Nature and Science, 1(1), 2003, Hongxia Li and Liming Cui, Empirical Study of Capital Structure

Empirical Study of Capital Structure on Agency Costs in

Chinese Listed Firms
Hongxia Li*, Liming Cui**
* School of Management, Harbin University of Science and Technology, Harbin, Heilongjiang 150080,
China; ** Aston Business School, Aston University, Birmingham, B4 7ET, UK.
lhx_2002@hotmail.com, Liming_cui@hotmail.com.

Abstract: This study examines the impact of capital structure on agency costs in 211 non-financial
Chinese listed firms for the period 1999-2001. There are two main findings. (1) Firms with high
debt to asset ratio have high ratio of annual sales to total assets and high ratio of return-on-equity. If
a firm has a high debt to asset ratio, creditors are much more concerned about the payment of
interest and repayment of principal and will have incentives to monitor the firm. Thus, a capital
structure with high debt decreases agency costs. (2) Positive and significant correlation is identified
between ownership concentration and the return-on-equity ratio. This is because the largest
shareholders have a strong interest in firm performance and therefore a high ability to reduce
agency costs. Our empirical results further illustrate that firms have inclination of refinancing
through stock market and harm small shareholders’ interest. [Nature and Science 2003;1(1):12-20].

Keywords: capital structure; agency costs; corporate governance

1. Introduction controlling power of the state shares. This led to the

The Chinese stock market was established in the early false placement of state property. “Inside control”
1990s. By listing on the stock exchange, state-owned problem is serious3. Secondly, state shares are uniquely
enterprises (SOEs) have improved their debt to asset big and there is serious impingement upon the interests
ratio, and promoted their development by directly of small shareholders. Thirdly, the board of directors is
financing from the stock market. After 10 years’ formed mainly by executive directors and controlling
development, there are more than 1200 listed firms on shareholders, directors lack integrity obligations, failing
the two stock exchanges now. The total market to perform their duties industriously 4 . Additionally,
capitalisation at the end of March 2001 was RMB there is not much pressure of dividend from
(Chinese dollar) 5 trillion, 54% of the last year’s GDP. shareholders, so re-financing of listed firms usually
This shows that listed firms play a significant role in the place the order of debt after additional or right shares.
national economy. On the other hand, China still has By pecking order pattern of financing observed in
much to do on reform in corporate governance 1 . For advanced country corporations, firms obtain capital by
historical reasons, the majority of China’s listed firms making greater use of internal finance followed by debt
were restructured and transformed from previous state- and turning to stock market finance only as a last resort.
owned enterprises or other government controlled Jensen’s (1986) free cash flow theory considers debt can
entities 2 and there are many problems left with the mitigate the agency problems between shareholders and
governance structure. The central problem with the managers of firms and motivate management to act in
governance structure is the ambiguous definition of the the interests of the shareholders. How would their
financing patterns differ from those of advanced country
The financial support provided by the Heilongjiang firms? How to decide the best strategy of refinancing?
Province Natural Science Fund G01-10. There has been much research conducted concerning
1
In its narrowest sense, corporate governance is about the agency problem in developed markets. There are
the relationship of the owners or shareholders of a firm three general ways in which to reduce the conflicts of
with its manager (Iskander and Chamlou, 2000), which interest between managers and the shareholders: 1)
is often characterized by economists as the “agency Increasing management ownership because high
problem”. management ownership aligns the interests of
2
About 75% of listed firms are formerly state-owned. management and shareholders (Jensen, 1993; Ang,
Another 10% are firms that mostly had significant
3
shares held by SOEs. Only less than 10% of listed firms The “inside control” viewpoint was aired in 1995 by
are formerly private-owned firms or foreign-invested Japanese scholar Masahiko Aoki.
4
firms, which in most cases had SOEs as their joint Wu Jinglian, June 8, 2001. The behaviour of control
venture partners. See the website of China Securities shareholders and corporate governance. Shanghai
Regulatory Commission: www.csrc.gov.cn security news.

http://www.sciencepub.net 12 editor@sciencepub.net
 Nature and Science, 1(1), 2003, Hongxia Li and Liming Cui, Empirical Study of Capital Structure

1999). 2) Monitoring management by large shareholders firm have potential conflicts of interest with the outside
(Shleifer, 1986). 3) Using debt financing to discipline shareholders, since they choose to reinvest the free cash
managers (Jensen, 1986; Stulz, 1990). The first option is rather than return it to investors (Jensen, 1976, 1986).
not the focus of this study. This is because the vast The conflict arises when there is moral hazard inside the
majority of China’s listed firms are formerly state- firm, which is called the agency costs of equity. This
owned or state-controlled firms. The management agency problem can be solved by increasing
holding of shares is extremely low, about 0.03% of the management ownership because high management
total shareholding5, and it is subordinate to controlling ownership aligns the interests of management and
shareholders. This situation will not change in the near shareholders (Jensen, 1976). Other possibilities include
future. Similarly, the second option is also excluded. monitoring of management by large shareholders
The state holds shares of most listed firms in great (Shleifer, 1986), and the use of debt financing to
concentration, listed firms, is not really separated from discipline managers (Jensen, 1986; Stulz, 1990).
controlling shareholders in personnel, financial affairs
and assets. Due to the complicated market and 2.1 Managerial Ownership and Agency Costs
administrative process associated with the reform of Managerial ownership has considered non-linear forms
Chinese property system, many constraints were present (Morck, 1988; McConnell, 1995; Kole, 1995). Jensen
on reducing agency costs through the first two options. (1993) “convergence of interest” hypothesis suggests
Thus our research concentrated on the third option that managerial shareholdings help align the interests of
mentioned: the impact of capital structure on agency shareholders and mangers, and as the proportion of
costs. managerial equity ownership increases, so does
The contribution of this study is two-dimensional. corporate performance. In contrast, Morck et al (1988)
Firstly, it contributes to the literature of the impact of argued that high level of managerial ownership could
capital structure on agency costs in Chinese listed firms. lead to ‘entrenchment’, as external shareholders find the
Plenty of research about Chinese corporate governance actions of such managers difficult. Kole’s (1995)
has focused mainly on ownership structure, Xu (2000), argument suggests that managerial ownership may
Zheng (2002), Shi (2000), Zhang (1996), etc. The focus impact large and small firms differently with respect to
of this research is to study the interrelationship between value. Ang examined the relationship between agency
the capital structure of listed firms and agency costs of costs and managerial ownership for small firms, and
equity. Secondly, among the limited research, Yang et Singh et al tested same work on the relationship for
al (2001) analyses the interrelationship between the large firms.
capital structure of listed firms and agency costs. No
empirical analysis has been conducted so far. This paper 2.2 Concentrated Ownership and Agency Costs
will fill the gap in this area. It provides evidence on firm An important line of agency costs literature relates to
capital structure and agency costs measured in terms of concentrated ownership. Stiglitz (1985) has argued that
ratio of sales to total assets and ratio of return on equity. one of the most important ways of value maximization
We wish to provide evidence through empirical study as by firms is through concentrated ownership of the firm’s
to how would capital structure influence agency costs. shares. Shome and Singh (1995) replicate this result and
This paper is structured as follows. Section 2 gives provide evidence that the large shareholder’s presence
previous relevant literature review. Section 3 describes improves accounting performance. Large shareholders
the current status of Chinese corporate governance, thus address the agency problem as that they both have
section 4 presents the methodology including empirical a general interest in profit maximisation, and enough
model and sample. Section 5 gives the empirical results, control over the assets of the firm to have their interests
and section 6 concludes. respected. Many scholars argued that outside large
shareholders reduce managerial entrenchment (Shleifer,
2. Literature Review 1986; Kang, 1995; Yosha, 1996; Porta, 1998, 1999;
Agency problems are caused by the separation of Park, 1995; Denis, 1996).
ownership from control in large firms (Berle, 1932). However, this does not exclude the possibility of rising
Jensen and Meckling (1976) apply agency theory to the concentration of share ownership to depreciate the
modern corporation and formally model the agency market value of the firm (Huddat, 1993; Admati, 1994).
costs of external equity6. Managers who own anything The control shareholders often have better access to
less than 100% of the residual cash flow rights of the information, hold more power in selecting management
and involve in key decision-makings. Especially when
5
The statistics get from: www.csrc.gov.cn. the manager holds fewer shares7 and is subordinate to
6
Jensen and Meckling model the agency costs of debt 7
in this paper. However, for our purposes in this paper, it Morck, Shleifer, Vishny (1988) showed in their
is the agency costs of equity that are relevant. empirical study that the proportion of equity held by

http://www.sciencepub.net 13 editor@sciencepub.net
 Nature and Science, 1(1), 2003, Hongxia Li and Liming Cui, Empirical Study of Capital Structure

controlling shareholders, control shareholders impinge ownership structure and agency costs measured in terms
upon the interests of small shareholders by way of non- of asset utilization and operating expenses. Ang used
division of dividends and diversion of profits. The data on small business in America to examine how
exploitation of small shareholders by controlling agency costs vary with a firm’s ownership structure.
shareholders constitutes ex ante an expropriation threat They find agency costs 1) are higher when an outsider
that reduces managerial initiative and non-contractible rather than an insider manages the firm; 2) are inversely
investments and may come into conflict with related to the manager’s ownership share; 3) increase
performance-based incentive schemes (Burkart, 1997). with the number of non-manager shareholders, and 4) to
a lesser extent, are lower with greater monitoring by
2.3 Debt and Agency Costs banks. In the second study, Singh and Davidson extend
Another strand of the agency literature has focused on the work of Ang’s analysis of relationship between
the role of debt as a means of disciplining managers. corporate ownership structure and agency costs to large
Grosseman and Hart (1982) were the first to argue that publicly traded corporations. Using slightly different
managers could pre-commit to work hard by using debt measures of agency costs9, they analysed multi-period
rather than equity. Similarly, Jensen’s (1986) free cash data for the year 1992 and 1994, and studied not only
flow theory considers additional debt beneficial since inside ownership structure as a determinant of agency
the firm attempts to improve the productivity of its costs, but also the role of outside large equity holders in
assets as a result of additional debt acquired. Debt not disciplining the management. They found outside large
only reduces the free cash flow but also provides shareholders’ ownership may only have a limited effect
discipline to management through the debt market. Debt on reducing agency costs and board size was negatively
monitoring hypothesis is formalised by Harris and related to asset turnover, and unrelated to discretionary
Raviv (1990) and Stulz (1990) and empirically expenditures.
demonstrated by Maloney et al. (1993). Shleifer and In this paper, following their example we use ratio of
Vishny (1997) provided extensive survey about the role sales to assets as one of the measures of agency costs10.
for debt in reducing the conflict of interests between Additionally, we use ROE as an alternative measure of
managers and shareholders 8 . On the other hand, agency costs to analyse the possible impact of variables
increased leverage also has costs. As leverage increases, on agency costs in Chinese listed firms. Although ROE
the usual agency costs of debt rise, including is a more manipulability measure in economic sense, the
bankruptcy cost (Jenson 1986). Myers (1977) pointed to regulating authorities in China use this particular ratio
the debt overhang problem where firms may forego as a standard to decide whether a firm is qualified for
good projects if they have significant debt outstanding. right shares or additional shares. It is one of the most
The reason is that for a firm facing financial distress, a strictly regulated ratios in China and is widely used for
large part of the returns to a good project go to comparative purpose.
bondholders. Therefore, in choosing their debt-equity
level, firms should trade off between the agency costs of 3. Chinese Corporate Governance
debt and the agency costs of equity. By appropriately 3.1 Ownership Structure and Corporate Governance
allocating refinance between equity and debt, capital The main characteristic of the Chinese corporate
structure can balance the conflicts between investors governance is the over concentration of equity structure.
and management as well as that between management Most of the listed firms in China are transformed from
and creditors. state-owned enterprises. Ownership structure displays
Finally, two previous studies most closely related to the phenomenon of the co-existence of control
this study are Ang et al (1999) and Singh (2002). In the shareholders, who are normally related to the state and
first case, Ang et al provided evidence on corporate many other small and comparatively weak shareholders.
State shares are uniquely large. Statistics show that the
managers and the valuation of the firm and proportion is state holds shares of most listed firms in great
in inverted U-shape. concentration. Of the listed companies, 54% of the
8
Several articles model the benefits and costs of debt, equities belong to the state or state-owned corporate
the benefit is usually the reduction in the agency cost,
9
such as preventing the manager from investing in Use the SG & an expense ration instead of operating
negative net present value projects, or forcing him to expenses to measure agency costs.
10
sell assets that are worth more in alternative use. The Singh found the SG&A expense ratio not
main costs of debt is firms may be prevented from significantly influenced by ownership. This is because
undertaking good projects because debt covenants keep governance variables are not as visibly related to cash
them from raising additional funds, or else they may be flows generated by firms that are sales revenues. We
forced by creditors to liquidate when it is not efficient to conducted a similar analysis and didn’t find any
do so. significance relationship either.

http://www.sciencepub.net 14 editor@sciencepub.net
 Nature and Science, 1(1), 2003, Hongxia Li and Liming Cui, Empirical Study of Capital Structure

persons 11 . Among the 1104 listed firms on Shenzhen from shareholders, so refinancing of listed firms in
and Shanghai stock markets, the proportion of shares of China usually place the order of debt after additional or
the number one shareholder had reached 45% and the right shares.
second largest shareholder made up 8%. The optimal debt-to-equity ratio is the point at which
The management level lacks long-term incentive and firm value is maximised, the point where the marginal
restraining mechanism. The board of directors is mainly costs of debt just offset the marginal benefits. The over
formed by control shareholders. The lack of low level of debt to asset ratio reflected the poor
independent directors makes it difficult to display their management of corporate financial gear of Chinese
regulating and balancing roles. Among all the directors listed companies. Refinancing through equity is not the
of the listed firms, 73.3% have the background of state optimal strategy to reduce their capital cost. It’s not a
shares (27.9%) or shares of state-owned corporate common phenomenon for a modern corporate to rely
persons (45.4%). Since the manager holds fewer shares, almost totally on it’s own capital, using none or merely
and is subordinate to controlling shareholders, the little debt. One of the most important reasons that
agency problem between shareholders of a firm and its Chinese listed companies don’t bother to use debt is the
manager has turned into the second dimension of the fact that they generally can obtain “free capital” easily
agency problem in a firm, the conflict between the from the equity market. In order to limit the “equity
controlling shareholders and small shareholders. financing thirst”, China Security Regulatory
Commission requires 12 that the debt to asset ratio of
3.2 Leverage and Corporate Governance listed firms who want to add shares on stock market
The central goal of corporation, including public listing, must have higher debt to asset ratio than the average
is to establish “a modern enterprise system” in China, level of the same industry (Table 1). Listed firms have
featuring the corporate governance structure that paid more attention to their capital structure since then,
separates the government from enterprises. Another and it helps to improve the capital structure of listed
objective is to raise capital for SOE’s and reduce their firms.
high level of debt to asset ratio by increasing direct
finance through selling equity to the public. The vast 4. Data and Methodology
majority of China’s listed firms are formerly state 4.1 The Data
owned or state controlled firms, mostly large and better The sample was a pool of several data of firms listed on
performing firms. Before initial public offering, they do the China (Shanghai and Shenzhen) Stock Exchanges
their best to dispose of the debt. So, the debt to asset from 1999 to 2001. 211 listed firms 13 were randomly
ratio of listed firms is lower during the first couple of chosen excluding finance and insurance industry, ST
years after initial public offering. (special treatment)14 and PT (particular transfer)15 firms
According to capital structure theory, the way to were not included in the sample either. The accounting
refinance is determined by the cost of capital. In data was obtained from listed firms’ annual reports from
developed capital market, the top managers are 1999 to 2001, which were published on the web site
restrained by shareholders and creditors, facing the (http://www.csrc.gov.cn) of China’s Securities
pressure of paying dividend and debt. The empirical Regulatory Commission (CSRC). The inside and
results show that listed firms obtain capital first from outside ownership information and board size
internal sources, then from debt, and last from equity.
Capital cost influences the style of financing. In China,
12
due to the special ownership structure of listed firms, In Table 1, the debt to asset ratio is the industry
state share is absolutely the largest among total shares average level of listed firms. The industry classification
and the representatives of state shares are usually absent. conforms to the first grade industry classification of
This reduces the restriction to management, and the Chinese Stock Exchange. This policy was announced on
managers would over pursue the control right of cash March 18th 2001.
13
flow. The consequence is that re-financing of listed There were 1160 listed firms on stock market in the
firms would have partiality for equity rather than debt. end of 2001. Sample =1160 - (43 ST firms +8 PT firms)
Additionally, there is not much pressure of dividend * 20% - 11 firms IPO after 1999 =211 firms.
14
Shanghai Stock Exchange and Shenzhen stock
11
State shares are held by government bodies such as exchange declared emerging abnormal phenomenon
state asset management agencies, or institutions from some listed firms’ financial statements. Such listed
authorised to hold shares on behalf of the state such as a firms’ stock was specially treated. The stock is called
wholly state-owned investment firm. Legal person ST stock. There are 43 ST firms in China’s two stock
shares are shares held by any entity or institution with a exchanges at the end of 2001.
15
legal person status, including an SOE or a firm Listed firms that have continuous 3-year loss are
controlled by an SOE. called PT firms. There are 8 PT firms at the end of 2001.

http://www.sciencepub.net 15 editor@sciencepub.net
 Nature and Science, 1(1), 2003, Hongxia Li and Liming Cui, Empirical Study of Capital Structure

Table 1. Listed Firms Industry Average Debt to Asset Ratio (%)

Industry
Classification (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (12) (11) (13)

Average Debt to asset

39.2 30.60 47.74 39.93 54.71 36.97 49.21 61.94 91.74 73.65 38.70 42.51 63.38
ratio
(1) Agriculture, Forestry, Animal Husbandry and Fishing. (2) Mining. (3) Manufacture. (4) Reduction and Distribution of Electricity, Gas and
water.(5) Construction. (6) Transport, Storage and Post. (7) Information transmission, Computer services and software. (8) Whole sale and
Retail sale.(9) Finance and Insurance. (10) Real estate. (11) Resident service and other service. (12) Disseminate and Culture. (13) Synthesis.
Material source: http://www.bigsun.com.cn/data/agora/20020730/125466.hml.

Table 2. Sample Descriptive Statistics

Variable 1999 2000 2001 Pooled
Mean Median Mean Median Mean Median Mean Median
Assets turnover 0.5335 0.4783 0.5500 0.4591 0.5634 0.4503 0.5575 0.4608
Return on equity ratio % 11.4210 10.4800 9.9839 9.6400 7.2498 7.8200 9.6150 9.5500
Debt to asset ratio % 40.4895 38.5374 39.6272 38.7600 41.4436 41.1300 40.4429 40.0496
Outside block ownership:
The largest shareholder % 48.7506 49.0200 46.3133 45.5000 46.3500 45.9700 46.6115 47.2100
The five largest shareholders % 61.6255 61.6600 59.2176 60.1000 58.3813 58.9900 59.6540 60.5100
Total sales (RMB 10000) 90894.29 46347.96 116239.10 65616.90 136408.90 73046.00 115986.3 62795.68
Board size 9.6540 9.0000 9.7678 9.0000 9.7915 9.0000 9.7300 9.0000
Managerial ownership % 0.0441 0.0191 0.0316 0.0154 0.0273 0.0116 0.0343 0.0150
The data in this table contain 211 non-financial companies listed on the Stock Exchange of China. Outside block ownership is defined as
percentage of total stock held by the largest shareholder and the five largest shareholders. Debt to asset ratio is debt divided by total assets in
the end of accounting year. The size of the board of directors measures board size by determining the number of board members. Managerial
ownership is the percentage of shares owned by managers.

information were obtained from the web site: purposes. Therefore, higher asset turnover has less
http://www.cnlist.com and http://www.cninfo.com.cn. agency conflict.
The sample was a pool of seven ratios of firms listed We use an additional measure of agency costs, the
on the China (Shanghai and Shenzhen) Stock ratio of return on equity (ROE), as a measure of
Exchanges from 1999 to 2001 (Table 2). The debt to profitability. This indicator measures profitability from
asset ratio includes data of firms’ assets turnover, return a different angle. In China, most listed firms were
on equity, debt to asset ratio, outside block ownership transformed from state owned enterprises. In order to
(the percentage of shareholding of the largest and five protect the value of state assets, fixed assets
largest shareholders), firm total sales, board size and depreciation rates are centrally determined and often are
managerial ownership. artificially low, thus leading to an upward bias in fixed
asset estimates. Current assets include some stockpiled
4.2 Methodology goods that either cannot be sold at their book value, or
The methodology we use is a system of simultaneous cannot be sold at all. ROE is clearly a more preferable
equations. indicator of profitability, matching the common usage
The system of two equations to be estimated is: of the market economics. Profit is the return to equity
Agency costs=β0+β1capitalstructure+β2Conc+β3Size holders; therefore higher turn on equity has less agency
+β4board+∑βjDumjt (1) conflict.
Capital structure=β5+β6Conc+β7ROE+β8Val (2) Independent variables were chosen mainly based on
(Conc – concentration) the existing agency literature. The first variable used is
We use two alternative measures for agency costs. The capital structure, measured by debt to asset ratio (total
first measure for agency costs is the ratio of annual sales debt divided by total book assets). The second variable
to total assets (asset utilization), following the research used is ownership concentrations. Ownership
of Ang et al (1999). This ratio measures management’s concentration is measured by the proportion of the
ability to employ assets efficiently. A high ratio of shares held by the largest shareholder to the total shares
annual sales to total assets shows a large amount of and the share proportion of the top five largest
sales and ultimately cash flows that are generated for a shareholders. The third categories of variables are
given level of assets. While a high asset turnover may control variables. They are included in the regressions
be identified with efficient asset management practices to control for other potential influences on the agency
and hence shareholders value creation, a low asset costs of firms. The variables included are the size of
utilisation reflects asset deployment for unproductive board of directors, firm size, and industry Dummies.

http://www.sciencepub.net 16 editor@sciencepub.net
 Nature and Science, 1(1), 2003, Hongxia Li and Liming Cui, Empirical Study of Capital Structure

The number of board members measures the size of the is used by Leech and Leahy (1991) and is calculated as
board of directors. The measure of a firm’s size is the the market value of the firm at the end of its accounting
logarithm of total sales. The industry dummy variable is year, divided by the book value of equity.
Chinese listed firms’ classified 13 industries, excluding Thus, with two equations, one determining agency
financial firms. This provides 11 industry dummy costs, and the other determining the debt to asset ratio,
variables in the multiple regression models. another exogenous variable is needed in the
To solve potential endogenous problems, whether the determination of the debt to asset ratio in order for
ratio of return on equity or the debt to asset ratio is equation (1) to be identified. This is outside ownership
simultaneous is tested. On one hand, the debt to asset concentration. When agency costs is measured by asset
ratio can affect return on equity. Firstly, an increase in turnover, using only equation (1) would suffice. When
the debt to asset ratio, through financial charges would agency costs are measured by ROE, the debt to asset
reduce profit. A high debt-asset ratio, finally, should ratio is an endogenous variable and two equations are
imply a high degree of external control as creditors, necessary.
concerned about the payment of interest and the
repayment of the principal. Creditors have incentives to 5. Empirical Results
monitor the enterprise. A higher degree of supervision Empirical results are presented in Table 3 (Panel A & B)
could lead to higher profitability. Secondly, by using and Table 4. Table 3 presents the OLS regression
debt, ROE would increase even though the profit analysis result that analysed agency costs, measured by
doesn’t increase, with a constant equity. On the contrary, asset utilisation and ROE respectively. Panel A gives
obtaining capital through equity would reduce ROE. On the result of agency costs measured by asset utilisation
the other hand, ROE affect the leverage, with a high and Panel B gives the result of agency costs measured
level of ROE, listed firms can get funds from newly by ROE. Table 4 presents the result of the simultaneous
accumulated profits or from stock market by additional equation of ROE.
or right shares. The debt to asset ratio may depend not
only on ROE but also on the ownership structure. If 5.1 Agency Costs Measured by Ratio of Annual Sales
there is high ownership concentration, the control to Total Assets
shareholders might want to reduce the dividend, In panel A of Table 3, dependent variable proxy for
implying a low debt to asset ratio. However, if the agency costs is the ratio of annual sales to total assets.
equity value growth is high, the firm can obtain plenty There are three groups of independent variables: capital
of cash flow, and impact on debt to asset ratio. We use structure variables, ownership concentration variables
Val to measure the equity value growth. This valuation and control variables. Rows 1 and 2 report the

Table 3. Multivariate Regression Analysis debt to Asset Ratio and Ownership Concentration to Agency Costs
Panel A: Agency costs as measured by the ratio of annual sales to total assets.
Panel B: agency costs as measured by ratio of return on equity.
Ownership concentration % Control variables
The five Adj-R-
Regress Debt to asset The largest Firm Size
Constant largest Board size Industries squared
-ion ratio shareholder (log -sales)
shareholders
Panel A:
Row 1 0.1411 0.0072 0.0026 0.0912
(2.2505) ** (7.8303)*** (2.9038)***
Row 2 0.2093 0.0069 0.0012 0.0806
(2.6118) *** (7.4990)*** (1.0415)
Row 3 -1.1820 0.0028 0.0002 -0.0094 0.1718 Yes 0.3986
(-8.0504)*** (3.4429)*** (0.2773) (-1.9838)** (14.4776)***
Row 4 -1.2294 0.0027 0.0010 -0.0088 0.1716 Yes 0.3995
(-7.9337)*** (3.3710)*** (0.9863) (-1.8553)** (14.8980)***

Panel B:

Row 5 5.0937 0.1287 0.0279 0.0684

(4.0120)*** (6.9215)*** (1.4925)
Row 6 2.7343 0.1251 0.0626 0.0762
(1.7017)* (6.8009)*** (2.7472)***
Row 7 8.9014 0.1416 0.02619 -0.5538 0.1217 Yes 0.1663
(2.5744)*** (7.4026)*** (1.3788) (-4.9447)*** (0.4356)
Row 8 7.2598 0.1364 0.0377 -0.5447 0.1817
(1.9893)** (7.1781)*** (1.6420)* (-4.8468)*** (0.6694) Yes 0.1674

Values in parentheses are t values. *10% level, ** 5% level ***1% level

Equation. (1): Agency costs =β0+β1capital structure +β2Conc +β3 Size +β4 Board +∑βj Dumj

http://www.sciencepub.net 17 editor@sciencepub.net
 Nature and Science, 1(1), 2003, Hongxia Li and Liming Cui, Empirical Study of Capital Structure

Table 4. Impact of Debt to Asset Ratio on the Ratio of Return on Equity (2SLS results)
Ownership Concentration Control Variables Equity
Regressi Constant Debt-to-asset The largest The five largest Firm size Indus Return on Market-to-
on % shareholder % shareholders% Board (log sale) tries Equity % Book Value
Eq. (1)
1999 -29.9728 0.9154 0.0585 -0.0186 0.1813 No
(-74.9606)*** (254.1369)*** (22.2788)*** (-1.6130) (6.7096)***
2000 -20.9198 0.6317 0.0345 -0.0140 0.3565 Yes
(-36.2628)*** (187.1876)*** (9.9195)*** (-0.8665) (9.0342)***
2001 -50.1593 0.8332 0.1354 -0.1589 1.5859 Yes
(-9.3704)*** (18.7996)*** (4.3807)*** (-1.0291) (4.3131)***
pool 1 -25.6625 0.6978 0.1198 -0.0179 0.1859 Yes
(114.1235)*** (428.5688)*** (102.6309)*** (-2.6085)*** (11.3209)***
pool 2 -26.2627 0.6868 0.0895 -0.0385 0.3490
(-61.7831)*** (235.0715)*** (35.6543)*** (-3.0672)*** (12.0077)*** Yes

Eq.(2)

1999 31.3633 -0.0577 1.0353 0.0795

(5.3222)*** (-0.6591) (2.4545)** (0.2693)
2000 26.1591 -0.06832 1.5447 0.3620
(3.5698)*** (-0.6374) (3.6285) *** (0.8032)
2001 41.1093 -0.2024 0.7980 1.4301
(6.7670)*** (-1.9635) ** (1.9841)** (4.0992)***
pool 1 34.7605 -0.1694 1.3532 0.1105
(12.1722)*** (-3.2660)*** (4.4911)*** (0.5409)
pool 2 34.7928 -0.13278 1.3031 0.2024
(9.1036)*** (-2.1735)** (4.4705)*** (1.0390)
Equation. (1): ROE=β0+β1 Capital structure +β2Conc +β3 Size +β4 Board +∑βjDumjt
Equation. (2): Capital structure= β5+β6Conc +β7ROE +β8Val
Values in parentheses are t values. *10% level, ** 5% level *** 1% level

regression result on capital structure together with the displayed in both Row 3 and Row 4. This shows that
largest shareholder concentration and the five largest large boards reduce asset utilization and they are
shareholders’ concentration respectively. Rows 3 and 4 detrimental to shareholders’ interest. This is because
report the regression result including the control the function and work procedures of board of directors
variables: board size, firm size and industry dummies. are not standardized. The amounts of shares held by
In all of the four rows mentioned above, a positive directors are extremely low (Table 1, managerial
relationship between agency costs and capital structure ownership). Many directors are appointed by the
was identified, all significant at a better than 1% level. government, and they are not paid by the listed
This proves that firms with a higher debt to asset ratio companies, but paid by some government institutes
are more efficient in their asset utilization. This result instead. This way they hardly find their own interest
supports Jensen’s (1986) theory of free cash flow, align with the company. In the absence of integrity
which considers additional debt beneficial as the firm obligation, directors fail to perform their duties
attempts to improve the productivity of its assets as a industriously to improve the firms’ asset utilization. The
result of additional debt acquired. Such positive negligence of large shareholders together with smaller
relationship between capital structure and asset shareholders’ lack of supervision incentive, the smaller
utilization is also identified by Gorman (2000), Ang shareholders would choose to ‘vote with feet’, which
(1999) and Singh (2001). would deteriorate the ‘insider control’ problem among
We found mixed result for the largest shareholder the management of listed companies. The firm size
concentration, a positive relationship with agency costs factor among the control variables shows a positive
displayed in Row 1, while an insignificant relationship relationship to asset utilization, significant at 1%. Hence
was found in Row 3 when regressed with other control the agency costs would be lower for a larger firm. Large
variables. For the five largest shareholders’ firms have more efficient corporate governance. There
concentration, no significant relationship was identified are five industries’ coefficients significant with asset
with agency costs. This proves that control shareholders turnover. They are electricity, transportation, wholesales,
don’t have much interest in improving their asset real estate and service industry. Except for wholesales,
utilization ratio. This result is found by Signh et al the coefficients of other industries are negatively related
(2001) too, as they report that the proportion of equity to asset turnover.
held by outside block owners does not relate to agency
costs as measured by asset utilization. 5.2 Agency Costs Measured by Ratio of Return on
Among the control variables, the coefficients for the Equity
control variables for board size are negative and In Panel B of Table 3, dependent variable proxy for
significant at 5% in relation to the asset turnover ratio as agency costs is ROE instead of asset utilization as in

http://www.sciencepub.net 18 editor@sciencepub.net
 Nature and Science, 1(1), 2003, Hongxia Li and Liming Cui, Empirical Study of Capital Structure

Panel A, with the independent variables identical to that they are only significant at the pooled level, except for
of Panel A. Row 5 and Row 6 report the regression year 2001, this shows the large shareholders prefer to
result on capital structure with largest shareholder refinance through equity than debt. The reason of an
ownership concentration and five largest ownerships’ increased significance of the third year and for the
concentration respectively. Row 7 and Row 8 result pooled result is because CSRC requires the average
including also the control variables: board size, firm return on equity level over the last three year (from
size and industry dummies. 2001) must be more than 10% to be qualified for
As expected, we found a positive relationship between additional share. Thus if a firm’s return on equity ratio
capital structure and ROE, significant at 1% level, is higher than this level, they like to finance from equity
confirmed the result in Panel A. Firms with a higher more than debt. This result conforms to Shleifer and
leverage level have a higher return on equity. This Vishny’s (1997) theory that large shareholders claim
conforms to the theory that creditors, concerned with they both have a general interest in profit maximization,
the repayment of the debt, would exert positive and enough control over the assets of the firm to have
influence on the management of the firm and thus their interests respected.
improve the firm’s profit return. The ownership Our results also provide new evidence for the second
concentration, however, displayed different result than dimension of agency problem: the conflict between
that of Panel A: they are positive with ROE at 1% and large shareholders and small shareholders. Since the
10% respectively in Row 6 and Row 8. Comparing to boards of directors are mainly constituted by large
the asset utilization result, our ROE result proves that shareholders in China, the boards’ decisions reflect
large shareholders concern about their profitability. The large shareholders’ will. The positive relationship
highly concentrated ownership would benefit the between large shareholder and ROE confirmed by our
operation of the business. For the control variables, the data (Table 4) proves that the large shareholders are
board size result is consistent with that of Panel B and very concerned about the agency problem to maximize
showed negative relationship with ROE, both their own benefits. Higher return on equity would mean
significant at 1%. Firm size displayed positive more profits for the large shareholders. However, the
relationship with ROE, but not significant. The industry non-significant positive relationship between larger
dummy variable for the 11 industries doesn’t show any shareholder and asset utilization found in Table 3 (panel
significance, except one at a 10% level. Together with A) illustrated they are not genuinely interested in
the industry result from Panel A, it proves the industry improving the firm utilization. Again their significant
factor doesn’t play any significant role in deciding the negative relation with debt to asset ratio, their prejudice
agency costs. against using debt the less costly capital, proves that
Since the debt to asset ratio could be an endogenous they are sacrificing the smaller shareholders’ interest for
variable, to solve the potential endogeneity problems we their own good. Although large investors can be very
need two equations, one determining the ratio of return effective in solving the agency problem, they may also
on equity, and the other determining the debt to asset inefficiently refinance the firm through equity while
ratio. Table 4 reports the two stage least-squares using debt could have maximize the firm value. Driven
regression results. by their own interest they may also redistribute the
For equation (1), both the pooled data result and the wealth of the firm from other small investors. Because
result of each individual year show that the debt to asset small shareholders unlike creditors, they are not
ratio has positive relationship with the ratio of return on promised any payments in return for their financial
equity at significant better than 1% level. This result investment in the firm, and have no claim to specific
supports Jensen’s (1986) debt monitoring hypothesis. assets of the firms.
This is examined by Harris and Raviv (1990) and Stulz Board size has a negative relation to agency costs, and
(1990) and empirically demonstrated by Maloney et al. its coefficient is not significant at each individual year,
(1993) and Gul and Tsui (1998). Highly leveraged firms but significant at a better than 1% level at pooled level.
should be subject to better supervision than those listed The individual level result is that the large shareholders
firms whose assets are primarily financed through are motivated to care for the company performance.
“free” equity that comes with little monitoring. This somehow counter-effects the negative effects of
The result for the largest and the five largest the overall board members’ ‘shirking’ behavior.
shareholders’ concentration is identical to each other. However, this kind behavior would reveal itself more
They are both positive to ROE at significant better than clearly through pooled year by members’ seeking more
1 percent level in equation (1) in all the results. This discrete ways to enhance their own interest at the firms’
finding supports the view that large shareholders play an cost. Large companies are more efficient in dealing with
active role in corporate governance (Shleifer, 1986; agency problems. All the results of ROE are positive to
Yafeh, 1996; Denis, 1996). However, all the results are the capital structure, significant at either better than 1% or
negative on debt to asset ratio (equation 2). Although 5%. The equity market-to-book value is not significant.

http://www.sciencepub.net 19 editor@sciencepub.net
 Nature and Science, 1(1), 2003, Hongxia Li and Liming Cui, Empirical Study of Capital Structure

6. Conclusion Berle A, Means G. The modern corporation and private property.

New York: MacMillan. 1932.
There have been many research conducted concerning Denis DK. Twenty-five years of corporate governance research and
agency costs in developed markets, however, not counting. Review of Financial Economics 2001;10:191-212.
enough attention has been paid to emerging market like Fama E, Jensen M. Agency problems and residual claims. Journal of
China. The contribution of this study is two-dimensional. Law and Economics 1983;26:327-49.
Filbech G, Gorman RF. Capital structure and asset utilization: the
Firstly, it contributes to the literature of the impact of case of resource intensive industries. Resources Policy 2000;26:211-8.
capital structure on agency cost in Chinese stock market. Glen J, Lee K, Singh A. Competition, corporate governance and
Secondly, among the limited research concerning financing of corporate growth in emerging markets. Working
capital structure, Yang et al (2002) analyses the paper/Accounting and Finance Discussion paper, RePEc: cam:
camafp:00af46.http://www.econ.cam.ac.uk/dae/repec/cam/pdf/afp46.p
interrelationship between the capital structure of listed df.
firms and agency costs in a descriptive way. However, Hart O. Firms contracts and financial structure. Oxford University
no empirical analysis has been conducted so far. This Press, London. 1995.
paper provides statistical evidence on firm capital Hirota S. Are corporate financing decisions different in Japan? An
empirical study on capital structure. Journal of Japanese and
structure and agency costs measured in terms of ratio of International Economies 1999;13:201-29.
sales to total assets and ratio of return on equity. Holz CA. The impact of liability-asset on profitability in China’s
The forces working on firms’ capital structure in other industrial state-owned enterprises. China Economic Review
countries also work in a quite similar way in China. The 2002;13:1-26.
Jensen M, Meckling W. Theory of the firm: Managerial behaviour,
results indicate that firms with a higher debt to asset agency costs and ownership structure. Journal of Financial Economics
ratio have a higher ratio of sales to assets and a higher 1976;3:305-60.
ratio of return on equity, and this relationship is Jensen M. Agency costs of free cash flow, corporate finance, and
statistically significant at better than 1 percent level. takeovers. American Economic Review 1986;76:323-9.
Jorge A, Chan-Lau. The impact of corporate governance structures
Capital structure theories suppose that managers make on the agency costs of debt. IMF Working Paper WP/01/204.
financing decisions so as to maximize value of equity Kevin HS. Managerial ownership and performance of firms:
for shareholders. This finding is supportive of the theory Evidence from the UK. Journal of Corporate Finance 1999;5:79-101.
put forth by Williams (1987) that additional debt La Porta R, Lopeze-de-Silanes F, Sheifer A. Corporate ownership
around the world. Journal of Finance 1999;54:471-517.
decreases agency costs and the theory by Jensen (1986) Laura M Cha. The future of China’s capital market and the role of
that debt can reduce the agency costs of free cash flow corporate governance. Luncheon speech at China business summit.
by reducing the cash flow available for spending at the www.csrc.gov.cn/csrcsite/enews/efi20010427.htm.
discretion of managers. The ownership structure also Lu T, Wang W. Corporate governance reform: China and World.
Economic Management Publishing House, Beijing. 2002.Page 42-58
affects capital structure. Firms with higher Ownership Maug E. Board of directors and capital structure: Alternative forms
Concentration tend to have lower debt to asset ratio. of corporate restructuring. Journal of Corporate Finance 1997;3:113-
Why do Chinese listed firms have such a low leverage? 39.
One possible reason is that Chinese firms prefer and Shleifer A, Vishny R. A survey of corporate governance. The Journal
of Finance 1997;52:737-83.
have access to equity financing once they go public as Singh M, Wallace ND. Agency costs, ownership structure and
most firms enjoy a favorable high stock price. corporate governance mechanisms. Journal of Banking & Finance
Our results also support Shleifer and Vishny’s (1997) 2003;27(5);793-816.
statement is that large shareholders claim that they both Sun Y, Huang Z. Ownership structure & efficiency of listed
companies in Chinese listed companies, Economic Research 1999;12.
have a general interest in profit maximization, and Tam OK. Ethical issues in the evolution of corporate governance in
enough control over the assets of the firm to have their China. Journal of Business Ethics 2002;37:303-20.
interests respected. However, we also find that large Wang X, Xu LC, Zhu T. Is public listing a way out for state-owned
shareholders are mostly concerned about their personal enterprises? The case of China. Working paper series No 200111.
http://www.ncer.tsinghua.edu.cn/research/paper/PaperContext.asp?yea
benefits, thus failing to improve the asset turnover. The rID=2001
absolute control of listed firms’ large shareholders Yupana. W. An empirical study on the determinants of the capital
makes it difficult for small shareholders to vote against structure of Thai firms. Pacific-basin Finance Journal 1999;7:371-403.
the board’s decision. This results in firms’ inclination of Xu X, Wang Y. Ownership structure and corporate governance in
Chinese stock companies. China Economic Review 2000;10:75-98.
refinancing through stock market and it harms the small Yan Y. An analysis of question between capital structure and agency
shareholders’ interest. costs in Chinese listed firms. Accounting research 2001;9
(www.e521.com).
Correspondence to: Hongxia Li Yishay Y, Yosha O. Large shareholders and banks: Who monitors
School of Management, Harbin University of Science and Technology and how? Manuscript, Hebrew University, Jerusalem, Israel, 1996.
Harbin, Heilongjiang 150080, China Zhang W. Ownership corporate governance and the agency
Telephone: 01186-451-664-2515 relationship. Economic research 1996;9.
Cellular phone: 01186-13030027575 http://www.jjxj.com.cn/news_detail.asp?keyno=189
E-mail: lhx_2002@hotmail.com Zheng J. Agency conflict, equity value and path dependence in
reducing state-hold shares. Conference paper: corporate governance
References reform: China and World, www.iwep.org.cn/cccg.
Ang J, Cole R, Lin J. Agency costs and ownership structure. Journal
of Finance 1999;55:81-106.

http://www.sciencepub.net 20 editor@sciencepub.net