we first provide univariate results for an event study in which we examine the stock returns of acquisitive POEs around the announcement of a cross-border M&A deal (see Table 10). We find that shareholders react positively to cross-border M&A announcements with statistically significant CAARs between 1% and 1.9%, depending on the event window. These findings are in line with Du and Boateng (2015) who find that shareholders react positively to cross-border M&A deals by Chinese acquirers.
Over similar event windows, their CAARs range from 0.45% to 0.64%. However, their study pools We complement these univariate findings with a multivariate analysis in which we control for deal characteristics as well as cultural differences between China and the country where the target company is domiciled (see Table 11). The results are consistent with the univariate analysis above.
The announcement returns are on average about 1.6% lower for politically connected POEs. 10,11 We interpret this as further support for Hypothesis 3, namely that investors may believe that a politically connected top management has other (e.g. political) motives when completing crossborder M&A transactions instead of focusing purely on shareholder wealth maximization. In additional analyses we tested indirectly if the acquisition is related to strategic asset seeking. To do 28 so, we compared the difference in Research Intensity between POEs that have completed a crossborder M&A and propensity-score-matched "control" firms that are not active in the acquisition market (see Proelss et al., 2017). We find that the average two-year Research Intensity of acquisitive POEs after deal completion is 1.3 percentage points higher than that of control firms (p-value = 0.008). We interpret this as evidence that POEs that gain access to "cutting-edge technology" through cross-border acquisitions need to increase their R&D spending to successfully employ that technology in China (see Wu, 2015).
The only deal characteristic that is statistically significantly related to the observed announcement returns is a deal payment by cash only (All Cash Deal). This positive relationship is well documented in the literature (see Travlos, 1987;Fishman, 1989;Brown and Ryngaert, 1991;Martin, 1996;and Fuller, Netter, and Stegemoller, 2002).

-Please insert Tables 10 and 11 about here-
To examine a POE's financial performance during the three-year period after it has completed a cross-border deal, we calculate the return on equity (ROE) for acquisitive and non-acquisitive POEs. Our main variable of interest is the interaction term × . The coefficients of Connection and the interaction term ( × ) are 0.054 and -0.198, respectively, and are both statistically significant at the 1% level (see Table 12). This indicates an underperformance of about 14 percentage points (0.054 -0.198), measured by ROE, of politically connected POEs relative to non-connected POEs during the three-year period after completing a cross-border M&A deal. Notwithstanding the potential endogeneity concern, this result is consistent with our political connection trade-off theory (and Hypothesis 3), under which politically connected top managers complete cross-border M&A deals largely as a means of 29 political empire building. This may occur at the expense of shareholder value, and may thus be associated with a decrease in the POE's firm value. The coefficient of is positive and significant at the 5%-level, which is consistent with our univariate evidence showing that, on average, investors react positively to cross-border M&A announcements. We also find that serial acquirers have statistically significantly higher accounting performance than one-time acquirers after completing a cross-border M&A deal, which could be explained by learning gains through serial acquisitions (cf., Aktas, Bodt and Roll, 2013). Table 12 about here-

Robustness Checks
Our first set of robustness checks focuses on an alternative explanation for the importance of  Table A2 in the appendix). Moreover, the coefficients for the top management political connection variables (Connection and PC Index) do not change substantially and remain statistically significant at least at the 5% level. Similarly, when we re-perform our quasi-experiment (see Table A3 in the appendix), we find that the replacement of a blockholder by a new blockholder with stronger ties is unrelated to the probability of completing a cross-border M&A transaction, unlike a political turnover of the top management (see Table 9). In sum, we find no evidence that ownership-based political connections increase the likelihood of a POE becoming acquisitive in foreign markets, while the political connections of top management continue to produce similar effects.
Furthermore, to address the potential influence of clustered or serial acquisitions by POEs, we checked for robustness by using a count model (i.e. panel and cross-sectional Poisson regressions) in which the dependent variable is the number of overseas targets acquired by POEs (see Table A4 in the appendix). Some of these model specifications also control for ownership-level political connections. We find that political connections (measured by Connection and the PC Index) are statistically significantly positively related to the number of completed cross-border M&A deals whereas ownership-level political connections show no association. Therefore, we do not find any evidence that the main results are driven by clustered or serial acquisitions.
Finally, we test for a potential interaction between political connections and a POE's corporate governance. Such an interaction might be expected if politically connected top managers tend to pursue cross-border M&A deals for reasons other than maximizing shareholder value, such as maximizing political capital. In this context, we conjecture that higher corporate governance standards within a company limit top management's propensity for political empire building at the cost of shareholder value. For example, we expect the interaction term (Connection x Gov Index) to have a positive coefficient in a regression of POEs' financial performance after a cross-border M&A. In unreported results, we include this interaction term in all previous analyses and find that is has no statistical significance, regardless of the dependent variable in question (e.g. the likelihood of a cross-border M&A or stock prices following cross-border M&A announcements). One possible explanation is provided by Claessens and Fan (2002) who argue that corporate governance 31 mechanisms have very limited effectiveness in systems with weak institutions. The arguably weak institutional environment in China seems to carry more weight than a sound corporate governance system; thus, the latter is neither able to block (value destroying) cross-border M&A deals nor deter politically connected top managers from prioritizing their political capital over shareholder interests.

Conclusion
This study investigates the factors that affect the likelihood and consequences ( (1) reports the results using Connection as a proxy for political connections; column (2) uses the PC Index as a measure for political connections. We report coefficient estimates with p-values in parentheses below. Industry and Year Fixed Effects are included in both regressions. In the last two rows, we report the maximum and mean variance inflation factors (VIF). ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.
(1)  This table reports the results of a propensity score matching (PSM) routine for acquiring and nonacquiring Chinese POEs from 2007 to 2016. We match firms using a nearest neighbor propensity score matching algorithm and an array of firm-specific characteristics (ROA, Leverage, Firm Size, Tobin's q, Tangibility) in the year the POE completes its cross-border deal. Panel A reports the univariate balanced test results for pairs of treatment and control firms after matching. Panel B reports parameter estimates for the probit model used in estimating the propensity scores of the treated and control groups (where the treatment is a cross-border acquisition). We match firms in the year before completing a cross-border M&A deal with non-acquiring POEs. The "Pre-Match" column contains the parameter estimates of the probit model estimated using the sample prior to matching. These estimates are then used to generate the propensity scores for matching acquiring and non-acquiring POEs. The "Post-Match" column contains the parameter estimates of the probit model estimated using the subsample of matched treatment-control pairs after matching. We match firms using a one-to-one nearest neighbor propensity score matching, without replacement. Definitions for all variables are provided in Panel A of Table A1 in the appendix. Industry and Year Fixed Effects are included in both regressions in Panel B. We report coefficient estimates with p-values in parentheses below. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

of Political Influence on Becoming Acquisitive-A Crosssectional Analysis
This table reports the results of a post-matching logit regression analysis for Chinese acquiring and nonacquiring POEs between 2007 and 2016. Acquiring POEs are defined as those with at least one crossborder M&A transaction within the sample period. Non-acquiring companies are the one-to-one nearest neighbors as defined in Table 6. The dependent variable is a dummy variable that equals 1 if the POE completes a cross-border M&A deal in a given year, and 0 otherwise. See equation (1) for details.
Columns (1) and (2) report the post-matching results using all cross-border deals; columns (3) and (4) report the post-matching results considering only the first cross-border deals for each acquiring POE. We report coefficient estimates with p-values in parentheses below. p-values are calculated using the clustered standard errors at the firm level for Columns (1) and (2). p-values are based on robust standard errors for columns (3) and (4). Industry and Year Fixed Effects are included in all regressions. In the last two rows, we report the maximum and mean variance inflation factors (VIF). ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.
(1)  This table reports the result of propensity score matching (PSM) for Chinese acquiring and non-acquiring POEs from 2007 to 2016. We match firms using a nearest neighbor propensity score matching algorithm and an array of firm-specific characteristics (ROA, Leverage, Firm Size, Tobin's q, Tangibility) plus the PC Index in the year before the acquiring POE completes its first cross-border merger. Panel A reports the univariate balanced test results for pairs of treatment and control firms after matching. Panel B reports parameter estimates for the probit model used in estimating the propensity scores of the treated and control groups (where the treatment is a cross-border M&A). We match firms in the year before completing a cross-border M&A transaction with non-acquiring POEs. The "Pre-Match" column contains the parameter estimates of the probit model estimated using the sample prior to matching. These estimates are then used to generate the propensity scores for matching acquiring and non-acquiring POEs. The "Post-Match" column contains the parameter estimates of the probit model estimated using the subsample of matched treatment-control pairs after matching. We match firms using a one-to-one nearest neighbor propensity score matching, without replacement. Definitions for all variables are provided in Panel A of   Panel A of this table reports the cumulative average abnormal returns (CAAR) around the announcement date of a cross-border M&A transaction by a Chinese POE for the event windows (-1, 1), (0, 1), (-2, 2), and (-3, 3). The CAARs are calculated using a one-factor market model (employing the value-weighted Shanghai and Shenzhen Composite Index as the market factor). The estimation period spans from 240 to 21 days before the announcement date (see, Du and Boateng, 2015). t-statistics and p-values are calculated using robust standard errors. Panel B shows the average difference in the cumulative abnormal returns for the event window (-1, 1) between politically connected and non-connected POEs. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

Observations 176
Event Window CAAR t-statistic p-value (-1, 1) 0.012*** 3.760 0.000 (0, 1) 0.010*** 3.560 0.000 (-2, 2) 0.016*** 4.150 0.000 (-3, 3) 0.019*** 3.780 0.000 Observations 226  (4)). The dependent variable is the CAR (Cumulative Abnormal Return) calculated using a one-factor market model (employing the value-weighted Shanghai and Shenzhen Composite Index as the market factor) over the event window (-1, 1). Specification (1) includes Connection and all control variables and specification (2) includes PC Index and all control variables; both models are for all cross-border M&A announcements. Specifications (1)' and (2)' are based on subsamples of the data and include only the CAR of the first cross-border M&A announcement for each POE. Therefore, the variable Multi Acquirer is not included in the model. The variable All Stock Deal is also not included, because there was no cross-border M&A transaction that was financed only with stocks within the subsample for first crossborder M&A announcement for each POE. Therefore, the variable Multi Acquirer is not included in the model. The variable All Stock Deal is also excluded, because none of the cross-border M&A transactions within the first-transaction subsamples were financed purely with stocks. All variables are defined in Table A1 in the appendix. We report coefficient estimates with p-values in parentheses below. p-values are calculated using robust standard errors at the firm level. Industry and Year Fixed Effects are included in all regressions. In the last two rows, we report the maximum and mean variance inflation factors (VIF). ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.
(1)   (5) for details. We report coefficient estimates with p-values in parentheses below. Industry and Year Fixed Effects are included in the regression. In the last two rows, we report the maximum and mean variance inflation factors (VIF). ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.
(1)    year t is less than the mean value of the sample in fiscal year t, and 0 otherwise (Waelchli and Zeller, 2013;Jiang and Kim, 2015). 15 Chairman tenure Number of years the company's chairman has been in office Equals 1 if the tenure of the chairman of firm i in fiscal year t is less than the mean value of the sample in fiscal year t, and 0 otherwise (Berger, Ofek, and Yermack, 1997;Jiang and Kim, 2015).

Board size Number of directors on the board of directors
Equals 1 if the board size of firm i in fiscal year t is less than the mean value of the sample in fiscal year t, and 0 otherwise (Yermack, 1996;Conyon and Peck, 1998;Core, Holthausen, and Larcker, 1999).

Board independence Number of independent directors on the board of directors
Equals 1 if the number of independent directors on the board of firm i in fiscal year t is greater than the mean value of the sample in fiscal year t, and 0 otherwise (Agrawal and Knoeber, 1996;Kim, Kitsabunnarat-Chatjuthamard, and Nofsinger, 2007).

Board meeting Number of annual meetings of the board of directors
Equals 1 if the number of annual meetings of the board of directors of firm i in fiscal year t is less than the mean value of the sample in fiscal year t, and 0 otherwise (Vafeas, 1999;Yi, Yu, and Jiang, 2011).

Supervisory board size
Number of supervisors on the supervisory board Equals 1 if the number of supervisors on the supervisory board of firm i in fiscal year t is greater than the mean value of the sample in fiscal year t, and 0 otherwise (Firth et al., 2007;Ding et al., 2010;Jia et al., 2009).

Ownership concentration
Percentage of shares held by the company's largest shareholder Equals 1 if the percentage of shares held by the company's largest shareholder of firm i in fiscal year t is greater than the mean value of the sample in fiscal year t, and 0 otherwise (Stiglitz, 1985;Rediker and Seth, 1995;Voulgaris, Stathopoulos, and Walker, 2010;Huang et al., 2011).

Foreign auditor
Hiring of a foreign auditor Equals 1 if firm i hires a foreign auditor in fiscal year t, and 0 otherwise (Gao and Kling, 2008;Peng, Wei, and Yang, 2011).
State shares State shares account for at least 5% of the firm's total shares Equals 0 if the state holds more than 5% of the shares in firm i in fiscal year t, and 1 otherwise (Bloom et al., 2012;Jiang, Huang, and Kim, 2013). 15 As Jiang and Kim (2015, pp 209) point out, using chairman age and tenure for constructing the corporate governance index for Chinese companies is appropriate because "the actual person who is actively in charge of the business is not the CEO. It is the board chairperson who actively controls and runs the firm. In China, this is common knowledge. However, based on the academic literature, it seems that many scholars are unaware of this."  (1) and (2)) and a postmatching cross-sectional logit regression analysis (columns (3) and (4)) for all cross-border acquiring and non-acquiring POEs between 2007 and 2016. Columns (5) and (6) repeat the post-matching cross sectional analyses, but only consider the first cross-border deal for each acquisitive POE. Acquiring POEs are defined as those with at least one cross-border M&A transaction within the sample period. The dependent variable is a dummy variable that equals 1 if the POE completes a cross-border M&A deal in a given year, and 0 otherwise. See equation (1) for details. To measure the political connections of blockholders, we employ the variables LBH Connection (the percentage ownership of the largest politically-connected blockholder) and Multi BH Connections (a dummy variable indicating whether a firm has multiple politically-connected blockholders). We report coefficient estimates with p-values in parentheses below. Industry and Year Fixed Effects are included in all regressions. In the last two rows, we report the maximum and mean variance inflation factors (VIF). ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. (1)

Acquisitive-A Quasi Experiment
This table reports the results of a logit regression analysis for acquiring and non-acquiring Chinese POEs between 2007 and 2016 after the change in the political connections of the blockholders within the five years before the firm's first cross-border M&A. The dependent variable is a dummy variable that equals 1 if the POE completes a cross-border M&A transaction, and 0 otherwise. Non-acquiring companies are identified via a one-to-one nearest neighbor propensity score matching algorithm. To measure the political connections of blockholders, we employ the variables LBH Connection (the percentage of the largest politically-connected blockholder) and Multi BH Connections (a dummy variable indicating whether a firm has multiple politically-connected blockholders). We report coefficient estimates with pvalues in parentheses below. p-values are calculated using robust standard errors. Industry and Year Fixed Effects are included in all regressions. In the last two rows, we report the maximum and mean variance inflation factors (VIF). ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.
(   (4)) and a post-matching Poisson regression analysis (columns (5) to (8)) for all crossborder deals. The sample period is between 2007 and 2016. Chinese acquiring POEs are defined as those with at least one cross-border M&A transaction within the sample period. The dependent variable is a count variable indicating the number of overseas targets acquired by Chinese POEs. See equation (2) for details. Columns (1) to (2) and (5) to (6) only use Connection or PC Index as the measure for political connections (of the management), while columns (3) to (4) and (7) to (8) also use LBH Connection and Multi BH Connections so as to also incorporate measures for political connections of the blockholders. We report incidence-rate ratios (IRR) together with p-values in parentheses below. Industry and Year Fixed Effects are included in all regressions. In the last two rows, we report the maximum and mean variance inflation factors (VIF). ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. (1)