Applied Economics

ISSN: 0003-6846 (Print) 1466-4283 (Online) Journal homepage: https://www.tandfonline.com/loi/raec20

How does political connection affect firm financial

distress and resolution in China?

Yu He, Lei Xu & Ron P. McIver

To cite this article: Yu He, Lei Xu & Ron P. McIver (2019) How does political connection affect
firm financial distress and resolution in China?, Applied Economics, 51:26, 2770-2792, DOI:
10.1080/00036846.2018.1558358

To link to this article: https://doi.org/10.1080/00036846.2018.1558358

Published online: 17 Dec 2018.

Submit your article to this journal

Article views: 369

View related articles

View Crossmark data

Citing articles: 5 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=raec20
APPLIED ECONOMICS
2019, VOL. 51, NO. 26, 2770–2792
https://doi.org/10.1080/00036846.2018.1558358

How does political connection affect firm financial distress and resolution in
China?
Yu Hea, Lei Xub and Ron P. McIver b

a
Research Center for Economy of Upper Reaches of the Yangtse River, Chongqing Technology and Business University, Chongqing, China;
b
School of Commerce, University of South Australia, Adelaide, Australia

ABSTRACT KEYWORDS
We examine the impact of political connections on firm performance, financial distress, and its Political connection;
resolution in China, a country where government influence over stock markets has been demon- financial distress; distress
strated to be considerable. Our findings suggest that over 1999 to 2015, although political recovery; Chinese stock
connections had limited impact on the emergence of financial distress, such connections assisted markets
distressed firms in gaining increments to debt financing and contributed to a higher likelihood of
JEL CLASSIFICATION
recovery. This indicates that Chinese authorities follow market economy principles, and only G33; G38
intervene in firms’ operations after they fall into financial distress. In addition, central and local
government political connections have different impacts on distress recovery. We conduct
additional analyses on differences in distress outcomes for various ownership (State-owned
enterprises, SOEs, and non–SOEs) and sample sub-periods (1999–2007 and 2008–2015). Our
results are robust to potential endogeneity issues and to alternative measures of financial
distress.

I. Introduction Fan, Huang, and Zhu 2013). This paper focuses on

an additional area, where China’s emerging market
Financial distress, both its causes and consequences,
differs from those in developed economies, and also
has received considerable academic attention over
one largely neglected field in the literature on firm-
recent decades.1 The importance of exit and bank-
level financial distress and its resolution: Political
ruptcy in resolving corporate financial distress has
connectedness.
been extensively studied, especially for developed
In this study, we examine the influence of political
economies.2 However, delisting and bankruptcy are
connections on firm performance, financial distress
less common features in emerging markets and, for
and distress resolution for China’s listed firms. To
China’s listed firms, rarely happen. For example,
investigate, we adopt a widely accepted measure of
from 1995 to 2012, more than 3,000 firms (6.2%)
firm political connection3 in our empirical analyses:
were delisted from the New York Stock Exchange
CEO or Chair of Board connectedness. Firms estab-
(NYSE), while only 78 firms (0.3%) were delisted
lish political connections by hiring politically con-
from China’s stock markets, with bankruptcy rates
nected senior executives to acquire resources and
in China being even lower. The foci of current
funds from the authorities. In addition, we identify
research on financial distress and its resolution for
whether these political connections are at central or
China’s listed firms also reflects such country-level
local government levels, and clarify differences in
differences, including corporate governance (Wang
impacts by levels of connections on financial distress
and Deng 2006; Wang and Li 2007), macroeconomic
and resolution, reflecting conflicts of interests
conditions (Bhattacharjee and Han 2014), and insti-
between central and local governments. In order to
tutional factors (Kam, Citron, and Muradoglu 2008;

CONTACT Yu He [email protected] Research Center for Economy of Upper Reaches of the Yangtse River,Chongqing Technology and Business
University, 19 Xuefu Road, Nan’an District, Chongqing 400067, China
1
See, for example, Opler and Titman (1994), Siegfried and Evans (1994), Caves (1998), Pakes and Ericson (1998), Jovanovic and Rousseau (2002), Liu (2004),
Campbell, Hilscher, and Szilagyi (2008), Bhattacharjee et al. (2009), Donker, Santen, and Zahir (2009), Tinoco and Wilson (2013), Rosenfeld (2014), Höwer
(2016) and Ivashina, Benjamin, and David (2016).
2
See, for example, Gilson (1997), Maksimovic and Phillips (1998) and Eckbo and Thorburn (2003).
3
See, for example, Fan, Wong, and Zhang (2007), Chen et al. (2011b), Wu et al. (2012), Su, Zhang, and Zhang (2013) and Fonseka et al. (2015).
© 2018 Informa UK Limited, trading as Taylor & Francis Group
 APPLIED ECONOMICS 2771

explore the relationship between political connec- Lin et al. 2015; Cull et al. 2015), there is a lack of
tions, financial distress and distress resolution, we discussion on the relationship between political con-
first discuss the influence of such connections on nections and financial distress. This study fills this
firm performance, in terms of firm profitability, research gap by investigating the influence of poli-
leverage and development capacity, based on tical connections on financial distress and resolution
a sample from 1999 to 2015. We then identify firm for China’s listed firms. Second, our empirical
financial distress using the Emerging Market (EM) results show a limited influence of political connect-
Z–score model (Altman, Hartzell, and Peck 1995) edness on firm operations when firm is in a sound
and explore the impacts of political connections on financial condition, but a strong impact on leverage
financial distress and distress resolution. Finally, we and distress resolution where firm is in financial
undertake additional tests to re-estimate the influ- distress. This indicates that Chinese authorities
ence of political connections on financial distress abide by market economy principles and assist
and resolution for various ownership (SOEs and firms in gaining debt financing after they fall into
non–SOEs) and sample sub-periods (1999–2007 financial distress, which comes to a major contribu-
and 2008–2015). tion of this study. Third, this study demonstrates
We reach several important findings in this that central and local government levels of political
paper. Politically connected CEO or Chair of connectedness affect financial distress and distress
Board, including central and local government con- resolution differently. Fourth, the concave relation-
nection, is associated with higher growth in lever- ship between firm leverage and distress recovery
age during periods of financial distress. Financially indicates that political connections assist firms in
distressed firms with political connections, espe- gaining incremental debt financing and contribute
cially CEO or Chair of Board connectedness with to a higher likelihood of distress recovery. This
central government, have a higher probability of identifies a mechanism for how political connections
recovering and with less time. The above indicates affect distress recovery via leverage. Fifth, this paper
that Chinese authorities assist politically connected contributes to the literature by suggesting that non–
firms in gaining debt financing after they fall into SOEs with political connections tend to have
financial distress, and contribute to a higher like- a higher likelihood of distress recovery, compared
lihood of distress recovery. We detect a convex with SOEs, implying potential incentives for non–
(concave) relationship between firm leverage and SOEs to establish political connections, especially
financial distress (distress recovery). Lastly, our during the financial distress. Finally, our results
additional tests show that political connection, show that political connectedness has a stronger
especially the connection with central government, influence on distress recovery after the Global
has a stronger influence on distress resolution for Financial Crisis (GFC). Such results imply that the
non–SOEs than that for SOEs, emphasizing the Four Trillion RMB Economic Stimulus Program4
magnitude of politically connected managers for effectively assisted the recovery of distressed firms,
private firms. Also, central government connected- which has been one practical contribution of this
ness has a stronger impact on distress resolution study. Given the unique institutional environment
during 2008–2015 than 1999–2007. Our results are in China, the above results enhance our understand-
robust when controlling for potential endogeneity ing of the impact of political connection on financial
issues and for alternative measures of financial dis- distress and resolution in the world’s largest emer-
tress, including Special Treatment (ST) status and ging market.
interest coverage ratio. The remainder of the paper is organized as
Our study makes several contributions. First, follows. In Section 2, we briefly discuss the
while previous studies discuss how political connec- Chinese institutional background and the litera-
tions affect firm performance and financing condi- ture on political connections and develop hypoth-
tions (Fan, Wong, and Zhang 2007; Wu et al. 2012; eses. We then describe data collection, sample

4
The State Council of China announced the Four Trillion RMB Economic Stimulus Program on 9 November 2008 as an attempt to minimize the impact of the
GFC on the national economy.
2772 Y. HE ET AL.

selection, measurement of financial distress and restructuring process. This reflects that central gov-
related methodology in Section 3. We report ernment is mainly concerned with national eco-
empirical analyses and findings in Section 4. nomic growth, whereas local government’s focus
Finally, we provide conclusions in Section 5. is on the local economy and employment.

Political connection and hypothesis

II. Institutional background and literature
review Literature on political connection
In light of the institutional background supporting
Institutional background
the development of political connections by
The institutional environment surrounding China’s listed firms, multiple scholars have
Chinese stock markets facilitates the study over explored the relationship between political con-
the influence of political connections on financial nections and financial performance (Fan, Wong,
distress and resolution. Several of this institutional and Zhang 2007; Li et al. 2008; Chen et al. 2011b;
environment’s unique features have resulted in Wu et al. 2012; Lin et al. 2015). The prior litera-
there being significant potential for political con- ture has led to a conflicting relationship being
nections to impact listed firms’ operations. identified between political connectedness and
The first feature concerns the complexities asso- firm performance.
ciated with China’s economic restructuring, from According to Wu et al. (2012), private firms with
a centrally–planned to a market-oriented economy. politically connected CEOs or Chairs of Board out-
Associated with this had been the emergence of large perform those without connections. Li et al. (2008)
numbers of private firms alongside significant state state that politically connected private firms have
ownership. Moreover, the unique re-establishment better performance than non–connected private
of China’s stock markets in the early 1990s assisted firms, as the connections assist firms to obtain
SOEs to undertake reforms to raise capital. This led loans from banks and other state institutions.
to the government becoming both a major stock- Moreover, Lin et al. (2015) argue that Chinese
holder, in terms of state ownership, and also the firms use corporate social responsibility (CSR) activ-
regulator of Chinese stock markets. This phenom- ities to establish government connections, and that
enon has implied evidence of political connections such actions increase following electoral replace-
between SOEs and government. The preferential ment of local mayors. They find that firms with
allocation of resources to SOEs has required that higher levels of investments in CSR outperform
private firms build political connections to attain those firms without such investments. The conclu-
additional resources and mitigate competitive dis- sion drawn from the literature is that, amongst all
advantages. For instance, Wu et al. (2012) find that listed firms, politically connected firms tend to out-
from 1999 to 2007, China’s listed firms built political perform those without connections.
connections through the recruitment of current or Nevertheless, a negative impact of political con-
former government officials as managers. Cull et al. nectedness on firm performance has also been
(2015) argue that firms tend to establish political detected in previous studies (Fan, Wong, and
connections with governments to improve financing Zhang 2007; Chen et al. 2011b; You and Du
conditions. 2012). Evaluating 790 privatized listed firms in
Furthermore, the separation between the central China over 1993 to 2001, Fan, Wong, and Zhang
and local levels of governments of administrative (2007) find politically connected firms underper-
power (formerly held by the central government) form non–connected firms by around 18% in the
has gradually resulted in an obstacle in the eco- three years period following Initial Public Offering
nomic restructuring process. Although designed to (IPO). Chen et al. (2011b) state that SOEs with
encourage local governments to play an important politically connected top executives have poorer
role in promoting their local economies, this investment efficiency than non–connected SOEs
separation has given rise to conflicts between cen- and private firms. You and Du (2012) find that
tral and local government, which inhibits the politically connected CEOs tend to benefit
 APPLIED ECONOMICS 2773

themselves with political resources rather than to resolution. In this case, we link such connection
benefit listed firms. The existence of a negative to firm financial distress and distress recovery, but
relationship between political connections and do not expect the direction of such influence.
firm performance has also been identified in Based on our expectations above, we establish
other countries; see, for example, Leuz and the following hypotheses:
Oberholzer–Gee (2006), Boubakri, Cosset, and
Saffar (2008) and Faccio (2010). H1. Political connection does relate to the like-
Other strands of literature explore political con- lihood of firm financial distress.
nections from perspectives of minimizing the IPO
underpricing effect (Francis, Hasan, and Sun H2a. After firms fall into financial distress, politi-
2009), expropriation from government (Chen cal connection does relate to the likelihood of
et al. 2011a), acquisition activities (Su, Zhang, distress recovery.
and Zhang 2013), improving financing conditions
(Cull et al. 2015) and private-equity placements H2b. After firms fall into financial distress, poli-
(Fonseka et al. 2015). The literature identifies that tical connection does relate to the length of recov-
political connections influence firm operations, ery time.
implying its potential effect on firm financial dis-
tress and resolution. This paper therefore contri-
butes to the literature by first establishing links
III. Methodology
with empirical evidence on the relationship
between political connections, financial distress Data collection and sample description
and financial resolution.
Our financial data are derived from the China Stock
Market and Accounting Research (CSMAR) database.
Development of hypotheses Macroeconomic data are collected from the China
Previous literature explores financial distress and Economic Monitoring Centre,5 and Federal Reserve
resolution for China’s listed firms from the per- Bank of St. Louis,6 respectively. In total, data are
spective of corporate governance (Wang and Deng collected for all 3,027 stocks listed on the Shanghai
2006; Wang and Li 2007), macroeconomic condi- Stock Exchange (SHSE) and Shenzhen Stock
tions (Bhattacharjee and Han 2014) and institu- Exchange (SZSE) from 1999 to 2015. This sample
tional factors (Kam, Citron, and Muradoglu 2008; period is chosen, as certain data, such as political
Fan, Huang, and Zhu 2013). There is a lack of connections, are not available until 1999.
study exploring the influence of political connec- Additionally, previous studies in this field generally
tion on financial distress and resolution. As dis- focus on financial distress and distress resolution
cussed above, the impact of political connection before 2006 (Kam, Citron, and Muradoglu 2008;
may either positively or negatively affect firm per- Fan, Huang, and Zhu 2013; Bhattacharjee and Han
formance (Fan, Wong, and Zhang 2007; Li et al. 2014). However, our selected sample allows investiga-
2008; Chen et al. 2011b; Wu et al. 2012; You and tion of the influence of political connections, macro-
Du 2012; Lin et al. 2015). Moreover, financially economic conditions, and financial fundamentals, on
distressed firms are inefficient producers with financial distress over the last decade, which incorpo-
poorer performance, higher financial leverage rates important events like the Non–Tradable Share
and more cash flow problems (Chan and Chen (NTS) reform7 in 2007 and GFC in 2008.
1991). Thus, the influence of political connection A number of firm-year observations in our
may further affect firm financial distress and sample were deleted, based on the following
5
China Economic Monitoring Center: http://www.cemac.org.cn/Ozsbz.html.
6
Federal Reserve Bank of St. Louis: https://www.stlouisfed.org/.
7
The NTS reform was started in 2005 and completed at the end of 2007. There were two kinds of shares in Chinese stock markets before the NTS reform-
tradable shares (TS) and non–tradable shares (NTS). Almost 70% of outstanding shares were NTS and most of these shares were held by block holders. The
trading of NTS was highly restricted. The NTS reform was divided into three stages. In the first stage, NTS holders were allowed to trade less than 5% of
a firm’s outstanding shares. In the second stage, NTS holders were allowed to trade 10% of a firm’s outstanding shares. In the final stage, all NTS become
tradable (Liu and Tian 2012).
2774 Y. HE ET AL.

three criteria: a) firms with less than two consecu- EMZ
score ¼ 6:56  A þ 3:26  B þ 6:72
tive years of financial information; b) firms in the  C þ 1:05  D þ 3:25 (1)
financial sector, due to differences in regulations;
and c) firm-year observations for the first year where A, working capital divided by total
when firm goes public, as data from previous assets, represents liquidity; B, retained earnings
time period is required to calculate change vari- divided by total assets, captures sustainable prof-
ables. Based on these conditions, we excluded 311 itability; C, operating income divided by total
firms with less than two firm–years of observa- assets, represents operating efficiency; and D,
tions, two firms with missing values, and 59 firms the book value of equity divided by total liabil-
in the finance sector. In total, there are 2,655 ities, measures leverage.
stocks in our final sample. The Z–score model is widely used in related stu-
dies; see, for example, Fan, Huang, and Zhu (2013)
Measuring distress for China, Richardson, Lanis, and Taylor (2015) for
the US and Almamy, Aston, and Ngwa (2016) for
The current literature on financial distress in
Britain. In this study, we define firms as financially
China utilizes a range of measures to identify
distressed if their Z–scores (Altman, Hartzell, and
financial distress. Wang and Deng (2006) and
Peck 1995) are below the cut–off point of 1.1 for
Wang and Li (2007), for example, use Special
two consecutive years.9 We further define firm
Treatment (ST) status as the indicator. The
emergence from financial distress as occurring
China Securities Regulatory Commission (CSRC)
where firm Z–scores exceed the cut–off point for
introduced the ST policy in 1998 to alert investors
two consecutive years. Based on these criteria, we
to the risk of poorly performing firms. Under the
observe a total of 179 (6.74% of 2,655) firms falling
ST policy, any listed firm suffering a net loss for
into financial distress, including 107 (59.78% of 179)
two consecutive years, where the ROE is below 0,
distressed firms that finally recovered during the
is categorized as an ST.8 Another common mea-
sample period. For robustness purposes, we then
sure is the interest coverage ratio (Kam, Citron,
employ ST status and interest coverage ratio as alter-
and Muradoglu 2008; Bhattacharjee and Han
native measures of financial distress. We list the
2014). Specifically, Kam, Citron, and Muradoglu
details for distressed firms in Table 1.10
(2008) adopt an interest coverage ratio below one
for two consecutive years to signal distress, while
Bhattacharjee and Han (2014) specify an interest
Variables
coverage ratio of less than 0.7.
A problem of using ST status and interest cov- Dependent variables
erage ratio as indicators of financial distress is that This study utilizes a two-step process to investigate
these indicators only capture one potential aspect the influence of political connections, macroeco-
of financial distress, such as profitability and sol- nomic conditions, and financial fundamentals on
vency, respectively. Therefore, our preference is to firm performance, financial distress and recovery
adopt a more comprehensive measure. Following from distress. In the first step, we investigate the
Altman, Hartzell, and Peck (1995) and Altman impacts of the above factors on firm performance,
(2005), we adopt the EM Z–score model, which given its role in preventing firm financial distress.
captures liquidity, profitability, operating effi- Following Fan, Huang, and Zhu (2013), we adopt
ciency and leverage characteristics, to measure three dependent variables, in change forms, to cap-
financial distress: ture three aspects of firm performance: Profitability
8
In fact, listed firm will be categorized as an ST under four conditions: a) a listed firm has a net loss (ROE < 0) in two consecutive years; b) the last annual
report shows the equity is lower than the registered capital; c) the auditor issues a negative opinion to the listed firm or even unable to issue an opinion;
d) there is evidence suggesting that the listed firm has stopped or will stop its business activities for more than three months. There were more than 80%
of listed firms which became STs due to net losses for two consecutive years.
9
Previous finance study (Fan, Huang, and Zhu 2013) introducing the Emerging Market Z–score model into the investigation of financial distress in China
employs the cut–off point of 0 to identify distress. In this paper, we make related adjustments to change the cut–off point from 0 to 1.1. We also adopt
alternative cut–off points, 0 and 2.6 (Altman, Hartzell, and Peck 1995; Fan, Huang, and Zhu 2013), to identify various stages of distress for robustness tests
and report the empirical results in Table A1.
10
Table 1 lists the number of firms falling into financial distress using alternative cut–off points and measures of financial distress.
 APPLIED ECONOMICS 2775

Table 1. The number of firms falling into financial distress.

Panel A: The number of financially distressed firms
Measures Z Score 0.00 Z Score 1.10 Z Score 2.60 ST ICR11
Number of all listed firms 2,655 2,655 2,655 2,655 2,655
Number of financially distressed firms 149 179 224 524 1,539
Proportion of financially distressed firms 5.61% 6.74% 8.44% 19.74% 57.97%
Panel B: The number of financially distressed firms by year
Financially Distressed Firms
Year Z Score 0.00 Z Score 1.10 Z Score 2.60 ST ICR
2000 9 16 29 30 124
2001 10 13 14 24 60
2002 13 13 20 51 95
2003 9 12 14 66 68
2004 12 13 14 43 48
2005 13 15 26 32 95
2006 20 28 30 51 84
2007 14 17 21 57 44
2008 10 9 14 22 41
2009 13 13 15 24 74
2010 7 9 7 30 52
2011 4 3 2 10 138
2012 4 4 7 20 243
2013 4 5 3 17 177
2014 3 5 3 25 95
2015 4 4 5 22 101

(EPS growth), leverage (Leverage growth) and devel- recovery.12 Similar measures have been widely utilized
opment capability (Assets growth). Uses of these vari- in related studies; see, for example, Whitaker (1999),
ables in change forms are able to mitigate issues Fan, Huang, and Zhu (2013), Bhattacharjee and Han
related to data heterogeneity (Opler and Titman (2014) and Koh et al. (2015).
1994; Lin, Zhang, and Zhu 2009; Fan, Huang, and
Zhu 2013). Independent variables
In the second step, we introduce three dependent We quantify firms’ political connections with
variables, Distress dummy, Recover dummy and a measure used widely in previous studies (Fan,
Recover time, to explore how listed firms fall into Wong, and Zhang 2007; Chen et al. 2011b; Wu
and emerge from financial distress using the dis- et al. 2012; Su, Zhang, and Zhang 2013; Fonseka
tressed firm subsample. Distress dummy takes et al. 2015): A firm is politically connected if its
a value of 1 if a firm falls into financial distress during CEO or Chair of Board is or was a government official
our sample period, and 0 otherwise. A firm falls into (PC).13 Fan, Wong, and Zhang (2007) define a firm as
financial distress when its Z–scores are below the cut– politically connected if the CEO is/was a government
off point of 1.1 for two consecutive years. Recover official, and find that firms with politically connected
dummy equals 1 if a distressed firm recovers during CEOs underperform those without politically con-
our sample period, and 0 otherwise. A distressed firm nected CEOs. Similarly, Chen et al. (2011b) and Wu
recovers when its Z–scores are above the cut–off point et al. (2012) use either CEO or Chair of Board who is/
of 1.1 for two consecutive years. Recover time is the was a government official to define political connect-
number of years from falling into distress to that of edness. Additionally, we divide political connections
recovery. If a firm does not recover during our sample of CEOs and Chairs of Board into two levels, central
period, Recover time refers to the number of years government (PC central) and local government (PC
from falling into distress until the end of our sample local), to investigate the influence of different levels of
period. For firms falling into distress repeatedly, our political connection on financial distress. As identi-
study considers the time of first distress and last fied in literature review, there is a conflict of interest
11
ICR is short for the Interest Coverage Ratio.
12
There are five firms that fall into financial distress twice during our sample period.
13
PC is short for political connection, which refers to a type of relationship with Chinese authorities, such as Chinese government, China Communist Party
Committee, Central Military Commission, Chinese People’s Political Consultative Conference (CPPCC) and National People’s Congress.
2776 Y. HE ET AL.

between central and local government. Central gov- Summary statistics

ernment is mainly concerned with national economic
Table 2 provides summary statistics for the vari-
growth, whereas local government focuses on the
ables discussed in Section 3.3. From Panel A, we
local economy and employment. Given this, we
find that, on average, sample firms experienced
expect these different levels of political connectedness
a decrease in profitability and an increase in
to affect firm performance, financial distress, and
both leverage and total assets during the sample
distress resolution, differently. With a dummy vari-
period, with mean growth in EPS of −0.0209,
able, Distress, we further estimate the influence of
changes in leverage of 0.0193 and growth in assets
political connectedness on firm performance during
of 0.2475, respectively. In addition, the statistics
the years of financial distress. Distress takes a value of
for most variables are based on firm-year observa-
1 for those years in which firm financial distress is
tions, excepting those for Distress dummy, Recover
observed, and 0 otherwise.
dummy, and Recover time, which are based on
Controlling for external and internal macroeco-
firm-level observations. The mean of Distress
nomic conditions, we employ two macroeconomic
dummy and Recover dummy are 0.0674 and
indicators: The World Consumer Price Inflation
0.5978, respectively, indicating that 6.74% (179)
(ICP world) and the Consistent Macroeconomic
of our sample of 2,655 firms fell into financial
Index for China (Consistent). The World Consumer
distress, and that 59.78% (107) of these 179 firms
Price Inflation (ICP world) is used to measure external
recovered during our sample period. Moreover,
macroeconomic conditions, and tends to have strong
firms experiencing financial distress spent, on
explanatory power on firm performance, in terms of
average, 6.0168 years prior to recover from it.
corporate defaults (Nickell, Perraudin, and Varotto
With respect to political connections, 37.25% of
2000; Bhattacharjee et al. 2009). Firms in China were
firm-year observations identify the presence of
highly active in global markets from 1999 to 2015, and
political connections, with political connectedness
so exposed to pricing in global input and output
to central government (22.22%) being more com-
markets. Thus, ICP world provides a suitable indicator
mon than to local government (21.41%). In some
to capture the impact of external macroeconomic
cases, firms have both central and local govern-
conditions on firm performance, financial distress
ment connections as the sum (43.45%) of PC
and distress recovery. In addition, internal macroeco-
central (22.22%) and PC local (21.41%) is larger
nomic conditions, such as economic downturns in
than PC (37.25%).
a particular country, impact firm probability of finan-
In Panel B, we report the results of t-tests of
cial distress (Tinoco and Wilson 2013; Rosenfeld
differences in mean between firms that experi-
2014). We suggest that macroeconomic conditions
enced financial distress and those that did not.
also apply to financial distress in China
We find that a greater proportion of firms that
(Bhattacharjee and Han 2014) and thus use the
did not experience financial distress have political
Consistent Macroeconomic Index for China
connections than those firms that experienced
(Consistent) to control for internal macroeconomic
financial distress. The means of PC, PC central
conditions.
and PC local for the former group of firms are
Previous studies suggest that financial fundamen-
statistically significantly larger than those of firms
tals also significantly affect distress; see, for example,
that experienced financial distress. Also, non–dis-
O’Hara and Shaw (1990), Opler and Titman (1994),
tressed firms have shorter time periods since going
Caves (1998) and Lennox (1999). Thus, we adopt firm
public, lower leverage ratios, and larger firm size
age (Age), leverage (Leverage) and size (Size) to con-
than those firms that experienced financial dis-
trol for financial fundamentals in our empirical ana-
tress. Table 3 identifies that, although statistically
lyses. Age is the logarithm of years since the firm went
significant in many cases, the correlations between
public; Leverage equals total debt divided by total
our main variables are generally small. Therefore,
assets; and Size equals the logarithm of firm’s total
we conclude that potential problems arising from
assets. Industry fixed effects are also considered in this
collinearity between variables in this study are
study. Definitions for the variables used in our study
unlikely to be an issue.
are summarized in Table A2.
 APPLIED ECONOMICS 2777

Table 2. Summary statistics.

Panel A: Full sample
Variables Observations Mean Median Std. Dev.
Dependent variables
EPS growth 28,499 −0.0209 −0.0057 0.6261
Leverage growth 28,499 0.0193 0.0104 1.8193
Assets growth 28,499 0.2475 0.0924 1.9840
Distress dummy 2,655 0.0674 0.0000 0.2507
Recover dummy 179 0.5978 1.0000 0.4903
Recover time 179 6.0168 5.0000 3.4080
Independent variables
PC 28,499 0.3725 0.0000 0.4835
PC central 28,499 0.2222 0.0000 0.4157
PC local 28,499 0.2141 0.0000 0.4102
Distress 28,499 0.0268 0.0000 0.1615
ICP world 28,499 3.7752 3.5548 1.6051
Consistent 28,499 98.6076 98.2400 3.0431
Age 28,499 2.0718 2.1972 0.6609
Leverage 28,499 0.5717 0.4823 2.0651
Size 28,499 21.5939 21.4634 1.3045
Panel B: Financially distressed firms vs. non financially distressed firms
Financially distressed sample Non financially distressed sample t–test of difference
Variables Obs Mean Obs Mean Difference t–stat.
EPS growth 2,575 −0.0067 25,924 −0.0223 0.0156 1.205
Leverage growth 2,575 0.0942 25,924 0.0118 0.0824** 2.193
Assets growth 2,575 0.4997 25,924 0.2225 0.2772*** 6.768
PC 2,575 0.2579 25,924 0.3839 −0.1261*** −12.655
PC central 2,575 0.1142 25,924 0.2330 −0.1188*** −13.873
PC local 2,575 0.1852 25,924 0.2170 −0.0317*** −3.745
ICP world 2,575 3.8898 25,924 3.7638 0.1260*** 3.800
Consistent 2,575 98.7573 25,924 98.5928 0.1645*** 2.617
Age 2,575 2.3477 25,924 2.0444 0.3033*** 22.404
Leverage 2,575 1.6615 25,924 0.4635 1.1981*** 28.474
Size 2,575 20.4072 25,924 21.7118 −1.3046*** −50.521
*** and ** refer to the significance at the 1%, and 5% levels, respectively.
This table presents the summary statistics of the main variables. The financial data is extracted from the CSMAR database and macroeconomic data is
collected from the China Economic Monitoring Centre and Federal Reserve Bank of St. Louis. In total, the sample contains 2,655 stocks in SHSE and SZSE
from 1999 to 2015, with 179 (6.74% of 2,655) financially distressed firms and 2,476 non-financially-distressed firms, respectively. There are 107 (59.78% of
179) financially distressed firms recovering at a given time during the sample period. Moreover, 5 firms fall into financial distress twice. The statistics for
most variables are based on firm-year observations, including EPS growth, Leverage growth, Assets growth, PC, PC central, PC local, Distress, ICP world,
Consistent, Age, Leverage and Size. The statistics for Distress dummy, Recover dummy and Recover time are based on firm-level observations. For the
definitions of variables refer to Table A2.

IV. Empirical results We include three dependent variables–changes in

EPS, leverage and assets–to measure firm perfor-
Firm performance
mance, and denote each as the growth in the variable
As mentioned in Section 3.3.1, this study utilizes (e.g. EPS growth). Variables identifying political con-
a two-step process for our empirical analysis. In nections include PC (a dummy variable which equals
the first step, we adopt the following Ordinary 1 if the CEO or Chair of Board is/was a government
Least Squares (OLS) regression model to investi- official, and 0 otherwise), PC central (a dummy vari-
gate the influence of the factors proposed on firm able which equals 1 if the CEO or Chair of Board is/
performance, which strongly affects firm move- was a central government official, and 0 otherwise)
ment into financial distress. and PC local (a dummy variable which equals 1 if the
CEO or Chair of Board is/was a local government
Performancei;t ¼ αi;t þ βi;t Political connectioni;t official and 0, otherwise). Macroeconomic variables
þ γi;t Political connectioni;t Distressi;t include ICP world and Consistent. Control variables
include Age, Leverage, and Size. Interaction tests are
þ δi;t Distressi;t þ ηt Macroeconomict
conducted with an additional variable, Distress (a
þ θi;t Controli;t þ μi;t Industryi;t þ εi;t dummy variable which equals 1 for years in which
(2) the firm experienced financial distress, and 0 other-
wise), to explore the influence of the above factors on
2778 Y. HE ET AL.

Table 3. Correlation matrix.

EPS Leverage Assets
growth growth growth PC PC central PC local Distress ICP world Consistent Age Leverage
Leverage −0.0385***
growth
(0.000)
Assets 0.0809*** −0.0887***
growth
(0.000) (0.000)
PC 0.0019 −0.0035 0.0145**
(0.745) (0.551) (0.014)
PC central −0.0056 −0.0024 0.0139** 0.6937***
(0.348) (0.687) (0.019) (0.000)
PC local 0.0087 −0.0026 0.0070 0.6774*** 0.0951***
(0.141) (0.659) (0.235) (0.000) (0.000)
Distress 0.1259*** 0.0057 0.0806*** −0.0511*** −0.0579*** −0.0120**
(0.000) (0.335) (0.000) (0.000) (0.000) (0.044)
ICP world −0.0160*** 0.0128** −0.0044 0.0282*** −0.0039 0.0496*** 0.0445***
(0.007) (0.031) (0.462) (0.000) (0.506) (0.000) (0.000)
Consistent 0.0829*** −0.0033 −0.0080 0.0457*** 0.0067 0.0646*** 0.0499*** 0.1808***
(0.000) (0.581) (0.180) (0.000) (0.259) (0.000) (0.000) (0.000)
Age 0.0582*** −0.0086 0.0216*** −0.0452*** −0.1409*** 0.0847*** 0.1163*** −0.0347*** 0.0133**
(0.000) (0.146) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.025)
Leverage 0.0139** 0.4619*** −0.0079 −0.0330*** −0.0317*** −0.0130** 0.2748*** 0.0408*** 0.0326*** 0.0673***
(0.019) (0.000) (0.182) (0.000) (0.000) (0.028) (0.000) (0.000) (0.000) (0.000)
Size −0.0103* −0.0330*** 0.0647*** 0.1059*** 0.0591*** 0.1037*** −0.2907*** −0.1021*** −0.0824*** 0.2253*** −0.1386***
(0.081) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
This table presents the correlations between the main variables. For the definitions of variables refer to Table A2. P–values are in parentheses.

firm performance during years in which financial performance of political connectedness is statisti-
distress is identified. Furthermore, we include indus- cally insignificant, with a negative impact of local
try fixed effects in our empirical analyses. government connectedness on firm development
Table 4 presents the OLS regression results for capacity (Assets growth) being the exception. The
firm performance. In general, the impact on firm results for interaction effects between political

Table 4. Firm performance.

EPS growth Leverage growth Assets growth
PC 0.0037 −0.0041 −0.0048
(0.589) (0.661) (0.804)
PC * Distress −0.1742 1.2012*** 1.4844
(0.139) (0.005) (0.150)
PC central 0.0024 −0.0050 0.0390
(0.800) (0.740) (0.153)
PC central * Distress 0.0314 0.6279* 1.9097
(0.881) (0.066) (0.327)
PC local 0.0020 −0.0008 −0.0632**
(0.802) (0.845) (0.022)
PC local * Distress −0.2234** 1.0571*** 1.2128
(0.045) (0.005) (0.279)
Distress 0.5616*** 0.5614*** −1.7265*** −1.6992*** 1.1126*** 1.0713***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.001)
ICP world −0.0128*** −0.0128*** 0.0005 0.0004 0.0043 0.0046
(0.000) (0.000) (0.962) (0.969) (0.675) (0.660)
Consistent 0.0173*** 0.0173*** −0.0070 −0.0071 −0.0023 −0.0019
(0.000) (0.000) (0.120) (0.119) (0.569) (0.641)
Age 0.0341*** 0.0341*** −0.0847*** −0.0851*** −0.0548*** −0.0488***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Leverage −0.0076 −0.0075 0.4527*** 0.4522*** −0.0186** −0.0179**
(0.444) (0.445) (0.003) (0.003) (0.018) (0.020)
Size 0.0101*** 0.0101*** 0.0089 0.0091 0.1622*** 0.1617***
(0.008) (0.008) (0.685) (0.679) (0.000) (0.000)
Industry Yes Yes Yes Yes Yes Yes
Observations 28,499 28,499 28,499 28,499 28,499 28,499
R–squared 0.028 0.028 0.236 0.236 0.024 0.026
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
This table presents the results of firm performance using the OLS regression models and all firm-year observations during the period of 1999–2015. Firm
performance is measured by three change variables: EPS growth (defined as: EPSt
EPSt
1 ), Leverage growth (defined as: Leveraget
Leveraget
1 ) and
Assets growth (defined as: (Assetst
Assetst
1 )/Assetst
1 ). For the definitions of variables refer to Table A2. P–values are in parentheses.
 APPLIED ECONOMICS 2779

connections and distress identify some significant ProbðRecover dummyÞi;t ¼ αi;t þ βi;t Political connectioni;t
impacts of these connections on firm performance þ ηt Macroeconomict
during periods of financial distress. Specifically, þ θi;t Controli;t þ μi;t Industryi;t
local government connectedness is negatively
þ εi;t
associated with firm profitability, in terms of EPS
(4)
growth, during periods of financial distress. With
respect to Leverage growth, the coefficient of PC TobitðRecover timeÞi;t ¼ αi;t þ βi;t Political connectioni;t
during periods of financial distress is positive and þ ηt Macroeconomict
significant at the 1% level. Our results highlight þ θi;t Controli;t þ μi;t Industryi;t
the vital role of the connected CEO or Chair of
þ εi;t
Board for firm performance, especially in periods
of financial distress. Regarding the level of govern- (5)
ment to which the CEO or Chair of Board is In Equation 3, Distress dummy is a dependent variable
connected, we find that central and local political that equals 1 if firm i falls into financial distress in
connections are positively related to Leverage a given year t during the sample period, and 0 other-
growth. These empirical results imply that the wise. A firm falls into financial distress when its Z–
Chinese authorities follow market economy prin- scores are below the cut–off point of 1.1 for two con-
ciples and only assist firms with debt financing secutive years. Recovery dummy in Equation 4 is
after they fall into financial distress. a dependent variable which takes a value of 1 if firm
Macroeconomic conditions and financial funda- i emerges from financial distress in a given year
mentals also significantly affect firm performance. t during the sample period, and 0 otherwise.
Firms that face falling global inflation (ICP world) A distressed firm recovers when its Z–scores are
and strong domestic macroeconomic conditions above the cut–off point of 1.1 for two consecutive
(Consistent) experience higher EPS growth. A one years. Recover time is the number of years from falling
unit decrease in ICP world (increase in Consistent) into financial distress to that of recovery. For those
leads to a 0.0128 (0.0173) unit increase in EPS growth. financially distressed firms that do not recover,
Firm with a longer time period since going public Recover time refers to the number of years from falling
(Age) generally have higher EPS growth, as well as into financial distress until the end of our sample
lower Leverage growth and Assets growth. Leverage is period. The other variables are as per Equation 2.
negatively related to development capacity (Assets Table 5 identifies a limited influence of political
growth), and Size is positively associated with firm connectedness on the occurrence of financial dis-
profitability (EPS growth). tress, due to its subtle impact on firms’ operations
before they fall into financial distress. With respect
to the resolution of financial distress, our results
provide empirical evidence on differences in the
Financial distress and resolution
influence of central and local government political
The previous subsection reports the results for the connections on distress recovery. We find political
first–step of our empirical analysis. In this subsection, connections, especially these connections with cen-
as the second-step of our analysis, we adopt the fol- tral government, are positively related to distress
lowing Probit and Tobit regression models to explore recovery and negatively associated with recovery
determinants of financial distress and recovery. time. Thus, such connections increase the likelihood
of recovery from financial distress and decrease the
ProbðDistress dummyÞi;t ¼ αi;t þ βi;t Political connectioni;t length of time over which recovery occurs. These
þ ηt Macroeconomict findings have been supported by empirical results in
þ θi;t Controli;t þ μi;t Industryi;t Table 4: Political connections have significant
þ εi;t impacts on firm performance during periods of
financial distress, implying that Chinese authorities
(3)
assist firms in debt financing only after they fall into
financial distress and further contribute to higher
2780 Y. HE ET AL.

Table 5. Financial distress and financial distress recovery.

Distress dummy Recover dummy Recover time
PC 0.1652 0.2766* −0.1265**
(0.293) (0.057) (0.019)
PC central −0.2783 0.6301*** −0.2591***
(0.178) (0.002) (0.002)
PC local 0.3090 −0.0994 0.0251
(0.106) (0.553) (0.681)
ICP world −0.0436 −0.0441 0.2009*** 0.2047*** 0.1143*** 0.1140***
(0.437) (0.431) (0.000) (0.000) (0.000) (0.000)
Consistent −0.0058 −0.0057 0.0846*** 0.0849*** 0.0813*** 0.0805***
(0.830) (0.835) (0.002) (0.001) (0.000) (0.000)
Age 1.0266*** 1.0202*** 0.6212*** 0.6326*** −0.7191*** −0.7225***
(0.000) (0.000) (0.002) (0.001) (0.000) (0.000)
Leverage 3.7658*** 3.7688*** −3.0624*** −3.1171*** −0.0021 −0.0020
(0.000) (0.000) (0.000) (0.000) (0.412) (0.436)
Size −0.8166*** −0.8228*** 0.5673*** 0.5729*** −0.1095*** −0.1104***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Industry Yes Yes Yes Yes Yes Yes
Observations 1,337 1,337 966 966 1,059 1,059
Pseudo R–squared 0.626 0.628 0.532 0.537 0.203 0.205
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
This table presents the Probit and Tobit regression results of entering financial distress and recovery from financial distress using the financially distressed
firm sample. The Distress dummy takes a value of 1 if a listed firm falls into financial distress during the sample period, and 0 otherwise. The Recover
dummy takes a value of 1 if a financially distressed firm recovers during the sample period and 0 otherwise. Recover time is the number of years from the
time a firm falls into financial distress to the time it recovers. If a firm does not recover eventually, we take the number of years from the time when this
firm falls into financial distress until the end of the sample period. If a firm falls into financial distress repeatedly, we study on the first period of distress
and last recovery. For the definitions of variables refer to Table A2. P–values are in parentheses.

likelihood of distress recovery and to a shorter this type of endogeneity issue for political connec-
length of recovery time. tions occurs when senior executives attempt to
With respect to macroeconomic conditions and establish these connections to improve poor firm
financial fundamentals, we find that macroeco- performance. In other words, firm performance
nomic conditions, proxied by ICP world and (Performancet ) affects incentives to create political
Consistent, contribute to higher likelihood of recov- connections in the present (PCt ) or future (PCtþ1 ),
ery from financial distress. Moreover, our empirical rather than in the past (PCt
1 ). To control for this
results show that smaller firms with higher leverage potential endogeneity issue, we utilize the first–
ratios have a higher likelihood of falling into finan- order lag of political connection (PCt
1 ) in our
cial distress, consistent with prior studies (O’Hara analyses and re-estimate its impacts on firm per-
and Shaw 1990; Opler and Titman 1994; Caves formance, financial distress and recovery.
1998). Similarly, larger firms with lower leverage As shown in Table 6, our results using the first–
ratios are more likely to emerge from financial order lag of political connection (PCt
1 ) are simi-
distress and over a shorter time period. lar as those in Tables 4 and 5. These results sug-
gest that the potential endogeneity issue has
limited influence on our empirical analyses, sup-
Endogeneity tests porting our findings in Sections 4.1 and 4.2.
Political connection: first–order lag
In previous sections, we proposed that political Selection bias
connectedness affects firm performance and Macroeconomic conditions and financial funda-
further has impacts on financial distress and mentals influence the likelihood of falling into finan-
recovery. Nevertheless, there may be a potential cial distress (Bris, Schwartz, and Welch 2005; Fan,
endogeneity issue between political connections Huang, and Zhu 2013; Tinoco and Wilson 2013;
and firm performance: Financially distressed Bhattacharjee and Han 2014), which may lead to
firms or those with unsatisfied performance may an endogeneity issue–selection bias for firm perfor-
attempt to establish political connections to solve mance. Therefore, we adopt the Heckman Two-
their difficulties, which will have an impact on stage regression procedure to control for potential
financial distress and recovery. It is logical that endogeneity arising from the inclusion of the above
 APPLIED ECONOMICS 2781

Table 6. Endogeneity – political connection: First–order lag.

Panel A: Firm performance
EPS growth Leverage growth Assets growth
PCt-1 0.0128* −0.0077 −0.0596***
(0.076) (0.469) (0.004)
PCt-1 * Distress −0.1132 1.1859*** 1.0774
(0.360) (0.004) (0.314)
PC centralt-1 0.0052 −0.0099 −0.0145
(0.615) (0.550) (0.565)
PC centralt-1 * Distress 0.0816 0.5260 1.1491
(0.750) (0.114) (0.598)
PC localt-1 0.0104 −0.0004 −0.0969***
(0.234) (0.935) (0.000)
PC localt-1 * Distress −0.2006* 1.0687*** 1.3591
(0.088) (0.004) (0.227)
Distress 0.5654*** 0.5728*** −1.7158*** −1.6896*** 1.2553*** 1.1579***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.001)
ICP world −0.0116*** −0.0116*** 0.0012 0.0013 0.0043 0.0046
(0.000) (0.000) (0.919) (0.913) (0.685) (0.671)
Consistent 0.0191*** 0.0191*** −0.0074 −0.0075 −0.0036 −0.0032
(0.000) (0.000) (0.128) (0.125) (0.391) (0.443)
Age 0.0189*** 0.0183*** −0.1030*** −0.1040*** −0.0352** −0.0294*
(0.004) (0.008) (0.000) (0.000) (0.035) (0.075)
Leverage −0.0074 −0.0075 0.4537*** 0.4533*** −0.0201** −0.0186**
(0.456) (0.450) (0.003) (0.003) (0.014) (0.020)
Size 0.0142*** 0.0144*** 0.0105 0.0107 0.1739*** 0.1735***
(0.000) (0.000) (0.657) (0.654) (0.000) (0.000)
Industry Yes Yes Yes Yes Yes Yes
Observations 25,822 25,822 25,822 25,822 25,822 25,822
R–squared 0.028 0.028 0.237 0.237 0.024 0.025
Panel B: Financial distress and financial distress recovery
Distress dummy Recover dummy Recover time
PCt-1 0.0276 0.4554*** −0.0881
(0.865) (0.002) (0.108)
PC centralt-1 −0.3243 0.5733*** −0.1882**
(0.152) (0.009) (0.036)
PC localt-1 0.2082 0.1283 0.0263
(0.271) (0.454) (0.670)
ICP world −0.0485 −0.0502 0.1940*** 0.1965*** 0.1167*** 0.1154***
(0.394) (0.382) (0.000) (0.000) (0.000) (0.000)
Consistent −0.0123 −0.0148 0.0784*** 0.0792*** 0.0838*** 0.0836***
(0.664) (0.604) (0.004) (0.004) (0.000) (0.000)
Age 1.1543*** 1.1510*** 0.4627** 0.4748** −0.6687*** −0.6722***
(0.000) (0.000) (0.031) (0.026) (0.000) (0.000)
Leverage 3.6578*** 3.6684*** −3.0589*** −3.0946*** −0.0014 −0.0014
(0.000) (0.000) (0.000) (0.000) (0.573) (0.582)
Size −0.8099*** −0.8079*** 0.5527*** 0.5592*** −0.0980*** −0.0996***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Industry Yes Yes Yes Yes Yes Yes
Observations 1,148 1,148 919 919 1,026 1,026
Pseudo R–squared 0.616 0.618 0.523 0.522 0.195 0.195
*** and ** refer to the significance at the 1%, and 5% levels, respectively.
The Panel A of this table presents the results of firm performance using the OLS regression models. We replace PCt with PCt
1 to control for the potential
endogeneity issue and re-estimate the influence of political connection on firm performance. We delete several observations due to data availability before
1999 or before IPO, such as those for PCt
1 . For the definitions of variables refer to Table A2. P–values are in parentheses.
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
Panel B of this table presents the results of factors of financial distress and recovery with Probit and Tobit regression models using the financially distressed
firm sample. We replace PCt with PCt
1 to control for the potential endogeneity issue and re-estimate the influence of political connection on financial
distress and financial distress recovery. We delete several observations due to data availability before 1999 or before IPO, such as those for PCt
1 . For the
definitions of variables refer to Table A2. P–values are in parentheses.

factors in explaining financial distress. More speci- are related to the likelihood of falling into financial
fically, we regard ICP world, Consistent, Age, distress. After we control for the potential endogene-
Leverage, Size and Industry fixed–effects as instru- ity issue, our results (Panel B) for the second-stage
ments for the likelihood of falling into financial regression are consistent with those in Table 4. It is
distress. From Panel A of Table 7, the results of the worthwhile noting that political connectedness
first–stage regression show that all these instruments remains a statistically significant influence on firm
2782 Y. HE ET AL.

Table 7. Endogeneity – selection bias.

Panel A: The first-stage regression
ICP world −0.0395***
(0.000)
Consistent −0.0350***
(0.000)
Age 0.4472***
(0.000)
Leverage 1.5428***
(0.000)
Size −0.5199***
(0.000)
Industry Yes
Observations 28,499
R–squared -
Panel B: The second–stage regression
EPS growth Leverage growth Assets growth
PC 0.0487 −0.0789 −0.3316
(0.505) (0.784) (0.193)
PC * Distress −0.3080** 1.3560** 1.1958**
(0.027) (0.013) (0.014)
PC central 0.0387 −0.0285 −0.5972*
(0.698) (0.943) (0.086)
PC central * Distress −0.0866 0.7243 2.2988***
(0.680) (0.379) (0.002)
PC local 0.0564 −0.0470 −0.0908
(0.501) (0.887) (0.756)
PC local * Distress −0.3413** 1.1930* 0.6435
(0.028) (0.050) (0.238)
Distress 1.1425*** 1.1419*** −1.7234*** −1.6899*** 2.6611*** 2.6049***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
ICP world −0.0277 −0.0278 0.0472 0.0458 0.1069* 0.1008
(0.119) (0.119) (0.497) (0.510) (0.088) (0.108)
Consistent 0.0245*** 0.0245*** −0.0468 −0.0470 −0.0243 −0.0239
(0.007) (0.007) (0.187) (0.185) (0.447) (0.454)
Age 0.0258 0.0260 −0.6116*** −0.6184*** 0.6300*** 0.6408***
(0.660) (0.657) (0.008) (0.007) (0.002) (0.002)
Leverage −0.0046 −0.0045 0.5276*** 0.5273*** −0.0025 −0.0014
(0.325) (0.328) (0.000) (0.000) (0.881) (0.931)
Size −0.0586* −0.0589* 0.7070*** 0.7080*** 0.0541 0.0535
(0.098) (0.096) (0.000) (0.000) (0.665) (0.668)
Mill’s ratio 0.4620*** 0.4609*** −0.9867*** −0.9868*** 2.0478*** 2.0460***
(0.000) (0.000) (0.001) (0.001) (0.000) (0.000)
Industry Yes Yes Yes Yes Yes Yes
Observations 28,499 28,499 28,499 28,499 28,499 28,499
R–squared - - - - - -
*** refers to the significance at the 1% level.
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
This table presents the Heckman regression results of firm performance using firm-year observations for the full sample during the period of 1999–2015. For
the first–stage, the dependent variable is a dummy variable which equals 1 if a firm falls into financial distress during 1999–2015. For the second-stage, the
Mill’s ratio is calculated from the first–stage regression. Panel A and Panel B show the empirical results of the first- and second-stage regressions,
respectively. For the definitions of variables refer to Table A2. P–values are in parentheses.

performance during periods of financial distress. those with small net losses. Considering the
Therefore, these results suggest that the potential potential impact of firm behaviour before dis-
endogeneity issue has not affected our analyses and tress on recovery, we therefore utilize a one–
support our finding that political connections are stage method (Heckman–probit Two-stage
more likely to affect firm operations after they fall regression model) to complete systematic
into financial distress. empirical analyses. Similar to Section 4.3.2, we
regard ICP world, Consistent, Age, Leverage, Size
Financial distress recovery and Industry fixed–effects as instruments for the
Firm behaviour before financial distress may likelihood of falling into financial distress (Bris,
affect the likelihood of recovery. Firms falling Schwartz, and Welch 2005; Fan, Huang,
into financial distress due to large net losses and Zhu 2013; Tinoco and Wilson 2013;
have a lower possibility of recovering than Bhattacharjee and Han 2014).
 APPLIED ECONOMICS 2783

The Panel A of Table 8 shows that these instru- (Altman, Hartzell, and Peck 1995; Altman 2005).
ments are related to financial distress, confirming Given the potential impacts of financial fundamen-
the significant impacts of macroeconomic condi- tals used to construct the Z–score on financial dis-
tions and financial fundamentals on financial dis- tress, we utilize alternative measures of financial
tress. With respect to the Panel B of Table 8, we distress, including ST status (Wang and Deng 2006;
find that the empirical results are similar to those Wang and Li 2007) and interest coverage ratio (Kam,
in Section 4.2, where political connections are Citron, and Muradoglu 2008; Bhattacharjee and Han
positively related to distress recovery, and nega- 2014). We re-estimate the impacts of our proposed
tively associated with recovery time. We conclude factors on financial distress as robustness tests.
that our analyses are robust to various methodol- We define a firm as financially distressed if the
ogies (both one – and two–stage methods). firm is categorized as having ST status in a given year,
while distress recovery is identified as occurring
when the firm is removed from ST status during
Robustness tests: alternative measures of financial
a given year within the sample period. In addition
distress
to ST status, we further define firm financial distress
As discussed in Section 3, we define financially dis- and recovery based on the interest coverage ratio.
tressed firms based on the EM Z–score model Firms fall into financial distress if their ratios are

Table 8. Endogeneity – financial distress recovery.

Panel A: The first-stage regression
ICP world −0.0377 −0.0385 −0.0637** −0.0642**
(0.194) (0.185) (0.025) (0.024)
Consistent −0.0379** −0.0375** −0.0512*** −0.0511***
(0.017) (0.018) (0.001) (0.001)
Age 2.0868*** 2.0769*** 1.8294*** 1.8290***
(0.000) (0.000) (0.000) (0.000)
Leverage 2.1069*** 2.1191*** 2.3232*** 2.3229***
(0.000) (0.000) (0.000) (0.000)
Size −0.5388*** −0.5379*** −0.5139*** −0.5142***
(0.000) (0.000) (0.000) (0.000)
Industry Yes Yes Yes Yes
Observations 2,038 2,038 2,038 2,038
R–squared - - - -
Panel B: The second–stage regression
Recover dummy Recover time
PC 0.2543** −0.1189***
(0.046) (0.009)
PC central 0.4841*** −0.1592**
(0.009) (0.024)
PC local 0.0105 −0.0491
(0.941) (0.329)
ICP world 0.1389*** 0.1377*** 0.0897*** 0.0898***
(0.000) (0.000) (0.000) (0.000)
Consistent 0.0682*** 0.0681*** 0.0637*** 0.0633***
(0.002) (0.002) (0.000) (0.000)
Age 0.3702 0.4466* −0.6427*** −0.6442***
(0.119) (0.057) (0.000) (0.000)
Leverage −1.8995*** −1.8668*** 0.0030 0.0031
(0.000) (0.000) (0.176) (0.165)
Size 0.5214*** 0.5070*** −0.0399** −0.0403**
(0.000) (0.000) (0.018) (0.017)
Mill’s ratio 0.1593 0.2428 −0.7442*** −0.7452***
(0.595) (0.420) (0.000) (0.000)
Industry Yes Yes No No
Observations 2,038 2,038 2,038 2,038
R–squared - - - -
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
This table presents the results of factors of distress recovery with Heckman regression models using the financially
distressed firm sample. For the first–stage, the dependent variable is a dummy variable which equals to 1 if a firm
falls into financial distress during 1999–2015. For the second-stage, the Mill’s ratio is calculated from the first–stage
regression. Panel A and Panel B show the empirical results of the first- and second-stage regressions, respectively.
For the definitions of variables refer to Table A2. P–values are in parentheses.
2784 Y. HE ET AL.

below the cut–off point of 1 for two consecutive TobitðRecover timeÞi;t ¼ αi;t þ βi;t Political connectioni;t
years, while they emerge from financial distress if þ ηt Macroeconomict
their ratios exceed the cut–off point for two conse- þ θi;t Agei;t þ γi;t Leveragei;t
cutive years. Based on the ST status (interest cover-
þ δi;t Leverage2i;t þ λi;t Sizei;t
age ratio) criterion, we observe a total of 524 (1,539)
firms falling into financial distress, and 420 (637) þ μi;t Industryi;t þ εi;t
financially distressed firms recovering during the (8)
sample period. We list the details on financially As per Table 10, we find that Leverage is positively
distressed firms in Table 1. and Leverage2 is negatively related to the probability
As shown in Table 9, our results almost remain of entering financial distress, evidence of a convex
the same when using alternative measures of relationship between firm leverage and the probabil-
financial distress. This confirms the robustness of ity of entering into financial distress. Conversely,
our previous analyses using the EM Z–score Leverage is negatively and Leverage2 is positively
model to identify distressed firms. associated with the probability of recovery, indicating
a concave relationship between firm leverage and the
probability of recovery from financial distress. These
Additional tests results show that increments to leverage contribute to
a lower (higher) likelihood of financial distress
Non–linearities in the impact of leverage
(recovery) after reaching the equilibrium point–
The empirical results in Tables 4 and Table 5 show
peak (bottom). These empirical results also support
that political connectedness is positively related to
our finding that political connectedness assists firm
Leverage growth during periods of financial distress,
in gaining higher increments to debt financing in the
and contributes to a higher likelihood of recovery
period of financial distress, and so contributes to
from financial distress. Thus, these results suggest
a higher likelihood of distress recovery.
a positive relationship between firm leverage and
distress recovery. Nevertheless, Table 5 shows that
leverage is positively related to the Distress dummy Subsample tests
and negatively associated with the Recover dummy. Subsample test: SOE vs. non–SOE. Previous stu-
The above results imply a non-linear relationship dies (Wu et al. 2012; Cull et al. 2015) report that
between firm leverage and financial distress (recov- political connections have different influences on
ery). We attempt to investigate these relationships firm operations between SOEs and non–SOEs but
with the following equations: neglect discussion of the different influences of
political connections on financial distress. In this
ProbðDistress dummyÞi;t ¼ αi;t þ βi;t Political connectioni;t
case, we divide our distressed firm sample into
þ ηt Macroeconomict SOEs14 and non–SOEs and re-estimate the impacts
þ θi;t Agei;t þ γi;t Leveragei;t of such connections on firm financial distress and
þ δi;t Leverage2i;t þ λi;t Sizei;t related resolution with these two sub–samples.
þ μi;t Industryi;t þ εi;t Table 11 shows that political connections, espe-
(6) cially the CEO or Chair of Board connectedness with
central government, have stronger influence on the
ProbðRecover dummyÞi;t ¼ αi;t þ βi;t Political connectioni;t recover likelihood for non–SOEs than that of SOEs.
þ ηt Macroeconomict We find that non–SOEs with political connectedness
þ θi;t Agei;t þ γi;t Leveragei;t spend less time to recover from financial distress
than SOEs, as the coefficients of PC are −0.2620
þ δi;t Leverage2i;t þ λi;t Sizei;t
and −0.1680, respectively. These results indicate
þ μi;t Industryi;t þ εi;t that non–SOEs achieve a higher likelihood of distress
(7) recovery with less time by hiring politically
14
We also consider splitting the SOEs into two subsamples: Central and local SOEs, to further explore the different influences of political connection on
central and local SOEs. However, we observe too few observations, especially for the subsample of central SOEs, to complete empirical analyses.
Table 9. Robustness tests.
ST Status Interest Coverage Ratio
Distress dummy Recover dummy Recover time Distress Dummy Recover Dummy Recover time
PC −0.0054 0.1833*** −0.0783** 0.0960*** 0.0949*** −0.0333**
(0.925) (0.003) (0.019) (0.004) (0.009) (0.035)
PC central −0.0384 0.4059*** −0.2304*** 0.2232*** 0.1518*** −0.1220***
(0.624) (0.000) (0.000) (0.000) (0.000) (0.000)
PC local 0.0140 0.0445 0.0122 −0.0978** 0.0027 0.0697***
(0.829) (0.514) (0.739) (0.018) (0.950) (0.000)
ICP world −0.0389* −0.0390* −0.0277* −0.0275* 0.1008*** 0.1007*** −0.0549*** −0.0548*** −0.0079 −0.0072 0.1890*** 0.1880***
(0.080) (0.080) (0.081) (0.084) (0.000) (0.000) (0.000) (0.000) (0.446) (0.490) (0.000) (0.000)
Consistent 0.0215** 0.0215** −0.0479*** −0.0474*** 0.0782*** 0.0776*** −0.0082 −0.0090 −0.0109* −0.0103* 0.1224*** 0.1216***
(0.028) (0.028) (0.000) (0.000) (0.000) (0.000) (0.141) (0.106) (0.051) (0.067) (0.000) (0.000)
Age 0.8290*** 0.8271*** 0.7448*** 0.7547*** −0.6873*** −0.6908*** −0.2694*** −0.2515*** 0.2605*** 0.2739*** −0.2270*** −0.2430***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Leverage 2.4751*** 2.4755*** −0.5754*** −0.5849*** −0.0029 −0.0027 0.0738** 0.0806*** 0.0031 0.0030 −0.0089*** −0.0086***
(0.000) (0.000) (0.007) (0.005) (0.347) (0.383) (0.012) (0.007) (0.645) (0.661) (0.000) (0.000)
Size −0.3523*** −0.3514*** 0.2012*** 0.1966*** −0.1044*** −0.1016*** −0.0019 0.0010 0.0813*** 0.0805*** −0.0735*** −0.0725***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.916) (0.953) (0.000) (0.000) (0.000) (0.000)
Industry Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 3,742 3,745 2,998 2,998 3,028 3,028 7,424 7,424 9,744 9,744 9,866 9,866
Pseudo R–squared 0.246 0.246 0.160 0.164 0.121 0.123 0.055 0.058 0.051 0.051 0.198 0.200
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
This table presents the results of using the ST status and Interest Coverage Ratio as alternative financial distress measures. For the definitions of variables refer to Table A2. P–values are in parentheses.
APPLIED ECONOMICS
2785
2786 Y. HE ET AL.

Table 10. Non–linearities in the impact of leverage.

Distress dummy Recover dummy Recover time
PC 0.1647 0.2766* −0.1186**
(0.295) (0.057) (0.029)
PC central −0.2796 0.6304*** −0.2524***
(0.176) (0.002) (0.003)
PC local 0.3094 −0.0997 0.0317
(0.106) (0.552) (0.604)
ICP world −0.0438 −0.0443 0.2009*** 0.2046*** 0.1134*** 0.1131***
(0.434) (0.428) (0.000) (0.000) (0.000) (0.000)
Consistent −0.0061 −0.0060 0.0848*** 0.0850*** 0.0801*** 0.0792***
(0.821) (0.826) (0.002) (0.001) (0.000) (0.000)
Age 1.0267*** 1.0204*** 0.6204*** 0.6318*** −0.7265*** −0.7304***
(0.000) (0.000) (0.002) (0.001) (0.000) (0.000)
Leverage 3.8844*** 3.8874*** −3.1067*** −3.1623*** 0.0056 −0.0060
(0.000) (0.000) (0.000) (0.000) (0.370) (0.330)
Leverage2 −0.0596*** −0.0597*** 0.0249*** 0.0253*** −0.0001 −0.0001
(0.000) (0.000) (0.000) (0.000) (0.177) (0.157)
Size −0.8173*** −0.8235*** 0.5676*** 0.5732*** −0.1034*** −0.1039***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Industry Yes Yes Yes Yes Yes Yes
Observations 1,337 1,337 966 966 1,059 1,059
Pseudo R–squared 0.626 0.628 0.532 0.537 0.204 0.205
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
This table presents the results of financial distress and distress recovery with Probit and Tobit regression models using the financially distressed firm
subsample. Leverage2 stands for the square of Leverage. For the definitions of variables refer to Table A2. P–values are in parentheses.

connected senior executives, emphasizing the mag- and generally spend more time on recovery. For
nitude of politically connected managers for private 2008–2015, we find that central government political
firms. connectedness also has a positive influence on dis-
tress recovery, with the influence being stronger
Sub–sample test: 1999–2007 vs. 2008–2015. than that in the period of 1999–2007. The coeffi-
Unlike previous studies (Kam, Citron, and cients of PC central on Recover dummy are 0.9521
Muradoglu 2008; Fan, Huang, and Zhu 2013; (2008–2015) and 0.7169 (1999–2007), respectively.
Bhattacharjee and Han 2014), this paper contributes This stronger influence may be a result of the Four
to the literature: a) by exploring the impacts of poli- Trillion RMB Economic Stimulus Program after the
tical connection and other proposed factors on finan- GFC, when firms with central government connec-
cial distress; b) with an updated sample covering tions were more likely to acquire funding support.
several important events in recent decade, such as
the NTS reform in 2007 and the GFC in 2008. These
V. Conclusion
events may impact on firm behaviours in relation to
financial distress. In this subsection, we therefore re- This study sheds light on the role of political con-
estimate the impact of our chosen variables on the nections on entry into and recovery from financial
likelihood of financial distress and recovery with two distress in this important transition economy. We
subsamples, ‘1999–2007ʹ and ‘2008–2015ʹ.15 The explore how political connections, macroeconomic
empirical results are presented in Table 12. conditions and financial fundamentals affect firm
From Table 12, we find that different levels of performance, financial distress and resolution in
political connectedness affect recovery from finan- China. Using a two-step empirical analysis, we first
cial distress differently during the 1999–2007 sample observe that political connections have significant
period. To be specific, firms with central govern- influence on firm performance during periods of
ment connections are more likely to recover from financial distress. We then find political connected-
financial distress over shorter time periods. ness has limited influence on entry into financial
Conversely, firms with local government connec- distress but does contribute to a higher likelihood
tions have a lower likelihood of distress recovery of recovery, indicating that Chinese authorities
15
We also complete the comparison test by separating the distressed firm sample into two other subsamples, ‘1999–2006’ and ‘2008–2015’, and the results
remain the same.
Table 11. Subsample tests: SOE vs. non–SOE.
SOE Non–SOE
Distress dummy Recover dummy Recover time Distress Dummy Recover Dummy Recover time
PC 0.2434 0.3558* −0.1280** −0.1259 0.5954** −0.2620**
(0.236) (0.053) (0.029) (0.736) (0.031) (0.018)
PC central −0.4870* 0.5647** −0.0612 −0.4995 0.7991** −0.5827***
(0.062) (0.029) (0.524) (0.225) (0.039) (0.000)
PC local 0.4993** −0.0311 −0.0852 0.2577 0.0698 0.0765
(0.041) (0.882) (0.193) (0.629) (0.819) (0.546)
ICP world −0.0621 −0.0684 0.2033*** 0.2058*** 0.1028*** 0.1023*** 0.0334 0.0408 0.2529*** 0.2462*** 0.1004*** 0.1006***
(0.314) (0.269) (0.000) (0.000) (0.000) (0.000) (0.811) (0.765) (0.001) (0.002) (0.000) (0.000)
Consistent −0.0587* −0.0642* 0.0916*** 0.0927*** 0.0680*** 0.0675*** 0.1297** 0.1330** 0.1228** 0.1099** 0.0756*** 0.0754***
(0.077) (0.061) (0.007) (0.007) (0.000) (0.000) (0.017) (0.017) (0.011) (0.023) (0.000) (0.000)
Age 1.6408*** 1.6481*** 0.7905** 0.7771** −1.0595*** −1.0604*** 0.7711** 0.7637** 1.0082*** 0.9596*** −0.5504*** −0.5873***
(0.000) (0.000) (0.013) (0.013) (0.000) (0.000) (0.021) (0.026) (0.001) (0.001) (0.000) (0.000)
Leverage 4.4034*** 4.4199*** −3.1919*** −3.2336*** −0.0049* −0.0049* 3.9324*** 3.9174*** −3.1617*** −3.2084*** 0.0044 0.0048
(0.000) (0.000) (0.000) (0.000) (0.058) (0.059) (0.000) (0.000) (0.000) (0.000) (0.402) (0.355)
Size −0.7795*** −0.7691*** 0.5669*** 0.5641*** −0.1452*** −0.1471*** −1.4444*** −1.4649*** 0.5195*** 0.5294*** 0.0222 0.0268
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.568) (0.483)
Industry Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 997 997 621 621 735 735 331 331 272 272 324 324
Pseudo R–squared 0.658 0.662 0.534 0.535 0.293 0.292 0.702 0.705 0.543 0.545 0.196 0.206
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
This table presents the results of subsample tests: SOE vs. non–SOE. We divide our financially distressed firm sample into two subsamples, SOEs and non–SOEs, and re-estimate the impacts of proposed variables on firm
financial distress and distress recovery for these two groups. For the definitions of variables refer to Table A2. P–values are in parentheses.
APPLIED ECONOMICS
2787
 2788
Y. HE ET AL.

Table 12. Sub–sample tests: 1999–2007 vs. 2008–2015.

Subsample 1999–2007 Subsample 2008–2015
Distress dummy Recover dummy Recover time Distress Dummy Recover Dummy Recover time
PC 0.1622 0.2707 0.0067 0.7799 0.0841 −0.1042*
(0.478) (0.386) (0.916) (0.247) (0.702) (0.082)
PC central 0.3302 0.7169** −0.3857*** 2.3889*** 0.9521*** −0.1280
(0.317) (0.044) (0.000) (0.003) (0.003) (0.162)
PC local 0.0136 −0.6193* 0.2329*** 0.6034 −0.2542 −0.0499
(0.959) (0.091) (0.001) (0.430) (0.306) (0.455)
ICP world 0.9748*** 0.9870*** −0.3174 −0.3930 −0.5976*** −0.5860*** −0.5506*** −0.5554*** 0.1012* 0.0804 0.1901*** 0.1919***
(0.000) (0.000) (0.190) (0.113) (0.000) (0.000) (0.000) (0.000) (0.060) (0.145) (0.000) (0.000)
Consistent 0.0794* 0.0784* −0.0005 0.0013 −0.0591*** −0.0614*** −0.3192*** −0.3241*** 0.0462 0.0361 0.1164*** 0.1171***
(0.053) (0.057) (0.994) (0.986) (0.000) (0.000) (0.000) (0.000) (0.236) (0.357) (0.000) (0.000)
Age 0.9236** 0.9441** 3.2078*** 3.4954*** −0.3191*** −0.3931*** −0.4396 −0.6144 −0.6355 −1.0307** −0.0284 −0.0085
(0.021) (0.020) (0.000) (0.000) (0.001) (0.000) (0.600) (0.493) (0.156) (0.025) (0.795) (0.939)
Leverage 4.9834*** 5.0015*** −3.7138*** −4.0189*** −0.0080 −0.0088 2.6904*** 2.6494*** −2.6148*** −2.7470*** 0.0009 0.0010
(0.000) (0.000) (0.000) (0.000) (0.164) (0.118) (0.000) (0.000) (0.000) (0.000) (0.673) (0.658)
Size −0.9055*** −0.8928*** 0.0465 0.0487 0.0167 −0.0051 −1.3015*** −1.2843*** 0.8299*** 0.8220*** −0.1161*** −0.1168***
(0.000) (0.000) (0.806) (0.805) (0.611) (0.874) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Industry Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 709 709 321 321 476 476 165 165 467 467 566 566
Pseudo R–squared 0.759 0.759 0.534 0.546 0.360 0.379 0.504 0.506 0.563 0.572 0.403 0.403
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
This table presents the results of subsample tests: 1999–2007 vs. 2008–2015. We divide our financially distressed firm sample into two subsamples, including the sample of 1999–2007 and 2008–2015, and re-estimate
the impacts of proposed variables on firm financial distress and recovery. For the definitions of variables refer to Table A2. P–values are in parentheses.
 APPLIED ECONOMICS 2789

intervene in business operations strongly after firm Campbell, J. Y., J. Hilscher, and J. Szilagyi. 2008. “In Search
financial distress. In addition, the likelihood of of Distress Risk.” Journal of Finance 63 (6): 2899–2939.
Caves, R. E. 1998. “Industrial Organization and New
emerging from financial distress is higher for firms
Findings on the Turnover and Mobility of Firms.”
with central government political connections, but is Journal of Economic Literature 36 (4): 1947–1982.
unaffected by the presence of local government con- Chan, K. C., and N. F. Chen. 1991. “Structural and Return
nectedness. Not surprisingly, financially distressed Characteristics of Small and Large Firms.” Journal of
firms have a higher likelihood of recovery in the Finance 46 (4): 1467–1484.
presence of favourable macroeconomic conditions. Chen, C. J. P., Z. Li, X. Sun, and Z. Sun. 2011a. “Rent–
Seeking Incentives, Corporate Political Connections, and
Regarding our additional tests, we find firms with
the Control Structure of Private Firms: Chinese Evidence.”
political connections are better able to access addi- Journal of Corporate Finance 17 (2): 229–243.
tional debt financing, which eventually contributes Chen, S., Z. Sun, S. Tang, and D. Wu. 2011b. “Government
to a higher likelihood of distress recovery. Non– Intervention and Investment Efficiency: Evidence from
SOEs with political connections have a higher like- China.” Journal of Corporate Finance 17 (2): 259–271.
lihood of distress recovery, compared to SOEs. Cull, R., W. Li, B. Sun, and L. C. Xu. 2015. “Government
Connections and Financial Constraints: Evidence from
Finally, central government connectedness exhibits
a Large Representative Sample of Chinese Firms.”
a more significant impact on distress resolution in Journal of Corporate Finance 32: 271–294.
the period following the 2008 GFC. Donker, H., B. Santen, and S. Zahir. 2009. “Ownership Structure
and the Likelihood of Financial Distress in the Netherlands.”
Applied Financial Economics 19 (21): 1687–1696.
Eckbo, B., and K. Thorburn. 2003. “Control Benefits and
Disclosure statement
CEO Discipline in Automatic Bankruptcy Auctions.”
No potential conflict of interest was reported by the authors. Journal of Financial Economics 69 (1): 227–258.
Faccio, M. 2010. “Differences between Politically Connected
and Nonconnected Firms: A Cross–Country Analysis.”
ORCID Financial Management 39 (3): 905–927.
Fan, J. P. H., J. Huang, and N. Zhu. 2013. “Institutions,
Ron P. McIver http://orcid.org/0000-0001-8754-1560
Ownership Structures, and Distress Resolution in China.”
Journal of Corporate Finance 23: 71–87.
Fan, J. P. H., T. J. Wong, and T. Zhang. 2007. “Politically
References Connected CEOs, Corporate Governance, and post–IPO
Performance of China’s Newly Partially Privatized Firms.”
Almamy, J., J. Aston, and L. N. Ngwa. 2016. “An Evaluation of Journal of Financial Economics 84 (2): 330–357.
Altman’s Z–Score Using Cash Flow Ratio to Predict Corporate Fonseka, M. M., X. Yang, G. L. Tian, and
Failure amid the Recent Financial Crisis: Evidence from the S. R. N. Colombage. 2015. “Political Connections,
UK.” Journal of Corporate Finance 36: 278–285. Ownership Structure and Private–Equity Placement
Altman, E. 2005. “An Emerging Market Credit Scoring Decision: Evidence from Chinese Listed Firms.” Applied
System for Corporate Bonds.” Emerging Markets Review Economics 47 (52): 5648–5666.
6 (4): 311–323. Francis, B. B., I. Hasan, and X. Sun. 2009. “Political Connections
Altman, E., J. Hartzell, and M. Peck. 1995. Emerging Markets and the Process of Going Public: Evidence from China.”
Corporate Bonds: A Scoring System. New York, NY: Journal of International Money and Finance 28 (4): 696–719.
Salomon Brothers. Gilson, S. 1997. “Transaction Costs and Capital Structure
Bhattacharjee, A., and J. Han. 2014. “Financial Distress of Chinese Choice: Evidence from Financially Distressed Firms.”
Firms: Microeconomic, Macroeconomic and Institutional Journal of Finance 52 (1): 161–196.
Influences.” China Economic Review 30: 244–262. Höwer, D. 2016. “The Role of Bank Relationships When
Bhattacharjee, A., C. Higson, S. Holly, and P. Kattuman. Firms are Financially Distressed.” Journal of Banking and
2009. “Macroeconomic Conditions and Business Exit: Finance 65: 59–75.
Determinants of Failures and Acquisitions of UK Firms.” Ivashina, V., I. Benjamin, and C. S. David. 2016. “The Ownership
Economica 76 (301): 108–131. and Trading of Debt Claims in Chapter 11 Restructurings.”
Boubakri, N., J. C. Cosset, and W. Saffar. 2008. “Political Journal of Financial Economics 119 (2): 316–335.
Connections of Newly Privatized Firms.” Journal of Jovanovic, B., and P. L. Rousseau. 2002. “The Q–Theory of
Corporate Finance 14 (5): 654–673. Mergers.” The American Economic Review 92 (2): 198–204.
Bris, A., A. Schwartz, and I. Welch. 2005. “Who Should Pay Kam, A., D. Citron, and G. Muradoglu. 2008. “Distress and
for the Bankruptcy Costs?” The Journal of Legal Studies 34 Restructuring in China: Does Ownership Matter?” China
(2): 295–341. Economic Review 19 (4): 567–579.
2790 Y. HE ET AL.

Koh, S., R. B. Durand, L. Dai, and M. Chang. 2015. “Financial Pakes, A., and R. Ericson. 1998. “Empirical Implications of
Distress: Lifecycle and Corporate Restructuring.” Journal of Alternative Models of Firm Dynamics.” Journal of
Corporate Finance 33: 19–33. Economic Theory 79 (1): 1–46.
Lennox, C. 1999. “Identifying Failing Companies: A Re– Richardson, G., R. Lanis, and G. Taylor. 2015. “Financial Distress,
Evaluation of the Logit, Probit and DA Approaches.” outside Directors and Corporate Tax Aggressiveness Spanning
Journal of Economics and Business 51 (4): 347–364. the Global Financial Crisis: An Empirical Analysis.” Journal of
Leuz, C., and F. Oberholzer–Gee. 2006. “Political Relationships, Banking and Finance 52: 112–129.
Global Financing, and Corporate Transparency.” Journal of Rosenfeld, C. M. 2014. “The Effect of Banking Relationship
Financial Economics 81 (2): 411–439. on the Future of Financially Distressed Firms.” Journal of
Li, H., L. Meng, Q. Wang, and L. –. A. Zhou. 2008. “Political Corporate Finance 25: 403–418.
Connections, Financing and Firm Performance: Evidence Siegfried, J. J., and L. B. Evans. 1994. “Empirical Studies of
from Chinese Private Firms.” Journal of Development Entry and Exit: A Survey of the Evidence.” Review of
Economics 87 (2): 282–299. Industrial Organization 9 (2): 121–155.
Lin, K. J., J. Tan, L. Zhao, and K. Karim. 2015. “In the Name Su, J., M. Zhang, and W. Zhang. 2013. “The Effect of Political
of Charity: Political Connections and Strategic Corporate Connections on Acquisition–Evidence from Chinese
Social Responsibility in a Transition Economy.” Journal of nonSOEs.” Applied Financial Economics 23 (24):
Corporate Finance 32: 327–346. 1871–1890.
Lin, X., Y. Zhang, and N. Zhu. 2009. “Does Bank Ownership Tinoco, M. H., and N. Wilson. 2013. “Financial Distress
Increase Firm Value? Evidence from China.” Journal of and Bankruptcy Prediction among Listed Companies
International Money and Finance 28 (4): 720–737. Using Accounting, Market and Macroeconomic
Liu, J. 2004. “Macroeconomic Determinants of Corporate Variables.” International Review of Financial Analysis
Failures: Evidence from the UK.” Applied Economics 36 (9): 30: 394–419.
939–945. Wang, Z., and H. Li. 2007. “Financial Distress Prediction
Liu, Q., and G. Tian. 2012. “Controlling Shareholder, of Chinese Listed Companies: A Rough Set
Expropriations and Firm’s Leverage Decision: Evidence Methodology.” Chinese Management Studies 1 (2):
from Chinese Non–Tradable Share Reform.” Journal of 93–110.
Corporate Finance 18 (4): 782–803. Wang, Z. J., and X. L. Deng. 2006. “Corporate Governance
Maksimovic, V., and G. Phillips. 1998. “Asset Efficiency and and Financial Distress: Evidence from Chinese Listed
Reallocation Decisions of Bankrupt Firms.” Journal of Companies.” The Chinese Economy 39 (5): 5–27.
Finance 53 (5): 1495–1532. Whitaker, R. 1999. “The Early Stages of Financial Distress.”
Nickell, P., W. Perraudin, and S. Varotto. 2000. “Stability of Journal of Economics and Finance 23 (2): 123–133.
Rating Transitions.” Journal of Banking and Finance 24 Wu, W. F., C. F. Wu, C. Y. Zhou, and J. Wu. 2012. “Political
(1–2): 203–227. Connections, Tax Benefits and Firm Performance:
O’Hara, M., and W. Shaw. 1990. “Deposit Insurance and Evidence from China.” Journal of Accounting and Public
Wealth Effects: The Value of Being “Too Big to Fail”.” Policy 31 (3): 277–300.
Journal of Finance 45 (5): 1587–1600. You, J., and G. Du. 2012. “Are Political Connections
Opler, T. C., and S. Titman. 1994. “Financial Distress and a Blessing or a Curse? Evidence from CEO Turnover in
Corporate Performance.” Journal of Finance 49 (3): China.” Corporate Governance: An International Review 20
1015–1040. (2): 179–194.
Appendices

Table A1. Robustness tests – different cut–off points.

Cut–off point 0 Cut–off point 2.6
Distress dummy Recover dummy Recover time Distress dummy Recover dummy Recover time
PC 0.1483 0.0505 −0.0847 0.2496** 0.4040*** −0.1345***
(0.424) (0.711) (0.148) (0.038) (0.001) (0.005)
PC central −0.1346 0.2130 −0.1381 0.1754 0.5728*** −0.2122***
(0.565) (0.392) (0.156) (0.341) (0.000) (0.004)
PC local 0.2764 −0.1349 −0.0085 0.1542 0.2303* −0.0660
(0.198) (0.496) (0.896) (0.226) (0.092) (0.219)
ICP world −0.1199* −0.1173* 0.2104*** 0.2127*** 0.1122*** 0.1128*** 0.0842* 0.0841* 0.1185*** 0.1177*** 0.1135*** 0.1138***
(0.077) (0.084) (0.000) (0.000) (0.000) (0.000) (0.063) (0.063) (0.000) (0.000) (0.000) (0.000)
Consistent −0.0559* −0.0562* 0.1159*** 0.1171*** 0.0838*** 0.0838*** 0.0001 0.0007 0.0545*** 0.0528*** 0.0817*** 0.0816***
(0.088) (0.086) (0.001) (0.001) (0.000) (0.000) (0.995) (0.972) (0.004) (0.005) (0.000) (0.000)
Age 1.2183*** 1.2144*** 0.4698* 0.4614* −0.8672*** −0.8661*** 0.8687*** 0.8602*** 0.5697*** 0.5898*** −0.7123*** −0.7174***
(0.000) (0.000) (0.063) (0.066) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Leverage 3.8610*** 3.8579*** −3.0961*** −3.1325*** −0.0016 −0.0015 3.6997*** 3.7083*** −2.8441*** −2.8507*** 0.0001 0.0001
(0.000) (0.000) (0.000) (0.000) (0.524) (0.545) (0.000) (0.000) (0.000) (0.000) (0.954) (0.964)
Size −1.0065*** −1.0090*** 0.7130*** 0.7118*** −0.1022*** −0.1019*** −0.6274*** −0.6216*** 0.4009*** 0.4008*** −0.0478*** −0.0491***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.007) (0.005)
Industry Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 1,151 1,151 769 769 888 888 1,526 1,526 1,343 1,343 1,431 1,431
Pseudo R-squared 0.710 0.711 0.603 0.604 0.245 0.246 0.538 0.537 0.417 0.420 0.164 0.164
***, ** and * refer to the significance at the 1%, 5% and 10% levels, respectively.
Results of Probit and Tobit regressions to explain financial distress and distress recovery using alternative cut–off points. Definitions of variables remain the same as those in Table A2. P–values are in parentheses.
APPLIED ECONOMICS
2791
2792 Y. HE ET AL.

We adopt alternative cut–off points of EM Z–score determinants of distress recovery, and find that political
model, 0 and 2.6 (Fan, Huang, and Zhu 2013; Altman, connection contributes to higher likelihood of distress
Hartzell, and Peck 1995), as stringent and lax standards to recovery with less time. Our results are robust to alter-
identify seriously and slightly distressed firms for robust- native cut–off points. Therefore, these robustness tests
ness purposes. With these alternative cut–off points and support our findings that political connection mainly
criteria mentioned,16 we observe 149 seriously and 224 impacts on firms operations after they fall into financial
slightly distressed firms, including 87 and 136 distressed distress. More specifically, political connection, especially
firms that recovered by the end of the sample period, the connection with central government, is positively
respectively. related to recovery likelihood (Recover dummy) and nega-
As per Table A1, we use the Probit (Equation 4) and tively associated with the length of recovery time (Recover
Tobit (Equation 5) regression models to investigate the time).

Table A2. Variable definition.

Variable Definition
Dependent variables
EPS growth ● The change of EPS, defined as EPSt
EPSt
1 . EPS is earnings per share.
Leverage ● The change of leverage, defined as Leveraget
Leveraget
1 . Leverage equals to total liabilities divided by total assets.
growth
Assets growth ● The change of total assets, defined as (Assetst
Assetst
1 )/Assetst
1 . Assets represent the total assets.
Distress ● Dummy variable, equals 1 if a listed firm falls into financial distress in a given year during the sample period, and 0 otherwise. If a firm
dummy falls into financial distress repeatedly, we base our study on the first period of financial distress.
Recover ● Dummy variable, equals 1 if a distressed firm recovers at a given time during the sample period and 0 otherwise. If a firm falls into
dummy distress repeatedly, we study on the last recovery.
Recover time ● The number of years from the time a firm falls into financial distress to the time it recovers. If a firm does not recover, we take the
number of years from the time when this firm falls into financial distress until the end of the sample period. If a firm falls into
financial distress repeatedly, we take into account the last round of recovery.
Independent variables
PC ● Dummy variable, equals 1 if chairman of board or CEO is or was a government official, and 0 otherwise.
PC central ● Dummy variable, equals 1 if chairman of board or CEO is or was a central government official, and 0 otherwise.
PC local ● Dummy variable, equals 1 if chairman of board or CEO is or was a local government official, and 0 otherwise.
Distress ● Dummy variable, equals 1 for years in which a firm is in financial distress, and 0 otherwise.
ICP world ● Global Consumer Price Inflation.
Consistent ● The Consistent Macroeconomic Index for China.
Age ● The logarithm of the number of years since the firm was listed.
Leverage ● Total liabilities divided by total assets.
Size ● The logarithm of the value of total assets.
Industry ● Industry dummy variables.

16
We define firms falling into financial distress if their Z–scores are below the cut-off points for two consecutive years, and firms recovering from financial
distress when their Z–scores are above the cut–off points for two consecutive years. The details of financially distressed firms are listed in Table 1.