Political connections with corrupt government bureaucrats and corporate
M&A decisions: A natural experiment from the anti-corruption cases in
China
Qigui Liu, Tianpei Luo and Gary Tian
School of Accounting, Economics and Finance, University of Wollongong
Abstract
Using the recent anti-corruption cases in China as a natural experiment, this study documents
that connections with the government officials through bribing and personal connecting
enable Chinese non-SOEs to increase their M&A activities, merge more local targets, pay
less M&A premium, and enjoy a better post-M&A performance before the corrupt
government official is removed in an anti-corruption case. However, after the event, their
M&A activities decrease significantly, i.e., they conduct less M&A, pay more M&A
premium, and have worse post-M&A performance. Our results are robust when we exclude
bribing firms and immune to the reputation concerns. Furthermore, the M&A decisions and
performance of corruption related non-SOEs vary across regions with different level of
corruption index and industries with different government support and competition. Overall,
our study provides direct evidence to the question: why firms seek to establish connections
with government officials through bribing or personal connecting, meanwhile, we reveal the
benefits and costs for such connections.
Key words: Political connections, anti-corruption, M&A, M&A performance
JEL classification: G32; G34; G38
Gary Gang Tian is corresponding author, his email address is [email protected]. Qigui Liu’s email address is [email protected]. Tianpei Luo’s email address is [email protected].
1
1. Introduction
Corruption is reported as a prevalent phenomenon around the world and becoming an
increasingly important concern for companies. Literature shows that corruption may bring
various benefits to corruption related firms such as improve firms’ efficiency (Beck and
Maher, 1986), reduce transaction costs (Cheung et al., 2012), better access to external
financial market (Fan et al., 2009) and thus increase firm value (Cheung et al., 2012; Karpoff
et al., 2010). It is not surprising that the benefits of corruption is more important in emerging
markets because corruption substitutes for a low quality of public governance and therefore
can reduce the negative effect of the low government quality (Meon and Sekkat, 2005). That
is why firms, especially those non-state owned enterprises (non-SOEs) in emerging markets
have strong incentive to bribe or connecting with government officials. In other word,
corruption is an important channel through which firms establish political connections and it
brings benefits to those corruption related firms. However, given the illegality of corruption,
the inherent uncertainty creates costs and may offset its value enhancing mechanisms
(Bardhan, 1997 and Meon and Sekkat, 2005). In particular, corruption related firms may lose
their connections with the corrupt bureaucrats after the corruption scandal is exposed,
resulting in an important political and operation risk to those firms. Therefore, an
investigation into how anti-corruption event has influence on corporate decisions of the
corruption related firms enable us better understand how political connections with corrupt
government bureaucrats through bribing or personal connecting influence corporate
decisions. Moreover, the anti-corruption event is exogenous to corporate decisions which
mean that it could be used as a natural experiment to avoid the potential endogeneity issue.
Literature on political connections shows that in emerging markets such as China where the
government still have controlling power on firms’ access to financial resources and
investment projects, manager’s political connections play an important role in helping firms
2
to get access to various financial resources such as bank loans and equity market accessing
(Faccio et al., 2006; Fan et al., 2008; Li et al., 2008 and Liu et al., 2013). However these
conventional literature always use the current or former working experience of managers as
government officials as a measure of firms’ political connections (Chen et al., 2011; Liu et
al., 2013) and implicitly conjecture that politically connected managers enable connected
firms to get benefits from the government through rent seeking behaviors. The drawbacks
are: (1) the definition of political connections suffers from endogeneity issue although it is
controlled; (2) it is argued that politically connected managers tend to bring benefits to
connected firms through rent seeking from the government regulations, but the channel
through which politically connected managers seek rent is not identified.
In order to solve these problems, this study hypothesizes that firms establish connections with
the government officials through bribing and personal connecting with bureaucrats, which
enable them to obtain good investment projects. To conduct our study, we identify the
corruption related firms and classify them into two types: (1) corruption related firms that
include: the listed firms whose senior managers or directors bribed the corrupt bureaucrats
(bribing firms), and the listed firms whose senior managers or directors are connected with
the misconducted bureaucrats through either former working experience or educational
experience (bureaucrats connected firms); and (2) unrelated firms that include: firms whose
senior managers or directors neither bribed nor connected with misconducted bureaucrats.
We expect that compared to unrelated firms, corruption related firms may receive benefits
from the M&A market by obtaining more quality targets especially from the local M&A
market that are under the control of the related government bureaucrats, paying lower M&A
premium and thus having a better post-M&A performance1. However, these benefits should
1 For instance, the controlling shareholder of Thaihot Group Co (000732), Huang Qisen, is identified as a briber and bribed the former secretary-general of the CPC in Fujian province, Chen Shaoyong, who was involved in a corruption scandal and was arrested in 2008. In 2004, Mr. Chen helped Mr. Huang to acquire the state-owned shares of Fujian Mindong Electric Power Limited Company. In return, Mr. Chen received a Patek Philippe watch as a gift from Mr. Huang which was worth for 80,000 Hong Kong dollars.
3
disappear after these firms lose their connections in an anti-corruption event. In particular we
would like to answer the question that how firms’ M&A decisions and performance change
after they lose their connections with the government in an anti-corruption event.
To avoid endogenous problems which are usually faced by cross section studies in M&As,
we follow Fan et al. (2008) and employ the high level bureaucrats’ corruption scandal cases
as an exogenous shock to test the effect of political connection on post-M&A performance in
China. Since the corruption scandals are not directly caused by corruption related firms, they
can effectively alleviate the potential endogeneity issues of political connections. To conduct
this natural experiment, we collect a sample of 29 provincial or higher level bureaucrats’
corruption cases in China during the period from 2005 to 2011 and investigate how the M&A
decisions and performance of corruption related firms are influenced by the anti-corruption
event compared to unrelated firms. This empirical design is less subject to endogeneity
problems than previous studies because the arrest of corrupt bureaucrats is exogenous to the
firms and is not directly caused by the firms (Fan et al., 2008). Moreover, the bureaucrats
connected firms may not actively participate in the corruption, while only get the connection
be chance. Therefore, any impacts on changes of post-M&A performance should not be
directly driven by the anti-corruption event but by losing their connections.
As the largest emerging market, the capital market in China provides an ideal institutional
environment to conduct our analysis. Firstly, China is recognized as a highly corrupt country.
According to the Corruption Perception Index (CPI) from Transparency International, China
ranks 75 out of 182 countries around the world in 2011. Moreover, La Porta et al. (2004)
document that China is among the worst countries in term of political freedom and the
protection of property rights. The more direct evidence form the official reports of the Central
Commission for Discipline Inspection of the Communist Party of China records that 643,759
corruption cases were under investigated, 639,068 corruption cases were concluded, 668,429
4
people were punished by the administrative and Party discipline agencies and 24,584 people
were transferred to the judicial organs for further investigations from the end of 2007 to 2012.
There were 81,391 commercial bribery cases which were under investigation and more than
20 billion Chinese Yuan were involved in these cases. This huge amount of corruption cases
not only provides opportunities to conduct our studies, but also reveals the importance and
prevalence of having connections with government officials in China. Furthermore, the
network (so-called ‘Guanxi’) is regarded as one of the most important and dominated factors
in business operations in China rather than formal contracts (Xin and Pearce, 1996). Under
this institutional environment, the impact of informal social connections such as political
connections is enhanced.
Secondly, Chinese government still has great direct intervention on firms’ M&A at the
various levels. For instance, while the Chinese economy has transformed from the
government controlled economy to the market-oriented economy, while local governments
still retain the rights to allocate various resources through licensing, granting the right of land
use and accessing to capital market (Cull and Xu, 2003; Firth et al., 2009; Li et al., 2009 and
Chen et al., 2011). Moreover, most of M&As conducted by Chinese listed firms must be
approved by the China Security Regulation Committee (CSRC) and even getting the
permissions from local governments if the M&A relate to a state owned enterprise (SOE).
Under this circumstance, Chinese firms would have strong incentive to build connections
with bureaucrats through bribing or hiring connected managers. These political connections
would facilitate firms getting the M&A approval and also create value when they conduct
M&A. Therefore, we expect that the firms which has bribed and connected with bureaucrats
would have superior post-M&A performance before misconducted bureaucrats get arrested
and their post-M&As performance would decline significantly after the event.
5
Lastly, in contrast to SOEs which in nature have a close relationship with the government,
non-SOEs are more likely to establish connections with the government through bribing due
to the continuing rule of Communist Party and ideological discrimination against non-SOEs.
Therefore, we expect that the anti-corruption event should have a more pronounced influence
on M&A decisions and performance of non-SOEs compared to SOEs.
Using a sample of 802 M&A announcements in Chinese public listed firms from 2005 to
2011, we conduct a series of empirical tests to provide evidence for our hypotheses. Our
results confirm that corruption related firms include bribing firms and bureaucrats connected
firms are more likely to conduct M&A than those unrelated firms before the anti-corruption
events and the likelihood of conducting M&A in these corruption related firms decline
significantly after the anti-corruption events. However, this phenomenon is only observed in
non-SOEs. Moreover, corruption related non-SOEs are more likely to conduct local M&A
(M&A under the corruption bureaucrats’ jurisdiction) and they pay less M&A premium
before the event, but they conduct less local M&A and pay more M&A premium after the
event. Consistently, the post-M&A performance of corruption related non-SOEs, as measured
by the short-term cumulative abnormal returns (CARs) around announcement date and long-
term buy-and-hold returns are significantly better than their unrelated counterparts before the
anti-corruption event. Nevertheless, the post-M&A performance of corruption related non-
SOEs decline significantly after the anti-corruption events. Overall, our results suggest that
Chinese non-SOEs do benefit from bribing and connecting with corrupt bureaucrats by
having better access to the M&A market, especially when they merger local target firms and
paying less M&A premium, which result in a good post-M&A performance.
Our robustness results further show that the impact that the anti-corruption event has on post-
M&A performance varies across regions and industries: the event has greater impact on
6
M&A in highly corrupt regions, government supported industries and low competitive
industries.
This study contributes to the literature in the following important ways. First, our study
extends the current literature on M&A decisions and performance. Previous studies in this
area mainly focus on the effect of firm specific and industry factors on M&A performance
(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). By using the anti-
corruption events in China as a natural experiment which greatly alleviate the endogeneity
issue, we provide direct evidence that in emerging markets with poor shareholder protection
and high government interventions, public sector governance has an important influence on
firms’ M&A decisions and performance. Therefore, we add new evidence to this strand of
literature.
Our study also contributes to the literature on political connections. In particular, previous
studies mainly focus on the effect of political connections on the firm value and the ability to
access to external financial market (Fisman, 2001; Johnson and Mitton, 2003; Khwaja and
Mian, 2005). We extend these studies by investigating how political connections affect firms’
M&A policy. More importantly, different from the traditional literature that measures
political connections using the managers who serve as a former or current government
bureaucrat, our study adopts a more direct measure of political connections, i.e., firms that
relate to the corrupted government officials either through direct bribing or personal
connecting. Furthermore, by showing that corruption related firms conduct more (less)
M&As, pay (lower) higher M&A premium, and have (better) worse post-M&A performance
before (after) the anti-corruption event, we provide direct evidence on the benefits and costs
of the political connections.
7
The remainder of the paper is organized as follows. Section 2 develops our hypotheses.
Section 3 introduces our data, sample, variables and empirical model employed. Section 4
presents the empirical results and interpretations. Section 5 summarizes and concludes this
paper.
2. Hypothesis development
2.1. The arrest of corrupt bureaucrats and corporate M&A decision
2.1.1. The arrest of corrupt bureaucrats and likelihood of conducting M&As
As discussed in the previous section, the Chinese capital market is known for its high degree
of government intervention and discrimination against private firms. Moreover, political
connections usually bring benefits to the connected firms, which create strong incentives for
non-SOEs to establish political connections with local government officials to obtain benefits
from the government and to overcome the ideological discrimination. Since Chinese M&A
market is highly regulated and intervened by the authorized government agencies, bribing or
connecting with the local government bureaucrats may help firms have better access to M&A
market and get the deals approved. However, when the related corrupt bureaucrats are under
investigation or arrested, such political connections would be broken. Thus, the preferential
access to M&A market would disappear. Therefore, we predict that the likelihood of
conducting M&As for corruption related non-SOEs would decrease significantly subsequent
to the arrest of misconducted bureaucrats.
Although we expect a negative effect of anti-corruption events on the likelihood of M&As in
non-SOEs, this effect is not clear in SOEs. Given the naturally close relationship between
SOEs and government, they do not need to bribe bureaucrats to seek rent from government.
Moreover, managers of SOEs usually do not hold shares of the company and their
compensation is not closely related to firm performance, so they do not have incentive to
8
bribe government officials for getting good M&A projects. Even they do so, it is more likely
for their personal benefits rather than shareholders’ interests2. Therefore, we argue that the
significant decrease in the likelihood of conducting M&A activities cannot be observed in
SOEs. Based on above analysis, we hypothesize that:
H1a: The likelihood of conducting M&A deals of corruption related (bribing or bureaucrats
connecting) non-SOEs would significantly decrease after the arrest of corrupt bureaucrats,
but this phenomenon cannot be observed in SOEs.
2.1.2. The arrest of corrupt bureaucrats and local M&A
As discussed above, corrupt bureaucrats help corruption related non-SOEs access to M&A
market. However the power of bureaucrats is often limited by the geographic distance, which
means that the corruption related firms would more likely to acquire the targets which are
allocated in their connected bureaucrats’ jurisdiction. Therefore, we expect that the
corruption related non-SOEs would more likely to conduct local M&As. However, their
abilities to access to local M&A market would reduce after the arrest of corrupt bureaucrats.
Therefore, we further hypothesize that:
H1b: The likelihood of conducting local acquisitions of corruption related (bribing or
connecting) non-SOEs would significantly decrease after the arrest of corrupt bureaucrats,
but this phenomenon cannot be observed in SOEs.
2.1.3. The arrest of corrupt bureaucrats and takeover premium
2 In 2003, the former governor of Shanghai, Chen Liangyu, pushed Huawen Media Investment Corporation (000793) which is a state owned enterprise to acquire Shanghai New Huang Pu Real Estate Co (600638) which is a SOE as well. This deal would benefit the chairman of Shanghai New Huang Pu Real Estate Co, Wu Minglei. Using the money form Shanghai New Huang Pu Real Estate Co, Mr. Wu helped Mr. Chen’s family to pay travelling expenses at 356,131.50 Chinese Yuan.
9
In the previous section, we argue that corruption related non-SOEs would have preference to
access M&A market. Another possible benefit that corruption related non-SOEs can obtain
through bribing or connecting with government bureaucrats is to acquire target firms by
paying a lower M&A premium. By bribing or connecting with local bureaucrats, the
acquirers may have the chance to know the true value of target firms and reduce the
information asymmetry in the deals. Moreover, through government interventions, the
corrupt bureaucrats may even influence the managers/shareholders of the target firms to
accept the offer from corruption related firms to merge their firms at a relatively lower price.
Thus, we argue that political connections with government bureaucrats would reduce the
takeover premium for corruption related non-SOEs. However, subsequent to the arrest of
misconducted local bureaucrats, corruption related non-SOEs would lose their connections
and have to pay a higher premium than before. Therefore, we hypothesize that the arrest of
corrupt bureaucrats would increase the takeover premium paid for non-SOEs, and similarly,
this impact would not affect takeover premium for SOEs. Thus our next hypothesis is:
H1c: The takeover premium for corruption related (bribing or connecting) non-SOEs would
significantly increase after the arrest of corrupt bureaucrats, but such outcome does not exist
in SOEs.
2.2. The arrest of corrupt bureaucrats and post-M&A performance
As discussed above, by relating to the government bureaucrats through either bribing or
connecting, non-SOEs may obtain good M&A opportunities but pay relative low M&A price,
and thus their M&A activities are more likely to be value-creating. But after corrupt
government bureaucrats are removed, the non-SOEs may conduct more value-destroying
M&A activities as they may lose the various benefits from the corrupt bureaucrats they
related. Therefore, it is expected that corruption related non-SOEs may have better post-
M&A performance before the anti-corruption event is announced because they are able to
10
avoid paying too much for the targets which is documented as the most common reason for
the acquirers’ losses (Cai and Sevilir, 2012). While such good performance should drop after
local bureaucrats was arrested. In SOEs, as discussed above, the purpose of bribing and
connecting with bureaucrats may not to maximize shareholders’ wealth, but to pursue
manager’s personal interests such as political promotion or career development. Therefore,
we expect that the M&A performance of corruption related non-SOEs would decrease
significantly after the arrest of corrupt bureaucrats, but this phenomenon cannot be observed
in SOEs. Therefore we develop our second hypothesis as follows:
H2: The post-M&A performance of corruption related (bribing or connecting) non-SOEs
would significantly drop after the arrest of corrupt bureaucrats, but this relationship cannot
be observed in SOEs.
3. Methodology and measurement of variables
3.1. The corrupt bureaucrat cases
To test our hypotheses, we manually collect a list of the corruption cases which include the
provincial level or higher bureaucrats in China and identify the listed firms whose senior
managers or directors have directly bribed or connected through past working and
educational experience with the corrupt government officials. As discussed earlier, we only
focus on the high level government officials’ corruption cases (provincial level and above) to
keep our study free from endogeneity issue. We employ the following process to collect the
corruption cases. First, we collet a list of corrupt government officers from 2005 to 2011
from the Central Commission for Discipline Inspection of the Communist Party of China and
three other government website: people.cn, huanqiu.com and xinhua.net. Those websites are
sponsored by the People’s Daily which is the official newspaper of the Chinese government.
We further make efforts to extend our list of corruption cases from other famous websites in
11
China, such as ifeng.com, sina.com.cn, and NetEase. As discussed, we only keep the high
profile government officers corruption cases in this study, which mostly include the
bureaucrats worked at provincial and central level governments. In order to identify the
bribing firms in our corruption cases, we only keep corruption cases which are reported in
detail by the Legislative Affairs Office of the State Council P. R. China or in the bulletin of
the supreme people’s procuratorate of the People’s Republic of China. These resources
usually report the detailed information about the people who have offered bribes to the
corrupt bureaucrats and the firms who have been illegally assisted by the corrupt bureaucrats.
Finally, we collect 29 high level corruption cases from 2005 to 2011.
For all of the 29 corruption cases, we go through the reports published by the Legislative
Affairs Office of the State Council P. R. China and the bulletin of the supreme people’s
procuratorate of the People’s Republic of China to identify any of the listed firms (bribing
firms) bribed the bureaucrats in question. We make more efforts to search through the above
three official government news agencies and other available news which discloses the
investigation about these cases to find out any other the listed firms’ senior managers or
directors involved in these corruption cases.
To identify the firms (bureaucrats connecting firms) that are connected with the corrupt
bureaucrats but are not bribers and involved in the corruption scandals, we search through
firms’ annual report and initial public offer prospectus and collect the profile of senior
managers and directors for each firm before the corruption event. Then we define the
connected firms which are the firms’ senior managers or directors were the family members
of or have previous job and education affiliations with the corrupt government officers. The
unrelated firms are the remaining listed firms which were neither bribed nor connected with
the corrupt bureaucrats.
12
We admit that we may miss some corruption cases and event firms in the procedures of
collection corruption cases and the processes to identify the bribing and connection firms.
However, we find that the information disclosure about the corrupt bureaucrats and firms
involved in these cases is similar among the corruption cases. Therefore, those missing data is
more likely to be random and less likely to influence our results. In Contrast to Fan et al.
(2008) we do not limit our research in the firms which locate in the corrupt bureaucrats’
jurisdiction, but focus on all listed firms in Shanghai and Shenzhen stock exchanges in order
to avoid missing some corruption related firms. However, we also construct a sample which
only includes firms located in the misconducted government officers’ jurisdictions. We find
our results still hold by using this sample. A list of the corruption events is reported in the
Appendix A.
3.2. Sample
The sample used in this paper includes the M&A deals conducted by public listed firms on
Shanghai and Shenzhen stock exchanges from 2005 to 2011. We use the CSMAR Chinese
Listed Firm M&A Database to obtain announcement dates, acquiring and target firms’
information 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. It includes
the Chinese listed firm annual report database from 2005 to 2011; the Chinese listed firm
corporate governance database from 2005 to 2011; the Chinese stock market trading database
from 2004 to 2011. The CSMAR database is one of the most important and widely used
databases in research on the Chinese capital market.
Following previous studies in M&A, we require M&A deals to meet 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
13
value is less than 1% of acquirer’s market capitalization (Cai and Sevilir, 2012; Masulis et al.,
2007). In order to make our calculation of M&A performance free from contamination from
other announcements, we further exclude the M&A announcements that the acquiring firms
announce two or more M&A deals within three months and the announcements that the
acquiring firms make other announcements during our event window. We require the
acquirer has annual financial statement information available (three years prior to acquisition
announcements and three years post to these announcements) and stock return data (250
trading days prior to M&A announcements) from CSMAR databases. Finally, we exclude the
deal which acquiring or target firm information, announcement date and financial data are
missing. After meeting these criteria, our final sample yields 802 M&A cases among totally
11,259 firm year observations.
3.3. Measurement of variables
3.3.1. Corruption
As discussed above, we measure the firms’ connections with government as the firms who
are briber or connected with corrupt bureaucrats. These corruption related firms (bribing
firms and bureaucrats connected firms) are defined as a dummy variable ‘CORRUPTION’
which equals one if the firms is a briber or connected firms with a corrupt government
officer. We further define a dummy variable ‘POST’ which equals to one if the M&A deal is
conducted after the corruption event. The emphasis of this study is on the interaction term
‘CORRP*POST’. This interaction term picks up the post corruption event changes in the
corruption related firm’s M&A decisions and performance relative to the unrelated firms. We
expected that this interaction would significantly negatively related to post-acquisition
performance and positively associated with takeover premium in non-SOEs, while it would
be insignificant in SOEs.
14
3.3.2. Other variables
A series of variables are constructed to measure firms’ M&A decision and post-M&A
performance. We first define a dummy variable ‘M&A’ which is equals to one if the firm
conducts an M&A in a given year to measure the likelihood of conducting an M&A.
The short term post-M&A performance is measured by the 1-, 2- and 5-days post-M&A
cumulative abnormal market-adjusted stock returns (CARs). We use Capital Asset Pricing
Model (CAPM) to find the announced firm predicted return during event windows for
adjustment in all our stock performance analyses. The market value weighted market index of
both Shanghai and Shenzhen stock exchanges are employed as market return in this model.
The estimate window is 250 trading days which start from 250 trading days prior to the
announcement. The long term performance is measured by the 1-, 2- and 3 years buy and
hold stock returns (BHARs).
We use two variables to measure the takeover premium. The ‘PREMIUM1’ is a relative
measure defined as the total dollar premium (measured by trading value minus fair value of
target firm’s equity value) relative to the fair value of target firm’s equity value. The
‘PREMIUM2’ is defined as the natural logarithm of the dollar premium.
We expect that connected firms are more likely to conduct local acquisitions which under the
corrupt bureaucrats control. Thus, we also collect the target firm location data and define a
dummy variable (LOCAL) which equals to 1 if acquiring and target firm are located in same
province.
To conduct our regression analysis, we also include various control variables in our
regression models to control factors which may affect M&A performance. The definitions of
these variables are reported in Appendix B in detail.
3.4. Regression model and research design
15
We use following equation as the baseline regression model to test our hypotheses:
M∧A /PREMIUM /LOCAL / Performance /¿ α0+β1× CORRUPTION i , t+β1× POST i , t+β1× CORP× POST i , t+β2× Control Variablesi , t +β3× Industry Dumm+ε ①
In the equation 1, the dependent variables are ‘M&A’ dummy variable; post-M&A
performance; M&A premium and Local M&A dummy. All control variables are as defined in
Appendix A. We also include a set of industry dummy variables to control for the industry
fixed effects.
To examine how corruption related firms’ M&A decisions and post-M&A performance
changes after the anti-corruption events, we firstly employ our full sample to run above
regression model. For the robustness check, we then repeat the main regression analyses, but
keep the sample firms which located in the corrupt government officers’ jurisdictions. In
order to eliminate the concern that our results are mainly driven by the bribing firms, we
thirdly remove the bribing firms from our sample and reanalysis the impact of anti-corruption
events on the M&A decisions and performance for bureaucrats connected firms only. Finally,
we limit our sample period to the years before the corruption scandals are exposed and test
the question whether corrupt government officers provide benefits to corruption related firms
when they conduct M&A.
4. Empirical results
4.1. Summary statistics and univariate tests
4.1.1 Summary statistics
Table 1 presents the summary statistics for our main variables. The results show that the
proportion of corruption related firms which bribed or was connected with corrupt
bureaucrats through personal relationship counts 22% of our firm year observations. The high
proportion of corruption related firms reveals the importance of being connected with local
governments in China. For the M&A performance of acquiring firms, our results show that
16
the acquiring firms’ shareholders earning slightly positive return (about 2%) around the
announcements, while relatively high long term post-acquisition performance after
conducting the acquisition deals. The proportion of local acquisition counts 54% out of the
802 acquisitions from 2005 to 2011. We also find that acquisition firms are willing to pay
high premium to get the controlling rights of the target which is demonstrated by the positive
value of PREMIUM1 (the average PREMIUM1 is 2.1, which means that firms pay on
average 2.1 times of the estimated fair value of target). In this study, the target value is
relatively small comparing to the bidding firms which is only about 30% of the acquirer’s
market value. In our sample, almost 87% of M&As in China are paid by cash which is much
higher than developed countries like U.S. Ishii and Xuan (2014) show that the pure cash deal
is only about 32% in U.S. from 1999 to 2007.
<Table 1>
4.1.2. Univariate test
The univariate test results are reported in table 2. The difference–in-difference method is used
to test the changes in M&A characteristics before and after the corruption events. The panel
A presents the results of the impact of corruption events on the likelihood of conducting
M&A. Panel B and C show the univariate test results for the effects of corruption events on
the takeover premium and the likelihood of conducting local M&A. The results on influences
of corruption events on post-M&A performance are reported in panel D. We report all
univariate test results in subsamples of SOEs and non-SOEs separately to illustrate the
different impact that the corruption event has on M&A decisions of SOEs and non-SOEs.
As shown in panel A, we find the likelihood of conducting M&As of corruption related non-
SOEs is significantly higher than their unrelated counterparts before the corruption scandals
are exposed, while the difference become insignificant after the arrest of corrupt bureaucrats.
17
Moreover, after the corruption events, corruption related non-SOEs significantly reduce the
likelihood of conducting M&As. However, corruption events do not significantly change the
probability of conducting M&As in SOEs. Overall, the results support our hypothesis H1a.
The results in panels B and C show that corruption related non-SOEs are more likely to
acquire local firms before the event and they pay lower takeover premium than unrelated
firms before corruption event (1.21 vs. 2.16). Moreover, we find after the arrest of corrupt
bureaucrats, the likelihood of corruption related non-SOEs to acquire a local firm drop
significantly and the takeover premium of corruption related non-SOEs is significantly higher
than unrelated non-SOEs. These results reveal that the corruption related non-SOEs lose the
ability to access local M&A market after they lose the connections with government officials
which confirm our hypothesis H1b and H1c.
In panel D of table 2, we find the difference of post-M&A performance between corruption
case related non-SOEs and unrelated non-SOEs are significant and positive before the arrest
of corrupt bureaucrats. After the corruption event, the post-M&A performance of related non-
SOEs reduces significantly compared to unrelated firms. These findings suggest that being
connected with corrupt bureaucrats, M&As conducted by corruption related firms perform
superiorly before the event, while the outstanding performance disappears following losing
their political connections after the event. The results in panel B confirms hypothesis H2.
<Table 2>
4.2 Regression analysis
We next conduct regression analysis to examine whether M&A decisions and post-M&A
performance of corruption related firms changes after the arrest of corrupt bureaucrats in our
expected manner when other factors that are identified to have influence on M&A decisions
are controlled.
18
4.2.1 Anti-corruption event and M&A decisions
4.2.1.1 Anti-corruption event and the likelihood of conducting M&A
Our regression analysis begins by investigating the impact of the anti-corruption events on
the likelihood of conducting M&A deals. Table 3 presents the result. We firstly focus on the
column 3 which reports the results for non-SOEs sample. As per our expectations, the
estimated coefficient of the corruption related firms ‘CORRUPTION’ is significantly
positively associated with the likelihood to conduct M&A for non-SOEs, suggesting that
corruption related non-SOEs conduct more M&As in general. We are particularly interested
in the interaction term ‘CORP*POST’ which demonstrates how the likelihood of corruption
related firms to conduct M&As changes after the arrest of corrupt bureaucrats. Our result
shows that the coefficient of interaction term, ‘CORP*POST’, is significant and negative in
non-SOE subsample. This result reveals that the likelihood of conducting M&A of corruption
related non-SOEs is substantially reduced after the corruption event.
In the SOEs subsample, we find that corruption related SOEs more actively participate in
M&A after the arrest of corrupt bureaucrats. One possible explanation for the results is that
SOE managers do not bribe for acquiring good M&A opportunity but for their personal
interest, such as political promotion or career development. Thus they do not conduct
significantly more M&As before the event than unrelated SOEs. After the arrest of corrupt
bureaucrats, these bribing managers in SOEs are usually investigated and replaced, which
reduces the managerial agency issues in those firms, and the newly appointed SOE managers
have strong incentive to conduct M&A and establish their personal reputation.
For robustness, we limit our sample to the firms which located in the corrupt bureaucrats’
jurisdictions around the anti-corruption events and the results are reported in column 4 to 6.
19
The results are consistent with our previous findings. Overall, our results in table 3 confirm
our hypothesis H1a.
<Table 3>
4.2.1.2 The impact of anti-corruption event on local acquisitions
Table 4 presents the results on the effect of corruption event on the corruption related firms’
M&A decisions. We expected that corruption related non-SOEs could more likely conduct
local M&A since the corrupt bureaucrats would have more power to influence commercial
activities under their jurisdictions. In columns 3 and 6 of table 4, we find in general
corruption related non-SOEs are more likely to conduct local M&As than unrelated
counterparts. More importantly, we demonstrate that corruption related non-SOEs’ ability to
acquire a local target is weakened after the arrest of corrupt bureaucrats. Given the acquirers
are more likely to take advantage of a nearby target’s resources and lead to a better post-
M&A performance. This result also supports our expectation on the better post-M&A
performance in corruption related non-SOEs.
Similarly the above results are not found in SOEs subsample. Therefore, the results in table 5
support our acquisition decision hypothesis H1b.
<Table 4>
4.2.1.3 The impact of anti-corruption event on takeover premium
In this subsection, we investigate the effect of corruption event on takeover premium. Our
aim is to shed the light on whether corrupt bureaucrats could help corruption related firms in
M&A deals by avoiding overpay to a target. Our expectation is that the anti-corruption events
would significantly increase the M&A premium for related firms, especially for non-SOEs.
The results are reported in Table 5.
20
The results in columns 3 and 6 of table 5 show that the estimated coefficient of
‘CORRUPTION’ is significantly negatively associated with takeover premium in non-SOEs
subsample. These results confirm that corruption related non-SOEs in general pay less to the
target in the M&As. In addition, the coefficient of interaction term, ‘CORP*POST’ in non-
SOEs subsample, is significantly and positively related to the takeover premium either we use
all listed firms (columns 1 to 3) or firms located in the bureaucrats’ jurisdictions (columns 4
to 6) in our tests. This result reveals that the capacity of corruption related firms to avoid
overpaying in M&A diminishes substantially after the anti-corruption event.
In the panel B of table 5, we redo our regression analysis by replacing the dependent variable
as the natural logarithm of the M&A premium. The regression results are similar to these in
Panel A. Overall, our results in table 5 demonstrate that by bribing or connecting with corrupt
bureaucrats, corruption related non-SOEs could avoid overpaying in acquisitions, while this
capacity would be weakened following the anti-corruption events. This results also support
the our argument that corruption related firms would have better performance in acquisitions
because paying too high for target is the major reason why acquirers usually lose in M&As
(Cai and Sevilir, 2012). Thus the results in table 5 are consistent with our hypothesis H1c.
<Table 5>
4.2.2 Anti-corruption event and post-M&A performance
Previous results show that the anti-corruption events significantly affect the M&A decisions
for corruption related non-SOEs. Their better capacity to avoid overpaying in M&A deals and
to access local M&A market could lead to a superior post-M&A performance. However, it is
not clear whether firms could have better M&A performance by bribing and connecting with
corrupt bureaucrats and how the anti-corruption events changes this relationship. The tables 6
21
and 7 present the regression results on the impact of anti-corruption event on the short term
and long term post-M&A performance, respectively.
Panel A of table 6 shows that the estimated coefficient of CORRUPTION is significantly
positively associated with post-M&A performance which is measure by the 3, 5 and 11 days
cumulative abnormal market adjusted return (CARs) in non-SOEs subsample. More
importantly, we find the interaction term, ‘CORP*POST’, is negative and statistically
significant in all non-SOEs regressions. These results reveal that the post-M&A performances
of corruption related non-SOEs are significantly negatively affected by the anti-corruption
event. Their post-M&A performances are substantially worse than unrelated non-SOEs after
they lose the connections with government bureaucrats. However, the impacts of anti-
corruption event on post-M&A performances do not have significant difference between
related SOEs and unrelated SOEs.
The results of long term post-M&A performances in Panel A of Table 7 confirm our short
term post-M&A performance results. We find that after the arrest of corrupt bureaucrats,
corruption related non-SOEs’ post-M&A long-term performance are significantly worse than
their counterparts. Again, we cannot make similar conclude from the SOEs subsample.
We repeat above regression analyses by using the corrupt bureaucrats’ jurisdiction sample
and are reported in Panel B of Tables 6 and 7. The results are similar with to those in Panel A
of these tables. Therefore, these results confirm our hypothesis H2.
<Table 6>
<Table 7>
4.3 Excluding the effect from bribing firms
So far, we have provided substantial evidence to support our main hypotheses that corruption
related non-SOEs conduct less M&A, pay higher M&A premium, are less likely to conduct
22
local M&A and have worse post-M&A performance after the anti-corruption event.
However, it is possible that the previous results are mainly driven by the bribing firms.
Therefore, it would be helpful to investigate whether our above findings can be explained by
losing political connections alone. In order to do so, we repeat the above regressions only
using the bureaucrats connecting firms who have connection with the corrupt bureaucrats, but
have not bribed the corrupt bureaucrats. These results are report in table 8.
In table 8, we present a set of regressions to test the impact of corruption event on the firms’
likelihood of conducting M&As (Panel A), likelihood to merge local target firms (Panel B),
takeover premium (Panel C), and the short-term and long-term post-M&A market
performance (Panels D and E). In this set of regressions, as bribing firms are excluded, the
dummy variable ‘CORRUPTION’ is defined as one if a firm’s senior managers or directors
were the family member of or have previous job and education affiliations with the corrupt
government officers, but did not bribe them. From the regression results reported in table 8,
we find in the non-SOEs subsample, the estimated coefficient of the interaction term,
‘CORP*POST’, is statistically significantly and negatively associated with the likelihood of
conducting M&As, the probability to acquire a local target and post-M&A performance,
while significantly and positively to the takeover premium. However, similar results cannot
be found in the SOEs subsample. These results demonstrate that by losing political
connections with the corrupt bureaucrats, bureaucrats related non-SOEs would conduct less
M&As, pay higher premium, less likely to acquire local targets and lose their preferential
performance in the deals. All of these results confirm our previous findings. As a robustness
check, we repeat above regression models in the sample which only include the listed firms in
the corrupt bureaucrats’ jurisdictions. The results are also consistent with above findings, but
we do not report these results to save the spaces. To provide more robust results, we also
23
limit our sample to those firms which have acquisition deals before and after the corruption
event, and similar results are found, the results are also not reported to save space.
<Table 8>
Overall, the results in above two sections confirm our main hypotheses that an anti-corruption
event has significant impact on corporate M&A decisions of Chinese non-SOEs in that
corruption related non-SOEs conduct less M&A, especially local M&A, they pay higher
M&A premium, and the M&A conducted by corruption related non-SOEs have worse post-
M&A performance after the arrest of bureaucrats. Our results hold even when the firms have
connections but do not bribe the corrupt bureaucrats, suggesting that the diminished M&A
performance in corruption related non-SOEs is not solely caused by the corruption events, but
it is a result of losing political connections with these corrupt bureaucrats.
4.4 Alternative interpretation
Although previous sections provide solid evidences for the impact of corruption event on the
M&A decision of corruption related firms, these results could have another interpretation.
After the arrest of corrupt bureaucrats, the market and regulatory government agencies would
have the detail information about the connected firms and bribers. Because of the reputation
concerns, the shareholders of quality target firms may refuse corruption related firms’ offers
and government agencies may be not willing to approve the M&A applications from
corruption related firms. Moreover, external financial market may also refuse to provide
financial supports for these firms with bad reputations. Therefore, the less likelihood of
conducting M&A and weaker post-M&A performance subsequent to the anti-corruption
event may be not driven by the loss of political connections, but the bad reputation.
To rule out the alternative interpretation, we further examine whether the connections with
government corrupt bureaucrats provide benefits to corruption related firms in M&As before
24
the misconducted bureaucrats were arrested (before our event). To conduct this test, we
repeat all our regression analyses but limit our sample period to the years before the
corruption scandal erupted. We regress our main dependent variables on the corruption
related firm dummy variable ‘CORRUPTION’ and control for other independent variables.
We expect that corruption related firms would have high probability to conduct M&As, are
more likely to conduct local M&A deals, pay lower M&A premium, and have better post-
M&A performance before the anti-corruption event. These results are reported in table 9.
As shown in table 9, in non-SOEs subsample, we find that the estimated coefficient of
‘CORRUPTION’ is significantly and positively associated with the likelihood of conduct
M&As (Panel A), the chances to conduct local M&A (Panel B) and short term and long term
post-M&A market performance (Panels D and E), while the estimated coefficient of
‘CORRUPTION’ is significantly negatively related to takeover premium in non-SOEs in
panel C. All of these results are consistent with our expectations and support our previous
arguments. Overall, these results suggest that corruption related non-SOEs are able to obtain
various benefits from the corrupt bureaucrats before the corruption event. Therefore, the
worse acquisition performance subsequent to corruption event is not entirely caused by the
alternative reputation interpretation.
<Table 9>
4.5 Robustness tests
In order to provide additional evidence to our results, we further conduct the following
robustness tests: (1) we investigate whether the impact of anti-corruption event on corporate
M&A decisions and performance differs between different regions and industries. China is
characterized as have great variation in terms of political and economic development in
different regions and industries. In particular, there is great variation in terms of corruption in
25
different provinces; and the level of government intervention on industry development and
market competition also differs in different provinces. Such variations provide us an ideal
setting to investigate whether corruption event has different impact on corporate M&A
decisions and performance of corruption related non-SOEs across different regions and
industries. (2) M&A is a specific type of corporate investment, thus we further investigate
how the anti-corruption event has impact on corporate investment policy and investment
efficiency in general. We only report our robustness results for non-SOEs to save space
because we have provided evidence that the anti-corruption event only have significant
impact on M&A decisions and performance of non-SOEs. In no reported results, we also
conduct all robustness tests on SOEs and no significant relationship is found.
4.5.1 The impact of anti-corruption event on corporate M&A performance across
regions
We first examine the effect of anti-corruption event on corporate M&A performance in
regions with different level of corruption. Following our expectation, corruption related non-
SOEs tend to obtain benefits from their related government bureaucrats through bribing or
connecting before the event and such benefits tend to disappear after the event, it is further
expected that the M&A performance in corruption related non-SOEs should decrease more
for firms operates in regions with high corruption. To conduct this test, we collect the number
of corruption cases from the Procuratorial Yearbook of China for each province from 2005 to
2011. Then we collect the provincial number of government officials who work in the public
administration and social organization departments form the China Statistical Yearbook and
China Population and Employment Statistics Yearbook from 2005 to 2011. Then we scale the
provincial number of corruption cases by the number of government officials who work in
the departments which related to public administration and social organization. We define
this ratio is the provincial corruption index. This index shows the level and importance of
26
corruption in each province. The higher corruption index means the more severe corruption in
a province. We thus rank the level of corruption in different provinces according to the
provincial corruption index. Then we divide our non-SOEs sample into two subsamples: non-
SOEs in regions with high (low) corruption index, and conduct regressions on the effect of
anti-corruption event on M&A decisions and performance in the two subsamples separately,
and repeat our regressions on the effect of anti-corruption event on post-M&A performance
of non-SOEs using the two subsamples. Table 10 presents our regression results. Not
surprisingly, our results show that the M&A performance of corruption related non-SOEs
operates in highly corrupt regions (regions with high corruption index) decrease significantly
more than the subsample of related non-SOEs operates in low corrupt regions (regions with
low corrupt regions), which supports our expectation.
<Table 10>
4.5.2 The impact of anti-corruption event on corporate M&A performance across
industries
In this subsection, we further investigate whether anti-corruption event have different impact
on corporate M&A performance of firms in different industries. We expect that the post-
M&A performance of corruption related non-SOEs should decrease more if the non-SOE
operates in: (1) government supported industries than non-government supported industries;
and (2) low competitive industries than high competitive industries. This is because (1) local
government officials tend to have more direct intervention on government supported
industries and low competitive industries. In particular, the Chinese government tends to
influence the industry development through five-year plans, and some industries are selected
to be supported industries in each plan and can obtain supports from local or central
government. Given that local government bureaucrats has the right to decide which firms
within the supported industries are eligible to receive government supports such as tax
27
refund, preference treatment in conducting M&A. The corruption related non-SOEs within
the supported industries should receive more benefits than their unrelated counterparts; (2)
another channel through which Chinese government exerts intervention on industry is issuing
license to firms and allowing them entering these restricted industries. This creates some
industries with very low competition since the government restriction and scare license
available. At the same time, other industries are highly competitive because firms are able to
freely access to these markets. Connecting with government through bribing or connecting
with bureaucrats may help corruption related non-SOEs access to some restricted industries
and allow them entering restricted industries through M&A, but such benefits may disappear
after firms lost such connections.
We therefore divide our non-SOEs into two pairs of subsamples: (1) government supported
industries versus non-government supported industries according to whether the firm operates
in a supported industry outlined in 11th five-year plan in a particular year; (2) low competitive
industries versus high competitive industries according to the industry competition. Then we
repeat our regressions on the effect of anti-corruption event on M&A performance using the
two pairs of subsamples.
Panels A and B of Table 11 present the regression results. As expected, the empirical results
support our expectation and suggesting that our main results are robust.
<Table 11>
4.5.3 Anti-corruption event, corporate investment and efficiency
In the previous sections, we show that the corruption case related non-SOEs are able to obtain
various benefits when conducting M&A such as better opportunities to access M&A market,
pay less M&A premium, are more likely to acquire a local target and better post-M&A
performance before the corruption event, but all those benefits attenuate after the government
28
bureaucrats are arrested. However, we still do not know how the anti-corruption event has
impact on corporate investment policy and investment efficiency. In order to provide a full
picture to our study, we further examine the impacts of corruption event on firms’ investment
and investment efficiency of Chinese non-SOEs from 2005 to 2011. Because of losing
connection with local government in the anti-corruption event, corruption related non-SOEs
would lose the preferential accessing to investment market and better investment projects.
Therefore, we expect that corruption related non-SOEs would have a significantly reduction
in their capital expenditure and investment efficiency after the arrest of corrupt. Tables 12
and 13 report the empirical results.
In table 12, corporate investment is measured by the variable ‘CAPEXTA’ which is cash
payments for fixed assets, intangible assets and other long term assets from the cash flow
statement less the cash receipts from selling these assets scaled by the total assets. We find
the interaction term ‘CORRUPTION t-1*POST t-1’ is significantly and negatively associated
with the corporate investment in non-SOEs in either full sample or the bureaucrats’
jurisdiction sample, and our results hold when we exclude bribing firms from our sample.
<Table 12>
In Table 13, corporate investment efficiency is measured by the sensitivity of corporate
investment and investment opportunity (Tobin’s Q). Not surprisingly, our results show that
the investment efficiency of corruption related non-SOEs decrease significantly after an anti-
corruption event.
<Table 13>
Overall, our robustness results provide further evidence that the impact of corruption event on
corporate M&A performance is greater in highly corrupt regions, government supported and
low competitive industries. We also demonstrate that the anti-corruption event has a negative
29
effect on corporate investment and investment efficiency. All those results support our main
argument that by bribing or connecting with local government bureaucrats, non-SOEs are
able to obtain various benefits when conducting M&A, while they lose all these benefits
when the related government official is arrested in an anti-corruption event.
5. Conclusion
This study examines how firms’ relationship with government bureaucrats through bribing or
connecting has impact on corporate M&A decisions and performance in Chinese listed firms.
By using Chinese bureaucrat corruption cases are used as a natural experiment, we manually
collect Chinese high level (provincial or higher level) government official corruption
scandals and identify the listed firms that bribed or connected with these bureaucrats. We find
that the corruption related (bribing or bureaucrats connecting) non-SOEs are able to obtain
various benefits such as have better access to M&A market, pay less M&A premium and
better access to local targets when conducting M&A, resulting in a better post-M&A
performance before the corruption scandals are exposure. However, all those benefits
attenuate subsequent to the arrest of the corrupt bureaucrats in an anti-corruption event. We
further provide evidence that the impact of the anti-corruption event on corporate M&A
decisions and performance of non-SOEs is greater in regions with high level of corruption
index, and in industries with government support and low competition. The anti-corruption
event is also identified to have negative effect on corporate investment and investment
efficiency of corruption related non-SOEs.
Overall, our findings suggest that in emerging markets with great government interventions
and ideological discrimination against non-SOEs, political connections through bribing or
personal connecting with government bureaucrats have great influence on M&A decisions
30
and post-M&A performance in non-SOEs. It provides direct evidence for the importance of
country level institutional factors on the determination of firms’ M&A decisions and post-
M&A performance without being influenced by endogeneity of political connections. In
particular, we document that firms in emerging markets tend to establish connections with the
government through bribing and connecting with government officials intently to maximize
shareholder value, but in the meantime such connections create risk to shareholders when the
connections are lost.
31
Reference
Agrawal, A., and Knoeber, C.R., 2001, Play a political role?, Journal of Law and Economics,
44, 179-198.
Cai, Y. and Sevilir, M., 2012, Board connections and M&A transactions, Journal of
Financial Economics, vol.103, pp.327-349
Chen, C.J., Li, Z., Su, X. and Sun, Z., 2011, Rent-seeking incentives, corporate political
connections, and the control structure of private firms: Chinese evidence, Journal of
Corporate Finance, vol.17, pp.229-243
Chen, S., Sun, Z., Tang, S. and Wu, D., 2011, Government intervention and investment
efficiency: Evidence from China, Journal of Corporate Finance, 17, 259-271
Cull, R., Xu, L., 2003, Who gets credit? The behaviour of bureaucrats and state banks in
allocating credit to Chinese state-owned enterprises, Journal of Development Economics,71
(2), 533–559.
Dong, M., Hirshleifer, D., Richardson, S., Teoh, S.H., 2006. Does investor misvaluation drive
the takeover market? Journal of Finance, 61, 725–762
Faccio, M. and Masulis, R.W., 2005, The choice of payment method in European mergers
and acquisitions, Journal of Finance, 60, 1345-1388.
Faccio, M., Masulis, R.W, and McConnell, J.J., 2006, Political connections and corporate
bailouts, The Journal of Finance, vol.61, no.6, pp.2597-2635
Faccio, M., 2006, Politically connected firms, The American Economic Review, vol.96, no.1,
pp.369-386
Fan, J.P.H., Rui, O.M. and Zhao, M., 2008, Public governance and corporate finance:
Evidence from corruption cases, Journal of Comparative Economics, vol.36, pp.343-364
32
Fan, J., Wei, J., Xu, X., 2011, Corporate finance and governance in emerging markets: a
selective review and an agenda for future research, Journal of Corporate Finance, 17, 207–
214.
Fisman, R., 2001, Estimating the value of political connections, American Economic Review,
91, 1095–1102.
Firth, M., Lin, C., Liu, P., Wong, S.M.L., 2009, Inside the black box: bank credit allocation
in China's private sector, Journal of Banking and Finance, 33 (6), 1144–1155.
Gaspar, J., Massa, M. and Matos, P., 2005, Shareholder investment horizons and the market
for corporate control, Journal of Financial Economics, 76, 135-165
Haveman, H.A., Jia, N., Shi, J. and Wang, Y., 2014, The dynamics of political embeddedness
in China: The case of publicly listed firms, working paper.
Ishii, J. and Xuan, Y., 2014, Acquirer-target social ties and merger outcomes, Journal of
Financial Economics, vol.112, pp.344-363
Johnson, S., Mitton, T., 2003, Cronyism and capital controls: evidence from Malaysia,
Journal of Financial Economics, 67, 351–382.
Khwaja, A., Mian, A., 2005, Do lenders favor politically connected firms? Rent provision in
an emerging financial market, The Quarterly Journal of Economics, 120 (4), 1371–1411.
La Porta, R., Lopez-de-Silanes, F., Pop-Eleches, C. and Shleifer, A., 2004, Judicial check and
balances, Journal of Political Economy, 112, 445-470
Li, H., Meng, L., Wang, Q. and Zhou, L., 2008, Political connections, financing and firm
performance: evidence from Chinese private firms, Journal of Development Economics,
vol.87, pp.283-299
33
Li, K., Yue, H., Zhao, L., 2009, Ownership, institutions, and capital structure: evidence from
China, Journal of Comparative Economics, 37 (3), 471–490.
Liu, Q., Tang, J. and Tian, G.G., 2013, Does political capital create value in the IPO market?
Evidence from China, Journal of Corporate Finance, 23, 395-413.
Masulis, R.W., Wang, C., and Xie, F., 2007, Corporate governance and acquirer returns,
Journal of Finance, 62 (4), 1851-1899.
Moeller, S.B., Scholingmann, F.P. and Stulz, R.M., 2004, Firm size and the gains from
acquisitions, Journal of Financial Economics, 73, 201-228.
Malmendier, U. and Tate, G., 2008, Who makes acquisitions? CEO overconfidence and the
market’s reaction, Journal of Financial Economics, vol.89, pp.20-43
Rhodes-Kropf, M., Robinson, D.T. and Viswanathan, S., 2005, Valuation waves and merger
activity: The empirical evidence, vol.77, pp.561-603
Rossi, S., and Volpin, P. F., 2004, Cross-country determinants of mergers and acquisitions,
Journal of Financial Economics, 74, 277-304
Xin, K.l., and Pearce, J.l., 1996, Guanxi: connections as substitutes for formal institutional
support, Academy of Management Journal, 39, 1641-1658.
34
Table 1 Summary statistics
Variables Name count Mean St. Dev. 25% 50% 75%CORRUPTION 802 0.221 0.415 0.000 0.000 0.000CAR (-1, 1) 802 0.021 0.098 -0.024 0.006 0.045CAR (-2, 2) 802 0.024 0.107 -0.032 0.012 0.057CAR (-5, 5) 802 0.023 0.139 -0.043 0.008 0.072BHAR1 770 0.212 0.813 -0.135 0.050 0.332BHAR2 591 0.469 2.095 -0.106 0.200 0.559BHAR3 433 0.763 2.141 -0.043 0.341 1.076LOCAL 802 0.544 0.498 0.000 1.000 1.000PREMIUM1 341 2.165 5.575 -0.013 0.000 0.111PREMIUM2 341 23.044 0.068 23.026 23.026 23.026RELATIVE SIZE 802 0.304 1.505 0.006 0.020 0.073CASH PAYMENT 802 0.863 0.344 1.000 1.000 1.000SIZE 802 21.901 1.418 21.010 21.822 22.637Q 802 1.923 1.417 1.104 1.407 2.119OPCFTA 802 0.048 0.098 -0.000 0.050 0.096LEVERAGE 802 0.553 0.337 0.395 0.550 0.681BOARDSIZE 802 9.122 1.940 9.000 9.000 9.000BOARDIND 802 0.365 0.056 0.333 0.333 0.375OWNERHIP 802 0.016 0.071 0.000 0.000 0.000
35
Table 2 Univariate tests
The ‘Corruption related firms’ (Unrelated firms) represents this firms which have (have not) bribed or connected with the local government officers. The ‘Before’ and ‘After’ represent the period before the arrest of corrupt bureaucrats and after the arrest of corrupt bureaucrats respectively. Definition of variables are detailed in Appendix B.
Panel A The impact of corruption event on the likelihood of conducting acquisition
Corruption related Firms Unrelated firms Difference test
Mean Median Mean Median t value z valueSOEs Before 0.058 - 0.060 - -0.002 -
After 0.112 - 0.065 - 0.047*** --0.054*** - -0.005 - -0.049** -
Non-SOES Before 0.157 - 0.064 - 0.093*** -
After 0.096 - 0.079 - 0.017 -0.061** - -0.015* - 0.076*** -
Panel B The impact of corruption event on the acquisition decisions (local M&A)
Corruption related Firms Unrelated firms Difference test
Mean Median Mean Median t value z valueSOEs Before 0.519 - 0.597 - -0.078 -
After 0.478 - 0.514 - -0.036 -0.041 - 0.083 - -0.042 -
Non-SOES Before 0.809 - 0.476 - 0.333*** -
After 0.471 - 0.534 - -0.063 -0.338*** - -0.058 - 0.396*** -
Panel C The impact of corruption event on the takeover premium
Corruption related Firms Unrelated firms Difference test
Mean Median Mean Median t value z valueSOEs Before -0.001 0.000 2.144 0.000 -2.145 0.000
After 1.756 0.000 1.068 0.000 0.688 0.000-1.757 0.000 1.076* 0.000 -2.833
Non-SOES Before 1.213 0.000 2.155 0.000 -0.942 0.000
After 7.416 0.000 2.912 0.000 4.504** 0.000*-6.203*** 0.000* -0.757 0.000 -5.446**
36
Panel D The impact of corruption event on post-acquisition performance
SOEs subsample
Corruption related Firms Unrelated firms Difference test
Mean Median Mean Median t value z valueCAR (-1, +1) Before 0.007 0.004 0.011 -0.001 -0.004 0.005
After 0.003 0.008 0.025 0.006 -0.022** 0.0020.004 -0.004 -0.014 -0.007 0.018
CAR (-2, +2) Before 0.019 0.031 0.011 0.004 0.008 0.027
After 0.002 0.006 0.028 0.013 -0.026* -0.0070.017 0.025 -0.017 -0.009 0.034
CAR (-5, +5) Before 0.028 0.022 0.001 0.002 0.027 0.020
After 0.001 0.001 0.027 0.008 -0.026 -0.0070.027 0.021 -0.026 -0.006 0.053
Non-SOEs subsample
Corruption related Firms Unrelated firms Difference
testMean Median Mean Median t value z value
CAR (-1, +1) Before 0.086 0.027 0.027 0.011 0.059** 0.016*
After -0.02 -0.001 0.021 0.009 -0.041*** -0.010***
0.106 0.028 0.006 0.002 0.100***CAR (-2, +2) Before 0.09** 0.039*** 0.03 0.005 0.060*** 0.034***
After -0.029 -0.021 0.031 0.021 -0.060*** -0.042***
0.119*** 0.06*** -0.001 -0.016 0.120***CAR (-5, +5) Before 0.042 0.066 0.033 0.003 0.009 0.063*
After -0.046 -0.02 0.044 0.014 -0.09*** -0.034***
0.088** 0.086*** -0.011 -0.011 0.099**
37
Table 3 Regression results for the impacts of corruption events on the likelihood of M&As This table presents logistic regression results on the impact of corruption events on the likelihood of conducting M&As by corruption case related firms. The ‘FULL’ sample includes all listed firms in Shenzhen and Shanghai stock exchange and the ‘JURISDICTION’ sample only include the firms whose under corrupt bureaucrats’ jurisdictions. The dependent variable ‘M&A’ is binary where 1 signifies that the firm makes at least one merger bid that is eventually successful in a given year. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. ‘SIZE’ is the natural logarithm of total assets. Control variables are defined in Appendix B. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
FULL JURISDICTIONALL SOEs Non-SOEs ALL SOEs Non-SOEsM&A M&A M&A M&A M&A M&A
CORRUPTION 0.53*** -0.00 1.05*** 0.59*** -0.22 1.27***(0.00) (0.99) (0.00) (0.00) (0.45) (0.00)
POST 0.05 -0.07 0.17 0.14 0.04 0.29*(0.57) (0.59) (0.17) (0.19) (0.78) (0.07)
CORP*POST -0.18 0.56** -0.94*** -0.21 0.72** -1.05***(0.31) (0.04) (0.00) (0.34) (0.04) (0.00)
SIZE 0.30*** 0.29*** 0.42*** 0.34*** 0.32*** 0.49***(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Q 0.19*** 0.12** 0.23*** 0.15*** 0.09 0.18***(0.00) (0.02) (0.00) (0.00) (0.18) (0.00)
OPCFTA 0.40 0.11 0.68 0.82* 0.19 1.26**(0.27) (0.86) (0.19) (0.07) (0.81) (0.04)
LEVERAGE 0.00 0.04 0.00* 0.03*** -0.45 0.04***(0.18) (0.84) (0.09) (0.00) (0.17) (0.00)
BOARDSIZE -0.06*** -0.03 -0.08** -0.09*** -0.06* -0.08**(0.01) (0.25) (0.03) (0.00) (0.07) (0.04)
BOARDIND -0.14 -0.71 -0.02 -1.44 -2.28* -0.54(0.85) (0.48) (0.99) (0.10) (0.08) (0.68)
OWNERSHIP -0.70 -11.13 -0.89 -0.35 -17.43 -0.50(0.19) (0.21) (0.11) (0.52) (0.19) (0.39)
INTERCEPT -9.08*** -8.61*** -11.87*** -9.58*** -8.43*** -14.24***(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
INDUSTRY YES YES YES YES YES YESN 11259 6327 4875 8376 4524 3826pseudo R-sq 0.04 0.03 0.07 0.04 0.03 0.08
Table 4 Regression results for the impacts of corruption events on the likelihood of conducting a local M&As
38
This table presents logistic regression results on the impact of corruption events on the likelihood of conducting local M&As. Annual observation of the corruption case related firms and control firms are included in the regressions. The corruption case related firms are defined as the firm whose senior managers or director bribed the bureaucrat in question or were connected with the bureaucrat through past job affiliations and educational experience. The control firms are firms whose neither bribed nor connected with the bureaucrat in question. The ‘FULL’ sample includes all listed firms in Shenzhen and Shanghai stock exchange and the ‘JURISDICTION’ sample only include the firms whose under corrupt bureaucrats’ jurisdiction. The dependent variable ‘LOCAL’ is binary where 1 signifies that the firm merger a local target. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Definition of control variables is in Appendix B. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
FULL JURISDICTIONALL SOEs Non-SOEs ALL SOEs Non-SOEsLOCAL LOCAL LOCAL LOCAL LOCAL LOCAL
CORRUPTION 0.64** -0.52 1.80*** 1.15*** 0.40 1.61***(0.02) (0.23) (0.00) (0.00) (0.53) (0.00)
POST -0.05 -0.40 0.37 0.01 -0.15 0.14(0.75) (0.12) (0.16) (0.96) (0.64) (0.65)
CORP*POST -0.84** 0.32 -1.76*** -1.27*** -0.59 -1.51**(0.02) (0.55) (0.00) (0.01) (0.42) (0.03)
SIZE -0.09 0.00 -0.22* -0.13 0.11 -0.27*(0.24) (1.00) (0.07) (0.17) (0.48) (0.05)
Q -0.10 -0.02 -0.17* -0.05 -0.17 -0.08(0.13) (0.85) (0.06) (0.60) (0.28) (0.51)
OPCFTA 0.84 -0.15 2.12* 0.95 1.16 1.46(0.31) (0.92) (0.05) (0.33) (0.52) (0.24)
LEVERAGE -0.27 -1.39** -0.04 -0.52 -2.38*** -0.41(0.33) (0.02) (0.66) (0.20) (0.01) (0.42)
BOARDSIZE 0.00 -0.01 0.07 0.02 -0.08 0.10(0.98) (0.89) (0.39) (0.71) (0.34) (0.23)
BOARDIND -0.09 0.83 -2.09 0.57 0.87 -1.36(0.95) (0.68) (0.36) (0.75) (0.72) (0.63)
OWNERSHIP 0.65 -34.23 1.21 0.63 -12.79 0.90(0.56) (0.22) (0.32) (0.60) (0.66) (0.50)
INTERCEPT 1.73 0.61 3.82 2.74 0.42 -8.23(0.30) (0.79) (0.20) (0.21) (0.89) (0.99)
INDUSTRY YES YES YES YES YES YESN 800 417 376 583 280 300pseudo R-sq 0.03 0.04 0.09 0.04 0.05 0.10
Table 5 Regression results for the impact of corruption events on M&A premium
39
This table presents the regression results for the impact of corruption events on firms’ takeover premium. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. The ‘CORP*POST’ is the interaction between ‘CORRUPTION’ and ‘POST’. Definition of control variables is in Appendix B. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
Panel A. The dependent variable ‘PREMIUM1’ is defined as the ratio of M&A premium to estimated fair value of equity of target firm.
FULL JURISDICTIONALL SOEs Non-SOEs ALL SOEs Non-SOEsPREMIUM1
PREMIUM1
PREMIUM1
PREMIUM1
PREMIUM1
PREMIUM1
CORRUPTION -2.720** -2.206 -2.475 -3.722*** -3.719 -3.784*(0.02) (0.27) (0.14) (0.01) (0.13) (0.08)
POST -1.141 -1.311 -0.168 -0.998 -0.800 -1.657(0.11) (0.14) (0.89) (0.28) (0.39) (0.36)
CORP*POST 4.415*** 3.315 6.477*** 5.438*** 2.969 7.789***(0.00) (0.15) (0.01) (0.00) (0.28) (0.01)
RELATIVE SIZE 0.002 2.150* 0.003 -0.001 2.208* 0.002
(0.35) (0.09) (0.50) (0.75) (0.08) (0.85)CASH PAYMENT -0.412 1.144 -0.928 -0.979 0.470 -0.033
(0.61) (0.33) (0.53) (0.29) (0.68) (0.99)SIZE 0.923*** 0.275 1.934*** 1.019*** 0.234 2.436***
(0.00) (0.49) (0.00) (0.01) (0.56) (0.00)Q 1.039*** -0.202 1.683*** 1.375*** 0.304 2.209***
(0.00) (0.64) (0.00) (0.00) (0.51) (0.00)OPCFTA 10.799*** 4.828 8.822* 9.626*** 2.021 9.353*
(0.00) (0.38) (0.05) (0.01) (0.72) (0.07)LEVERAGE -1.204 -0.205 -1.276 0.763 3.683 -0.722
(0.37) (0.92) (0.52) (0.74) (0.11) (0.86)BOARDSIZE -0.310 -0.108 -0.415 -0.503** -0.218 -0.426
(0.10) (0.66) (0.25) (0.02) (0.33) (0.31)BOARDIND -10.810* -4.772 -21.064* -10.280 -2.264 -13.566
(0.05) (0.46) (0.06) (0.11) (0.71) (0.32)OWNERSHIP 7.470* 4.795 10.097** 7.844* 69.185 9.573*
(0.09) (0.97) (0.05) (0.08) (0.58) (0.10)
INTERCEPT -12.405* -3.242 -29.074** -9.791 -0.284 -45.795***
(0.06) (0.72) (0.02) (0.29) (0.98) (0.01)INDUSTRY YES YES YES YES YES YESN 341 184 157 243 117 126adj. R-sq 0.16 0.01 0.32 0.23 0.10 0.33
40
Panel B. The ‘PREMIUM2’ is defined as the natural logarithm of the difference between trading value of the target and the target estimated value.
FULL JURISDICTIONALL SOEs Non-SOEs ALL SOEs Non-SOEsPREMIUM2
PREMIUM2
PREMIUM2
PREMIUM2
PREMIUM2
PREMIUM2
CORRUPTION -0.045*** -0.034* -0.041* -0.060*** -0.083** -0.043(0.00) (0.08) (0.05) (0.00) (0.01) (0.11)
POST -0.012 -0.025*** 0.011 -0.011 -0.032** 0.012(0.15) (0.01) (0.49) (0.37) (0.01) (0.61)
CORP*POST 0.044** 0.033 0.078** 0.073*** 0.089** 0.083**(0.02) (0.15) (0.01) (0.00) (0.02) (0.03)
RELATIVE SIZE 0.000 0.046*** -0.000 0.000 0.063*** -0.000
(0.39) (0.00) (0.71) (0.72) (0.00) (0.51)CASH PAYMENT -0.043*** -0.020* -0.042** -0.045*** 0.008 -0.046*
(0.00) (0.09) (0.02) (0.00) (0.60) (0.06)SIZE 0.012*** 0.007* 0.023*** 0.013*** 0.005 0.027***
(0.00) (0.09) (0.00) (0.01) (0.34) (0.00)Q 0.013*** -0.007 0.024*** 0.012** -0.003 0.023***
(0.00) (0.11) (0.00) (0.01) (0.67) (0.00)OPCFTA 0.092** 0.098* 0.071 0.098* 0.134* 0.098
(0.02) (0.07) (0.21) (0.05) (0.07) (0.13)LEVERAGE -0.012 -0.048** 0.010 -0.010 -0.013 0.036
(0.43) (0.02) (0.67) (0.75) (0.66) (0.49)BOARDSIZE 0.001 0.002 0.001 -0.001 -0.003 -0.000
(0.79) (0.44) (0.91) (0.74) (0.30) (0.97)BOARDIND -0.067 0.021 -0.271** -0.026 0.083 -0.205
(0.31) (0.74) (0.05) (0.76) (0.29) (0.23)OWNERSHIP 0.057 0.094 0.112* 0.064 -0.057 0.143*
(0.27) (0.94) (0.08) (0.29) (0.97) (0.05)INTERCEPT 22.846*** 22.921*** 22.684*** 22.862*** 23.004*** 22.478***
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)INDUSTRY YES YES YES YES YES YESN 341 184 157 243 117 126adj. R-sq 0.21 0.27 0.34 0.17 0.40 0.29
41
Table 6 Regression results for the impact of corruption events on Firms’ short term post-M&A performance. This table present the regression results for the impact of corruption events on firm’s short term post-M&A performance. Panel A include full sample of observations, while panel B only include M&A cases that are under the corrupt bureaucrats’ jurisdiction. The dependent variable is post-M&A performance measured by CARs for three-day, five-day and eleven-day window (CAR (-1, 1), CAR (-2, 2) and CAR (-5, 5)). The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. The ‘CORP*POST’ is the interaction between ‘CORRUPTION’ and ‘POST’. Definition of control variables is in Appendix B. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
Panel A. Full sample ALL SOEs Non-SOEsCAR (-1, 1) CAR (-2, 2) CAR (-5, 5) CAR (-1, 1) CAR (-2, 2) CAR (-5, 5) CAR (-1, 1) CAR (-2, 2) CAR (-5, 5)
CORRUPTION 0.038*** 0.043*** 0.014 0.005 0.021 0.025 0.053*** 0.059*** 0.015(0.00) (0.00) (0.40) (0.70) (0.26) (0.30) (0.00) (0.00) (0.49)
POST 0.012 0.015* 0.017 0.012 0.014 0.008 0.009 0.014 0.025(0.10) (0.06) (0.10) (0.15) (0.19) (0.55) (0.45) (0.23) (0.12)
CORP*POST -0.065*** -0.081*** -0.066*** -0.017 -0.032 -0.035 -0.120*** -0.139*** -0.108***(0.00) (0.00) (0.00) (0.30) (0.15) (0.22) (0.00) (0.00) (0.00)
RELATIVE SIZE 0.010*** 0.015*** 0.022*** 0.045*** 0.075*** 0.062*** 0.012*** 0.017*** 0.023***(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
CASH PAYMENT -0.076*** -0.095*** -0.096*** -0.081*** -0.091*** -0.095*** -0.041** -0.061*** -0.068***(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.04) (0.00) (0.01)
SIZE 0.005 0.008** 0.012*** 0.008** 0.009** 0.009 -0.000 0.007 0.023***(0.13) (0.01) (0.01) (0.03) (0.05) (0.13) (0.94) (0.13) (0.00)
Q 0.008** 0.010*** 0.019*** 0.007* 0.007 0.011* 0.006 0.011** 0.026***(0.01) (0.00) (0.00) (0.07) (0.18) (0.08) (0.22) (0.01) (0.00)
OPCFTA 0.034 0.051 -0.020 0.061 0.059 0.061 0.027 0.045 -0.079(0.34) (0.17) (0.68) (0.19) (0.33) (0.44) (0.60) (0.35) (0.23)
LEVERAGE -0.012 -0.026** -0.056*** -0.013 -0.030 -0.119*** -0.012 -0.023* -0.035**(0.22) (0.01) (0.00) (0.50) (0.23) (0.00) (0.37) (0.07) (0.04)
BOARDSIZE -0.004** -0.004** -0.000 -0.004* -0.005* -0.004 -0.006 -0.004 0.004
42
(0.02) (0.05) (0.94) (0.06) (0.07) (0.27) (0.10) (0.26) (0.40)BOARDIND -0.024 -0.044 -0.077 -0.118* -0.115 -0.101 0.044 0.027 0.074
(0.69) (0.50) (0.37) (0.06) (0.17) (0.34) (0.68) (0.79) (0.59)OWNERSHIP 0.060 0.044 -0.026 -0.052 -0.322 -1.164 0.043 0.031 -0.006
(0.19) (0.36) (0.69) (0.94) (0.72) (0.32) (0.45) (0.56) (0.93)INTERCEPT 0.159** 0.049 -0.191* 0.026 0.026 0.046 0.355*** 0.043 -0.708***
(0.02) (0.51) (0.05) (0.73) (0.79) (0.72) (0.01) (0.73) (0.00)INDUSTRY YES YES YES YES YES YES YES YES YESN 802 802 802 419 419 419 383 383 383adj. R-sq 0.20 0.25 0.22 0.25 0.23 0.14 0.24 0.32 0.33
43
Panel B. Bureaucrats’ jurisdiction sampleALL SOEs Non-SOEsCAR (-1, 1)
CAR (-2, 2)
CAR (-5, 5)
CAR (-1, 1)
CAR (-2, 2)
CAR (-5, 5)
CAR (-1, 1)
CAR (-2, 2)
CAR (-5, 5)
CORRUPTION 0.004 0.005 -0.007 -0.016 -0.022 -0.013 0.019 0.025 0.014(0.74) (0.72) (0.72) (0.38) (0.36) (0.67) (0.16) (0.16) (0.57)
POST -0.002 0.000 0.004 -0.008 -0.008 -0.026* 0.006 0.010 0.031*(0.81) (0.99) (0.76) (0.43) (0.52) (0.09) (0.54) (0.43) (0.09)
CORP*POST -0.026* -0.028 -0.022 0.015 0.034 0.030 -0.059*** -0.077*** -0.078**(0.06) (0.13) (0.36) (0.48) (0.23) (0.38) (0.00) (0.00) (0.03)
RELATIVE SIZE 0.000 0.000 0.000 0.077*** 0.119*** 0.145*** 0.000 0.000 -0.000(0.29) (0.61) (0.83) (0.00) (0.00) (0.00) (0.12) (0.42) (0.57)
CASH PAYMENT -0.106*** -0.129*** -0.134*** -0.081*** -0.096*** -0.088*** -0.092*** -0.108*** -0.112***(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
SIZE 0.004 0.006 0.004 0.008* 0.009 0.006 0.001 0.006 0.008(0.22) (0.13) (0.39) (0.08) (0.14) (0.36) (0.86) (0.33) (0.30)
Q 0.011*** 0.017*** 0.024*** 0.006 0.007 0.014* 0.014*** 0.023*** 0.034***(0.00) (0.00) (0.00) (0.22) (0.23) (0.07) (0.00) (0.00) (0.00)
OPCFTA 0.044 0.074* 0.027 0.107** 0.141** 0.122 0.018 0.038 -0.016(0.15) (0.07) (0.61) (0.05) (0.05) (0.15) (0.63) (0.43) (0.81)
LEVERAGE -0.013 -0.010 -0.007 0.026 0.032 -0.012 -0.023 -0.017 0.013(0.28) (0.56) (0.73) (0.33) (0.35) (0.78) (0.11) (0.39) (0.64)
BOARDSIZE -0.002 -0.002 0.002 -0.006** -0.005 -0.002 0.001 0.001 0.004(0.17) (0.35) (0.44) (0.02) (0.10) (0.66) (0.76) (0.82) (0.36)
BOARDIND 0.004 -0.016 -0.048 -0.106 -0.111 -0.011 0.088 0.052 -0.089(0.94) (0.83) (0.63) (0.16) (0.26) (0.93) (0.30) (0.65) (0.57)
OWNERSHIP 0.073** 0.076 0.021 0.489 1.117 0.646 0.059 0.064 0.051(0.05) (0.12) (0.74) (0.57) (0.32) (0.64) (0.12) (0.21) (0.47)
INTERCEPT 0.062 0.028 0.015 0.013 0.019 -0.013 0.012 -0.163 -0.254
44
(0.37) (0.76) (0.90) (0.89) (0.88) (0.93) (0.92) (0.31) (0.25)INDUSTRY YES YES YES YES YES YES YES YES YESN 585 585 585 282 282 282 303 303 303adj. R-sq 0.27 0.23 0.18 0.30 0.30 0.24 0.34 0.32 0.27
45
Table 7 Regression results for the impact of corruption events on event firms’ long term post-M&A performance. This table present the regression results for the impact of corruption events on firm’s long term post-M&A performance. Annual observation of the corruption case related firms and control firms are included in the regressions. Panel A include full sample of observations, while panel B only include M&A cases that are under the corrupt bureaucrats’ jurisdiction. The corruption case related firms are defined as the firms whose senior managers or director bribed the bureaucrat in question or were connected with the bureaucrat through past job affiliations and educational experiences. The control firms are listed firms whose neither bribed nor connected with the bureaucrat in question. The dependent variable is post-M&A long term market performance measured by buy and hold stock return for one year, two years and three years window (BHAR1, BHAR 2 and BHAR3). The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. The ‘CORP*POST’ is the interaction between ‘CORRUPTION’ and ‘POST’. Definition of control variables is in Appendix B. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
Panel A. Full sample ALL SOE Non-SOEBHAR1 BHAR2 BHAR3 BHAR1 BHAR2 BHAR3 BHAR1 BHAR2 BHAR3
CORRUPTION 0.115 0.074 0.257 -0.158 -0.032 0.363 0.283** 0.409** 0.478*(0.26) (0.79) (0.38) (0.31) (0.95) (0.49) (0.04) (0.03) (0.06)
POST -0.077 -0.083 -0.067 0.017 0.020 -0.084 -0.107 0.074 0.062(0.26) (0.69) (0.79) (0.86) (0.96) (0.83) (0.29) (0.64) (0.79)
CORP*POST -0.098 -0.017 -0.339 0.241 0.091 -0.117 -0.278 -0.546* -1.273**(0.47) (0.97) (0.47) (0.21) (0.90) (0.87) (0.18) (0.10) (0.01)
RELATIVE SIZE -0.001*** -0.002*** -0.128 -0.319** -0.790* 0.329 -0.000 0.000 0.054(0.00) (0.00) (0.23) (0.02) (0.09) (0.73) (0.45) (0.82) (0.49)
CASH PAYMENT -0.413*** -0.577** -0.315 -0.575*** -0.935** -0.272 -0.332** -0.264 0.110(0.00) (0.03) (0.32) (0.00) (0.04) (0.60) (0.03) (0.22) (0.73)
SIZE -0.051* -0.361*** -0.495*** -0.163*** -0.689*** -0.741*** 0.037 -0.212*** -0.288**(0.09) (0.00) (0.00) (0.00) (0.00) (0.00) (0.44) (0.01) (0.01)
Q -0.001 -0.154** -0.150 0.024 0.025 0.166 0.027 -0.074 -0.276***(0.98) (0.04) (0.15) (0.58) (0.88) (0.36) (0.42) (0.16) (0.00)
OPCFTA 0.264 -0.928 -0.914 1.260** -3.930* -3.300 -0.257 -0.372 -0.238(0.42) (0.35) (0.43) (0.02) (0.05) (0.14) (0.53) (0.56) (0.80)
LEVERAGE 0.420*** 1.037*** 2.833*** 1.454*** 5.446*** 7.364*** 0.094 -0.062 0.176
46
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.47) (0.74) (0.65)BOARDSIZE 0.004 0.042 0.060 -0.002 0.105 0.108 0.041 0.015 -0.006
(0.83) (0.43) (0.30) (0.92) (0.25) (0.25) (0.15) (0.72) (0.91)BOARDIND -0.826 0.198 -0.420 -0.379 1.435 -0.631 -0.661 0.012 -0.804
(0.14) (0.91) (0.83) (0.60) (0.64) (0.84) (0.44) (0.99) (0.68)OWNERSHIP -0.498 -1.049 0.769 -2.051 13.504 9.733 -0.528 -1.642* 1.269
(0.23) (0.52) (0.83) (0.79) (0.62) (0.77) (0.24) (0.06) (0.59)INTERCEPT 2.118*** 8.192*** 10.187*** 4.300*** 12.070*** 12.205*** -0.520 4.997*** 7.155***
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.65) (0.01) (0.01)INDUSTRY YES YES YES YES YES YES YES YES YESN 769 590 432 401 324 243 368 266 189adj. R-sq 0.06 0.04 0.12 0.16 0.15 0.25 0.04 0.03 0.04
47
Panel B. Bureaucrats’ jurisdiction sampleALL SOE Non-SOEBHAR1 BHAR2 BHAR3 BHAR1 BHAR2 BHAR3 BHAR1 BHAR2 BHAR3
CORRUPTION 0.196* 0.498*** 0.469** -0.197 0.277 0.003 0.326* 0.549** 0.523(0.08) (0.00) (0.01) (0.19) (0.21) (0.99) (0.06) (0.02) (0.11)
POST -0.130* 0.033 -0.056 -0.105 0.041 -0.066 -0.126 0.038 0.016(0.08) (0.77) (0.69) (0.20) (0.76) (0.66) (0.32) (0.85) (0.95)
CORP*POST -0.231 -0.546** -0.682** 0.170 -0.410 -0.079 -0.314 -0.595 -1.292**(0.11) (0.01) (0.02) (0.33) (0.13) (0.80) (0.20) (0.13) (0.02)
RELATIVE SIZE -0.001* -0.000 -0.018 -0.081 -0.043 -0.246 -0.000 -0.000 0.026(0.07) (0.79) (0.89) (0.52) (0.82) (0.47) (0.28) (0.91) (0.89)
CASH PAYMENT -0.283*** -0.155 0.097 -0.284** -0.036 -0.009 -0.335* -0.227 0.243(0.00) (0.27) (0.62) (0.01) (0.83) (0.97) (0.06) (0.42) (0.54)
SIZE -0.006 -0.199*** -0.296*** -0.002 -0.055 -0.156** 0.037 -0.345*** -0.452***(0.84) (0.00) (0.00) (0.96) (0.36) (0.02) (0.50) (0.00) (0.00)
Q 0.068** -0.034 -0.116 -0.002 -0.004 -0.047 0.093** -0.106 -0.315**(0.02) (0.47) (0.12) (0.97) (0.96) (0.57) (0.04) (0.19) (0.03)
OPCFTA -0.192 -0.210 -0.501 0.810* 0.667 0.595 -0.493 -0.465 -0.496(0.55) (0.68) (0.46) (0.08) (0.37) (0.49) (0.29) (0.54) (0.67)
LEVERAGE 0.243* 0.033 0.380 0.129 -0.327 -0.191 0.209 0.008 0.501(0.08) (0.87) (0.34) (0.56) (0.37) (0.67) (0.29) (0.98) (0.50)
BOARDSIZE -0.010 0.014 0.021 -0.026 -0.024 -0.002 0.019 0.034 0.015(0.57) (0.63) (0.55) (0.20) (0.49) (0.95) (0.55) (0.49) (0.83)
BOARDIND -0.969 0.557 0.266 -1.376** -0.984 -1.362 -0.527 1.219 0.911(0.11) (0.60) (0.85) (0.03) (0.41) (0.32) (0.63) (0.53) (0.77)
OWNERSHIP -0.485 -1.506* 1.898 -0.497 1.315 3.746 -0.474 -1.960* 0.289(0.20) (0.07) (0.32) (0.94) (0.90) (0.74) (0.33) (0.06) (0.91)
INTERCEPT 0.699 3.977*** 6.613*** 1.068 1.853 4.820*** -0.483 7.067*** 9.985***(0.34) (0.00) (0.00) (0.18) (0.14) (0.00) (0.75) (0.00) (0.00)
INDUSTRY YES YES YES YES YES YES YES YES YES
48
Table 8 Anti-corruption event and M&A decisions: excluding the effect from bribing firms
Panel A The likelihood of conducting M&AThis table presents logistic regression results on the impact of corruption events on the likelihood of conducting M&As by corruption connected firms. Bribing firms are excluded in all regression models. The dependent variable ‘M&A’ is binary where 1 signifies that the firm makes at least one merger bid that is eventually successful in a given year. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a connected with the corrupt bureaucrat through past job affiliations and educational experiences, but never bribe the corrupt bureaucrat, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; P-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOEs Non-SOEsM&A M&A M&A
CORRUPTION 0.74*** 0.14 1.27***(0.00) (0.53) (0.00)
POST 0.04 -0.07 0.18(0.62) (0.54) (0.16)
CORP*POST -0.29 0.52* -1.10***(0.13) (0.06) (0.00)
Control variables YES YES YESINDUSTRY YES YES YESN 10916 6108 4751pseudo R-sq 0.04 0.03 0.08
Panel B Local M&A This table presents the regression results for the impact of corruption events on the likelihood for conducting local M&A. The dependent variable ‘LOCAL’ is binary where 1 signifies that the firm merger a local target. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a connected with the corrupt bureaucrat through past job affiliations and educational experiences, but never bribe the corrupt bureaucrat, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOEs Non-SOEsLOCAL LOCAL LOCAL
CORRUPTION 0.80*** -0.21 1.91***(0.01) (0.64) (0.00)
POST -0.05 -0.41 0.40(0.78) (0.10) (0.13)
CORP*POST -0.89** 0.31 -1.95***(0.02) (0.58) (0.00)
INDUSTRY YES YES YESN 777 403 367pseudo R-sq 0.03 0.04 0.09
50
Panel C Takeover premiumThis table presents the regression results for the impact of corruption events on firms’ takeover premium. The dependent variable ‘PREMIUM1’ is defined as the ratio of M&A trading premium to fair value of equity of target firms. The ‘PREMIUM2’ is defined as the logarithm of the total M&A premium measured by the difference between sum of trading value of the target and the estimated value. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a connected with the corrupt bureaucrat through past job affiliations and educational experiences, but never bribe the corrupt bureaucrat, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space . Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOEs Non-SOEs ALL SOEs Non-SOEsPREMIUM1
PREMIUM1
PREMIUM1
PREMIUM2
PREMIUM2
PREMIUM2
CORRUPTION -2.871** -2.233 -2.432 -0.048*** -0.039* -0.044***
(0.01) (0.30) (0.14) (0.00) (0.07) (0.01)POST -1.141 -1.324 -0.133 -0.011 -0.025*** 0.007
(0.11) (0.14) (0.91) (0.13) (0.01) (0.55)CORP*POST 4.979*** 3.512 6.570*** 0.047*** 0.035 0.076***
(0.00) (0.15) (0.01) (0.00) (0.15) (0.00)Control variables YES YES YES YES YES YES
INDUSTRY YES YES YES YES YES YESN 334 180 154 334 180 154adj. R-sq 0.17 0.01 0.32 0.36 0.27 0.49
51
Panel D Short term post-M&A performanceThis table presents the regression results for the impact of anti-corruption event on firm’s short-term post-M&A performance of corruption connected firms. Bribing firms are excluded in all regression models. The dependent variable is post-M&A performance measured by CARs for three-day, five-day and eleven-day window (CAR (-1, 1), CAR (-2, 2) and CAR (-5, 5)). The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a connected with the corrupt bureaucrat through past job affiliations and educational experiences, but never bribe the corrupt bureaucrat, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients. P-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOEs Non-SOEsCAR (-1, 1) CAR (-2, 2) CAR (-5, 5) CAR (-1, 1) CAR (-2, 2) CAR (-5, 5) CAR (-1, 1) CAR (-2, 2) CAR (-5, 5)
CORRUPTION 0.042*** 0.043*** 0.011 0.004 0.014 0.011 0.058*** 0.060*** 0.017(0.00) (0.00) (0.48) (0.77) (0.48) (0.65) (0.00) (0.00) (0.45)
POST 0.013* 0.018** 0.022** 0.013 0.016 0.010 0.012 0.017 0.030*(0.07) (0.03) (0.03) (0.11) (0.13) (0.44) (0.36) (0.16) (0.06)
CORP*POST -0.072*** -0.087*** -0.065*** -0.023 -0.039* -0.036 -0.127*** -0.138*** -0.097***(0.00) (0.00) (0.00) (0.20) (0.10) (0.23) (0.00) (0.00) (0.01)
Control variables YES YES YES YES YES YES YES YES YESINDUSTRY YES YES YES YES YES YES YES YES YESN 779 779 779 405 405 405 374 374 374adj. R-sq 0.21 0.27 0.25 0.26 0.25 0.15 0.25 0.34 0.36
52
Panel E Long term post-M&A performanceThis table presents the regression results for the impact of anti-corruption event on firm’s long-term post-M&A performance of corruption connected firms. Bribing firms are excluded in all regression models. The dependent variable is post-M&A long term market performance measured by buy and hold stock return for one year, two years and three years window (BHAR1, BHAR 2 and BHAR3). The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a connected with the corrupt bureaucrat through past job affiliations and educational experiences, but never bribe the corrupt bureaucrat, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients. P-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOE Non-SOEBHAR1 BHAR2 BHAR3 BHAR1 BHAR2 BHAR3 BHAR1 BHAR2 BHAR3
CORRUPTION 0.102 0.084 0.308 -0.063 0.116 0.592 0.209 0.365* 0.442*(0.33) (0.77) (0.31) (0.70) (0.84) (0.30) (0.13) (0.06) (0.08)
POST -0.065 -0.067 -0.009 0.030 0.054 -0.035 -0.092 0.087 0.101(0.34) (0.76) (0.97) (0.75) (0.88) (0.93) (0.35) (0.58) (0.67)
CORP*POST -0.099 -0.010 -0.454 0.144 0.016 -0.410 -0.262 -0.586* -1.216**(0.49) (0.98) (0.36) (0.48) (0.98) (0.60) (0.23) (0.08) (0.02)
Control variables YES YES YES YES YES YES YES YES YESINDUSTRY YES YES YES YES YES YES YES YES YESN 746 570 420 387 310 233 359 260 187adj. R-sq 0.07 0.05 0.12 0.17 0.16 0.26 0.08 0.09 0.03
53
Table 9 Regression results for the impacts of corruption events on M&A decision before the corruption event
Panel A The likelihood of conducting M&AsThis table presents logistic regression results on the impact of corruption events on the likelihood of conducting M&As before the corruption events. The dependent variable ‘M&A’ is binary where 1 signifies that the firm makes at least one merger bid that is eventually successful in a given year. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. Observations only include samples before the corruption event. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOEs Non-SOEsM&A M&A M&A
CORRUPTION 0.60*** 0.03 1.24***(0.00) (0.88) (0.00)
Control variables YES YES YESINDUSTRY YES YES YESN 5737 3497 2228pseudo R-sq 0.05 0.02 0.12
Panel B Local M&AsThis table presents the regression results for the impact of corruption events on firms’ M&A decisions before the corruption events. The dependent variable ‘LOCAL’ is binary where 1 signifies that the firm merger a local target. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. Observations only include samples before the corruption event. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOEs Non-SOEsLOCAL LOCAL LOCAL
CORRUPTION 0.63** -0.78* 1.78***(0.03) (0.09) (0.00)
Control variables YES YES YESINDUSTRY YES YES YESN 378 208 163pseudo R-sq 0.04 0.10 0.15
54
Panel C Takeover premiumThis table presents the regression results for the impact of corruption events on firms’ takeover premium before the corruption events. The dependent variable ‘PREMIUM1’ is defined as the ratio of total takeover premium to fair value of equity of the target firm. The ‘PREMIUM2’ is defined as the log of the total takeover premium measured by the difference between sum of trading value of the target and the estimated fair value. Observations only include samples before the corruption event. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOEs Non-SOEs ALL SOEs Non-SOEsPREMIUM1
PREMIUM1
PREMIUM1
PREMIUM2
PREMIUM2
PREMIUM2
CORRUPTION -2.268** -2.257 -2.557* -0.041*** -0.016 -0.041*(0.03) (0.32) (0.06) (0.00) (0.53) (0.07)
Control variables YES YES YES YES YES YES
INDUSTRY YES YES YES YES YES YESN 148 88 60 148 88 60adj. R-sq 0.30 0.14 0.48 0.35 0.27 0.44
55
Panel D Short term post-M&A performanceThis table present the regression results for the impact of corruption events on firm’s short term post-M&A performance before the corruption events. The dependent variable is post-M&A performance measured by CARs for three-day, five-day and eleven-day window (CAR (-1, 1), CAR (-2, 2) and CAR (-5, 5)). The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. Observations only include samples before the corruption event. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space . Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOEs Non-SOEsCAR (-1, 1) CAR (-2, 2) CAR (-5, 5) CAR (-1, 1) CAR (-2, 2) CAR (-5, 5) CAR (-1, 1) CAR (-2, 2) CAR (-5, 5)
CORRUPTION 0.038*** 0.045*** 0.021 0.003 0.022 0.023 0.053** 0.054** 0.017(0.01) (0.00) (0.23) (0.82) (0.22) (0.33) (0.04) (0.01) (0.53)
Control variables YES YES YES YES YES YES YES YES YESINDUSTRY YES YES YES YES YES YES YES YES YESN 379 379 379 208 208 208 171 171 171adj. R-sq 0.19 0.28 0.20 0.26 0.28 0.22 0.23 0.31 0.31
Panel E Long term post-M&A performanceThis table present the regression results for the impact of corruption events on firm’s long term post-M&A performance before the corruption events. The dependent variable is post-M&A long term market performance measured by buy and hold stock return for one year, two years and three years window (BHAR1, BHAR2 and BHAR3). The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. Observations only include samples before the corruption event. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
ALL SOE Non-SOEBHAR1 BHAR2 BHAR3 BHAR1 BHAR2 BHAR3 BHAR1 BHAR2 BHAR3
CORRUPTION 0.104 0.049 0.236 -0.188 0.091 0.287 0.241 0.377* 0.598**(0.43) (0.89) (0.51) (0.32) (0.89) (0.67) (0.21) (0.09) (0.02)
Control variables YES YES YES YES YES YES YES YES YESINDUSTRY YES YES YES YES YES YES YES YES YESN 365 332 269 197 177 140 168 155 129adj. R-sq 0.05 0.05 0.12 0.29 0.31 0.33 -0.03 0.01 0.06
56
Table 10 The impact of corruption event on short-term post-M&A performance of non-SOEs in regions with high and low corruption indexThis table present the regression results for the impact of corruption events on firm’s short term post-M&A performance in regions with high and low corruption index. The dependent variable is post-M&A performance measured by CARs for three-day, five-day and eleven-day window (CAR (-1, 1), CAR (-2, 2) and CAR (-5, 5)). The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
High Low High Low High LowCAR (-1, 1)
CAR (-1, 1)
CAR (-2, 2)
CAR (-2, 2)
CAR (-5, 5)
CAR (-5, 5)
CORRUPTION 0.068** 0.020 0.068*** 0.034 -0.000 0.031(0.02) (0.23) (0.01) (0.11) (0.99) (0.24)
POST 0.010 0.003 0.028 -0.001 0.042 0.013(0.67) (0.82) (0.15) (0.93) (0.12) (0.46)
CORP*POST -0.154*** -0.081*** -0.171*** -0.099*** -0.104* -0.087**(0.00) (0.00) (0.00) (0.00) (0.06) (0.01)
F-test: equal CORP*POST -0.073* -0.072** -0.017
(0.07) (0.04) (0.25)Control variables YES YES YES YES YES YESINDUSTRY YES YES YES YES YES YESN 196 187 196 187 196 187adj. R-sq 0.24 0.35 0.35 0.38 0.34 0.44
57
Table 11 The impact of corruption event on short-term post-M&A performance of non-SOEs across industryThis table present the regression results for the impact of corruption events on firm’s short term post-M&A performance of non-SOEs that operate in supported or non-supported industries (panel A), high competitive or low competitive industries (panel B). The dependent variable is post-M&A performance measured by CARs for three-day, five-day and eleven-day window (CAR (-1, 1), CAR (-2, 2) and CAR (-5, 5)). The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
Panel A Government supported versus non-government supported industriesSupported
Non-supported
Supported
Non-supported
Supported
Non-supported
CAR (-1, 1)
CAR (-1, 1)
CAR (-2, 2)
CAR (-2, 2)
CAR (-5, 5)
CAR (-5, 5)
CORRUPTION 0.262*** 0.014 0.222*** 0.029* 0.013 0.014(0.01) (0.25) (0.00) (0.08) (0.87) (0.55)
POST 0.072 0.006 0.043 0.015 0.004 0.026(0.30) (0.45) (0.32) (0.19) (0.95) (0.11)
CORP*POST -0.356*** -0.077*** -
0.302*** -0.114*** -0.023 -0.123***
(0.00) (0.00) (0.00) (0.00) (0.82) (0.00)F-test: equal CORP*POST -0.279*** -0.188*** -0.100
(0.00) (0.00) (0.53)Control variables YES YES YES YES YES YESINDUSTRY YES YES YES YES YES YESN 76 307 76 307 76 307adj. R-sq 0.16 0.38 0.26 0.36 0.14 0.36
Panel B High competitive versus low competitive industriesHigh Low High Low High LowCAR (-1, 1)
CAR (-1, 1)
CAR (-2, 2)
CAR (-2, 2)
CAR (-5, 5)
CAR (-5, 5)
CORRUPTION 0.015 0.115*** 0.024 0.111*** 0.024 -0.008(0.30) (0.00) (0.18) (0.00) (0.32) (0.83)
POST 0.018 0.004 0.031** 0.003 0.028 0.023(0.10) (0.88) (0.02) (0.90) (0.12) (0.38)
CORP*POST -0.069*** -0.170*** -0.088*** -0.203*** -0.088** -0.133**(0.00) (0.00) (0.00) (0.00) (0.01) (0.02)
F-test: equal CORP*POST 0.101** 0.115*** 0.045
(0.03) (0.01) (0.38)Control variables YES YES YES YES YES YESINDUSTRY YES YES YES YES YES YESN 206 175 206 175 206 175adj. R-sq 0.31 0.28 0.34 0.37 0.43 0.39
58
Table 12 Regression results for the impact of corruption events on corporate investment of non-SOEsThis table present the regression results for the impact of corruption events on corporate investment expenditure of non-SOEs. The ‘FULL’ sample includes all listed firms in Shenzhen and Shanghai stock exchange and the ‘JURISDICTION’ sample only include the firms whose under corrupt bureaucrats’ jurisdiction. The dependent variable ‘CAPEXTA’ is the ratio of capital expenditure to total assets. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
Corruption related sample Bureaucrats connecting sampleFULL JURISDICTION FULL JURISDICTIONCAPEXTA CAPEXTA CAPEXTA CAPEXTA
CORRUPTIONt-1 0.012*** 0.015*** 0.013*** 0.016***(0.00) (0.00) (0.00) (0.00)
POST t-1 0.005** 0.005** 0.005** 0.005**(0.02) (0.01) (0.02) (0.01)
CORRUPTION t-1*POST t-1 -0.011** -0.017*** -0.017*** -0.020***(0.02) (0.00) (0.00) (0.00)
Control variables YES YES YES YESINDUSTRY YES YES YES YESN 3845 2933 3728 2883Adj. R-sq 0.21 0.25 0.21 0.25
59
Table 13 Regression results for the impact of corruption events on corporate investment efficiency of non-SOEsThis table present the regression results for the impact of corruption events on corporate investment efficiency of non-SOEs. The ‘FULL’ sample includes all listed firms in Shenzhen and Shanghai stock exchange and the ‘JURISDICTION’ sample only include the firms whose under corrupt bureaucrats’ jurisdiction. ‘Before’ and ‘Post’ refer to the sample period before and after the corruption event. The dependent variable ‘CAPEXTA’ is the ratio of capital expenditure to total assets. The ‘CORRUPTION’ is a dummy variable that is equal to one if firm is a corruption case related firm, and zero otherwise. The ‘POST’ is a dummy variable that is equal to one if the observation is obtained after the year of the corruption event, otherwise zero. Q is Tobin’s Q measuring firms’ investment opportunity. Control variables are the same with table 3, the coefficients for those control variables are not reported to save space. Standardized beta coefficients; p-values in parentheses. *, **, *** represent significant at 10%, 5% and 1% level significance respectively.
Corruption related sample Bureaucrats connecting sampleFULL JURISDICTION FULL JURISDICTIONBefore Post Before Post Before Post Before PostCAPEXTA CAPEXTA CAPEXTA CAPEXTA CAPEXTA CAPEXTA CAPEXTA CAPEXTA
CORRUPTIONt-1 0.003 0.013** 0.005 0.013** 0.028*** 0.013 0.029*** 0.014*(0.73) (0.02) (0.59) (0.03) (0.01) (0.11) (0.01) (0.07)
Q t-1 0.007*** 0.003*** 0.007*** 0.003*** 0.015*** 0.003** 0.015*** 0.003***(0.00) (0.01) (0.00) (0.01) (0.00) (0.01) (0.00) (0.01)
CORPt-1*Q t-1 0.008 -0.004** 0.008 -0.005*** -0.006 -0.006* -0.008 -0.005*(0.16) (0.04) (0.19) (0.01) (0.40) (0.05) (0.25) (0.07)
Control variables YES YES YES YES YES YES YES YESINDUSTRY YES YES YES YES YES YES YES YESN 1821 2024 1774 1954 821 1106 840 1128Adj. R-sq 0.18 0.25 0.18 0.25 0.27 0.28 0.26 0.28
60
Appendix A. A list of high level bureaucrats’ corruption events in China from 2005 to 2011Name Announcement
dateProvince Duty
Zhang Enzhao 15 March 2005 Bank The Chairman of China Construction Bank, The member of Centre Committee for Discipline Inspection
Wang Youjie 28 August 2005 Henan The vice-chairman of NPC in Henan provinceHuang Songyou 15 October 2005 Central The vice-president of the Supreme People's CourtZhu Zuoyong 19 December 2005 Gansu The vice-chairman of the CPPCC in Gansu provinceLiu Zhihua 9 June 2006 Beijing The vice-governor of BeijingHe Minxu 22 June 2006 Anhui The vice governor of Anhui provinceWang Wulong 13 July 2006 Jiangsu The vice-CPC secretary of Jiangsu ProvincePang Jianyu 15 September 2006 Shanxi The vice-chairman of the CPPCC in Shanxi provinceChen Liangyu 24 September 2006 Shanghai The secretary of CPC in ShanghaiLiu Weiming 22 January 2007 Guangdong The vice-governor of Guangdong province, The vice-chairman of the CPPCC in Guangdong
ProvinceDu Shicheng 18 April 2007 Shandong The vice-CPC secretary of Shandong ProvinceSong Pingshun 3 June 2007 Tianjin The vice-CPC secretary of TianjinChen Shaoyong 8 August 2008 Fujian The secretary-general of the CPC in Fujian provinceWang Huayuan 16 April 2009 Zhejiang The secretary of CPC in Zhejiang provinceChen Shaoji 24 April 2009 Guangdong The Chairman of CPPCC in Guangdong provincePi Qiansheng 17 June 2009 Tianjin The member of Standing Committee of CPC in TianjinSong Yong 13 October 2009 Liaoning The vice-chairman of NPC in Liaoning provinceLi Tangtang 15 October 2009 Ningxia The vice-governor of Ningxia municipalityHuang Yao 24 October 2009 Guizhou The Chairman of the CPPCC in Guizhou provinceSun Shuyi 17 December 2009 Shandong The Chairman of the CPPCC in Shandong provinceLiu Jiameng 3 April 2010 Zhejiang The vice-chairman of NPC in Zhejiang provinceZhang Jiameng 3 April 2010 Zhejiang The vice-chairman of NPC in Zhejiang provinceSong Chenguang
10 July 2010 Jiangxi The vice-chairman of the CPPCC in Jiangxi province
Liu Zhuozhi 15 December 2010 Neimenggu The vice-governor of Neimengguo municipality
61
oTian Xueren 5 November 2011 Jilin The member of Standing Committee of CPC Jilin, The executive vice-governor of Jilin ProvinceHuang Sheng 24 November 2011 Shandong The vice-governor of Shandong provinceSun Yu November 2007 Guangxi The vice-governor of Guangxi Zhuang Autonomous RegionZhu Zhigang October 2008 Central The deputy Finance ministerWang Zhaoyao September 2005 Anhui The vice-CPC secretary of Anhui Province
62
Appendix B Variable definition
Variable name Detailed definitionCorruption case related firm dummy (CORRUPTION) The dummy variable s equal to 1 if firm is a corruption related firm and 0 otherwise.
Post event dummy (POST) The dummy equals to 1 if the observation is obtained after the year of the corruption event, otherwise 0.Merger and acquisition dummy (MA) The dummy equals to 1 if the firm announced a merger and acquisition, and 0 otherwise.
3-days CARs (CAR_1) The cumulative abnormal return over a three-day event window from the one day precious to the M&A announcement to the one day post the announcement.
5-days CARs (CAR_2) The cumulative abnormal return over a five-day event window from the two days precious to the M&A announcement to the two days post the announcement.
11-days CARs (CAR_5) The cumulative abnormal return over a eleven-day event window from the five days precious to the M&A announcement to the five days post the announcement.
One year buy and hold return (BHAR1) The one year post-M&A buy and hold return
Two years buy and hold return (BHAR2) The two year post-M&A buy and hold return
Three years buy and hold return (BHAR3) The three year post-M&A buy and hold return
Takeover premium(PREMIUM1) The ratio of trading value of the target on the estimated value minus oneTakeover premium(PREMIUM2) The natural logarithm of the difference between trading value of the target and the target estimated valueLocal M&As dummy (LOCAL) The dummy equals to 1 if the firm acquire a local target, and 0 otherwise
Capital expenditure (CAPEXTA) it is the ratio of cash payments for fixed assets, intangible assets and other long term assets from the cash flow statement less the cash receipts from selling these assets to the total assets
Firm size (SIZE) The natural logarithm of book value of total assets.Tobin's Q (Q) Market value/replacement value.Operation cash flow (OPCFTA) Total operation cash flow scaled by total assets.Leverage ratio (LEVERAGE) The ratio of total debt to total assetsBoard size (BOARDSIZE) The total number of board of directorsBoard independent (BOARDIND) The ratio of the number of independent board of directors to total board of directors.Manager's ownership (OWNERSHIP) The ratio of managers shareholding to the total share outstanding
63