Predicting the Risk of Financial Distress using

Corporate Governance Measures

Zhiyong Lia, b, Jonathan Crookb, Galina Andreevab, Ying Tanga

a
School of Finance, Southwestern University of Finance and Economics, China
b
Business School, University of Edinburgh, UK

Zhiyong Li: School of Finance, Southwestern University of Finance and Economics, 555

Liutai Avenue, Wenjiang, Chengdu 611130, China. Email: [email protected]

Jonathan Crook: University of Edinburgh Business School, 29 Buccleuch Place, Edinburgh

EH8 9JS, UK. Email: [email protected]

Galina Andreeva: University of Edinburgh Business School, 29 Buccleuch Place, Edinburgh

EH8 9JS, UK. Email: [email protected]

Ying Tang: School of Finance, Southwestern University of Finance and Economics, 555

Liutai Avenue, Wenjiang, Chengdu 611130, China. Email: [email protected]

Corresponding author: Ying Tang

Address: School of Finance, SWUFE, 555 Liutai Avenue, Wenjiang, Chengdu, Sichuan, China

Email: [email protected]

Tel: +86 18508226040

 Predicting the Risk of Financial Distress using Corporate

Governance Measures

Abstract

Corporate governance is an important determinant of enterprise performance. Poor corporate

governance can damage the interests of shareholders, and may lead to company collapse. This

paper expands the literature on financial risk management by assessing the effectiveness of

aspects of corporate governance for predicting financial distress in a dynamic discrete-time

survival analysis model. It is the most comprehensive and thorough study to date, using a large

selection of corporate governance measures, financial ratios and macroeconomic variables in a

panel data structure over a ten-year period. Furthermore, the paper addresses the association of

government ownership with the risk of financial distress in the largest emerging market in the

world - China. The results suggest that although corporate governance alone is not sufficient to

accurately predict financial distress, it adds to predictive power of financial ratios and

macroeconomic factors. In addition, the model provides insights into the role of state ownership,

independent directors and some personal characteristics of the Chair/CEO.

Keywords: Corporate Governance; Credit Risk; Financial Risk Management; Financial

Distress; Ownership

JEL classification: G32, G33, G34, G21, C25, C41

1
1. Introduction

Predicting corporate bankruptcy or financial distress has been a vibrant topic in banking,

business and finance because of its importance to creditors such as banks. For corporate debtors,

management quality is a key factor in their performance. Firm’s bankruptcy or financial

performance will affect the investment and debt repayment, and therefore, needs to be predicted.

It is not surprising that this topic has received a lot of attention in academic and practical work,

for example Altman (2005). Risk-taking decisions of creditors will depend on their ability to

analyze or predict the risk involved. There is a vast body of literature on bankruptcy prediction

models that can be classified into accounting based models using financial ratios (e.g. Altman

(1968) and Bonfim (2009)) and market based models using share prices (e.g. Milne (2014) and

Campbell, Hilscher, and Szilagyi (2008)) respectively. Corporate governance measures are less

common in bankruptcy prediction literature as they are not hard information like financial ratios

but soft information, although behaviors such as default on debt, financial distress and

bankruptcy have been found to be linked to corporate governance (see e.g. Daily, Dalton, and

Cannella (2003)). The research that aims to understand the role of corporate governance and

subsequent company performance is summarized in the next section of this paper. However,

we would like to take a different perspective of risk management and in addition to determining

measures of corporate governance that are statistically significant in explaining the financial

distress (which is the main focus of previous studies), we instead assess their predictive value

rather than testing hypotheses. We also take into account the findings from Shumway (2001)

and Campbell, Hilscher, and Szilagyi (2008) who argue that cross-sectional static models miss

important details in structures that can vary across time. Therefore, this paper applies a dynamic

prediction model to assess the relationships between various corporate governance measures

and distress risk. It is the most comprehensive and thorough study to date, using a large selection

of corporate governance variables in a panel data structure over a ten-year period. Furthermore,

2
this paper addresses the association of government ownership with the risk of financial distress

in the largest emerging market in the world. Since the data covers the period of the recent global

financial crisis and we incorporate macroeconomic variables, it is believed that the established

statistical relationships are robust over very different macroeconomic conditions, which is

required by the Basel Accords for risk management (BCBS 2011). In this way, we go beyond

very few studies that used the dynamic approach to explore the role of a limited number of

corporate governance measures in modelling financial distress (Wilson, Wright, and Scholes

2013, Wilson, Wright, and Altanlar 2014).

We find that aspects of board composition, ownership structure, management compensation

and personal characteristics can impact on the risk of financial distress of a company and so

can be used to predict it. However, we establish that using corporate governance measures alone

does not lead to sufficiently accurate predictions. If, however, they are assisted by financial

ratios, models can generate satisfactory predictions in advance. The best predictive model

combines corporate governance measures, financial ratios and macroeconomic factors.

This paper adds to the literature on corporate governance in three ways. First, we link

corporate governance to risk management and examine the role and predictive power of a list

of corporate governance measures from a different perspective of many previous governance

studies. In the credit risk management paradigm, we focus on the predictive power rather than

causality. Our business failure prediction model captures not only the symptoms but also the

causes of business distress/failure rooted in the governance, so predicting more accurately.

Second, we expand the empirical analysis into a new dimension - 35 governance variables in

four groups and 1688 companies over 10 years, which indicates robustness in statistics. Basel

Accord recommends that stress testing covers an economic cycle. Our data cover the recent

financial crisis and the out-of-sample and out-of-time validation has been applied. Third, we

built a dynamic model than which proves to be theoretically better than static models (Shumway,

3
2001). The governance in a company is not stationary but change over time. The dynamic model

can catch the time effect.

In Section 2, the main findings from previous research on aspects of corporate governance

in the prediction of bankruptcy or financial distress are reviewed. In Section 3 the econometric

method including the model specification, the sample, corporate governance measures and

other variables are presented. In Section 4 we present the results including the parameter

estimates and predictive accuracy of four panel models. In Section 5 we discuss the empirical

conclusions from our results and their implications for company owners and managers,

practitioners and especially policy makers.

2. Literature review

Predicting corporate bankruptcy has a long history since Altman (1968) introduced multiple

discriminant analysis to this area and various methods have been proposed to prevent potential

losses for banks and detect financial crisis caused by financial risks. Although financial ratios

have played a major role in the modelling, scholars such as Shumway (2001) and Bonfim (2009)

keep looking for new methods and information to improve model performance. In recent years,

the market price is seen as a forward looking indicator and frequently used to calculate the

distance to default (Milne 2014). The influence of macroeconomic level factors on the

performance of bank loan portfolios are also established (Memmel, Gündüz, and Raupach 2015)

and addressed by the New Basel Accord (BCBS 2011). Credit risks at the individual level can

also be assessed by soft information related to corporate governance (Daily and Dalton 1994b,

Wilson, Wright, and Scholes 2013).

In this section we discuss the literature on credit management from the perspectives of board

composition, ownership structure, management compensation and personal characteristics as

generally corporate governance involves a set of relationships between a company’s

4
management, its board, its shareholders and its stakeholders. It should be noted that corporate

governance theories mainly relate to private companies; in State-Owned Enterprises (SOEs) the

situation is likely to be different. Therefore, SOE issues are discussed separately.

2.1 Board Composition

The board of directors represent the top decision makers of a company and the CEO takes

care of daily operations. In some companies the CEO and the Chair of the Board may be the

same person (described as duality), even though their roles are very different. Daily and Dalton

(1994b) studied 50 pairs of bankrupt and non-bankrupt firms in three and five year horizons

and found that the interaction of CEO/Chair duality and independent directors is positively

related to bankruptcy. The separation of the Chair and the CEO can reduce the risk of

bankruptcy (De Maere, Jorissen, and Uhlaner 2014). On the other hand, Simpson and Gleason

(1999) found duality can reduce the risk of financial distress and in a survey of boards of

directors, Abdullah (2006) provided Malaysian evidence that board independence and CEO

duality are not associated with financial distress status.

One can distinguish between inside directors (executive directors), grey directors (non-

independent non-executive directors) and outside directors (independent directors) on the board

(Hsu and Wu 2014). Some studies (Chaganti, Mahajan, and Sharma 1985, Fich and Slezak 2008,

Salloum, Azoury, and Azzi 2013, Hsu and Wu 2014, Santen and Soppe 2009) have discussed

the influence of various directors on corporate bankruptcy/financial distress. The role of

independent or outside directors on the board has received considerable attention over many

years since they are believed to strengthen the monitoring of firm performance and helpfully

increase diversity. In the research of Li, Wang, and Deng (2008), independent directors turned

out to be negatively associated with the probability of financial distress. Daily (1996) explained

that an outside director in the negotiation process may assist the firm in “convincing creditor

5
groups to agree to a proposed reorganization plan prior to the formal bankruptcy filing”.

However, Chaganti, Mahajan, and Sharma (1985) argued that the influence of outside directors

is not significant in corporate failure. Hsu and Wu (2014) even found that outside directors are

unfavorable to firm survival and increase the likelihood of business failure and, instead, grey

directors can do better at monitoring the board. Santen and Soppe (2009), in a case study

relating to the Netherlands showed that distressed firms have a higher percentage of

independent directors in general. In summary, the previous studies fail to reach a consensus on

whether independent directors have a positive or negative effect on financial distress.

The effect of board size has been explored by Daily and Dalton (1994a) and Jensen (1993),

who suggested that small boards are more efficient and have lower productivity costs during

any coordination process. This argument was later supported by Simpson and Gleason (1999)

and Santen and Soppe (2009) in their empirical results. However, in the retail sector, non-failed

companies tend to have bigger boards (Chaganti, Mahajan, and Sharma 1985). Darrat et al.

(2016) found a mixed effect of board size: having a larger board reduces the risk of bankruptcy

for complex firms with diverse business segments but not for less diversified or single market

oriented firms. In new IPO firms, Chancharat, Krishnamurti, and Tian (2012) using survival

analysis found that either a small or a large board outperform those in a middle in their survival

time. It seems the effect of board size is nonlinear.

2.2 Ownership structure

Ownership structure is a crucial aspect when judging corporate governance because it

addresses the relationship between inside and outside investors. A great deal of research has

addressed issues in ownership structure, for example the type of controller and institutional

investor holding.

6
 Lee and Yeh (2004) in a Taiwanese case suggested that a concentrated ownership

environment such as family ownership will lead to a greater chance of distress. In Taiwan,

family control is very common and this is also true in many other East Asian countries

especially in emerging markets. Claessens, Djankov, and Lang (2000) and Salloum, Azoury,

and Azzi (2013) have also addressed the issue of family control. In contrast to Lee and Yeh

(2004), Wilson, Wright, and Scholes (2013) in their UK study documented that family

businesses are more likely to survive than nonfamily companies. In Mainland China, family

controlled companies do exist, but there is not enough information to determine whether a

company is a family business or not. State control is more relevant here and will be discussed

separately in Section 2.5.

In addition to the type of ownership, the role of institutional shareholders has received

attention, with mixed results. Lee and Yeh (2004) and Ting, Yen, and Chiu (2008) found that

institutional shareholding is lower in distressed companies than in healthy ones. This was

confirmed by Campbell, Hilscher, and Szilagyi (2008), who also found distress risk to be

negatively linked to institutional ownership, though Fich and Slezak (2008), Daily (1996) and

Donker, Santen, and Zahir (2009) found institutional ownership has no relationship with

bankruptcy. However, Liu, Uchida, and Yang (2012) found in China institutional shareholding

has a nonlinear relationship to corporate performance and financial constraints.

From a deep view of how institutional shareholder participate in the board, Manzaneque,

Priego and Merino (2016) found that directors appointed by pressure-resistant institutional

shareholders have a negative impact on the likelihood of business failure. Institutional owners’

impact on directorships can correct mistakes to prevent firms going wrong. This effect is strong

particularly if they have concentrated ownership.

Apart from institutional shareholding, the shareholding of insider and block holders has also

drawn academic attention. Abdullah (2006) found that outsider ownership had a negative

7
association with financial distress while Simpson and Gleason (1999) and Salloum, Azoury,

and Azzi (2013) found no definite link between insider ownership and the probability of

financial distress.

2.3 Management compensation

Salary, bonus and options are three common forms of compensation for managers. Mann

(2005) investigated the relationship between CEO compensation and credit risk. He found that

large, unexplained bonuses and option awards increased credit risk of a company while salary

did not seem to have the same effect. He argued that executives tend to pursue short term profits

rather than longer term financial benefits and have an incentive to adopt high risk business

strategies.

Gilson and Vetsuypens (1993) found that in financially distressed firms a considerable

number of CEOs were replaced or paid less than in normal times. Management compensation

was suggested to be a potentially significant variable to predict financial distress. Li, Wang,

and Deng (2008) also found that the administrative expense ratio was positively related to the

likelihood of financial distress. However, using an equilibrium model, Cyert, Kang, and Kumar

(2002) reported that CEO compensation including base salary, equity and discretionary

compensation was negatively associated with default risk.

2.4 Personal characteristics

Santen and Soppe (2009) incorporated the personal characteristics of directors in their

prediction models in six dimensions: workload, nationality, dependency, interlinked

directorships, age and education. From another perspective, Wilson, Wright, and Altanlar (2014)

described director characteristics from networks, proximity and involvement. Their survival

model on a large dataset of six million observations provided evidence of strong links between

8
director characteristics and new business survival. In their data, having female board directors

reduces the likelihood of insolvency because companies with female directors tend to have

better cash flow and less debt. Khaw et al. (2016) added a comment that men were more likely

to take excessive risks while women were more conservative.

Ruigrok, Peck, and Tacheva (2007) concluded that a foreigner on the board brought the

positive influence of different perspectives, skills and knowledge on the one hand, but the

negative influences of different values, norms and understanding on the other. Santen and

Soppe (2009) found that a foreigner on the board would increase the probability of financial

distress but the accumulated effect of these two aspects of foreign nationality remains to be

fully investigated.

Educational background is likely to affect managerial performance. Higher education

indicates a certain level of aptitude. Holding an MBA degree is evidence of both theoretical and

practical experience in business management. D'Aveni (1990), Daily and Dalton (1994a) and

Ruigrok, Peck, and Tacheva (2007) used education to partly represent the quality of a board.

They agreed that business education might affect the prestige of a company but no study has

yet linked education to the probability of financial distress directly.

Experience is hard to measure since it is personal and unique. Even so some results can be

seen. Wilson, Wright, and Altanlar (2014) concluded that directors with previous insolvency

experience or recent resignations have a higher insolvency risk while Salloum, Azoury, and

Azzi (2013) found insufficient evidence to suggest that shortage of experience in terms of years

served had any such effect.

For obvious reasons age is often used as a proxy for experience. Zahra and Pearce (1989)

used age as one of the relevant characteristics in their study and found that it was linked to

financial performance. Platt and Platt (2012) found an increase in both the CEO’s age and the

average age of the board decreased the chance of bankruptcy, but Fich and Slezak (2008) found

9
that only the CEO’s age is positively significant in one of their four bankruptcy prediction

models.

2.5 State Owned Enterprises

Among the Chinese studies focused on corporate governance (e.g. Tang and Wang (2011)

and Jiang, Feng, and Zhang (2012)), there are some which have addressed the issue of state

ownership. Under a central planning system, for example in Mainland China, SOEs have

dominated the economy in many important sectors such as banking, energy and transportation.

SOEs have some inherent advantages: they do not have to fully cover expenses from sales and

income; unprofitable SOEs and losses are subsidized; they receive funds from state-owned

banks regardless of risks (Lin and Tan 1999). Particularly state ownership in China can mitigate

financial constraints when suffering crisis (Liu, Uchida, and Yang 2012). While they reap

advantages from being part of a planned economy and so rarely go bankrupt, agency theory

implies that the interests of many levels of agents conflict with each other because the state is

both the regulator and the manager. Khaw et al. (2016) found state controlled companies are

less willing to take risk, which may lead to less chance of bankruptcy. Zeitun and Gang Tian

(2007) suggested government ownership could be used as a predictor of probable default.

However, their empirical results also showed that reducing government ownership could cause

the bankruptcy of some companies in the short term.

Maximizing the value of shareholder benefits is the ultimate goal for most companies, and,

therefore, appropriate corporate governance can ensure investors receive a return on their

investment (Shleifer and Vishny 1997). Shleifer and Vishny (1997) also noticed that the agency

problems in large companies in many countries were not only between investors and managers

but also between outside investors and concentrated shareholders who have dominant full

control over the managers. In state-owned companies where the government has large

10
concentrated shares, state ownership leads to problems of corruption, tax cheating and fraud

(Tam 2002). Therefore, state ownership is a double-edged sword: advantages and disadvantages

interact to influence firm performance. Further empirical evidence is required to establish the

relationship between state ownership and financial distress.

From the above we can see that although previous research has examined the relationships

between corporate governance measures and financial distress, no consensus has been reached

as to whether, which and how corporate governance variables affect the chance of financial

distress. Different countries have different regulatory systems of company structure, increasing

the complexity of analysis. In contrast to the Chinse study by Wang and Deng (2006) which is

limited to small samples, a few variables and a cross-sectional analysis, this paper reinvestigates

the relationship between corporate governance measures and the risk of financial distress, with

a large panel dataset of 1,688 companies over 10 years covering the recent financial crisis

ensuring robustness of the modelling results. A wide range of corporate governance measures

from board composition, ownership structure, management compensation and personal

characteristics is represented by 35 potential predictive variables. The case of China provides

an opportunity to address the issue of the role of state ownership, which has great impact on the

access to finance and conflicts of agents.

3. Method and data

In terms of Econometric methodology, one can classify past studies into those that have used

static cross-sectional models and those that have used survival analysis. Studies that have used

cross sectional models and have include Platt and Platt (2012) who compared means of

governance attributes between bankrupt and non-bankrupt companies. Zeitun and Gang Tian

(2007) used linear regression to investigate the relationship between default risk and

governance structure. But the majority, for example Ciampi (2015), Daily and Dalton (1994a),

11
Donker, Santen, and Zahir (2009), Hsu and Wu (2014), Lee and Yeh (2004), not surprisingly,

have applied logistic regression which has worked well in this context.

However as noted by Shumway (2001), dynamic models such as survival analysis are

superior to static models because dynamic models are able to employ multiple period data and

time varying covariates (TVCs) and so enable the prediction of the probability of an event in a

chosen future time period. Similar points are made by Bonfim (2009). Amendola, Restaino, and

Sensini (2015) used survival analysis to predict the probability of financial distress and firm

exits. In studies of corporate governance measures, De Maere, Jorissen, and Uhlaner (2014),

Chancharat, Krishnamurti, and Tian (2012) and Parker, Peters, and Turetsky (2002) employ

Cox Proportional Hazard models. But these studies do not make predictions and they treat time

as continuous when data relating to the covariates is available only yearly and so discrete time

survival modelling would be more appropriate. None of these studies relates to China.

In the literature, it is noted that generally predictive accuracy is improved by the

incorporation of corporate governance measures (Fich and Slezak 2008, Lee and Yeh 2004).

However, research has been inconsistent in the empirical findings or even controversial

regarding whether a variable is positively or negatively, and to what degree, associated with the

probability of financial distress. In a more practical way, this research considers predictive

value of new corporate governance variables, in addition to reporting regression parameters. In

this way the findings are more relevant for credit risk assessment (Reisz and Perlich 2007). To

avoid the problem of overfitting, only the most significant and informative variables including

financial ratios and macroeconomic factors are retained in the final model.

3.1 Model specification

Covariates can be time varying across multiple periods but most of them can only be

observed at specific time when economic and financial reports are disclosed. In this sense, Cox

12
Proportional hazard model in Parker, Peters, and Turetsky (2002) may be not suitable. We

follow Shumway (2001) and assume a discrete time setting in modelling. Shumway (2001)

proved that parameter estimate is the same as multi-period logistic regression, which is the

maximum likelihood method. Unlike Bonfim (2009) who assumed covariates act in the same

period of the dependent variable, a horizon of two years in advance is applied in this research

as it is more suitable for the context of prediction. The regression model uses covariates from

year t - 2 to predict whether a company is distressed in year t , marked as d = 1 . Therefore, the

form of the survival model is specified as follows:

=d 1 α + β 0T h0 (t ) + β1T xig,t − 2 + βT2 xir,t − 2 + βT3 xim,t − 2 ,

logit(h (t )) = (1)

where t is the survival time;

hd =1 (t ) is the probability of distress at time t ;

h0 (t ) is the baseline hazard in duration at time t and β 0 is its coefficient;

xig,t − 2 is a column vector of corporate governance variables for company i at time t − 2 ;

xir,t − 2 is a column vector of financial ratios of predictive power;

xim,t − 2 is a column vector of macroeconomic factors;

β1 , β 2 , β3 are vectors of coefficients;

α is the constant.

It should be noted that, unlike an econometric approach, credit risk prediction models do not

necessarily have to control other influences so no control variables are included in the

regression equation of survival analysis. All the independent variable in Equation (1) are

regarded as potentially predictive variables.

In the analytical process, first, we include four groups of corporate governance measures into

the regression without any other covariates. In this way, significant corporate governance

measures are identified and then enter the first prediction model (Model 1). The second model

13
uses financial ratios only (Model 2) and the third model combines both significant corporate

governance measures and financial ratios (Model 3). Model 4 further incorporates

macroeconomic factors. The predictive accuracy is assessed by Area Under the Receiver

Operation Curve (AUC) (equivalent to the Gini coefficient to assess the inequality of sample

distribution) which is commonly used in predictive modelling and credit risk management and

ranged from 0 to 1 with higher value indicating better results (Reisz and Perlich 2007). Four

groups of results of both in-sample and out-of-sample predictions are given for comparison

(Table 1).

Table 1
Model specification

Model Specification
Model 1 Survival model with corporate governance measures only
Model 2 Survival model with financial ratios only
Model 3 Survival model with governance measures and financial ratios
Model 4 Survival model with governance measures, financial ratios and macroeconomic
variables.

3.2 Sample

‘Special Treatment’ is imposed by the regulator China Securities Regulatory Commission

(CSRC) to give investors notice of potential risks. Therefore, it is an official indicator of

financial distress of listed companies. A listed company can be filed in Special Treatment for

any of these reasons: 1) negative net profit in the most recent two consecutive years; 2) failure

to disclose its annual report; 3) likelihood of being dissolved; 4) reorganisation, settlement or

bankruptcy liquidation. In over 80% of our cases, the companies in Special Treatment suffer

net losses in two consecutive years. So it is a popular indicator of financial distress as in Zhang,

Altman, and Yen (2010), Geng, Bose, and Chen (2015), Lin, Lo, and Wu (2016). Databases

Wind and GTA provide access to annual statements including accounting and governance

information. The original dataset contains 2,477 companies listed in China since 1991. Due to

14
the late disclosure of governance information only after 2002, the data is restricted to 2003

onwards. The final sample consists of 1,688 companies over 10 years between 2003 and 2012.

Predictions for financial distress are commonly verified in out-of-sample and out-of-time

datasets (Lin, Lo, and Wu 2016, Shumway 2001). Therefore, the whole sample is randomly

divided into two sub-samples in a 2:1 ratio (Sample One and Sample Two). Applying a stratified

sampling strategy to both samples and distress/non-distress groups, it is broken down into the

four panel datasets, as shown in Table 2. We use covariates in 2003-2008 to predict financial

distress in 2005-2010 (Sample One Panel A and Sample Two Panel B) and covariates in 2009-

2010 to predict financial distress in 2011-2012 (Sample One Panel C and Sample Two Panel

D). In the time dimension, the first six years are the in-time period and the following two years

are the out-of-time period. The average proportion of distressed cases across all years is very

similar: 2.58% in Sample One and 2.57% in Sample Two.

Table 2
Subsamples

Sample One Sample Two

Year Distress Total Distress Year Distress Total Distress
0 1 rate 0 1 rate
2003 744 28 772 3.63% 2003 379 11 390 2.82%
2004 793 21 814 2.58% 2004 396 14 410 3.41%
2005 785 18 803 2.24% 2005 392 8 400 2.00%
2006 787 44 831 5.29% 2006 390 14 404 3.47%
Panel A 2007 827 33 860 3.84% Panel B 2007 407 22 429 5.13%
and C 2008 866 14 880 1.59% and D 2008 425 8 433 1.85%
2009 907 17 924 1.84% 2009 452 10 462 2.16%
2010 911 26 937 2.77% 2010 459 13 472 2.75%
2011 904 7 911 0.77% 2011 455 4 459 0.87%
2012 889 15 904 1.66% 2012 448 7 455 1.54%
Total 8413 223 8636 2.58% Total 4203 111 4314 2.57%

3.3 Corporate governance measures

Argenti (1976) summarized six structural defects indicated by the experts: one-man rule,

non-participating board, unbalanced top team, lack of management depth, weak finance

15
function and combined chairman-chief executive. For instance, one-man rule is to describe a

CEO who dominates their colleagues rather than lead them in making decisions or hear their

advice. At some occasions, some of the functional directors who sit on main boards do not take

their responsibilities. The ‘top team' includes directors, senior executives and advisers may be

not balanced in their backgrounds or abilities. These situations are rooted in the management

of a company and we source the proxies of governance from them to describe the board, the

ownership and the senior management team.

At last, as discussed in the literature review and according to availability of data in the

database, corporate governance variables are classified into four groups and explained in Table

3. Lee and Yeh (2004) discussed the issue of ultimate control, which is very common in the

emerging markets where highly concentrated shares are held by a family or the state. Claessens,

Djankov, and Lang (2000) suggested that the controlling shareholder needs to be considered in

bankruptcy prediction models. In our study, the ultimate controller is determined according to

the CSRC regulations. Therefore, the ultimate controller is the indicator to denote whether a

company is a SOE. We also consider the connection between large shareholders. According to

Platt and Platt (2012), interlinked directorship provides benefits for the company.

Table 3

Corporate governance measures

Variable Definition
Board composition (6)
Board size Number of total directors
Independent director Proportion of independent board directors
Number of supervisors Number of supervisors
Number of senior managers Number of senior managers
Duality of Chair and CEO 1 if the Chair and the CEO is the same person
Independent director monitoring 1 if most independent directors work at the company address
Ownership Structure (10)
State ownership Proportion of state owned shares to total shares
SOE 1 if the ultimate controller is the state
Board shares Shares held by the board to total shares

16
 Supervisor shares Shares held by the supervision board to total shares
Top 10 shareholders Total shares of largest 10 shareholders to total shares
Institutional share holding Total shares of institutional shares to total shares
Average share holding Average shareholding to total shares
Listing somewhere else 1 if the company is listed on other exchanges outside China
Share capital change 1 if it changes from previous year
Connected top 10 shareholders 1 if any two top 10 shareholders are related
Management Compensation (5)
Salary of senior management Salary of directors, supervisors and senior managers to total salary cost
Salary of top 3 directors Salary of top 3 directors to total salary cost
Salary of top 3 independent
Salary of top 3 independent directors to total salary cost
directors
Salary of top 3 senior managers Salary of top 3 senior managers to total salary cost
Number of non-paid seniors Number of non-paid directors, supervisors and senior managers
Personal characteristics (14)
Chair age Age in the year
Chair gender 1 if female, 0 otherwise
Chair education 4 dummies (college, undergraduate, master, doctorate)
Chair professional qualification 1 if holding any professional qualification
Chair nationality 1 if not Chinese
Chair paid 1 if paid
Chair concurrent post 1 if holding a position in other companies
CEO age Age in the year
CEO gender 1 if female, 0 otherwise
CEO education 4 dummies (college, undergraduate, master, doctorate)
CEO professional qualification 1 if holding any professional qualification
CEO nationality 1 if not Chinese
CEO paid 1 if paid
CEO concurrent post 1 if holding a position in other companies

Regarding management compensation, constrained by availability of data, the data source

only provides a small fraction of option incentive information and the quality of very low. No

option incentive is considered here. We have only access to the salaries of the management

team. In China, most of bonuses are included in salaries in financial statements.

Whilst some papers (Fich and Slezak 2008, Platt and Platt 2012) are interested in the CEO

and some (Santen and Soppe 2009) are interested in the board directors, this research takes both

into account. Generally, the CEO is authorized by the board and is responsible for the overall

management, decision making, execution and the daily operation of the company. Therefore,

the personality and characteristics of the CEO will be reflected in the development of the

business. In the situation that the Chair of the board has control of the company and is more

17
involved in the management and decision making, the Chair will have more influence on

performance, which should not be ignored.

Personal information concerning both the Chair and the CEO for each company is recorded

in the database including four types of personal demographic information: age, gender,

nationality and education, and another three types of information regarding their professions:

whether they have professional qualifications, whether they get paid by the company and

whether they possess a position in any other organization.

3.4 Financial ratios and macroeconomic variables

As the BASEL Accord II addressed, macroeconomy as a systemic factor has an impact on

the business cycle so banks are necessary to consider it in their Probability of Default models

in Internal Rating-Based Approach (IRB). Though our focus is on governance variables, we

still incorporate financial ratios and macroeconomic factors in analysis because they are not

neglectable to credit risk management. For the selection of potential financial ratios and

macroeconomic factors, out of a range of potential ones, we consider recommendation in

literature, significance in preliminary analysis and correlation in collinearity diagnostics in the

selection process.

The first group of TVCs are financial ratios covering different aspects of a company. In the

literature, popular aspects to be assessed in financial ratio analysis are profitability, liability,

gearing, operations etc. Therefore Return on Assets, Current liabilities / Total Liabilities,

Tangible Assets / Total Assets, Cash Flow from Operating / Total Liabilities, Receivables

Turnover and Total Assets Growth are selected with reference to literature and predictive power

in preliminary analysis. In accordance with Shumway (2001) which also involves listed

companies in the sample, duration time in survival analysis is determined as the time since

18
listing on the exchange and the natural logarithm of the duration is chosen to be the baseline

function.

A series of macroeconomic factors are the other group of TVCs. However, unlike firm-

specific covariates which affect individual cases, macroeconomic factors are systematic

components that vary over time. For all companies existing in a period, macroeconomic

conditions have the same impact on each and have been major driver of credit risk for banks

(Memmel, Gündüz, and Raupach 2015). We incorporate lagged annualized values of GDP

growth, the unemployment rate, the inflation rate and interest rates, which are extracted from

World Databank, the database of the World Bank.

3.5 Data description

Table 4 shows descriptive statistics of the data. On average, there are 9.46 directors on the

board, of which 3.31 (or 35%) are independent directors. There are on average 4.08 supervisors

and 6.29 senior managers in each listed company. The government holds about one quarter of

the total shares, which indicates the substantial influence of the government on Chinese listed

companies is notable. Supervisors still own relatively small proportions of the shares (0.14%)

because some of them are shareholder and employee representatives. On average, the top 10

shareholders own over half of total shares (58%) and so are often block holders who make

important decisions. Institutional shareholders hold a large percentage of all shares, in some

cases up to 93% of total shares.

Table 4
Description of corporate governance measures 1

Variable N Mean Std. Dev. Min Max

Board composition
Board size 10221 9.46 2 3 19
Independent director 10221 0.35 0.05 0 0.8
Number of supervisors 10221 4.08 1.37 0 13
Number of senior managers 10221 6.29 2.37 1 45
Ownership structure

19
 State ownership 10221 0.25 0.25 0 0.86
Board shares 10221 0.13 0.11 0.03 0.75
Supervisor shares 10221 0.0014 0.01 0 0.27
Top 10 shareholders 10221 0.58 0.15 0.07 0.99
Institutional share holding 10221 0.14 0.18 0 0.93
Average share holding 10221 0.0004 0.0004 0 0.005
Management compensation
Salary of senior management 10221 0.5 0.11 0.03 0.85
Salary of top 3 directors 10221 0.17 0.07 0 0.51
Salary of top 3 independent directors 10221 0.05 0.04 0 0.33
Salary of top 3 senior managers 10221 0.2 0.07 0 0.65
Number of non-paid seniors 10221 4.29 3.17 0 19
Personal characteristics
Chair Age 10221 50.08 7.28 28 84
CEO Age 10221 46.27 6.47 24 75

For categorical governance variables, we present only their frequencies and percentages in

Table 5. It should be noted that the incidence in Table 5 is counted by company year but not

company case, but it is still surprising to find that in over two thirds of observations (company-

year), companies are state controlled. This may be because the Chinese exchanges were

originally in service for SOEs.

Table 5
Description of corporate governance measures 2

Distress 0 (% of 1 (% of
Variable N
0 1 total) total)
Board composition
Duality of Chair and CEO 10221 8835 1386 86.4 13.6
Independent director monitoring 10221 6390 3831 62.5 37.5
Ownership structure
SOE 10221 3226 6995 31.6 68.4
Listing somewhere else 10221 9368 853 91.7 8.3
Share capital change 10221 3731 6490 36.5 63.5
Connected top 10 shareholders 10221 6058 4163 59.3 40.7
Personal characteristics
Chair gender 10221 9833 388 96.2 3.8
Chair college 10221 8937 1284 87.4 12.6
Chair undergraduate 10221 7470 2751 73.1 26.9
Chair masters 10221 6936 3285 67.9 32.1
Chair doctorate 10221 9673 548 94.6 5.4
Chair qualification 10221 4374 5847 42.8 57.2

20
 Chair nationality 10221 10130 91 99.1 0.9
Chair paid 10221 3487 6734 34.1 65.9
Chair concurrent position 10221 3951 6270 38.7 61.3
CEO gender 10221 9757 464 95.5 4.5
CEO college 10221 9036 1185 88.4 11.6
CEO undergraduate 10221 7367 2854 72.1 27.9
CEO masters 10221 6922 3299 67.7 32.3
CEO doctorate 10221 9830 391 96.2 3.8
CEO professional qualification 10221 4750 5471 46.5 53.5
CEO nationality 10221 10120 101 99 1
CEO paid 10221 321 9900 3.1 96.9
CEO concurrent position 10221 6589 3632 64.5 35.5

Financial ratios and macroeconomic factors (Table 6) are transformed into percentages for

ease of interpretation and comparison. Generally, if only looking at the means, Chinese listed

companies have been achieving positive returns and growing in the past few years. The Chinese

economy has been growing comparatively quickly for decades while keeping inflation and

unemployment rates at relatively low levels.

Collinearity between explanatory variables could lead to potential problems in testing the

significance of covariates. In this study, there is no pair of variables with high correlation over

0.7, and between corporate governance and financial ratios all Variance Inflation Factors (VIFs)

are smaller than 3, with an average of 1.57.

Table 6
Description of financial ratios and macroeconomic factors

Variable N Mean SD Min Max

Financial ratios
Return on Assets 10221 4.98 7.01 -18.94 28.64
Tangible Assets / Total Assets 10221 39.65 20.00 -35.64 96.53
Current Liabilities / Total Liabilities 10221 79.81 13.52 28.29 100.00
Net cash flow from operation / Total Liabilities 10221 9.64 24.18 -106.98 101.45
Receivables Turnover 10221 57.14 52.18 -8.42 244.66
Total Assets Growth 10221 15.45 31.00 -93.33 126.21
Macroeconomic factors
GDP 10221 10.92 1.61 9.20 14.20
Inflation 10221 2.70 2.04 -0.70 5.86
Unemployment 10221 4.17 0.10 4.00 4.30
Interest 10221 5.81 0.68 5.31 7.47

21
4. Results

4.1 Model results

Measures of different aspects of corporate governance were entered into the models step by

step and assessed by their significance. Only variables significant at the 5% level were retained.

Model 1 consists of seven corporate governance measures, one from the board composition

category, two from ownership structure, one from management compensation, and three from

personal characteristics (Table 7). Model 2 includes six financial ratios and all appear to be

significant in predicting financial difficulty and have their expected signs. In Model 3, all

significant corporate governance measures and financial ratios are combined and all remain

significant with the same signs as in Model 1 and Model 2. In Model 4, macroeconomic factors

are added, and we find that the consumer price inflation rate is significant in the model.

We find that the monitoring of independent directors affects corporate performance. If they

present on site and serve their duties well, the risk of poor managerial decisions can be reduced.

This finding is similar to that in Wilson, Wright, and Scholes (2013), who found that if directors

live close to companies, they are better able to monitor management. Long distance indicates

loose control and monitoring. We also find that if the company is state controlled, it has a lower

chance of becoming distressed. This may be taken as evidence that the government has provided

abundant resources to support the company.

Table 7
Model results

Model 1 Model 2 Model 3 Model 4

ln (duration) 1.670** 0.0466* 1.048** 0.482
(5.49) (1.62) (3.02) (1.20)
Independent director monitoring -0.726** -0.718** -0.649**
(-3.32) (-3.08) (-2.76)
SOE -0.969** -0.877** -0.877**
(-5.35) (-4.50) (-4.48)
Institutional share holding -12.916** -6.782** -7.861**
(-6.00) (-3.38) (-3.74)

22
 Salary of top 3 independent directors 3.866** 3.649** 3.915**
(6.42) (5.53) (5.79)
Chair age -0.064** -0.061** -0.061**
(-5.32) (-4.61) (-4.58)
CEO Masters -0.827** -0.836** -0.863**
(-3.02) (-2.83) (-2.91)
CEO concurrent position -1.119** -0.881** -0.802**
(-4.45) (-3.26) (-2.95)
Return on assets -0.069** -0.065** -0.066**
(-4.28) (-3.70) (-3.76)
Tangible assets / total assets -0.028** -0.028** -0.029**
(-5.85) (-5.70) (-5.79)
Current liabilities / total liabilities 0.078** 0.069** 0.069**
(6.72) (5.96) (5.39)
Net cash flow from operation / -0.025* -0.020** -0.019**
Total Liabilities (-5.22) (-3.86) (-3.81)
Receivables turnover 0.007** 0.006** 0.006**
(4.67) (3.69) (3.73)
Total assets growth -0.026** -0.022** -0.022**
(-4.68) (-4.04) (-4.09)
Inflation 0.197**
(2.81)
Constant -3.277** -8.938** -7.510** -7.341**
(-4.33) (-7.64) (-5.61) (-5.46)
Log likelihood -542.607 -505.959 -438.580 -434.551
Number of observations 4635 4635 4635 4635
LR Chi-sq 245.89 319.19 453.95 462.01
Prob > Chi-sq 0 0 0 0
Pseudo R2 0.185 0.240 0.341 0.347
* p-value<0.05 ** p-value<0.01

The results also suggest that when the institutional investor has a stake in a listed company,

the chances of distress are lower. The institutional investors have expertise and skills in

detecting companies worthy of investment. According to Ting, Yen, and Chiu (2008), the

existence of institutional investors particularly foreign institutional investors apply pressure to

auditors so the auditing reports will release signals of creditworthiness. Further, if the salary

cost of independent directors is large, the company has a high risk of financial distress. There

may be two reasons for this. On one hand, the salary cost for an independent director places a

burden on a company’s financial condition. On the other hand, and more importantly, when

independent directors are highly paid, they tend not to speak up when finding problems or

disagree with management decisions. Of the fourteen characteristics for both the Chair and the

CEO, only three are significant: the Chair’s age, the CEO’s education if they have a master

23
degree, and whether the CEO has a position in other organizations. As the Chair grows older,

their experience increases and they become more cautious than young entrepreneurs. When the

CEO has a Master’s degree, our data shows it is beneficial for chief executives to improve their

corporate performance. When the CEO has another position in other organizations, they

presumably possess more social relationships and resources and so can bring extra benefits for

the company.

4.2 Predictive accuracy

As discussed previously, predictive accuracy is the true focus of credit management and its

measurement is presented in Table 8. Four panels are compared and Panel A gives results for

the model training sample. Unsurprisingly, in-sample prediction produces the best results as

compared to both out-of-sample and out-of-time predictions, followed by the within time out-

of-sample predictions. Understandably, results in Panel D show the least accuracy because

neither the sample nor the time is the same as in the model training.

AUC measures the ability of discriminant between the distress and non-distressed groups,

which is equivalent to the Gini coefficient. The performance of Model 1 with only governance

measures is acceptable in the in-time validation – Panel B (AUC=0.811). However, predictive

accuracy declines dramatically when it is applied to the other period (Panel C). In Panels C and

D, The AUC is only 0.694 and 0.768 respectively, which means using only corporate

governance measures to predict financial distress is not practical. When combined with

financial ratios in Model 3, the predictive accuracy is much improved.

The best performance comes from Model 4 where corporate governance measures, financial

ratios and macroeconomic factors are all used in the model. In the within-time predictions

(Panel A and B), the differences between Models 3 and 4 are trivial (AUC 0.915 to 0.916 and

0.902 to 0.903). But in the out-of-time predictions, macroeconomic factors give a significant

24
improvement to the predictive accuracy (AUC 0.868 to 0.876 and 0.852 to 0.860). Particularly,

when considering the economic conditions, out-of-time predictions are significantly improved.

In a further analysis, the out-of-sample predictive accuracy (Panel B and D) in separate years

is examined, taking AUC as an example (Figure 1). Model 3 & 4 are consistently better than

Model 1 & 2 across all years. The power of extra information is evident in our empirical results.

Table 8
Predictive accuracy of models

Panel A Panel B Panel C Panel D

(training) (in-time, out-of- (in-sample, out- (out-of-time,
sample) of-time) out-of-sample)
Model 1 0.829 0.811 0.694 0.768
Model 2 0.871 0.833 0.826 0.785
Model 3 0.915 0.902 0.868 0.852
Model 4 0.916 0.903 0.876 0.860

Figure 1
Model performance (AUC) across years

5. Discussion and conclusion

Over the past 20 years corporate governance has attracted wide academic attention in many

disciplines to find that certain aspects of the corporate governance of a company are linked to

25
its corporate performance or its financial position. From the perspective of bankruptcy/distress

prediction, this paper has tested a wide range of corporate governance measures as predictors

of corporate credit risk, using four panels of 10 years for 1,688 companies using survival models.

In search of the causes of corporate failure, Argenti (1976) did an in-depth survey and found

a universal comment that bad management was the prime cause of failure. We regard what was

described by ‘bad management’ as ‘poor governance’ due to the term ‘governance’ was not

popular at the time of the book. Behaviors of bad governance include one-man rule, non-

participating board, unbalanced top team, lack of management depth, weak finance function

and combined chairman-CEO. These behaviors are captured by our corporate governance

measures and results show consistent evidence. Argenti (1976) explained the channel from

corporate governance to financial distress that poor governance leads to inability of the

management team to correct the mistakes (due to one-man rule, chair-CEO duality, problematic

board/management team etc.), and so finally cause the company fail. In the process of distress

to bankruptcy, some symptoms are observed. For example, financial ratios behave worse

compared to others. However financial ratios as symptoms are delayed in disclosure. In order

to give early warning in advance, we have to go to the root – governance of a company. In our

empirical result, though we do not focus on the hypothesis, the channel is well established as

many empirical studies shown, for example, Daily and Dalton (1994) and Fich and Slezak

(2008), etc.

In the dynamic prediction model, thirty-five corporate governance measures are considered

that cover four aspects of a company management: board composition, ownership structure,

management compensation, and director and manager characteristics. First, our results show

that regarding the board composition, the monitoring of independent directors is significantly

associated with the risk of financial distress. Independent directors are expected to carry out

their duties so can effectively provide suggestions and improve performance based on

26
knowledge of other companies. Second, state ownership and institutional ownership reduce the

risk of a company becoming financially distressed. Active investors such as institutional

shareholders have the ability to detect potential risks to a company in which they have large

investments. This is consistent with the literature in Campbell, Hilscher, and Szilagyi (2008),

Lee and Yeh (2004) and Ting, Yen, and Chiu (2008). Third, in terms of management

compensation, risks are greater with independent directors being more paid, indicating that

independent directors are reluctant to comment negatively on their employers. Fourth,

regarding personal characteristics, when the Chair is older, and when the CEO has a Masters

degree or holds other positions in other organizations, the risk of distress is lower. Furthermore,

conditional on six financial ratios, the macroeconomic factor affects the risk of distress.

In terms of predictive accuracy, corporate governance measures alone have limited capacity

to detect financial distress. Financial ratios alone can do relatively better. However, when

combining the two together, the predictive accuracy is significantly improved. The best

predictive model comes from the combination of corporate governance measures, financial

ratios and the macroeconomic factors. This outperforms the other three models in both out-of-

sample and out-of-time predictions. In the performance separated by individual years, the

ranking of models in most years remains the same.

In identifying measures of corporate governance which are significantly linked to financial

distress these empirical results directly address issues of effective monitoring, business

prosperity and prevention of corporate collapse and thus have important implications for

financial stability in practice. Such information is helpful, first of all, for creditors in preventing

potential losses, and also for owners and managers in identifying problems and implementing

changes accordingly. It is also relevant to the corporate governance responsibilities of

shareholders and stakeholders and those of regulators who supervise listed and other forms of

companies, specifically regarding aspects of state ownership and independent directors. Finally,

27
corporate governance is closely linked to government policies and legal requirements that

ensure financial prudence and stable economic performance so our results should also be of

interest to policy makers and governments in the formulating of enterprise development

strategy and its enforcement.

At last, China as the largest emerging market in the world is focused in this study. And China

is highly unequal in economic development. Its stock market is relatively young with a history

of less than 30 years. For this reason and it shares many common patterns in the capital market

with many other emerging markets. It is believed that our findings will be insightful for the

developing countries, particularly for those in Asia where some of the cultural and political

issues are shared. For example, the SOEs in Vietnam also contribute to a significant portion to

its economy. Our model can be also developed fit other developing countries.

References

Abdullah, Shamsul Nahar. "Board structure and ownership in Malaysia: the case of distressed
listed companies." Corporate Governance: The International Journal of Business in Society
6 (5).2006:582-594.
Altman, Edward I. "Financial ratios, discriminant analysis and the prediction of corporate
bankruptcy." Journal of Finance 23 (4).1968:589-609.
Altman, Edward I. "An emerging market credit scoring system for corporate bonds." Emerging
Markets Review 6 (4).2005:311-323.
Amendola, Alessandra, Marialuisa Restaino, and Luca Sensini. "An analysis of the
determinants of financial distress in Italy: A competing risks approach." International Review
of Economics & Finance 37 2015:33-41.
Argenti, J. 1976. Corporate Collapse: The Causes and Symptoms. London: McGraw-Hill.
Basu, Sudipta, Lee-Seok Hwang, Toshiaki Mitsudome, and Joseph Weintrop. "Corporate
governance, top executive compensation and firm performance in Japan." Pacific-Basin
Finance Journal 15 (1).2007:56-79.
Basel Committee on Banking Supervision (BCBS).2011. Basel III: A Global Regulatory
Framework for More Resilient Banks and Banking Systems.
Bonfim, Diana. "Credit risk drivers: Evaluating the contribution of firm level information and
of macroeconomic dynamics." Journal of Banking & Finance 33 (2).2009:281-299.
Campbell, John Y., Jens Hilscher, and J. A. N. Szilagyi. "In Search of Distress Risk." Journal
of Finance 63 (6).2008:2899-2939.

28
Chaganti, Rajeswararao S., Vijay Mahajan, and Subhash Sharma. "Corporate board size,
composition and corporate failures in retailing industry." Journal of Management Studies 22
(4).1985:400-417.
Chancharat, Nongnit, Chandrasekhar Krishnamurti, and Gary Tian. "Board Structure and
Survival of New Economy IPO Firms." Corporate Governance: An International Review 20
(2).2012:144-163.
Ciampi, Francesco. "Corporate governance characteristics and default prediction modeling for
small enterprises. An empirical analysis of Italian firms." Journal of Business Research 68
(5).2015:1012-1025.
Claessens, Stijn, Simeon Djankov, and Larry H. P. Lang. "The separation of ownership and
control in East Asian Corporations." Journal of Financial Economics 58 (1–2).2000:81-112.
Cyert, Richard M., Sok-Hyon Kang, and Praveen Kumar. "Corporate Governance, Takeovers,
and Top-Management Compensation: Theory and Evidence." Management Science 48
(4).2002:453-469.
D'Aveni, Richard A. "Top managerial prestige and organizational bankruptcy." Organization
Science 1 (2).1990:121-142.
Daily, C. M. "Governance patterns in bankruptcy reorganizations." Strategic Management
Journal 17 (5).1996:355-375.
Daily, C. M., and D. R. Dalton. "Bankruptcy and corporate governance: the impact of board
composition and structure." Academy of Management Journal 37 (6).1994a:1603-1617.
Daily, C. M., and D. R. Dalton. "Corporate governance and the bankrupt firm: an empirical
assessment." Strategic Management Journal 15 (8).1994b:643-654.
Daily, C. M., D. R. Dalton, and A. A. Cannella. "Introduction to special topic forum corporate
governance: Decades of dialogue and data." Academy of Management Review 28
(3).2003:371-382.
Darrat, Ali F., Stephen Gray, Jung Chul Park, and Yanhui Wu. "Corporate Governance and
Bankruptcy Risk." Journal of Accounting, Auditing & Finance 31 (2).2016:163-202.
De Maere, Jonas, Ann Jorissen, and Lorraine M. Uhlaner. "Board Capital and the Downward
Spiral: Antecedents of Bankruptcy in a Sample of Unlisted Firms." Corporate Governance:
An International Review 22 (5).2014:387-407.
Donker, Han, Bernard Santen, and Saif Zahir. "Ownership structure and the likelihood of
financial distress in the Netherlands." Applied Financial Economics 19 (21).2009:1687-1696.
Fich, EliezerM, and SteveL Slezak. "Can corporate governance save distressed firms from
bankruptcy? An empirical analysis." Review of Quantitative Finance and Accounting 30
(2).2008:225-251.
Geng, Ruibin, Indranil Bose, and Xi Chen. "Prediction of financial distress: An empirical study
of listed Chinese companies using data mining." European Journal of Operational Research
241 (1).2015:236-247.
Gilson, S. C., and M. R. Vetsuypens. "CEO compensation in financially distressed firms - an
empirical-analysis." Journal of Finance 48 (2).1993:425-458.
Hsu, Hwa-Hsien, and Chloe Yu-Hsuan Wu. "Board composition, grey directors and corporate
failure in the UK." The British Accounting Review 46 (3).2014:215-227.
Jensen, M. C. "The modern industrial-revolution, exit, and the failure of internal control-
systems." Journal of Finance 48 (3).1993:831-880.
Jiang, Chunxia, Genfu Feng, and Jianhua Zhang. "Corporate governance and bank performance
in China." Journal of Chinese Economic and Business Studies 10 (2).2012:131-146.
Khaw, Karren Lee-Hwei, Jing Liao, David Tripe, and Udomsak Wongchoti. "Gender diversity,
state control, and corporate risk-taking: Evidence from China." Pacific-Basin Finance
Journal 39 2016:141-158.

29
Lee, T. S., and Y. H. Yeh. "Corporate governance and financial distress: Evidence from
Taiwan." Corporate Governance: An International Review 12 (3).2004:378-388.
Li, Hong‐Xia, Zong‐Jun Wang, and Xiao‐Lan Deng. "Ownership, independent directors,
agency costs and financial distress: evidence from Chinese listed companies." Corporate
Governance: The international journal of business in society 8 (5).2008:622-636.
Lin, Hsiou-Wei William, Huai-Chun Lo, and Ruei-Shian Wu. "Modeling default prediction
with earnings management." Pacific-Basin Finance Journal 40, Part B 2016:306-322.
Lin, J. Y. F., and G. F. Tan. "Policy burdens, accountability, and the soft budget constraint."
American Economic Review 89 (2).1999:426-431.
Liu, Chunyan, Konari Uchida, and Yufeng Yang. "Corporate governance and firm value during
the global financial crisis: Evidence from China." International Review of Financial Analysis
21 2012:70-80.
Mann, Christopher. 2005. CEO compensation and credit risk Special Comment,: New York:
Moody's Investors Service.
Manzaneque, M., Merino, E., & Priego, A. M. "The role of institutional shareholders as owners
and directors and the financial distress likelihood. Evidence from a concentrated ownership
context". European Management Journal, 34(4), 2016:439-451.
Memmel, Christoph, Yalin Gündüz, and Peter Raupach. "The common drivers of default risk."
Journal of Financial Stability 16 2015:232-247.
Milne, Alistair. "Distance to default and the financial crisis." Journal of Financial Stability 12
2014:26-36.
Parker, Susan, Gary F. Peters, and Howard F. Turetsky. "Corporate governance and corporate
failure: a survival analysis." Corporate Governance: The international journal of business
in society 2 (2).2002:4-12.
Platt, Harlan, and Marjorie Platt. "Corporate board attributes and bankruptcy." Journal of
Business Research 65 (8).2012:1139-1143.
Reisz, Alexander S., and Claudia Perlich. "A market-based framework for bankruptcy
prediction." Journal of Financial Stability 3 (2).2007:85-131.
Ruigrok, Winfried, Simon Peck, and Sabina Tacheva. "Nationality and gender diversity on
Swiss corporate boards." Corporate Governance: An International Review 15 (4).2007:546-
557.
Salloum, CharbelC, NehmeM Azoury, and TarekM Azzi. "Board of directors’ effects on
financial distress evidence of family owned businesses in Lebanon." International
Entrepreneurship and Management Journal 9 (1).2013:59-75.
Santen, Bernard, and Aloy Soppe. "NED characteristics, board structure and management
turnover in the netherlands in times of financial distress: a theoretical and empirical survey."
Corporate Ownership and Control 7 (1).2009:17.
Shleifer, A., and R. W. Vishny. "A survey of corporate governance." Journal of Finance 52
(2).1997:737-783.
Shumway, T. "Forecasting bankruptcy more accurately: A simple hazard model." Journal of
Business 74 (1).2001:101-124.
Simpson, W. Gary, and Anne E. Gleason. "Board structure, ownership, and financial distress
in banking firms." International Review of Economics & Finance 8 (3).1999:281-292.
Tam, On Kit. "Ethical Issues in the Evolution ofCorporate Governance in China." Journal of
Business Ethics 37 (3).2002:303-320.
Tang, Ke, and Changyun Wang. "Corporate Governance and Firm Liquidity: Evidence from
the Chinese Stock Market." Emerging Markets Finance and Trade 47 (sup1).2011:47-60.
Ting, Wei, Sin-Hui Yen, and Chien-Liang Chiu. "The Influence of Qualified Foreign
Institutional Investors on the Association between Default Risk and Audit Opinions:

30
 Evidence from the Chinese Stock Market." Corporate Governance: An International Review
16 (5).2008:400-415.
Wang, Zong-Jun, and Xiao-Lan Deng. "Corporate Governance and Financial Distress:
Evidence from Chinese Listed Companies." Chinese Economy 39 (5).2006:5-27.
Wilson, Nick, Mike Wright, and Ali Altanlar. "The survival of newly-incorporated companies
and founding director characteristics." International Small Business Journal 32
(7).2014:733-758.
Wilson, Nick, Mike Wright, and Louise Scholes. "Family Business Survival and the Role of
Boards." Entrepreneurship Theory and Practice 37 (6).2013:1369-1389.
Zahra, S. A., and J. A. Pearce. "Boards of directors and corporate financial performance - a
review and integrative model." Journal of Management 15 (2).1989:291-334.
Zeitun, Rami, and Gary Gang Tian. "Does ownership affect a firm's performance and default
risk in Jordan?" Corporate Governance: The International Journal of Business in Society 7
(1).2007:66-82.
Zhang, Ling, Edward I. Altman, and Jerome Yen. "Corporate financial distress diagnosis model
and application in credit rating for listing firms in China." Frontiers of Computer Science in
China 4 (2).2010:220-236.

31