- Email: [email protected]
The impact of political connections and business groups on cash holdings: Evidence from Chinese listed firms
Accepted Manuscript The impact of political connections and business groups on cash holdings: Evidence from Chinese listed firms
Tsui-Jung Lin, Hai-Yen Chang, Hui-Fun Yu, Ching-Pao Kao PII: DOI: Reference:
S1044-0283(18)30084-X doi:10.1016/j.gfj.2018.10.001 GLOFIN 451
To appear in:
Global Finance Journal
Received date: Revised date: Accepted date:
28 March 2018 1 October 2018 1 October 2018
Please cite this article as: Tsui-Jung Lin, Hai-Yen Chang, Hui-Fun Yu, Ching-Pao Kao , The impact of political connections and business groups on cash holdings: Evidence from Chinese listed firms. Glofin (2018), doi:10.1016/j.gfj.2018.10.001
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Title: The impact of political connections and business groups on cash holdings:
PT
Evidence from Chinese listed firms
Authors:
NU
SC
RI
Tsui-Jung Lin Professor, Department of Banking and Finance, Chinese Culture University, Taiwan E-mail: [email protected] No.55, Hwa Kang Rd., Yang Ming Shan, Taipei 11114, Taiwan
MA
Hai-Yen Chang Associate Professor, Department of Banking and Finance, Chinese Culture University, Taiwan E-mail: [email protected]
D
No.55, Hwa Kang Rd., Yang Ming Shan, Taipei 11114, Taiwan
CE
PT E
Hui-Fun Yu* Corresponding Author and Associate Professor, Department of Banking and Finance, Chinese Culture University, Yang Ming Shan, Taipei 11114, Taiwan E-mail: [email protected]; [email protected] No.55, Hwa Kang Rd., Yang Ming Shan, Taipei 11114, Taiwan
AC
Ching-Pao Kao
Department of Banking and Finance, Chinese Culture University, Taiwan E-mail: [email protected] No.55, Hwa Kang Rd., Yang Ming Shan, Taipei 11114, Taiwan
1
ACCEPTED MANUSCRIPT
ABSTRACT
We investigated the impact of political connections and business group affiliation on
PT
the cash holdings of firms listed on two main Chinese stock exchanges. We applied
RI
panel data regression analysis on a dataset comprising 10,832 observations for these
SC
companies from 2008 to 2015. We found political connections to be positively correlated with cash holdings, while business group affiliation is negatively correlated.
NU
Firms with both political connections and business group affiliations hold more cash.
JEL classifications:
CE
G3
PT E
D
motives of corporate managers.
MA
These findings are consistent with the tenets of agency theory and the speculative
AC
G30 Keywords:
Political connections Cash holdings Business groups
2
ACCEPTED MANUSCRIPT
1. Introduction
Cash holdings play an important role in emerging markets because they affect
PT
companies’ ability to raise capital for business operations and their investment
RI
decisions (Francis, Hasan, Song, & Waisman, 2013). Although firms in both developed
SC
and emerging markets need to make decisions about the appropriate level of cash holdings, companies in emerging markets have greater concerns regarding alternative,
NU
external sources of funding because of lower liquidity and less access to bank loans and
MA
financial markets for their bonds and common equity (Nasdaq, 2017).
D
Keynes (1936) proposed three main motives for holding cash: transactional (to
PT E
cover current transactions); precautionary (to deal with unforeseen future events); and speculative (to invest). Three theories have been developed to explain corporate cash
CE
holdings and capital structure. According to Myers’s (1977) trade-off theory, firms
AC
incur opportunity costs when holding excess cash but must pay the cost of debt if they do not have cash on hand. Myers and Majluf (1984), who developed the pecking order theory, explained that when firms see promising investment opportunities, they prefer internal (cash) to external (debt, equity) funding because internal sources of funding require the lowest costs. In the free cash flow theory, which extends agency theory, Jensen (1986) proposed that control over cash empowers corporate managers, who thus 3
ACCEPTED MANUSCRIPT
may hold onto larger-than-optimal levels of cash in conflict with the interests of shareholders or may squander company cash in their own interests. Although these theories have been more or less verified in empirical studies, it
PT
should be remembered that theories advanced in developed markets, such as the United
RI
States and the United Kingdom, may not be applicable to emerging markets (Al-Najjar
SC
& Belghitar, 2011; Brenes, Madrigal, & Requena, 2011; Tahir, Alifiah, Arshad, & Saleem, 2016) because of differences in economic and financial contexts. Emerging
NU
markets do not enjoy established capital markets with the high levels of liquidity,
MA
regulatory requirements, and market capitalization that are common in developed
D
countries (Nasdaq, 2017). The OECD (2017) stresses that institutional differences
PT E
such as the political system, equity market, banks, regulatory agencies, and legal framework affect the financial decisions of firms in developed countries and emerging
CE
markets. Businesses and investors must consider these factors.
AC
In developed countries, firm size is negatively related to cash holdings because larger companies have greater access to funding than smaller ones, and therefore less need to hold onto cash (Akben-Selcuk & Altiok-Yilmaz, 2016; Hill, Fuller, Kelly, & Washam, 2013). For example, Cai, Zeng, Lee, and Ozkan (2016) found that group-affiliated firms tend to hold significantly less cash than their unaffiliated counterparts. In emerging markets, firm size is positively related to cash because firms 4
ACCEPTED MANUSCRIPT
are more diversified and a parent company may have several subsidiaries (Al-Najjar & Belghitar, 2011; Shabbir, Hashmi, & Chaudhary, 2016). Further, business groups, which are common in many emerging markets, serve as a nonmarket mechanism used
PT
by firms to overcome a weak institutional context (Khanna & Palepu, 2000b). La
RI
Porta, Lopez-de-Silanes, and Shleifer (1999) also found pyramidal structures in the
economies with weak shareholder protection.
SC
control of firms and participation in management by families or the state, especially in
NU
Studies have also offered differing opinions regarding the connections between
MA
political connections and cash holdings. In emerging markets, political connections
D
can help firms secure bank loans or become publicly listed (Bao, Johan, & Kutsuna,
PT E
2016; Yang, Lian, & Liu, 2012). Some have claimed that political connections reduce the need to hold onto cash (Claessens, Feijen, & Laeven, 2008; Faccio, 2006, 2010;
CE
Hill et al., 2013). Political connections are particularly common in China, the world’s
AC
largest emerging market country, and one in which the Communist Party dominates. Most firms in China are owned by business groups, and the majority of these are politically connected to either the central or local government (Khanna & Palepu, 2000a; Yang et al., 2012). Our study investigates the effects of an internal force (the corporate structure of the business group) and an external force (political connections) on the cash holdings of Chinese firms. 5
ACCEPTED MANUSCRIPT
This study makes a threefold contribution to the literature. First, it extends our understanding of the impact of political connections on the level of cash holdings. The uniqueness of China’s capital market (e.g., market inefficiency and firms’ lack of
PT
capital) and its specific network characteristics (e.g., ties with government as a critical
RI
resource) make it a different setting in which to examine this influence. Second, this
SC
paper argues that the business group (BG) may moderate the relation between political connectedness and cash holdings. Finally, we also attempt to compare the
NU
theoretical predictions with empirical evidence. The results of this study may help
D
cash holdings in Chinese firms.
MA
investors understand how political connections and business group structure affect
PT E
We chose China as the sample country for this study because of its high population, rapid economic growth, and large trade surplus, which make it the largest
CE
emerging market. With 1.3 billion people, China accounts for 17.6% of the total 7.4
AC
billion world population (World Bank, 2016b). From 1995 to 2015, it averaged a gross domestic product (GDP) growth of 9.48% (World Bank, 2016a). Moreover, given that China has had the largest trade surplus of any nation in the world since 2000, Chinese firms have had ample opportunity to accumulate a great amount of cash (FRED, 2016). China’s unique economic environment is particularly desirable for examining 6
ACCEPTED MANUSCRIPT
cash holding theories because of the prevalence of political connections and the BG corporate structure (parent companies with subsidiaries), which could affect a firm’s financial decisions. By the end of 2009, 95.3% of all listed companies in China
PT
operated within BGs (Zhang, Lu, Zhang, & Jiang, 2015). Moreover, in terms of
RI
political connectedness, four large state-owned-enterprise (SOE) banks—Bank of
SC
China, Construction Bank of China, Agriculture Bank of China, and International Commercial Bank of China—dominate the country’s financial markets (Allen, Qian,
NU
& Qian, 2005). SOE companies can easily obtain funding from SOE banks, and most
MA
large companies rely on bank loans (Megginson, Ullah, & Wei, 2014). In short, political connection is also associated with ease of financing. Additionally, both firm
PT E
D
size and corporate structure affect firms’ access to loans, and thus their cash holdings (Boubakri, El Ghoul, & Saffar, 2013). To the best of our knowledge, this is the first
CE
study to examine empirical evidence demonstrating the importance of political
AC
connections in determining the ease of acquisition and amount of cash holdings for business groups in China. The remainder of this paper is organized as follows. Section 2 includes the literature review and hypothesis development. The data collection and measurement of variables are discussed in section 3. Section 4 presents the empirical results. Section 5 concludes the paper. 7
ACCEPTED MANUSCRIPT
2. Related literature and hypotheses development 2.1. Political connections and cash holdings
PT
Faccio (2006) found that corporate political connections are widespread around
RI
the world, especially in countries with emerging economies where regulations
SC
regarding political conflicts of interest are relaxed. Since larger firms tend to build higher levels of political connectedness than smaller ones (Agrawal & Knoeber, 2001;
NU
Faccio, 2006; Johnson & Mitton, 2003), large firms in emerging markets often find it
MA
easier to get external capital, such as bank loans (Allen et al., 2005; Bao et al., 2016).
D
Boubakri et al.(2013) and Chaney, Faccio, and Parsley (2011) have suggested that
PT E
agency problems are more likely to occur in politically connected (PC) firms than non–politically connected ones in emerging countries. The influx of funds raises the
CE
level of liquid assets and increases the number of managers who act at the
AC
shareholders’ expense. Further, Caprio, Faccio, and McConnell (2013) proposed that in emerging countries with weaker legal property protection, firms adjust their asset holdings to protect their assets from government seizure. They found a negative relation between corporate holdings of liquid assets and political corruption. However, scholars hold varying opinions about the effect of political connection on firms’ cash holdings. Some researchers have argued that political connection has a 8
ACCEPTED MANUSCRIPT
negative relationship with cash level. Faccio (2010) found that PC companies in developed countries hold onto less cash because they have political support and relationships with government agencies. In emerging countries, Al-Najjar (2013)
PT
found that companies with significant political support hold less cash since they have
RI
government support in financially critical periods.
SC
Others, however, have contended that political connection has a positive correlation with cash holdings. Kusnadi and Wei (2012) found that managers of PC
NU
firms in emerging markets, as well as firms in countries with weak legal protection of
MA
investors and prevalent corruption, are inclined to hoard more cash than nonpolitical
D
firms. Given China’s inefficient legal environment and unique business practices, we
PT E
inferred that politically-connected firms suffer severe agency problems and managers hold more cash to act self-interest. Therefore, we propose the first hypothesis as
AC
CE
follows:
Hypothesis 1. In China, political connection is positively correlated with corporate cash holdings.
2.2. Business groups and cash holdings Khanna and Palepu (2000a) proposed that in emerging markets, companies form 9
ACCEPTED MANUSCRIPT
business groups as a defense against inadequate investor protection, legal enforcement, and information disclosure. In effect, the BG serves as an internal financial market in which funds can move from parent company to subsidiaries or among the affiliated
PT
firms for operational or investment purposes (Bena & Ortiz-Molina, 2013). Business
RI
groups may thus reduce their members’ need for cash on hand (Schiantarelli &
SC
Sembenelli, 2000).
The rapid growth and rise in importance of China as an emerging market
NU
economy provides a natural platform for analyzing issues related to the role of the
MA
business group (BG). To become a business group in China, the parent company must
D
provide capital of over 50 million yuan (US$6 million) and must have at least five
PT E
affiliated companies. Moreover, the total capital of the parent and its affiliated companies must exceed 100 million yuan (US$12 million). He, Mao, Rui, and Zha
CE
(2013) noted rapid growth in the total number of business groups over the past few
AC
decades. According to the State Statistical Bureau (2002A–2007A), as of 2006, there were 2,856 officially recognized business groups in China, with 27,950 directly owned first-tier subsidiaries, employing around 30 million people and accounting for almost 60% of national industrial output (He et al., 2013). The ecology of business groups in China is substantially different from those of the so-called Big Six keiretsus in Japan or the top 30 chaebols in South Korea (Chang & Choi, 1988; Chang & Hong, 10
ACCEPTED MANUSCRIPT
2000). The size distribution of the groups and the identity of their ultimate controlling shareholders are different from those in other emerging markets or developed economies. China has a number of state-dominated business groups, established for
PT
purposes of economic development and social distribution, to carry out national goals.
RI
However, because of market failures and the lack of a well-developed legal and
SC
regulatory system in China, family- or natural person-dominated business groups have also been becoming more common.
NU
However, the evidence regarding the effects of BGs on cash holding is
MA
inconsistent. He et al. (2013) found that in China, BG-affiliated firms have a lower
D
risk of incurring cash shortages when faced with investment opportunities because
PT E
they are able to obtain funding within the BG. Furthermore, Cai et al. (2016) found that group-affiliated Chinese firms held significantly less cash than their nonaffiliated
CE
counterparts. This is consistent with both the trade-off theory and the pecking order
AC
theory. Firms with quicker access to funding have the freedom to hold less cash. However, less financially constrained state-owned enterprises (SOEs) are slower to respond with decreased cash holdings than non-SOEs. Studies also show varying assessments of the relationships among firm size, liquidity, and corporate cash holdings. Al-Najjar (2013) asserted that larger firms in emerging countries tend to hold less cash in general, that firms with leverage also 11
ACCEPTED MANUSCRIPT
hold onto less cash, and that cash holdings are lower when firms have easy access to funding. Shabbir et al. (2016) found that, in emerging markets, company size was positively correlated with increased cash holdings, and leverage (bank loans) with
PT
decreased cash holdings. In contrast, Uyar and Kuzey (2014) observed that firms with
RI
higher liquidity and larger firm size held less cash. Given all this evidence, we
SC
propose the following hypothesis:
NU
Hypothesis 2. In China, business group affiliation is negatively correlated with
D
MA
corporate cash holdings.
PT E
2.3. Relationship among political connections, business groups, and cash holdings Studies have offered two explanations as to why business groups seek to
CE
maintain cash holdings. The predominant view is based on the free cash flow
AC
hypothesis: the pyramidal structure of business groups worsens information asymmetry between managers and shareholders and exacerbates the agency problem, giving managers an incentive to control more cash. The other argument claims that business groups can take advantage of internal capital markets to overcome the financial constraints characteristic of underdeveloped financial markets, so managers in such groups hold less cash. The trade-off between these forces may partly depend 12
ACCEPTED MANUSCRIPT
on other institutional factors, for example political connections. He et al. (2013) found that, in China, privately owned business groups are more likely to use internal capital markets than are groups controlled by the central government. Cai et al. (2016) also
PT
found that state ownership moderates the benefit received from such an internal
RI
capital market, and Firth, Lin, and Wong (2008) found that, in China, the distribution
SC
of loans is skewed toward firms with political connections. Chen, Sun, Tang, and Wu (2011) found that political connections significantly reduced investment efficiency in
NU
SOEs but not in non-SOEs. Furthermore, Chen, Ding, and Kim (2010), analyzing
MA
firm-level data from 17 jurisdictions from 1997 to 2001, found that it was harder to
D
predict the earnings of firms with political connections than those of firms without
PT E
such connections. Apparently, having government ties deepens the problems of adverse selection and information asymmetry between managers and shareholders. In
CE
other words, the cost of political connectedness overcomes the benefit of the internal
AC
capital market to the business group, pushing managers to strengthen cash holdings. It should be noted that China has been criticized for its weak investor protection and internal corporate governance structure (Jiang & Kim, 2015). This means that, for firms in business groups with government ties, managers can use their own discretion to hold onto higher levels of cash holdings even to the point of harming the firm’s value. We therefore conclude that firms are more likely to suffer from agency cost 13
ACCEPTED MANUSCRIPT
problems when they are members of a business group holding higher levels of cash and having political connections:
PT
Hypothesis 3. In China, the combination of political connections and business group
SC
RI
affiliation is positively correlated with corporate cash holdings.
3. Research method
NU
We analyzed data on a sample of listed companies in China between 2008 and
MA
2015 to test the relationships among political connection, BG affiliation, and cash
D
holdings.
PT E
3.1. Data sample
We collected data on cash amount, other financial information, and corporate
CE
structure, as well as the names and positions of executives and managers, from the
AC
China Stock Market and Accounting Research (CSMAR) database. We removed companies in the financial sector because financial institutions do not have cash and funding issues, and excluded firms with missing financial information in order to avoid potential statistical errors. The remaining panel contains 9,777 observations.
3.2. Research model 14
ACCEPTED MANUSCRIPT
The key variables in this study are the level of cash holdings, PC, and BG. The data have both longitudinal and cross-sectional properties; therefore, following Opler, Pinkowitz, Stulz, and Williamson (1999), we apply panel data regression analysis to
of 200.794 (significant at 1%),
RI
(Greene, 2008, chapter 9) yielded a value of χ2
PT
measure cash holdings against the two other variables, PC and BG. The Hausman test
SC
ruling out the random effect assumption, so we used the fixed effect assumption in the regression model. Endogeneity is another potential issue: cash is a function of
NU
leverage and vice versa, and cash might also be endogenous with dividend payouts
MA
(DIVD)1. To mitigate this concern, following Harford, Mansi, and Maxwell (2008)
D
and Chen, Leung, and Goergen (2017), we lag some independent variables. Following
PT E
Al-dhamari and Ku Ismail (2015), we measure cash holdings by taking the logarithm of the proportion of cash and cash equivalents to net assets.2 The equation is
AC
CE
formulated as follows3:
1
We thank the reviewer for these suggestions.
2
When the proportion takes the logarithm, the value of logCash may be negative.
3
The correlations between leverage and net working capital were high, so we revised the original model and used leverage in the model. 15
ACCEPTED MANUSCRIPT
𝐿𝑜𝑔𝐶𝑎𝑠ℎ𝑖,𝑡 = 𝛽0 + 𝛽1 𝑃𝐶𝑖,𝑡 + 𝛽2 𝐵𝐺𝑖,𝑡 + 𝛽3 (𝑃𝐶𝑖,𝑡 ∗ 𝐵𝐺𝑖,𝑡 ) + 𝛽4 𝐶𝑎𝑠ℎ𝑖,𝑡−1 + 𝛽5 𝑆𝑖𝑧𝑒𝑖,𝑡−1 + 𝛽6 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑖,𝑡−1 + 𝛽7 𝑀⁄𝐵
𝑖,𝑡
+ 𝛽8 𝐶𝐹𝑖,𝑡
+ 𝛽9 𝐶𝑎𝑝𝑒𝑥𝑖,𝑡−1 + 𝛽10 𝐷𝐼𝑉𝐷𝑖,𝑡−1 + 𝛽11 𝐶𝐹𝑉𝑜𝑙𝑖,𝑡−1 + 𝛿 + 𝜃 (1)
RI
PT
+ 𝜖𝑖,𝑡 ,
SC
where i is the ith company; t -1is the t-1th year; δ indicates firm effects; θ is the
NU
year effects; and ∈ is the error term.
MA
3.3. Definition of variables
D
3.3.1. Cash holdings
PT E
LogCashi,t represents the logarithm of the amount of cash held by company i at the end of year t. The level of cash holdings is calculated as the sum of cash and cash
AC
equivalents).
CE
equivalents divided by net assets (total assets minus the sum of cash and cash
𝐿𝑜𝑔𝐶𝑎𝑠ℎ𝑖,𝑡 = log
(𝐶𝑎𝑠ℎ + Cash equivalents ) , 𝑛𝑒𝑡 𝑎𝑠𝑠𝑒𝑡𝑠
(2)
where Cash = cash on hand; Cash equivalents = assets quickly convertible into a known cash amount; Net assets = total assets – (cash + cash equivalents). 16
ACCEPTED MANUSCRIPT
3.3.2. Political connection Faccio (2006) defined a company as politically connected if “at least one of its
PT
large shareholders (anyone controlling at least 10 percent of voting shares) or one of
RI
its top officers (CEO, president, vice-president, chairman, or secretary) is a member
SC
of parliament, a minister or the head of state, or closely related to a top politician” (p. 370). Previous studies have used a dummy variable to capture political connections.
NU
However, simply using a dummy variable does not allow consideration of whether a
MA
firm has one or many connections, nor does it capture the nature of specific connections. In our case, it is important to know whether members of the board are
PT E
D
members of the Chinese People’s Congress (CPC) or the Chinese People’s Political Consultative Conference (CPPCC). We calculate political connection as the sum of
CE
the number of board members who are members of the CPC or the CPPCC divided by
AC
the board size.
3.3.3. Business group He et al. (2013) derived their definition of a BG from an annual survey conducted by the Chinese Securities Regulatory Commission (CSRC) covering the period from 1990 to 2006. The authors regarded a Chinese company as a BG if the firm’s ultimate 17
ACCEPTED MANUSCRIPT
controlling entity possessed more than one firm in a particular year. We adopt the same definition in our study, and use a dummy variable (BG) that is equal to 1 if the
PT
firm is a BG, and 0 if not.
RI
3.3.4. Control variables
SC
For other variables that influence the level of cash holdings we follow Opler et al. (1999), defining them as follows:
NU
Cash i, t-1 = (cash + cash equivalents) / net assets in the previous term;
MA
Size i, t-1 = natural logarithm of net assets in the previous term;
D
Leverage i, t-1 = (current liabilities + long-term liabilities) / net assets in the previous
PT E
term;
M/B (market-to-book ratio) i, t = market value of a firm / book value of a firm;
CE
CF (cash flow) i, t = earnings before interest, taxes, depreciation and amortization /
AC
net assets;
Capex (capital expenditure) i, t-1 = capital expenditure / net assets in the previous term; DIVD (dividend payout) i, t-1 = a dummy variable that is equal to 1 if a firm pays out cash dividends in the previous term, and 0 otherwise; CFVol i, t-1 = standard deviation of the cash flow in the previous term.
18
ACCEPTED MANUSCRIPT
4.
Results We first tested the relationship between PC and cash holdings and the correlation
between BG and cash holdings. Then we examined the effect of both PC and BG on
RI
PT
cash holdings, and compared the three outcomes.
SC
4.1. Descriptive statistics
The descriptive statistics for the variables (see Table 1) indicated significant
NU
differences between the maximum and minimum amounts. Table 2 shows the Pearson
MA
correlation coefficients (ρ) among the variables, which are all between 0.7 and -0.7
D
(-0.7＜ρ＜+0.7), indicating that the regression model does not suffer from
PT E
multicollinearity. In short, the low correlations among variables reflect the reliability of the regression model.
CE
We divided the firms into two groups, those in business groups (BG) and those
AC
not in business groups (NBG), and compared the means and medians of the variables between the two groups. The results are shown in Table 3. The means of cash holdings of the BG firms and NBG firms are -0.387 and -0.117, respectively, and the difference is significant at the 1% level. Similarly, the medians of cash holdings of the BG and NBG firms are -0.422 and -0.172, respectively, and this difference is also significant at the 1% level. In sum, BG firms hold significantly less cash than NBG 19
ACCEPTED MANUSCRIPT
firms.
Table1 Descriptive statistics for the variables. Observation Mean s
Media n
Maximum
Minimum
Std. Dev.
9777
-0.389
-0.379
1.676
-2.659
0.450
PC BG Casht-1 Sizet-1 Leveraget-1 M/B CF DIVDt-1 Capext-1
9777 9777 9777 9777 9777 9777 9777 9777 9777
0.047 0.851 0.006 6.352 0.569 2.310 0.086 0.561 0.029
0.000 1.000 -0.017 6.329 0.552 1.596 0.070 0.091 0.000
0.250 1.000 2.225 8.883 21.112 101.461 4.677 371.317 1.000
0.000 0.000 -1.733 3.613 0.017 0.113 -2.705 -0.999 0.000
0.058 0.356 0.337 0.588 0.413 3.030 0.126 8.086 0.167
CFVolt-1
9777
0.045
0.016
18.617
0.000
0.337
SC
NU
D
MA
RI
logCash
PT
Variable
PT E
Notes: Here, “logCash” means cash holdings measured by taking the natural logarithm of the sum of cash and cash equivalents divided by net assets. “PC”
CE
represents political connection, and is calculated as the sum of the number of members of the board with connections to the CPC or the CPPCC divided by board
AC
size. The total number of CPC/CPPCC-connected members ranges from 0 to 3. “BG” means business group; the dummy is equal to 1 if the firm is affiliated with a business group, and 0 if not. “Cash t-1” means cash + cash equivalents/net assets (total assets – cash –cash equivalents) in the previous term; “Size t-1” means firm size measured by taking the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as the sum of current liabilities and long-term liabilities, divided by net assets in the previous term. “M/B” means market-to-book ratio 20
ACCEPTED MANUSCRIPT measured as the firm’s market value divided by its book value. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by net assets. “Capex t-1” represents capital expenditure measured as capital expenditure divided by net assets in the previous term. “DIVD t-1” means dividend payout; the dummy is equal to1 if the firm has paid out cash dividends in the previous term, and 0
PT
if not. “CFVolt-1” means the standard deviation of cash flow in the previous term. ***
RI
represents the 1% significance level, ** represents the 5% significance level, and *
AC
CE
PT E
D
MA
NU
SC
represents the 10% significance level.
21
ACCEPTED MANUSCRIPT
Table 2 Correlations.
Variable
logCash
PCt-1
BGt-1
Casht-1
Sizet-1
Leveraget- M/B
CF
1.000 .111** -.231**
1.000 -.101**
1.000
Casht-1
.127**
-.029**
.075**
1.000
**
**
**
Sizet-1 -.423 Leveraget -.141**
I R
DIVDt-1 Capext-1 CFVolt-
SC
1
logCash PC BG
T P
1
U N
.290 .215**
-.001 .086**
-.238** -.131** -.166** -.057**
.020* .111** -.133** .046**
-.650** -.366** -.041** -.130**
-.425** -.246** -.090** .007
1.000 .401** .035** .051**
1.000 .133** .161**
1.000 .074**
1.000
-.082**
-.059**
-.287**
-.120**
.232**
.140**
.060**
.034**
.358** .393** .138** .098**
.048** .040** .086** .042**
CFVolt-1
.104**
.013
E C
C A
D E
PT
-1
M/B CF DIVDt-1 Capext-1
1.000 .319**
A M
-.050 -.059**
1.000
1.000
Notes: “logCash” means cash holdings measured by taking the natural logarithm of the sum of cash and cash equivalents divided by net assets. “PC” represents political connection, calculated as the sum of the number of board members with connections to the CPC or the CPPCC, divided
22
ACCEPTED MANUSCRIPT
by board size. “BG” means business group; the dummy is equal to 1 if the firm is affiliated with a business group, and 0 if not. “Cash t-1” means cash + cash equivalents/ net assets (total assets – cash –cash equivalents) in the previous term; “Size t-1” means size of firm measured by taking
T P
the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as the sum of current liabilities and
I R
long-term liabilities, divided by net assets in the previous term. “M/B” means market-to-book ratio measured as the firm’s market value divided
C S U
by its book value. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by net assets. “Capex t-1” represents capital expenditure measured as capital expenditure divided by net assets in the previous term. “DIVD t-1” means dividend
N A
payout; the dummy is equal to1 if the firm has paid out cash dividends in the previous term, and 0 if not. “CFVolt-1” means the standard deviation
M
of cash flow in the previous term. *** represents the 1% significance level, ** represents the 5% significance level, and * represents the 10%
D E
significance level.
T P E
C C
A
23
ACCEPTED MANUSCRIPT Table 3 Test results for means and medians of variables of business-group- and non-business-group-affiliated firms. Panel A: Mean
Panel B: Median
Variable
All samples
Variable
All samples
difference
difference
NBG
BG
logCash
-0.387
-0.117
-0.270***
-0.422
PC Casht-1 Sizet-1 Leveraget-1 M/B CF DIVDt-1 Capext-1 CFVolt-1
0.044
0.060
-0.017***
0.000
0.005
-0.027
0.032***
6.362
5.947
0.414***
0.579
0.433
0.146***
2.195
3.254
-1.058***
0.090
0.119
0.531
NBG
-0.172
-0.250***
0.074
-0.074***
-0.010
0.000
-0.010***
6.383
-0.071
6.454***
0.573
5.960
-5.386***
1.480
0.409
1.072***
-0.029***
0.067
2.296
-2.228***
0.630
-0.099
0.074
0.000
0.074***
0.029
0.055
-0.026***
0.000
0.088
-0.088***
0.047
0.045
0.002
0.015
0.200
-0.185***
MA
NU
SC
RI
PT
BG
D
Notes: “logCash” means cash holdings measured by taking the natural
PT E
logarithm of the sum of cash and cash equivalents divided by net assets. “PC” represents political connection, calculated as the sum of the number of
CE
members of the board with connections to the CPC or the CPPCC, divided by
AC
board size in the previous term. “BG” means business group; the dummy is equal to 1 if the firm is affiliated with a business group, and 0 if not, in the previous term. “Cash t-1” means cash + cash equivalents/net assets (total assets – cash –cash equivalents) in the previous term; “Size t-1” means firm size measured by taking the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as the sum of current liabilities and long-term liabilities divided by net assets in the previous term.
24
ACCEPTED MANUSCRIPT “M/B” means market-to-book ratio measured as the firm’s market value divided by its book value. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by net assets. “Capex t-1”
represents capital expenditure measured as capital expenditure divided by
net assets in the previous term. “DIVD t-1” means dividend payout; the
PT
dummy is equal to1 if the firm has paid out cash dividends in the previous
RI
term, and 0 if not. “CFVolt-1” means the standard deviation of cash flow in the
SC
previous term. *** represents the 1% significance level, ** represents the 5%
NU
significance level, and * represents the 10% significance level.
MA
4.2. Multivariate analysis
4.2.1. The relationship between political connections and cash holdings
PT E
D
Table 4 shows the results of our panel data analysis. In model 1, the coefficient of political connections (PC) is 0.759 and is significant at the 1% level, showing that
CE
companies whose board members held official positions in the government
AC
maintained a higher level of cash. The results are consistent with Kusnadi and Wei’s (2012) finding that in emerging markets, managers of PC firms are inclined to hoard more cash than managers of nonpolitical firms, and support our hypothesis 1, that political connection is positively correlated with corporate cash holdings in China.
Table 4 25
ACCEPTED MANUSCRIPT The effect of political connectedness and business group affiliation on cash holdings.
C PC
Model 1
Model 2
Model 3
1.137*** (25.389) 0.759*** (11.744)
1.705*** (39.419)
1.411*** (26.972) 0.478*** (3.031)
-0.123*** (-14.861)
-0.157*** (-10.442) 0.292*
PT
BG
Leveraget-1 M/B
0.755*** (24.813) DIVDt-1 0.187*** (8.191) Capext-1 -0.001 (-1.419) CFVolt-1 -0.018 (-1.470) Year effect Yes Firm effect Yes Ajusted R2 0.262 F-statistic 262.374*** 11012 Obs.
SC
AC
CE
PT E
D
CF
0.261*** (27.217) -0.318*** (-46.292) 0.015* (1.731) -0.010*** (-6.998)
NU
Sizet-1
0.281*** (24.421) -0.253*** (-37.341) -0.018** (-2.177) -0.007*** (-4.831)
MA
Casht-1
(1.681) 0.281*** (23.916)
RI
PC* BG
0.276*** (-34.464) 0.010 (1.075) 0.010*** (-6.243)
0.825*** (31.268) 0.125*** (6.976) -0.001* (-1.569) -0.047*** (-4.004) Yes Yes
0.744*** (23.454) 0.157*** (6.790) -0.001* (-1.646) -0.029** (-2.389) Yes Yes
0.341 470.132*** 13574
0.301 249.186*** 9777
Notes: “logCash” means cash holdings measured by taking the natural logarithm of the sum of cash and cash equivalents divided by net assets. “PC” represents political connection, calculated as the sum of the number of board members with connections to the CPC or the CPPCC, divided by board size. “BG” means business group; the 26
ACCEPTED MANUSCRIPT dummy is equal to 1 if the firm is affiliated with a business group, and 0 if not. “Cash t-1”
means cash + cash equivalents/net assets (total assets – cash –cash equivalents) in
the previous term; “Size t-1” means size of firm measured by taking the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as the sum of current liabilities and long-term liabilities, divided by net
PT
assets in the previous term. “M/B” means market-to-book ratio measured as the firm’s
RI
market value divided by its book value in the previous term. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by
SC
net assets. “Capex t-1” represents capital expenditure measured as capital expenditure
NU
divided by net assets in the previous term. “DIVD t-1” means dividend payout; the dummy is equal to1 if the firm paid out cash dividends in the previous term, and 0 if
MA
not. “CFVolt-1” means standard deviation of cash flow in the previous term. *** represents the 1% significance level, ** represents the 5% significance level, and *
PT E
D
represents the 10% significance level.
CE
4.2.2. The relationship between business group affiliation and cash holdings
AC
Model 2 in Table 4 shows the effect of BG on cash holdings. The coefficient of business group (BG) is -0.123 and is significant at the 1% level, indicating that business group affiliation is negatively correlated with cash holdings, as our hypothesis 2 predicts.
4.2.3. The relationship among political connection, business group affiliation, and
27
ACCEPTED MANUSCRIPT
cash holdings In Table 4, model 3, the coefficient of the interaction term PC*BG is 0.292 and is significant at the 10% level, suggesting that cash holdings are higher for those firms
PT
with ties to both government and a business group. As we explain in section 1, such
RI
ties are important ways to get outside resources, because most Chinese listed
SC
companies are not able to obtain capital from the market. However, as we also explain above, agency problems are more serious in China than elsewhere, so that affiliation
NU
with a business group does not significantly reduce cash holding by politically
D
MA
connected firms.
PT E
4.3. Robustness tests
To check the robustness of our model, we take into consideration another
CE
definition of political connection. Fan (2016) found that the presence of independent
AC
directors with political background may bring extra profit to a firm. We therefore redefine a company as politically connected if its independent directors have a political background, as measured by the sum of the number of connected independent directors divided by board size, and retest the model. The results are presented in Table 5. All three previous results—concerning the positive relationship between political connections and cash holdings, the negative relationship between 28
ACCEPTED MANUSCRIPT
BG and cash holdings, and the positive effect of the combination of political connections and group business affiliation on cash holdings—are found to be robust.
Table 5
1.371*** (26.835) 0.025 (0.965)
1.797*** (42.024)
0.289***
1.548*** (28.348) -0.066 (-0.961) -0.076*** (-5.106) 0.131* (1.793) 0.304***
(58.948) -0.332*** (-50.303) 0.018*** (2.870) -0.009*** (-10.551) 0.378*** (21.338)
(50.071) -0.296*** (-36.223) 0.014* (1.953) -0.008*** (-7.381) 0.337*** (16.253)
0.052*** (5.229) 0.000 (0.322 ) -0.006 (-0.861) Yes Yes 0.406 1033.769***
0.053*** (4.104) 0.000*** (0.328) -0.004 (-0.477) Yes Yes 0.361 503.702***
-0.058*** (-5.899)
NU
BG
MA
PC* BG
Leveraget-1 M/B CF
0.058*** (4.565) Capext-1 0.000 (0.517) CFVolt-1 0.026*** (3.659) Year effect Yes Firm effect Yes Ajusted R2 0.337 F-statistic 625.014***
AC
DIVDt-1
PT E
Sizet-1
(51.522) -0.277*** (-35.745) -0.020*** (-3.370) -0.005*** (-6.004) 0.364*** (18.782)
D
0.300***
CE
Casht-1
Model 3
RI
PC
Model 2
SC
C
Model 1
PT
Results of robustness tests.
29
ACCEPTED MANUSCRIPT Obs.
11012
13574
9777
Notes: “logCash” means cash holdings measured by taking the natural logarithm of the sum of cash and cash equivalents divided by net assets. “PC” represents political connection, calculated as the sum of the number of independent directors with political connections, divided by board size. “BG” means business group; the dummy
PT
is equal to 1 if the firm is affiliated with a business group, and 0 if not. “Cash t-1”
RI
means cash + cash equivalents/net assets (total assets – cash –cash equivalents) in the
SC
previous term; “Size t-1” means size of firm measured by taking the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as
NU
the sum of current liabilities and long-term liabilities, divided by net assets in the previous term. “M/B” means market-to-book ratio measured as the firm’s market
MA
value divided by its book value. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by net assets. “Capex t-1”
D
represents capital expenditure measured as capital expenditure divided by net assets in
PT E
the previous term. “DIVD t-1” means dividend payout; the dummy is equal to1 if the firm paid out cash dividends in the previous term, and 0 if not. “CFVolt-1” means the
CE
standard deviation of cash flow in the previous term. *** represents the 1% significance level, ** represents the 5% significance level, and * represents the 10%
AC
significance level.
5.
Discussion and conclusions Our panel data regression analysis found evidence that Chinese firms with
business group affiliations hold less cash than firms without. This outcome is consistent with internal capital market views. We deduce that BG-affiliated firms are 30
ACCEPTED MANUSCRIPT
able to get funding from internal sources, such as parent companies or other affiliated firms, rather than external sources, such as banks, so they have less need to hold onto excess cash. Further, Chinese firms with political connections hold more cash than
PT
firms without. This contradicts trade-off theory and pecking order theory, but aligns
RI
with agency theory (Jensen, 1986; Myers, 1977; Myers & Majluf, 1984). Politically
SC
connected firms have easier access to bank loans, but, as agency theory predicts, they are likely to hold onto more cash because corporate managers associate control of
NU
liquid assets with power (Jensen, 1986). We infer that managerial self-interest
MA
dominates precautionary and speculative motives in decisions regarding levels of cash
D
holdings. In other words, a cash cushion may tempt managers to waste company
PT E
money to further their own interests (e.g., buy a jet for their personal use or go on luxurious business trips).
CE
These findings suggest that PC and BG firms in emerging markets, particularly
AC
in China, may adopt different financial practices from their counterparts in developed countries such as the United States and Great Britain. There is no optimal cash level in these firms. We caution investors and financial analysts that when measuring the performance of politically connected and business-group-affiliated Chinese firms, they must understand that it is common for these firms to keep a high level of cash that may earn no interest. We have looked only at the level of cash holdings. It 31
ACCEPTED MANUSCRIPT
remains unclear how corporate managers make decisions regarding when and how excess cash should be used. It would be interesting for future researchers to identify the determinants of using excess cash in politically connected and non–politically
PT
connected firms in China.
RI
Acknowledgements
SC
We would like to acknowledge the financial support provided by the Minister of Science and Technology (MOST 106-2410-H-034-013).
NU
References
MA
Agrawal, A., & Knoeber, C. (2001). Do some outside directors play a political role?
D
Journal of Law and Economics, 44(1), 179–198.
PT E
Akben-Selcuk, E., & Altiok-Yilmaz, A. (2016). Determinants of corporate cash holdings: Firm level evidence from emerging markets. In: Ü. Hacioğlu, H.
CE
Dinçer, & N. Alayoğlu. (Eds.), Global business strategies in crisis:
AC
Contributions to management science (pp. 417–428). New York: Springer. Al-dhamari, R., & Ku Ismail, K. N. I. (2015). Cash holdings, political connections, and earnings quality: Some evidence from Malaysia. International Journal of Management Finance, 11(2), 215–231. Allen, F., Qian, J., & Qian, M. (2005). Law, finance, and economic growth in China. Journal of Financial Economics, 77(1), 57–116. 32
ACCEPTED MANUSCRIPT
Al-Najjar, B. (2013). The financial determinants of corporate cash holdings: Evidence from some emerging markets. International Business Review, 22(1), 77–88. Al-Najjar, B., & Belghitar, Y. (2011). Corporate cash holdings and dividend payments:
PT
Evidence from simultaneous analysis. Managerial and Decision Economics,
RI
32(4), 231–241.
SC
Bao, X., Johan, S., & Kutsuna, K. (2016). Do political connections matter in accessing capital markets? Evidence from China. Emerging Markets Review, 29,
NU
24–41.
MA
Bena, J., & Ortiz-Molina, H. (2013). Pyramidal ownership and the creation of new
D
firms. Journal of Financial Economics, 108(3), 798–821.
PT E
Boubakri, N., El Ghoul, S., & Saffar, W. (2013). Cash holdings of politically connected firms. Journal of Multinational Financial Management, 23(4), 338–
CE
355.
AC
Brenes, E. R., Madrigal, K., & Requena, B. (2011). Corporate governance and family business performance. Journal of Business Research, 64(3), 280–285. Cai, W., Zeng, C., Lee, E., & Ozkan, N. (2016). Do business groups affect corporate cash holdings? Evidence from a transition economy. China Journal of Accounting Research, 9(1), 1–24. Caprio, L., Faccio, M., & McConnell, J. J. (2013). Sheltering corporate assets from 33
ACCEPTED MANUSCRIPT
political extraction. Journal of Law and Economic Organization, 29(2), 332–354. Chaney, P. K., Faccio, M., & Parsley, D. (2011). The quality of accounting information in politically connected firms. Journal of Accounting and Economics,
PT
51(1), 58–76.
RI
Chang, S. J., & Choi, U. (1988). Strategy, structure and performance of Korean
SC
business groups: A transactions cost approach. Journal of Industrial Economics, 37, 141–158.
NU
Chang, S. J., & Hong, J. (2000). Economic performance of group-affiliated companies
MA
in Korea: Intragroup resource sharing and internal business transactions.
D
Academy of Management Journal, 43, 429–448.
PT E
Chen, C. J., Ding, Y., & Kim, C. F. (2010). High-level politically connected firms, corruption, and analyst forecast accuracy around the world. Journal of
CE
International Business Studies, 41(9), 1505–1524.
AC
Chen, J., Leung, W. S., & Goergen, W. (2017). The impact of board gender composition on dividend payouts. Journal of Corporate Finance, 43, 86–105. Chen, S., Sun, Z., Tang, S., & Wu, D. (2011). Government intervention and investment efficiency: Evidence from China. Journal of Corporate Finance, 17(2), 259–271. Claessens, S., Feijen, E., & Laeven, L. (2008). Political connections and preferential 34
ACCEPTED MANUSCRIPT
access to finance: The role of campaign contributions. Journal of Financial Economics, 88(3), 554–580. Faccio, M. (2006). Politically connected firms. American Economic Review, 96(1),
PT
369–386.
RI
Faccio, M. (2010). Differences between politically connected and nonconnected firms:
SC
A cross-country analysis. Financial Management, 39(3), 905–928. Fan, J. (2016). The value of political connections in China: Government officials on
NU
the board of directors. Retrieved from papers.ssrn.com.
MA
Federal Reserve Bank of St. Louis (FRED). (2016). China's trade surplus since 2000.
D
Retrieved from
PT E
https://fredblog.stlouisfed.org/2016/05/chinas-trade-surplus-since-2000/ Firth, M., Lin, C., & Wong, S. M. L. (2008). Leverage and investment under a
CE
state-owned bank lending environment: Evidence from China. Journal of
AC
Corporate Finance, 14, 642–653. Francis, B., Hasan, I., Song, L., & Waisman, M. (2013). Corporate governance and investment–cash flow sensitivity: Evidence from emerging markets. Emerging Markets Review, 15, 57–71. Greene, W. H. (2008). Econometric analysis (6th ed.). Upper Saddle River, NJ: Prentice Hall. 35
ACCEPTED MANUSCRIPT
Harford, J., Mansi, S. A., & Maxwell, W. F. (2008). Corporate governance and firm cash holdings in the U.S. Journal of Financial Economics, 87, 535–555. He, J., Mao, X., Rui, O. M., & Zha, X. (2013). Business groups in China. Journal of
PT
Corporate Finance, 22, 166–192.
RI
Hill, M. D., Fuller, K. P., Kelly, G. W., & Washam, J. O. (2013). Corporate cash
SC
holdings and political connections. Review of Quantitative Finance and Accounting, 42(1), 123–142.
NU
Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance, and
MA
takeovers. American Economic Review, 76(2), 323–329.
D
Jiang, F., & Kim, K. A. (2015). Corporate governance in China; A modern perspective.
PT E
Journal of Corporate Finance, 32, 190–216. Johnson, S., & Mitton, T. (2003). Cronyism and capital controls: Evidence from
CE
Malaysia. Journal of Financial Economics, 67(2), 351–382.
AC
Keynes, J. M. (1936). The general theory of employment, interest and money. London: Macmillan and Co. Khanna, T., & Palepu, K. (2000a). Is group affiliation profitable in emerging markets? An analysis of diversified Indian business groups. Journal of Finance, 55(2), 867–891. Khanna, T., & Palepu, K. (2000b). The future of business groups in emerging markets: 36
ACCEPTED MANUSCRIPT
Long-run evidence from Chile. Academy of Management Journal, 43, 268–285. Kusnadi, Y., & Wei, K. C. J. (2012). The effects of political connections on the level and value of cash holdings: International evidence. Working paper, Singapore
PT
Management University.
RI
La Porta, R., Lopez‐De‐Silanes, F., & Shleifer, A. (1999). Corporate ownership
SC
around the world. Journal of Finance, 54(2), 471–517.
Megginson, W. L., Ullah, B., & Wei, Z. (2014). State ownership, soft-budget
NU
constraints, and cash holdings: Evidence from China's privatized firms. Journal
MA
of Banking and Finance, 48, 276–291.
D
Myers, S. C. (1977). Determinants of corporate borrowing. Journal of Financial
PT E
Economics, 5(2), 147–175.
Myers, S. C., & Majluf, N. S. (1984). Corporate financing and investment decisions
CE
when firms have information that investors do not have. Journal of Financial
AC
Economics, 13(2), 187–221. Nasdaq. (2017). What is the difference between a developed, emerging, and frontier market? Retrieved from http://www.nasdaq.com/article/what-is-the-difference-between-a-developed-emer ging-and-frontier-market-cm140649 Opler, T., Pinkowitz, L., Stulz, R., & Williamson, R. (1999). The determinants and 37
ACCEPTED MANUSCRIPT
implications of corporate cash holdings. Journal of Financial Economics, 52(1), 3–46. Organization for Economic Cooperation and Development. (2017). Chinese economic
PT
performance in the long run: Institutional differences between Europe and China.
RI
Retrieved from http://www.oecd.org/dev/
SC
chineseeconomicperformanceinthelongruninstitutionaldifferencesbetweeneurope andchina.htm
NU
Schiantarelli, F., & Sembenelli, A. (2000). Form of ownership and financial
D
Empirica, 27(2), 175–192.
MA
constraints: Panel data evidence from flow of funds and investment equations.
PT E
Shabbir, M., Hashmi, S. H., & Chaudhary, G. M. (2016). Determinants of corporate cash holdings in Pakistan. International Journal of Organizational Leadership,
CE
5(1), 50–62.
AC
State Statistical Bureau. (2002–2007). China's big business group yearbook. Beijing: China Tongji Press (in Chinese). Tahir, M. S., Alifiah, M. N., Arshad, M. U., & Saleem, F. (2016). Financial theories with a focus on corporate cash holding behavior: A comprehensive review. International Journal of Economics and Financial Issues, 6(3S), 215–219. Uyar, A., & Kuzey, C. (2014). Determinants of corporate cash holdings: Evidence 38
ACCEPTED MANUSCRIPT
from the emerging market of Turkey. Applied Economics, 46(7–9), 1035–1048. World Bank (WB). (2016a). China's GDP growth annual percentage. Retrieved from https://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG?locations=CN
RI
https://data.worldbank.org/indicator/SP.POP.TOTL
PT
World Bank (WB). (2016b). Population of China. Retrieved from
SC
Yang, J., Lian, J., & Liu, X. (2012). Political connections, bank loans and firm value. Nankai Business Review International, 3(4), 376–397.
NU
Zhang, H., Lu, Z., Zhang, R., & Jiang, G. (2015). Insider ownership, subsidiary cash
MA
holdings, and economic consequences: Evidence from listed Chinese companies.
AC
CE
PT E
D
Emerging Markets Finance and Trade, 51(Suppl. 1), S174–S195.
39
Tsui-Jung Lin, Hai-Yen Chang, Hui-Fun Yu, Ching-Pao Kao PII: DOI: Reference:
S1044-0283(18)30084-X doi:10.1016/j.gfj.2018.10.001 GLOFIN 451
To appear in:
Global Finance Journal
Received date: Revised date: Accepted date:
28 March 2018 1 October 2018 1 October 2018
Please cite this article as: Tsui-Jung Lin, Hai-Yen Chang, Hui-Fun Yu, Ching-Pao Kao , The impact of political connections and business groups on cash holdings: Evidence from Chinese listed firms. Glofin (2018), doi:10.1016/j.gfj.2018.10.001
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Title: The impact of political connections and business groups on cash holdings:
PT
Evidence from Chinese listed firms
Authors:
NU
SC
RI
Tsui-Jung Lin Professor, Department of Banking and Finance, Chinese Culture University, Taiwan E-mail: [email protected] No.55, Hwa Kang Rd., Yang Ming Shan, Taipei 11114, Taiwan
MA
Hai-Yen Chang Associate Professor, Department of Banking and Finance, Chinese Culture University, Taiwan E-mail: [email protected]
D
No.55, Hwa Kang Rd., Yang Ming Shan, Taipei 11114, Taiwan
CE
PT E
Hui-Fun Yu* Corresponding Author and Associate Professor, Department of Banking and Finance, Chinese Culture University, Yang Ming Shan, Taipei 11114, Taiwan E-mail: [email protected]; [email protected] No.55, Hwa Kang Rd., Yang Ming Shan, Taipei 11114, Taiwan
AC
Ching-Pao Kao
Department of Banking and Finance, Chinese Culture University, Taiwan E-mail: [email protected] No.55, Hwa Kang Rd., Yang Ming Shan, Taipei 11114, Taiwan
1
ACCEPTED MANUSCRIPT
ABSTRACT
We investigated the impact of political connections and business group affiliation on
PT
the cash holdings of firms listed on two main Chinese stock exchanges. We applied
RI
panel data regression analysis on a dataset comprising 10,832 observations for these
SC
companies from 2008 to 2015. We found political connections to be positively correlated with cash holdings, while business group affiliation is negatively correlated.
NU
Firms with both political connections and business group affiliations hold more cash.
JEL classifications:
CE
G3
PT E
D
motives of corporate managers.
MA
These findings are consistent with the tenets of agency theory and the speculative
AC
G30 Keywords:
Political connections Cash holdings Business groups
2
ACCEPTED MANUSCRIPT
1. Introduction
Cash holdings play an important role in emerging markets because they affect
PT
companies’ ability to raise capital for business operations and their investment
RI
decisions (Francis, Hasan, Song, & Waisman, 2013). Although firms in both developed
SC
and emerging markets need to make decisions about the appropriate level of cash holdings, companies in emerging markets have greater concerns regarding alternative,
NU
external sources of funding because of lower liquidity and less access to bank loans and
MA
financial markets for their bonds and common equity (Nasdaq, 2017).
D
Keynes (1936) proposed three main motives for holding cash: transactional (to
PT E
cover current transactions); precautionary (to deal with unforeseen future events); and speculative (to invest). Three theories have been developed to explain corporate cash
CE
holdings and capital structure. According to Myers’s (1977) trade-off theory, firms
AC
incur opportunity costs when holding excess cash but must pay the cost of debt if they do not have cash on hand. Myers and Majluf (1984), who developed the pecking order theory, explained that when firms see promising investment opportunities, they prefer internal (cash) to external (debt, equity) funding because internal sources of funding require the lowest costs. In the free cash flow theory, which extends agency theory, Jensen (1986) proposed that control over cash empowers corporate managers, who thus 3
ACCEPTED MANUSCRIPT
may hold onto larger-than-optimal levels of cash in conflict with the interests of shareholders or may squander company cash in their own interests. Although these theories have been more or less verified in empirical studies, it
PT
should be remembered that theories advanced in developed markets, such as the United
RI
States and the United Kingdom, may not be applicable to emerging markets (Al-Najjar
SC
& Belghitar, 2011; Brenes, Madrigal, & Requena, 2011; Tahir, Alifiah, Arshad, & Saleem, 2016) because of differences in economic and financial contexts. Emerging
NU
markets do not enjoy established capital markets with the high levels of liquidity,
MA
regulatory requirements, and market capitalization that are common in developed
D
countries (Nasdaq, 2017). The OECD (2017) stresses that institutional differences
PT E
such as the political system, equity market, banks, regulatory agencies, and legal framework affect the financial decisions of firms in developed countries and emerging
CE
markets. Businesses and investors must consider these factors.
AC
In developed countries, firm size is negatively related to cash holdings because larger companies have greater access to funding than smaller ones, and therefore less need to hold onto cash (Akben-Selcuk & Altiok-Yilmaz, 2016; Hill, Fuller, Kelly, & Washam, 2013). For example, Cai, Zeng, Lee, and Ozkan (2016) found that group-affiliated firms tend to hold significantly less cash than their unaffiliated counterparts. In emerging markets, firm size is positively related to cash because firms 4
ACCEPTED MANUSCRIPT
are more diversified and a parent company may have several subsidiaries (Al-Najjar & Belghitar, 2011; Shabbir, Hashmi, & Chaudhary, 2016). Further, business groups, which are common in many emerging markets, serve as a nonmarket mechanism used
PT
by firms to overcome a weak institutional context (Khanna & Palepu, 2000b). La
RI
Porta, Lopez-de-Silanes, and Shleifer (1999) also found pyramidal structures in the
economies with weak shareholder protection.
SC
control of firms and participation in management by families or the state, especially in
NU
Studies have also offered differing opinions regarding the connections between
MA
political connections and cash holdings. In emerging markets, political connections
D
can help firms secure bank loans or become publicly listed (Bao, Johan, & Kutsuna,
PT E
2016; Yang, Lian, & Liu, 2012). Some have claimed that political connections reduce the need to hold onto cash (Claessens, Feijen, & Laeven, 2008; Faccio, 2006, 2010;
CE
Hill et al., 2013). Political connections are particularly common in China, the world’s
AC
largest emerging market country, and one in which the Communist Party dominates. Most firms in China are owned by business groups, and the majority of these are politically connected to either the central or local government (Khanna & Palepu, 2000a; Yang et al., 2012). Our study investigates the effects of an internal force (the corporate structure of the business group) and an external force (political connections) on the cash holdings of Chinese firms. 5
ACCEPTED MANUSCRIPT
This study makes a threefold contribution to the literature. First, it extends our understanding of the impact of political connections on the level of cash holdings. The uniqueness of China’s capital market (e.g., market inefficiency and firms’ lack of
PT
capital) and its specific network characteristics (e.g., ties with government as a critical
RI
resource) make it a different setting in which to examine this influence. Second, this
SC
paper argues that the business group (BG) may moderate the relation between political connectedness and cash holdings. Finally, we also attempt to compare the
NU
theoretical predictions with empirical evidence. The results of this study may help
D
cash holdings in Chinese firms.
MA
investors understand how political connections and business group structure affect
PT E
We chose China as the sample country for this study because of its high population, rapid economic growth, and large trade surplus, which make it the largest
CE
emerging market. With 1.3 billion people, China accounts for 17.6% of the total 7.4
AC
billion world population (World Bank, 2016b). From 1995 to 2015, it averaged a gross domestic product (GDP) growth of 9.48% (World Bank, 2016a). Moreover, given that China has had the largest trade surplus of any nation in the world since 2000, Chinese firms have had ample opportunity to accumulate a great amount of cash (FRED, 2016). China’s unique economic environment is particularly desirable for examining 6
ACCEPTED MANUSCRIPT
cash holding theories because of the prevalence of political connections and the BG corporate structure (parent companies with subsidiaries), which could affect a firm’s financial decisions. By the end of 2009, 95.3% of all listed companies in China
PT
operated within BGs (Zhang, Lu, Zhang, & Jiang, 2015). Moreover, in terms of
RI
political connectedness, four large state-owned-enterprise (SOE) banks—Bank of
SC
China, Construction Bank of China, Agriculture Bank of China, and International Commercial Bank of China—dominate the country’s financial markets (Allen, Qian,
NU
& Qian, 2005). SOE companies can easily obtain funding from SOE banks, and most
MA
large companies rely on bank loans (Megginson, Ullah, & Wei, 2014). In short, political connection is also associated with ease of financing. Additionally, both firm
PT E
D
size and corporate structure affect firms’ access to loans, and thus their cash holdings (Boubakri, El Ghoul, & Saffar, 2013). To the best of our knowledge, this is the first
CE
study to examine empirical evidence demonstrating the importance of political
AC
connections in determining the ease of acquisition and amount of cash holdings for business groups in China. The remainder of this paper is organized as follows. Section 2 includes the literature review and hypothesis development. The data collection and measurement of variables are discussed in section 3. Section 4 presents the empirical results. Section 5 concludes the paper. 7
ACCEPTED MANUSCRIPT
2. Related literature and hypotheses development 2.1. Political connections and cash holdings
PT
Faccio (2006) found that corporate political connections are widespread around
RI
the world, especially in countries with emerging economies where regulations
SC
regarding political conflicts of interest are relaxed. Since larger firms tend to build higher levels of political connectedness than smaller ones (Agrawal & Knoeber, 2001;
NU
Faccio, 2006; Johnson & Mitton, 2003), large firms in emerging markets often find it
MA
easier to get external capital, such as bank loans (Allen et al., 2005; Bao et al., 2016).
D
Boubakri et al.(2013) and Chaney, Faccio, and Parsley (2011) have suggested that
PT E
agency problems are more likely to occur in politically connected (PC) firms than non–politically connected ones in emerging countries. The influx of funds raises the
CE
level of liquid assets and increases the number of managers who act at the
AC
shareholders’ expense. Further, Caprio, Faccio, and McConnell (2013) proposed that in emerging countries with weaker legal property protection, firms adjust their asset holdings to protect their assets from government seizure. They found a negative relation between corporate holdings of liquid assets and political corruption. However, scholars hold varying opinions about the effect of political connection on firms’ cash holdings. Some researchers have argued that political connection has a 8
ACCEPTED MANUSCRIPT
negative relationship with cash level. Faccio (2010) found that PC companies in developed countries hold onto less cash because they have political support and relationships with government agencies. In emerging countries, Al-Najjar (2013)
PT
found that companies with significant political support hold less cash since they have
RI
government support in financially critical periods.
SC
Others, however, have contended that political connection has a positive correlation with cash holdings. Kusnadi and Wei (2012) found that managers of PC
NU
firms in emerging markets, as well as firms in countries with weak legal protection of
MA
investors and prevalent corruption, are inclined to hoard more cash than nonpolitical
D
firms. Given China’s inefficient legal environment and unique business practices, we
PT E
inferred that politically-connected firms suffer severe agency problems and managers hold more cash to act self-interest. Therefore, we propose the first hypothesis as
AC
CE
follows:
Hypothesis 1. In China, political connection is positively correlated with corporate cash holdings.
2.2. Business groups and cash holdings Khanna and Palepu (2000a) proposed that in emerging markets, companies form 9
ACCEPTED MANUSCRIPT
business groups as a defense against inadequate investor protection, legal enforcement, and information disclosure. In effect, the BG serves as an internal financial market in which funds can move from parent company to subsidiaries or among the affiliated
PT
firms for operational or investment purposes (Bena & Ortiz-Molina, 2013). Business
RI
groups may thus reduce their members’ need for cash on hand (Schiantarelli &
SC
Sembenelli, 2000).
The rapid growth and rise in importance of China as an emerging market
NU
economy provides a natural platform for analyzing issues related to the role of the
MA
business group (BG). To become a business group in China, the parent company must
D
provide capital of over 50 million yuan (US$6 million) and must have at least five
PT E
affiliated companies. Moreover, the total capital of the parent and its affiliated companies must exceed 100 million yuan (US$12 million). He, Mao, Rui, and Zha
CE
(2013) noted rapid growth in the total number of business groups over the past few
AC
decades. According to the State Statistical Bureau (2002A–2007A), as of 2006, there were 2,856 officially recognized business groups in China, with 27,950 directly owned first-tier subsidiaries, employing around 30 million people and accounting for almost 60% of national industrial output (He et al., 2013). The ecology of business groups in China is substantially different from those of the so-called Big Six keiretsus in Japan or the top 30 chaebols in South Korea (Chang & Choi, 1988; Chang & Hong, 10
ACCEPTED MANUSCRIPT
2000). The size distribution of the groups and the identity of their ultimate controlling shareholders are different from those in other emerging markets or developed economies. China has a number of state-dominated business groups, established for
PT
purposes of economic development and social distribution, to carry out national goals.
RI
However, because of market failures and the lack of a well-developed legal and
SC
regulatory system in China, family- or natural person-dominated business groups have also been becoming more common.
NU
However, the evidence regarding the effects of BGs on cash holding is
MA
inconsistent. He et al. (2013) found that in China, BG-affiliated firms have a lower
D
risk of incurring cash shortages when faced with investment opportunities because
PT E
they are able to obtain funding within the BG. Furthermore, Cai et al. (2016) found that group-affiliated Chinese firms held significantly less cash than their nonaffiliated
CE
counterparts. This is consistent with both the trade-off theory and the pecking order
AC
theory. Firms with quicker access to funding have the freedom to hold less cash. However, less financially constrained state-owned enterprises (SOEs) are slower to respond with decreased cash holdings than non-SOEs. Studies also show varying assessments of the relationships among firm size, liquidity, and corporate cash holdings. Al-Najjar (2013) asserted that larger firms in emerging countries tend to hold less cash in general, that firms with leverage also 11
ACCEPTED MANUSCRIPT
hold onto less cash, and that cash holdings are lower when firms have easy access to funding. Shabbir et al. (2016) found that, in emerging markets, company size was positively correlated with increased cash holdings, and leverage (bank loans) with
PT
decreased cash holdings. In contrast, Uyar and Kuzey (2014) observed that firms with
RI
higher liquidity and larger firm size held less cash. Given all this evidence, we
SC
propose the following hypothesis:
NU
Hypothesis 2. In China, business group affiliation is negatively correlated with
D
MA
corporate cash holdings.
PT E
2.3. Relationship among political connections, business groups, and cash holdings Studies have offered two explanations as to why business groups seek to
CE
maintain cash holdings. The predominant view is based on the free cash flow
AC
hypothesis: the pyramidal structure of business groups worsens information asymmetry between managers and shareholders and exacerbates the agency problem, giving managers an incentive to control more cash. The other argument claims that business groups can take advantage of internal capital markets to overcome the financial constraints characteristic of underdeveloped financial markets, so managers in such groups hold less cash. The trade-off between these forces may partly depend 12
ACCEPTED MANUSCRIPT
on other institutional factors, for example political connections. He et al. (2013) found that, in China, privately owned business groups are more likely to use internal capital markets than are groups controlled by the central government. Cai et al. (2016) also
PT
found that state ownership moderates the benefit received from such an internal
RI
capital market, and Firth, Lin, and Wong (2008) found that, in China, the distribution
SC
of loans is skewed toward firms with political connections. Chen, Sun, Tang, and Wu (2011) found that political connections significantly reduced investment efficiency in
NU
SOEs but not in non-SOEs. Furthermore, Chen, Ding, and Kim (2010), analyzing
MA
firm-level data from 17 jurisdictions from 1997 to 2001, found that it was harder to
D
predict the earnings of firms with political connections than those of firms without
PT E
such connections. Apparently, having government ties deepens the problems of adverse selection and information asymmetry between managers and shareholders. In
CE
other words, the cost of political connectedness overcomes the benefit of the internal
AC
capital market to the business group, pushing managers to strengthen cash holdings. It should be noted that China has been criticized for its weak investor protection and internal corporate governance structure (Jiang & Kim, 2015). This means that, for firms in business groups with government ties, managers can use their own discretion to hold onto higher levels of cash holdings even to the point of harming the firm’s value. We therefore conclude that firms are more likely to suffer from agency cost 13
ACCEPTED MANUSCRIPT
problems when they are members of a business group holding higher levels of cash and having political connections:
PT
Hypothesis 3. In China, the combination of political connections and business group
SC
RI
affiliation is positively correlated with corporate cash holdings.
3. Research method
NU
We analyzed data on a sample of listed companies in China between 2008 and
MA
2015 to test the relationships among political connection, BG affiliation, and cash
D
holdings.
PT E
3.1. Data sample
We collected data on cash amount, other financial information, and corporate
CE
structure, as well as the names and positions of executives and managers, from the
AC
China Stock Market and Accounting Research (CSMAR) database. We removed companies in the financial sector because financial institutions do not have cash and funding issues, and excluded firms with missing financial information in order to avoid potential statistical errors. The remaining panel contains 9,777 observations.
3.2. Research model 14
ACCEPTED MANUSCRIPT
The key variables in this study are the level of cash holdings, PC, and BG. The data have both longitudinal and cross-sectional properties; therefore, following Opler, Pinkowitz, Stulz, and Williamson (1999), we apply panel data regression analysis to
of 200.794 (significant at 1%),
RI
(Greene, 2008, chapter 9) yielded a value of χ2
PT
measure cash holdings against the two other variables, PC and BG. The Hausman test
SC
ruling out the random effect assumption, so we used the fixed effect assumption in the regression model. Endogeneity is another potential issue: cash is a function of
NU
leverage and vice versa, and cash might also be endogenous with dividend payouts
MA
(DIVD)1. To mitigate this concern, following Harford, Mansi, and Maxwell (2008)
D
and Chen, Leung, and Goergen (2017), we lag some independent variables. Following
PT E
Al-dhamari and Ku Ismail (2015), we measure cash holdings by taking the logarithm of the proportion of cash and cash equivalents to net assets.2 The equation is
AC
CE
formulated as follows3:
1
We thank the reviewer for these suggestions.
2
When the proportion takes the logarithm, the value of logCash may be negative.
3
The correlations between leverage and net working capital were high, so we revised the original model and used leverage in the model. 15
ACCEPTED MANUSCRIPT
𝐿𝑜𝑔𝐶𝑎𝑠ℎ𝑖,𝑡 = 𝛽0 + 𝛽1 𝑃𝐶𝑖,𝑡 + 𝛽2 𝐵𝐺𝑖,𝑡 + 𝛽3 (𝑃𝐶𝑖,𝑡 ∗ 𝐵𝐺𝑖,𝑡 ) + 𝛽4 𝐶𝑎𝑠ℎ𝑖,𝑡−1 + 𝛽5 𝑆𝑖𝑧𝑒𝑖,𝑡−1 + 𝛽6 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑖,𝑡−1 + 𝛽7 𝑀⁄𝐵
𝑖,𝑡
+ 𝛽8 𝐶𝐹𝑖,𝑡
+ 𝛽9 𝐶𝑎𝑝𝑒𝑥𝑖,𝑡−1 + 𝛽10 𝐷𝐼𝑉𝐷𝑖,𝑡−1 + 𝛽11 𝐶𝐹𝑉𝑜𝑙𝑖,𝑡−1 + 𝛿 + 𝜃 (1)
RI
PT
+ 𝜖𝑖,𝑡 ,
SC
where i is the ith company; t -1is the t-1th year; δ indicates firm effects; θ is the
NU
year effects; and ∈ is the error term.
MA
3.3. Definition of variables
D
3.3.1. Cash holdings
PT E
LogCashi,t represents the logarithm of the amount of cash held by company i at the end of year t. The level of cash holdings is calculated as the sum of cash and cash
AC
equivalents).
CE
equivalents divided by net assets (total assets minus the sum of cash and cash
𝐿𝑜𝑔𝐶𝑎𝑠ℎ𝑖,𝑡 = log
(𝐶𝑎𝑠ℎ + Cash equivalents ) , 𝑛𝑒𝑡 𝑎𝑠𝑠𝑒𝑡𝑠
(2)
where Cash = cash on hand; Cash equivalents = assets quickly convertible into a known cash amount; Net assets = total assets – (cash + cash equivalents). 16
ACCEPTED MANUSCRIPT
3.3.2. Political connection Faccio (2006) defined a company as politically connected if “at least one of its
PT
large shareholders (anyone controlling at least 10 percent of voting shares) or one of
RI
its top officers (CEO, president, vice-president, chairman, or secretary) is a member
SC
of parliament, a minister or the head of state, or closely related to a top politician” (p. 370). Previous studies have used a dummy variable to capture political connections.
NU
However, simply using a dummy variable does not allow consideration of whether a
MA
firm has one or many connections, nor does it capture the nature of specific connections. In our case, it is important to know whether members of the board are
PT E
D
members of the Chinese People’s Congress (CPC) or the Chinese People’s Political Consultative Conference (CPPCC). We calculate political connection as the sum of
CE
the number of board members who are members of the CPC or the CPPCC divided by
AC
the board size.
3.3.3. Business group He et al. (2013) derived their definition of a BG from an annual survey conducted by the Chinese Securities Regulatory Commission (CSRC) covering the period from 1990 to 2006. The authors regarded a Chinese company as a BG if the firm’s ultimate 17
ACCEPTED MANUSCRIPT
controlling entity possessed more than one firm in a particular year. We adopt the same definition in our study, and use a dummy variable (BG) that is equal to 1 if the
PT
firm is a BG, and 0 if not.
RI
3.3.4. Control variables
SC
For other variables that influence the level of cash holdings we follow Opler et al. (1999), defining them as follows:
NU
Cash i, t-1 = (cash + cash equivalents) / net assets in the previous term;
MA
Size i, t-1 = natural logarithm of net assets in the previous term;
D
Leverage i, t-1 = (current liabilities + long-term liabilities) / net assets in the previous
PT E
term;
M/B (market-to-book ratio) i, t = market value of a firm / book value of a firm;
CE
CF (cash flow) i, t = earnings before interest, taxes, depreciation and amortization /
AC
net assets;
Capex (capital expenditure) i, t-1 = capital expenditure / net assets in the previous term; DIVD (dividend payout) i, t-1 = a dummy variable that is equal to 1 if a firm pays out cash dividends in the previous term, and 0 otherwise; CFVol i, t-1 = standard deviation of the cash flow in the previous term.
18
ACCEPTED MANUSCRIPT
4.
Results We first tested the relationship between PC and cash holdings and the correlation
between BG and cash holdings. Then we examined the effect of both PC and BG on
RI
PT
cash holdings, and compared the three outcomes.
SC
4.1. Descriptive statistics
The descriptive statistics for the variables (see Table 1) indicated significant
NU
differences between the maximum and minimum amounts. Table 2 shows the Pearson
MA
correlation coefficients (ρ) among the variables, which are all between 0.7 and -0.7
D
(-0.7＜ρ＜+0.7), indicating that the regression model does not suffer from
PT E
multicollinearity. In short, the low correlations among variables reflect the reliability of the regression model.
CE
We divided the firms into two groups, those in business groups (BG) and those
AC
not in business groups (NBG), and compared the means and medians of the variables between the two groups. The results are shown in Table 3. The means of cash holdings of the BG firms and NBG firms are -0.387 and -0.117, respectively, and the difference is significant at the 1% level. Similarly, the medians of cash holdings of the BG and NBG firms are -0.422 and -0.172, respectively, and this difference is also significant at the 1% level. In sum, BG firms hold significantly less cash than NBG 19
ACCEPTED MANUSCRIPT
firms.
Table1 Descriptive statistics for the variables. Observation Mean s
Media n
Maximum
Minimum
Std. Dev.
9777
-0.389
-0.379
1.676
-2.659
0.450
PC BG Casht-1 Sizet-1 Leveraget-1 M/B CF DIVDt-1 Capext-1
9777 9777 9777 9777 9777 9777 9777 9777 9777
0.047 0.851 0.006 6.352 0.569 2.310 0.086 0.561 0.029
0.000 1.000 -0.017 6.329 0.552 1.596 0.070 0.091 0.000
0.250 1.000 2.225 8.883 21.112 101.461 4.677 371.317 1.000
0.000 0.000 -1.733 3.613 0.017 0.113 -2.705 -0.999 0.000
0.058 0.356 0.337 0.588 0.413 3.030 0.126 8.086 0.167
CFVolt-1
9777
0.045
0.016
18.617
0.000
0.337
SC
NU
D
MA
RI
logCash
PT
Variable
PT E
Notes: Here, “logCash” means cash holdings measured by taking the natural logarithm of the sum of cash and cash equivalents divided by net assets. “PC”
CE
represents political connection, and is calculated as the sum of the number of members of the board with connections to the CPC or the CPPCC divided by board
AC
size. The total number of CPC/CPPCC-connected members ranges from 0 to 3. “BG” means business group; the dummy is equal to 1 if the firm is affiliated with a business group, and 0 if not. “Cash t-1” means cash + cash equivalents/net assets (total assets – cash –cash equivalents) in the previous term; “Size t-1” means firm size measured by taking the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as the sum of current liabilities and long-term liabilities, divided by net assets in the previous term. “M/B” means market-to-book ratio 20
ACCEPTED MANUSCRIPT measured as the firm’s market value divided by its book value. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by net assets. “Capex t-1” represents capital expenditure measured as capital expenditure divided by net assets in the previous term. “DIVD t-1” means dividend payout; the dummy is equal to1 if the firm has paid out cash dividends in the previous term, and 0
PT
if not. “CFVolt-1” means the standard deviation of cash flow in the previous term. ***
RI
represents the 1% significance level, ** represents the 5% significance level, and *
AC
CE
PT E
D
MA
NU
SC
represents the 10% significance level.
21
ACCEPTED MANUSCRIPT
Table 2 Correlations.
Variable
logCash
PCt-1
BGt-1
Casht-1
Sizet-1
Leveraget- M/B
CF
1.000 .111** -.231**
1.000 -.101**
1.000
Casht-1
.127**
-.029**
.075**
1.000
**
**
**
Sizet-1 -.423 Leveraget -.141**
I R
DIVDt-1 Capext-1 CFVolt-
SC
1
logCash PC BG
T P
1
U N
.290 .215**
-.001 .086**
-.238** -.131** -.166** -.057**
.020* .111** -.133** .046**
-.650** -.366** -.041** -.130**
-.425** -.246** -.090** .007
1.000 .401** .035** .051**
1.000 .133** .161**
1.000 .074**
1.000
-.082**
-.059**
-.287**
-.120**
.232**
.140**
.060**
.034**
.358** .393** .138** .098**
.048** .040** .086** .042**
CFVolt-1
.104**
.013
E C
C A
D E
PT
-1
M/B CF DIVDt-1 Capext-1
1.000 .319**
A M
-.050 -.059**
1.000
1.000
Notes: “logCash” means cash holdings measured by taking the natural logarithm of the sum of cash and cash equivalents divided by net assets. “PC” represents political connection, calculated as the sum of the number of board members with connections to the CPC or the CPPCC, divided
22
ACCEPTED MANUSCRIPT
by board size. “BG” means business group; the dummy is equal to 1 if the firm is affiliated with a business group, and 0 if not. “Cash t-1” means cash + cash equivalents/ net assets (total assets – cash –cash equivalents) in the previous term; “Size t-1” means size of firm measured by taking
T P
the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as the sum of current liabilities and
I R
long-term liabilities, divided by net assets in the previous term. “M/B” means market-to-book ratio measured as the firm’s market value divided
C S U
by its book value. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by net assets. “Capex t-1” represents capital expenditure measured as capital expenditure divided by net assets in the previous term. “DIVD t-1” means dividend
N A
payout; the dummy is equal to1 if the firm has paid out cash dividends in the previous term, and 0 if not. “CFVolt-1” means the standard deviation
M
of cash flow in the previous term. *** represents the 1% significance level, ** represents the 5% significance level, and * represents the 10%
D E
significance level.
T P E
C C
A
23
ACCEPTED MANUSCRIPT Table 3 Test results for means and medians of variables of business-group- and non-business-group-affiliated firms. Panel A: Mean
Panel B: Median
Variable
All samples
Variable
All samples
difference
difference
NBG
BG
logCash
-0.387
-0.117
-0.270***
-0.422
PC Casht-1 Sizet-1 Leveraget-1 M/B CF DIVDt-1 Capext-1 CFVolt-1
0.044
0.060
-0.017***
0.000
0.005
-0.027
0.032***
6.362
5.947
0.414***
0.579
0.433
0.146***
2.195
3.254
-1.058***
0.090
0.119
0.531
NBG
-0.172
-0.250***
0.074
-0.074***
-0.010
0.000
-0.010***
6.383
-0.071
6.454***
0.573
5.960
-5.386***
1.480
0.409
1.072***
-0.029***
0.067
2.296
-2.228***
0.630
-0.099
0.074
0.000
0.074***
0.029
0.055
-0.026***
0.000
0.088
-0.088***
0.047
0.045
0.002
0.015
0.200
-0.185***
MA
NU
SC
RI
PT
BG
D
Notes: “logCash” means cash holdings measured by taking the natural
PT E
logarithm of the sum of cash and cash equivalents divided by net assets. “PC” represents political connection, calculated as the sum of the number of
CE
members of the board with connections to the CPC or the CPPCC, divided by
AC
board size in the previous term. “BG” means business group; the dummy is equal to 1 if the firm is affiliated with a business group, and 0 if not, in the previous term. “Cash t-1” means cash + cash equivalents/net assets (total assets – cash –cash equivalents) in the previous term; “Size t-1” means firm size measured by taking the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as the sum of current liabilities and long-term liabilities divided by net assets in the previous term.
24
ACCEPTED MANUSCRIPT “M/B” means market-to-book ratio measured as the firm’s market value divided by its book value. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by net assets. “Capex t-1”
represents capital expenditure measured as capital expenditure divided by
net assets in the previous term. “DIVD t-1” means dividend payout; the
PT
dummy is equal to1 if the firm has paid out cash dividends in the previous
RI
term, and 0 if not. “CFVolt-1” means the standard deviation of cash flow in the
SC
previous term. *** represents the 1% significance level, ** represents the 5%
NU
significance level, and * represents the 10% significance level.
MA
4.2. Multivariate analysis
4.2.1. The relationship between political connections and cash holdings
PT E
D
Table 4 shows the results of our panel data analysis. In model 1, the coefficient of political connections (PC) is 0.759 and is significant at the 1% level, showing that
CE
companies whose board members held official positions in the government
AC
maintained a higher level of cash. The results are consistent with Kusnadi and Wei’s (2012) finding that in emerging markets, managers of PC firms are inclined to hoard more cash than managers of nonpolitical firms, and support our hypothesis 1, that political connection is positively correlated with corporate cash holdings in China.
Table 4 25
ACCEPTED MANUSCRIPT The effect of political connectedness and business group affiliation on cash holdings.
C PC
Model 1
Model 2
Model 3
1.137*** (25.389) 0.759*** (11.744)
1.705*** (39.419)
1.411*** (26.972) 0.478*** (3.031)
-0.123*** (-14.861)
-0.157*** (-10.442) 0.292*
PT
BG
Leveraget-1 M/B
0.755*** (24.813) DIVDt-1 0.187*** (8.191) Capext-1 -0.001 (-1.419) CFVolt-1 -0.018 (-1.470) Year effect Yes Firm effect Yes Ajusted R2 0.262 F-statistic 262.374*** 11012 Obs.
SC
AC
CE
PT E
D
CF
0.261*** (27.217) -0.318*** (-46.292) 0.015* (1.731) -0.010*** (-6.998)
NU
Sizet-1
0.281*** (24.421) -0.253*** (-37.341) -0.018** (-2.177) -0.007*** (-4.831)
MA
Casht-1
(1.681) 0.281*** (23.916)
RI
PC* BG
0.276*** (-34.464) 0.010 (1.075) 0.010*** (-6.243)
0.825*** (31.268) 0.125*** (6.976) -0.001* (-1.569) -0.047*** (-4.004) Yes Yes
0.744*** (23.454) 0.157*** (6.790) -0.001* (-1.646) -0.029** (-2.389) Yes Yes
0.341 470.132*** 13574
0.301 249.186*** 9777
Notes: “logCash” means cash holdings measured by taking the natural logarithm of the sum of cash and cash equivalents divided by net assets. “PC” represents political connection, calculated as the sum of the number of board members with connections to the CPC or the CPPCC, divided by board size. “BG” means business group; the 26
ACCEPTED MANUSCRIPT dummy is equal to 1 if the firm is affiliated with a business group, and 0 if not. “Cash t-1”
means cash + cash equivalents/net assets (total assets – cash –cash equivalents) in
the previous term; “Size t-1” means size of firm measured by taking the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as the sum of current liabilities and long-term liabilities, divided by net
PT
assets in the previous term. “M/B” means market-to-book ratio measured as the firm’s
RI
market value divided by its book value in the previous term. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by
SC
net assets. “Capex t-1” represents capital expenditure measured as capital expenditure
NU
divided by net assets in the previous term. “DIVD t-1” means dividend payout; the dummy is equal to1 if the firm paid out cash dividends in the previous term, and 0 if
MA
not. “CFVolt-1” means standard deviation of cash flow in the previous term. *** represents the 1% significance level, ** represents the 5% significance level, and *
PT E
D
represents the 10% significance level.
CE
4.2.2. The relationship between business group affiliation and cash holdings
AC
Model 2 in Table 4 shows the effect of BG on cash holdings. The coefficient of business group (BG) is -0.123 and is significant at the 1% level, indicating that business group affiliation is negatively correlated with cash holdings, as our hypothesis 2 predicts.
4.2.3. The relationship among political connection, business group affiliation, and
27
ACCEPTED MANUSCRIPT
cash holdings In Table 4, model 3, the coefficient of the interaction term PC*BG is 0.292 and is significant at the 10% level, suggesting that cash holdings are higher for those firms
PT
with ties to both government and a business group. As we explain in section 1, such
RI
ties are important ways to get outside resources, because most Chinese listed
SC
companies are not able to obtain capital from the market. However, as we also explain above, agency problems are more serious in China than elsewhere, so that affiliation
NU
with a business group does not significantly reduce cash holding by politically
D
MA
connected firms.
PT E
4.3. Robustness tests
To check the robustness of our model, we take into consideration another
CE
definition of political connection. Fan (2016) found that the presence of independent
AC
directors with political background may bring extra profit to a firm. We therefore redefine a company as politically connected if its independent directors have a political background, as measured by the sum of the number of connected independent directors divided by board size, and retest the model. The results are presented in Table 5. All three previous results—concerning the positive relationship between political connections and cash holdings, the negative relationship between 28
ACCEPTED MANUSCRIPT
BG and cash holdings, and the positive effect of the combination of political connections and group business affiliation on cash holdings—are found to be robust.
Table 5
1.371*** (26.835) 0.025 (0.965)
1.797*** (42.024)
0.289***
1.548*** (28.348) -0.066 (-0.961) -0.076*** (-5.106) 0.131* (1.793) 0.304***
(58.948) -0.332*** (-50.303) 0.018*** (2.870) -0.009*** (-10.551) 0.378*** (21.338)
(50.071) -0.296*** (-36.223) 0.014* (1.953) -0.008*** (-7.381) 0.337*** (16.253)
0.052*** (5.229) 0.000 (0.322 ) -0.006 (-0.861) Yes Yes 0.406 1033.769***
0.053*** (4.104) 0.000*** (0.328) -0.004 (-0.477) Yes Yes 0.361 503.702***
-0.058*** (-5.899)
NU
BG
MA
PC* BG
Leveraget-1 M/B CF
0.058*** (4.565) Capext-1 0.000 (0.517) CFVolt-1 0.026*** (3.659) Year effect Yes Firm effect Yes Ajusted R2 0.337 F-statistic 625.014***
AC
DIVDt-1
PT E
Sizet-1
(51.522) -0.277*** (-35.745) -0.020*** (-3.370) -0.005*** (-6.004) 0.364*** (18.782)
D
0.300***
CE
Casht-1
Model 3
RI
PC
Model 2
SC
C
Model 1
PT
Results of robustness tests.
29
ACCEPTED MANUSCRIPT Obs.
11012
13574
9777
Notes: “logCash” means cash holdings measured by taking the natural logarithm of the sum of cash and cash equivalents divided by net assets. “PC” represents political connection, calculated as the sum of the number of independent directors with political connections, divided by board size. “BG” means business group; the dummy
PT
is equal to 1 if the firm is affiliated with a business group, and 0 if not. “Cash t-1”
RI
means cash + cash equivalents/net assets (total assets – cash –cash equivalents) in the
SC
previous term; “Size t-1” means size of firm measured by taking the natural logarithm of net assets in the previous term. “Leverage t-1” means borrowed funds, measured as
NU
the sum of current liabilities and long-term liabilities, divided by net assets in the previous term. “M/B” means market-to-book ratio measured as the firm’s market
MA
value divided by its book value. “CF” means cash flow measured as earnings before interest, taxes, depreciation, and amortization, divided by net assets. “Capex t-1”
D
represents capital expenditure measured as capital expenditure divided by net assets in
PT E
the previous term. “DIVD t-1” means dividend payout; the dummy is equal to1 if the firm paid out cash dividends in the previous term, and 0 if not. “CFVolt-1” means the
CE
standard deviation of cash flow in the previous term. *** represents the 1% significance level, ** represents the 5% significance level, and * represents the 10%
AC
significance level.
5.
Discussion and conclusions Our panel data regression analysis found evidence that Chinese firms with
business group affiliations hold less cash than firms without. This outcome is consistent with internal capital market views. We deduce that BG-affiliated firms are 30
ACCEPTED MANUSCRIPT
able to get funding from internal sources, such as parent companies or other affiliated firms, rather than external sources, such as banks, so they have less need to hold onto excess cash. Further, Chinese firms with political connections hold more cash than
PT
firms without. This contradicts trade-off theory and pecking order theory, but aligns
RI
with agency theory (Jensen, 1986; Myers, 1977; Myers & Majluf, 1984). Politically
SC
connected firms have easier access to bank loans, but, as agency theory predicts, they are likely to hold onto more cash because corporate managers associate control of
NU
liquid assets with power (Jensen, 1986). We infer that managerial self-interest
MA
dominates precautionary and speculative motives in decisions regarding levels of cash
D
holdings. In other words, a cash cushion may tempt managers to waste company
PT E
money to further their own interests (e.g., buy a jet for their personal use or go on luxurious business trips).
CE
These findings suggest that PC and BG firms in emerging markets, particularly
AC
in China, may adopt different financial practices from their counterparts in developed countries such as the United States and Great Britain. There is no optimal cash level in these firms. We caution investors and financial analysts that when measuring the performance of politically connected and business-group-affiliated Chinese firms, they must understand that it is common for these firms to keep a high level of cash that may earn no interest. We have looked only at the level of cash holdings. It 31
ACCEPTED MANUSCRIPT
remains unclear how corporate managers make decisions regarding when and how excess cash should be used. It would be interesting for future researchers to identify the determinants of using excess cash in politically connected and non–politically
PT
connected firms in China.
RI
Acknowledgements
SC
We would like to acknowledge the financial support provided by the Minister of Science and Technology (MOST 106-2410-H-034-013).
NU
References
MA
Agrawal, A., & Knoeber, C. (2001). Do some outside directors play a political role?
D
Journal of Law and Economics, 44(1), 179–198.
PT E
Akben-Selcuk, E., & Altiok-Yilmaz, A. (2016). Determinants of corporate cash holdings: Firm level evidence from emerging markets. In: Ü. Hacioğlu, H.
CE
Dinçer, & N. Alayoğlu. (Eds.), Global business strategies in crisis:
AC
Contributions to management science (pp. 417–428). New York: Springer. Al-dhamari, R., & Ku Ismail, K. N. I. (2015). Cash holdings, political connections, and earnings quality: Some evidence from Malaysia. International Journal of Management Finance, 11(2), 215–231. Allen, F., Qian, J., & Qian, M. (2005). Law, finance, and economic growth in China. Journal of Financial Economics, 77(1), 57–116. 32
ACCEPTED MANUSCRIPT
Al-Najjar, B. (2013). The financial determinants of corporate cash holdings: Evidence from some emerging markets. International Business Review, 22(1), 77–88. Al-Najjar, B., & Belghitar, Y. (2011). Corporate cash holdings and dividend payments:
PT
Evidence from simultaneous analysis. Managerial and Decision Economics,
RI
32(4), 231–241.
SC
Bao, X., Johan, S., & Kutsuna, K. (2016). Do political connections matter in accessing capital markets? Evidence from China. Emerging Markets Review, 29,
NU
24–41.
MA
Bena, J., & Ortiz-Molina, H. (2013). Pyramidal ownership and the creation of new
D
firms. Journal of Financial Economics, 108(3), 798–821.
PT E
Boubakri, N., El Ghoul, S., & Saffar, W. (2013). Cash holdings of politically connected firms. Journal of Multinational Financial Management, 23(4), 338–
CE
355.
AC
Brenes, E. R., Madrigal, K., & Requena, B. (2011). Corporate governance and family business performance. Journal of Business Research, 64(3), 280–285. Cai, W., Zeng, C., Lee, E., & Ozkan, N. (2016). Do business groups affect corporate cash holdings? Evidence from a transition economy. China Journal of Accounting Research, 9(1), 1–24. Caprio, L., Faccio, M., & McConnell, J. J. (2013). Sheltering corporate assets from 33
ACCEPTED MANUSCRIPT
political extraction. Journal of Law and Economic Organization, 29(2), 332–354. Chaney, P. K., Faccio, M., & Parsley, D. (2011). The quality of accounting information in politically connected firms. Journal of Accounting and Economics,
PT
51(1), 58–76.
RI
Chang, S. J., & Choi, U. (1988). Strategy, structure and performance of Korean
SC
business groups: A transactions cost approach. Journal of Industrial Economics, 37, 141–158.
NU
Chang, S. J., & Hong, J. (2000). Economic performance of group-affiliated companies
MA
in Korea: Intragroup resource sharing and internal business transactions.
D
Academy of Management Journal, 43, 429–448.
PT E
Chen, C. J., Ding, Y., & Kim, C. F. (2010). High-level politically connected firms, corruption, and analyst forecast accuracy around the world. Journal of
CE
International Business Studies, 41(9), 1505–1524.
AC
Chen, J., Leung, W. S., & Goergen, W. (2017). The impact of board gender composition on dividend payouts. Journal of Corporate Finance, 43, 86–105. Chen, S., Sun, Z., Tang, S., & Wu, D. (2011). Government intervention and investment efficiency: Evidence from China. Journal of Corporate Finance, 17(2), 259–271. Claessens, S., Feijen, E., & Laeven, L. (2008). Political connections and preferential 34
ACCEPTED MANUSCRIPT
access to finance: The role of campaign contributions. Journal of Financial Economics, 88(3), 554–580. Faccio, M. (2006). Politically connected firms. American Economic Review, 96(1),
PT
369–386.
RI
Faccio, M. (2010). Differences between politically connected and nonconnected firms:
SC
A cross-country analysis. Financial Management, 39(3), 905–928. Fan, J. (2016). The value of political connections in China: Government officials on
NU
the board of directors. Retrieved from papers.ssrn.com.
MA
Federal Reserve Bank of St. Louis (FRED). (2016). China's trade surplus since 2000.
D
Retrieved from
PT E
https://fredblog.stlouisfed.org/2016/05/chinas-trade-surplus-since-2000/ Firth, M., Lin, C., & Wong, S. M. L. (2008). Leverage and investment under a
CE
state-owned bank lending environment: Evidence from China. Journal of
AC
Corporate Finance, 14, 642–653. Francis, B., Hasan, I., Song, L., & Waisman, M. (2013). Corporate governance and investment–cash flow sensitivity: Evidence from emerging markets. Emerging Markets Review, 15, 57–71. Greene, W. H. (2008). Econometric analysis (6th ed.). Upper Saddle River, NJ: Prentice Hall. 35
ACCEPTED MANUSCRIPT
Harford, J., Mansi, S. A., & Maxwell, W. F. (2008). Corporate governance and firm cash holdings in the U.S. Journal of Financial Economics, 87, 535–555. He, J., Mao, X., Rui, O. M., & Zha, X. (2013). Business groups in China. Journal of
PT
Corporate Finance, 22, 166–192.
RI
Hill, M. D., Fuller, K. P., Kelly, G. W., & Washam, J. O. (2013). Corporate cash
SC
holdings and political connections. Review of Quantitative Finance and Accounting, 42(1), 123–142.
NU
Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance, and
MA
takeovers. American Economic Review, 76(2), 323–329.
D
Jiang, F., & Kim, K. A. (2015). Corporate governance in China; A modern perspective.
PT E
Journal of Corporate Finance, 32, 190–216. Johnson, S., & Mitton, T. (2003). Cronyism and capital controls: Evidence from
CE
Malaysia. Journal of Financial Economics, 67(2), 351–382.
AC
Keynes, J. M. (1936). The general theory of employment, interest and money. London: Macmillan and Co. Khanna, T., & Palepu, K. (2000a). Is group affiliation profitable in emerging markets? An analysis of diversified Indian business groups. Journal of Finance, 55(2), 867–891. Khanna, T., & Palepu, K. (2000b). The future of business groups in emerging markets: 36
ACCEPTED MANUSCRIPT
Long-run evidence from Chile. Academy of Management Journal, 43, 268–285. Kusnadi, Y., & Wei, K. C. J. (2012). The effects of political connections on the level and value of cash holdings: International evidence. Working paper, Singapore
PT
Management University.
RI
La Porta, R., Lopez‐De‐Silanes, F., & Shleifer, A. (1999). Corporate ownership
SC
around the world. Journal of Finance, 54(2), 471–517.
Megginson, W. L., Ullah, B., & Wei, Z. (2014). State ownership, soft-budget
NU
constraints, and cash holdings: Evidence from China's privatized firms. Journal
MA
of Banking and Finance, 48, 276–291.
D
Myers, S. C. (1977). Determinants of corporate borrowing. Journal of Financial
PT E
Economics, 5(2), 147–175.
Myers, S. C., & Majluf, N. S. (1984). Corporate financing and investment decisions
CE
when firms have information that investors do not have. Journal of Financial
AC
Economics, 13(2), 187–221. Nasdaq. (2017). What is the difference between a developed, emerging, and frontier market? Retrieved from http://www.nasdaq.com/article/what-is-the-difference-between-a-developed-emer ging-and-frontier-market-cm140649 Opler, T., Pinkowitz, L., Stulz, R., & Williamson, R. (1999). The determinants and 37
ACCEPTED MANUSCRIPT
implications of corporate cash holdings. Journal of Financial Economics, 52(1), 3–46. Organization for Economic Cooperation and Development. (2017). Chinese economic
PT
performance in the long run: Institutional differences between Europe and China.
RI
Retrieved from http://www.oecd.org/dev/
SC
chineseeconomicperformanceinthelongruninstitutionaldifferencesbetweeneurope andchina.htm
NU
Schiantarelli, F., & Sembenelli, A. (2000). Form of ownership and financial
D
Empirica, 27(2), 175–192.
MA
constraints: Panel data evidence from flow of funds and investment equations.
PT E
Shabbir, M., Hashmi, S. H., & Chaudhary, G. M. (2016). Determinants of corporate cash holdings in Pakistan. International Journal of Organizational Leadership,
CE
5(1), 50–62.
AC
State Statistical Bureau. (2002–2007). China's big business group yearbook. Beijing: China Tongji Press (in Chinese). Tahir, M. S., Alifiah, M. N., Arshad, M. U., & Saleem, F. (2016). Financial theories with a focus on corporate cash holding behavior: A comprehensive review. International Journal of Economics and Financial Issues, 6(3S), 215–219. Uyar, A., & Kuzey, C. (2014). Determinants of corporate cash holdings: Evidence 38
ACCEPTED MANUSCRIPT
from the emerging market of Turkey. Applied Economics, 46(7–9), 1035–1048. World Bank (WB). (2016a). China's GDP growth annual percentage. Retrieved from https://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG?locations=CN
RI
https://data.worldbank.org/indicator/SP.POP.TOTL
PT
World Bank (WB). (2016b). Population of China. Retrieved from
SC
Yang, J., Lian, J., & Liu, X. (2012). Political connections, bank loans and firm value. Nankai Business Review International, 3(4), 376–397.
NU
Zhang, H., Lu, Z., Zhang, R., & Jiang, G. (2015). Insider ownership, subsidiary cash
MA
holdings, and economic consequences: Evidence from listed Chinese companies.
AC
CE
PT E
D
Emerging Markets Finance and Trade, 51(Suppl. 1), S174–S195.
39