Ownership Structure and Dividend Policy Evidence from China

Ownership Structure and Dividend Policy

Evidence from China

Chenxi Zhong 1752243

MSc International Business & Management Specialization International Financial Management Department of Economics and Business

University of Groningen

Thesis supervisor: Dr. ing. N. Brunia

Table of Content

Abstract ...3

1. Introduction...4

2. Literature Review...7

2.1 Controlling shareholder and dividend policy ...7

2.2 Institutional background in China and hypotheses development ...8

2.3 Other factor of dividend policy ...13

3. Data and Methodology...14

3.1 Data ...14 3.2 Method ...15 3.2.1 Dependent variables ...16 3.2.2 Independent variables...16 3.2.3 Control variables ...17 3.3 Descriptive Statistics ...18

3.3.1 Characteristics of the ownership structure ...18

3.3.2 Simple statistics of the relationship between dividend policy and ownership structure...21

3.3.3 Summary statistics of the dataset ...22

4. Results...25

4.1 Voting rights and dividend payout...26

4.2 Cash-flow-to-voting ratio and dividends...27

4.3 Type of controlling shareholder and dividends ...28

Abstract

1. Introduction

What determines the dividend policy has long been a topic under intensive discussion among theoreticians and practitioners. Economists have offered a range of explanations of dividend payment, such as the effect of taxes, dividend signaling, agency conflict issues and transaction costs. The purpose of this study is to investigate the dividend policy in China from the agency perspective. Particularly, I present evidence that the ownership structure of the firm affects its dividend policy.

Recent empirical research indicates that in most parts of the world except some common law countries where ownership structure is widely-held, many publicly traded firms are controlled by large shareholders (La Porta et al., 1999). The agency conflict in these countries is thus between controlling shareholders and minority shareholders. Jensen and Meckling (1976) and La Porta et al. (2000) consider the dividend policy as either the outcome or the substitute of agency conflicts. Viewed as an outcome, dividend policy is a way for the controlling shareholders to expropriate the minority shareholders if the legal protection of minority shareholders is poor. Viewed as a substitute, dividends are paid to build reputation to attract the minority shareholders.

shareholders significantly negatively affects the dividend payout.

China’s corporate sector evolves from a unique institutional background, which makes the ownership structure in China differ broadly from that in other countries. The ownership structure creates a unique opportunity to analyze whether the ownership structure has an impact on dividend policy or not in China. The controlling shareholder plays a predominant role in China. Most Chinese firms listed in Shanghai and Shenzhen stock exchanges are highly concentrated to the hands of the state (Xu and Wang, 1997, 1999). Since the legal protection of minority shareholders in China is poor, the controlling shareholders are able to expropriate the minority shareholders by dividend policy. Furthermore, a large fraction of the stocks of listed firms are not fully circulated in the Chinese stock market, but held by state-owned enterprises (SOEs) as non-tradable shares at much lower prices than the tradable shares (Lee and Xiao, 2003; Bradford et al., 2006). The non-tradable shareholding by the SOEs is an important factor behind the dividend policy in China. For the other firms controlled by non-state entities, building reputation to attract the minority shareholders becomes a more important reason for paying dividends.

The controlling shareholder can hold the firm directly or through a control chain. However, previous studies on dividend policy in China only focus on the shareholders with direct control (Yuan, 2001; Lee and Xiao, 2003; Lv and Zhou, 2005), which do not show a clear picture of the ownership structure in China. This thesis analyzes the control of the controlling shareholder and the relationship between ownership structure and dividend payout in China in a more appropriate way. It attempts not only to contribute to the dividend literature in general, but also to help the investors, especially the minority shareholders, to understand the dividend behavior more deeply.

hypotheses. The main research question explored in this study is:

Does ownership structure determine the dividend policy of Chinese listed companies?

2. Literature Review

2.1 Controlling shareholder and dividend policy

Based on the investigation of the ownership structure of 27 economies, La Porta et al. (1999) indicate that the existence of controlling shareholders (i.e., the shareholders holding the largest proportion of the votes) is pervasive except in some rich common countries with good legal protection of minority shareholders. Previous research on corporate control in China documents that the controlling shareholders, who can effectively influence the decisions of the managers, control most Chinese listed firms (e.g., Xu and Wang, 1999; Yuan, 2001; Lee and Xiao, 2003). In countries where the legal protection of minority shareholders is poor, the controlling shareholder is the relevant insider to focus on and the agency conflict consequently is the one between the controlling shareholders and the minority shareholders (Shleifer and Vishny, 1997; Maury and Paujuste, 2002; Gugler and Yurtoglu, 2003).

The controlling shareholder can directly control the voting rights or maintain the voting rights indirectly through other firms, i.e. a control chain is formed with the controlling shareholder sitting on the apex. In the latter situation, the controlling shareholder may hold large proportion of votes with relatively small proportion of equity. The controlling shareholders can gain a control position by employing multiple layers of ownership while at the same time sharing the cash flow rights with other shareholders at each intermediate ownership tier (Da Silva et al., 2004).

benefits not shared with minority shareholders by for example allowing other companies under his control to trade with the firm at favorable terms or implementing an investment policy in line with his private preferences (Wallgren, 2006). The larger proportion of voting rights the controlling shareholder holds, ceteris paribus, the larger his influence over the dividend policy and other corporate decision is (Maury and Paujuste, 2002; Gugler and Yurtoglu, 2003).

The power of the controlling shareholder is limited by laws and regulations aimed at protecting minority shareholders. La Porta et al. (2000) suggest that dividend policy can be considered as either an outcome or a substitute of the legal protection of minority shareholders. Viewed as an outcome, the minority shareholders are assumed to be able to put pressure on the firms to pay out dividends if their legal power is sufficiently strong, thus preventing the insiders from generating their own benefits. If the legal protection of minority shareholders is poor, the controlling shareholders have incentives and opportunities to extract cash from the minority shareholders. Viewed as a substitute, if the legal protection of minority shareholders is poor, the controlling shareholder is assumed to raise dividends to establish reputation to signal that no rent extraction of the minority shareholders is taking place; and vice versa. Gomes (2000) predicts that controlling shareholders can implicitly commit not to expropriate minority shareholders; furthermore, paying out dividends may help to develop a reputation for treating minority shareholders well.

2.2 Institutional background in China and hypotheses development

the financial resources under their control without paying dividends to the minority shareholders (Maury and Paujuste, 2002). According to this view, Maury and Paujuste (2002) find a significantly negative relationship between the voting rights of the largest shareholder and dividend payout ratios in Finnish listed companies. Gugler and Yurtoglu (2003) find that controlling shareholder’s voting rights and dividend payments are negatively related in large German firms. Harada and Nguyen (2006) find similar negative relationship in Japan.

In China, the legal protection of minority shareholders is rather weak (Xu and Wang, 1997; Tian, 2001). Due to the predominant role of the controlling shareholders, the minority shareholders do not have enough power to affect important corporate matters (Yuan, 2001; Lee and Xiao, 2003). Dividend policy in China is used as a way not to protect the minority shareholders but to extract cash from them (Lee and Xiao, 2003). However, the institutional background in China makes the incentives to pay dividends different from those in Europe and other Asian countries.

The liquidation of non-tradable shares is strictly restricted. They can only be bought and sold through private placement with special permission from the government (Lee and Xiao, 2003). The non-tradable shares are acquired at the book value, which is far below the market price the tradable shareholders pay at the IPO. According to the investigation of Lee and Xiao (2003), the price of non-tradable shares through private placement is usually less than 20% of the market price of tradable shares. Yet, the non-tradable shares have equal right to dividends as the tradable shares. As for the same dividend per share, ceteris paribus, the investment return ratios of the non-tradable shares held by the state are much higher than those of the tradable shares. Since the state shareholders are not be able to get capital gains in the open market by holding non-tradable shares, Lee and Xiao (2003) predict that liquidation through dividends makes economic sense for the state shareholders in China, and it is a form of expropriation from the non-tradable shareholders on the tradable shareholders. Another form of expropriation is reflected in rights offering. The price of rights offering is always below the market price, but above the book value. Hence, all tradable shareholders would subscribe, but most non-tradable shareholders would not do so. Lee and Xiao (2004) find that non-tradable shareholders only subscribe about 30% of the shares in rights offering whereas tradable shareholders do 100%. Using the receipt from rights offering to pay dividends is equivalent to ask tradable shareholders to pay non-tradable shareholders (Lee and Xiao, 2004). Hence, for the state as the controlling shareholders, dividend policy is a vehicle for them to expropriate the minority shareholders. The controlling shareholders would rather transfer the financial resources from the listed firm by dividends even though these resources left in the firm have higher profitability (Bradford et al., 2006).

the state-controlled counterparts. Building reputation for good performance and not expropriating minority shareholders become more important for the controlling shareholders to decide the dividend policy. Otherwise, they would find it more difficult to raise external funds, and they would lose the minority shareholders’ confidence (Brandt and Li, 2003). Hence, for the firms controlled by non-state entities, dividend policy is a substitute of protection of the minority shareholders.

For both the state and non-state entities as the controlling shareholders, even though the incentives of paying dividends are different, high dividend payout enables them to gain benefits in accordance with their interests, which make the dividend policy resulted from the voting rights of controlling shareholders different from that of other countries. The higher the voting rights the controlling shareholders hold, the more they are likely to keep the dividends high. Thus, the first hypothesis is formulated as follows:

Hypothesis 1: The voting rights of the controlling shareholders determine the

dividend policy of listed firms in China. The voting rights of the controlling shareholders are positively related to the dividend payout.

efficiency of decision making (Jensen and Meckling, 1992) which is important in the highly competitive market environment. The incentives of the managers focus on the profitability and performance of the firms. They prefer to keep the financial resources in the firms instead of paying dividends. As a result, the dividends are lower in these firms than in those directly controlled by the state.

Similarly, for the non-state controlling shareholders holding the voting rights with relatively small proportion of cash flow rights, since they are able to enjoy all private benefits of control by less equity, they have strong incentives to expropriate the minority shareholders and keep the dividends low.

Fan et al. (2007) predict that the more the divergence of cash flow rights and voting rights, the more the controlling shareholders can extract from the minority shareholders. This prediction holds for both the state and non-state entities. Gugler and Yurtoglu (2003) find empirical evidence in Germany that if the controlling shareholder obtains voting rights with limited cash flow rights and the legal protection of minority shareholders is poor, the dividends is expected to be low. Thus, the second hypothesis is formed as follows:

Hypothesis 2: The divergence between cash flow rights and voting rights of the

controlling shareholder is negatively related to the dividends in Chinese listed firms.

The state can easily expropriate the minority shareholders by dividend payout. However, for firms controlled by non-state entities, more dividends also result in less financial resources under the control of the controlling shareholder and more capital the firm needs for more investment opportunities. It seems that non-state entities take more risk and the state has more incentives to pay dividends than non-state entities. Thus, hypothesis 3 is formed as follows:

Hypothesis 3: Dividends are higher in firms controlled by the state than firms

2.3 Other factor of dividend policy

Growth opportunity is a key determinant of the dividend policy (La Porta et al., 2000; Faccio et al., 2001). La Porta et al. (2000) find that with good shareholder protection, companies with rapid growth pay significantly lower dividends than companies with slow growth because the minority shareholders know that when this company’s investments pay off, they could extract high dividends, thus they don’t force the company to pay high dividends. In China where the legal protection of minority shareholders is poor, however, for the firms controlled by the state, the bright prospect makes the controlling shareholder to have more incentive to expropriate the minority shareholders. For firms held by non-state entities, rapid growth requires better reputation the firm has to establish for sufficient external funds. Chen et al. (2007) find that sales growth is significantly positively related to dividends in China.

Based on the empirical study of nine Eastern Asian countries, Claessens et al. (2000) predict that the larger the firm, the more incentives the controlling shareholder can expropriate the minority shareholders. In line with this view, previous research also finds firm size a factor that influences the dividend policy. Faccio et al. (2001) find that dividends are higher in larger companies in Western Europe and Eastern Asia. Lee and Xiao (2003) find similar results in Chinese listed firms.

3. Data and Methodology

3.1 Data

The sample is limited to the period of 2005-2007. For each year, all of the firms listed in the two stock exchanges in mainland China, Shanghai Stock Exchange and Shenzhen Stock Exchange, are included. The construction of the sample is summarized in table 1.

Table 1: Construction of the sample

4391 Chinese listed firms from 2005 to 2007 -135 -49 -22 -296 -53 -5 -15 -18 -1

Firms issuing B-share or H-share Financial institutions No controlling shareholders Negative earnings Dividends > earnings Dividends > sales Growth of sales > 1000% Return on equity < -100% Leverage ratio > 500% 3797 Total sample

with negative earnings or dividends exceeding earnings or dividends exceeding sales are omitted from the sample. Negative dividend ratios may result from other incentives. These incentives may not fit the discussion in previous sections, which are beyond the scope of this paper. Moreover, following Chen et al. (2007), I drop outliers including firms whose GS>1000%, ROE<-100% or LEV>500%. Consequently, the total sample consists of 3797 observations. The sum of market capitalization of sample firms at the end of each year exceeds 36247 billion RMB (about 4869.9 billion USD1).

Data of ownership structure in this study are mainly obtained from firms’ annual reports. Chinese publicly traded companies are required by China Securities Regulation Commition (CSRC) to disclose the large shareholders and the control of these shareholders at each tier since 2001. This regulation offers the possibility to identify the controlling shareholder and his voting rights and cash flow rights. For the companies which do not disclose enough information about the relationships among the large shareholders, especially the natural persons, I search the internet for alternative sources.

Data of dividends and leverage ratio are derived from Cninfo Database (http://www.cninfo.com.cn/). I consult firms’ annual reports for missing data in Cninfo Database and other accounting data for the empirical analyses in this study.

3.2 Method

The hypotheses of this study are regarded to the relationship between dividend policy and ownership structure. The ownership structure of most firms is relatively stable in China, especially for the controlling shareholders (Lee and Xiao, 2003). As a result, although the data of this study cover a period of three years, following Faccio et al.

1

(2001) and Lee and Xiao (2003), models estimated on this dataset possess a cross-sectional dimension, but no time-series dimension. Models based on cross-sectional data can be estimated by Ordinary Least Squares (OLS) if several assumptions are satisfied, which are discussed in the next section.

3.2.1 Dependent variables

Consist with earlier research (e.g., La Porta et al, 2000; Faccio et al, 2001; Maury and Pajuste, 2002), this study chooses dividend-to-earning ratio and dividend-to-sales ratio as the dependent variables to be used as proxies for firms’ dividend payout policy.

Dividend-to-earning (DTE) is the most commonly used measure of dividends payout.

It is the percentage of earnings paid to shareholders in dividends, measured by dividend per share divided by earnings per share. Dividends are the total cash dividends paid to the shareholders. Earnings are measured after taxes and interest but before extraordinary items, i.e. net earnings. This ratio is the indicator of the amount of earnings that have been ploughed back in the business. The lower the ratio, the higher will be the amount of earnings ploughed back in the business and vice versa.

However, dividend-to-earning ratio depends on accounting conventions and can be easily manipulated by accounting tricks (La Porta et al, 2000). Since sales are more difficult to manipulate or smooth, dividend-to-sales (DTS) is used to deal with the drawbacks (La Porta et al, 2000). It is calculated by dividend per share divided by sales per share.

3.2.2 Independent variables

controlling shareholder, following Claessen et al. (1999, 2000), for firms where there are two or more shareholders belonging to the same family, I assume that the family owns and votes collectively, and use the family group as a unit of analysis rather than distinguish among individual family members. The voting rights and cash flow rights of the family are the sum of the family members. However, for the shareholders belonging to different government departments, following Fan et al. (2007), I assume these departments are relatively independent to each other and calculate their voting rights and cash flow rights individually.

Voting (V) is used to test hypothesis 1, which stand for the voting rights of the

controlling shareholder. In line with the measurement of Claessen et al. (1999, 2000) and Faccio et al. (2001), voting is calculated by the sum of the weakest link along the control chain2. The cash flow rights of the controlling shareholder are measured by the sum of the product of ownership stakes along the control chain3 (Claessen et al., 1999, 2000).

Cash-flow-to-voting ratio (CTV) is employed to represent the divergence between

cash flow rights and voting rights of the controlling shareholder. The lower the ratio, the higher the divergence is. It is used to test hypothesis 2.

State is a variable to identify the type of the controlling shareholder, which equals one

if the controlling shareholder is the state and zero if the controlling shareholder is a non-state entity. It is used to test hypothesis 3.

3.2.3 Control variables

Furthermore, I acknowledge that other factors may influence the dividend policy and

2

Suppose, for example, investor X owns 15% of firm A, which in turn has 12% of the shares of firm B. At the same time, investor X owns 10% of firm C, which in turn has 8% of the shares of firm B. According to the definition, the voting rights of investor X in firm B are 22% (12%+10%).

therefore I choose additional relevant variables to control for effects on the dividend policy not captured by the ownership structure.

Growth rate of sales (GS) measures the growth opportunity of firm, calculated by the

average annual percentage growth in net sales over the last three years.

Size is used as a control variable, which equals the natural logarithm of the book value

of total assets at the end of the year.

Return on equity (ROE), measured by net income divided by total shareholders’

equity at the end of the year, is a variable to control the effect of profitability.

Leverage ratio (LEV) is measured by the book value of debt divided by total assets at

the end of the year.

3.3 Descriptive Statistics

3.3.1 Characteristics of the ownership structure

63.95% of the sample firms are controlled by the state. These firms accounts for 79.62% of the market capitalization, which is 28860 billion RMB. In line with the literature, the state-controlled firms play predominant roles in Chinese listed firms. According to Faccio et al (2001), the state plays much less predominant role in West Europe and East Asia, accounting for less than 5% of the ultimate shareholders. For the other firms controlled by non-state entities, the controlling shareholders are dispersed widely. Only six of the entities own more than one listed firms.

shareholder; otherwise, the firm is widely held. They find that the control of ultimate shareholder is prevalent in most of the world, especially in countries where the protection of minority shareholders is poor. Faccio et al (2001) find that over 60% of the companies in West Europe and East Asia are held by the ultimate shareholders. This proportion is even higher in China. As presented in Table 2, the control of ultimate shareholder exists in a large majority of Chinese listed firms. 3404 out of 3797 of the sample firms are controlled by the ultimate shareholders, accounting for 89.65% of the total. The proportion of firms with ultimate shareholder is slightly higher for those controlled by the state than those controlled by non-state entities (90.90% vs. 87.44%). 23.81% of the controlling shareholders control the firms with the voting rights between 20% and 30%, which is the range of the most frequent proportion of votes they hold to control the firms. Specifically, more non-state entities prefer this range of proportion. 32.94% of them control the firms with the voting rights within it. Moreover, the state is more likely to control more than 60% of the voting rights than non-state entities (10.38% vs. 6.21%).

Table 2: Distribution of voting rights by type of the controlling shareholders Number of firms (Proportion of the total)

Total sample State controlled Non-state entities controlled

Table 3 presents the distribution of cash flow rights of the controlling shareholders. Compared to the voting rights, the controlling shareholders prefer to hold the cash flow rights at a relatively lower level. The most prevalent proportion is below 20%. 26.57% of the controlling shareholders hold the voting rights at this level. For the non-state controlling shareholders, this preference is even higher. 48.43% of them hold the cash rights at this level. Whereas, there are only 18.37% of the state shareholders whose cash flow rights are less than 20%. Generally speaking, the state holds more cash flow rights than non-state entities. Over 40% of the state shareholders hold more than 30% of the cash flow rights. While less than 30% of the non-state entities do so.

Table 3: Distribution of cash flow rights by type of the controlling shareholders Number of firms (Proportion of the total)

Cash flow rights Total sample State controlled Non-state entities controlled

<20% 1009 (26.57%) 446(18.37%) 563 (48.43%) 20%-30% 870 (22.91%) 528 (21.75%) 342 (24.98%) 30%-40% 693 (18.25%) 506 (20.84%) 187 (13.66%) 40%-50% 595 (15.67%) 420 (17.30%) 175 (12.78%) 50%-60% 436 (11.48%) 366 (15.07%) 70 (5.11%) 60-70% 127 (3.35%) 107 (4.41%) 20 (1.46%) >70% 67 (1.77%) 55 (2.26%) 12 (0.88%) Total 3797 (100%) 2428 (100%) 1369 (100%)

Table 4: Distribution of cash-flow-to-voting ratio by type of the controlling shareholders

Cash-flow-to-voting Total sample State controlled Non-state entities controlled

<0.2 79 (2.08%) 18 (0.74%) 61 (4.46%) 0.2-0.4 282 (7.43%) 76 (3.13%) 206 (8.48%) 0.4-0.6 455 (11.98%) 237 (9.76%) 218 (15.92%) 0.6-0.8 448 (11.80%) 236 (9.72%) 212 (15.49%) 0.8-1 482 (12.69%) 179 (7.37%) 303 (22.13%) 1 2051 (54.02%) 1682 (69.28%) 369 (26.95%) Total 3797 (100%) 2428 (100%) 1369 (100%)

3.3.2 Simple statistics of the relationship between dividend policy and ownership structure

Table 5 presents the means of the dividend-to-earning ratio and the means of the

dividend-to-sales ratio for all sample firms, for state controlled firms and for

non-state controlled firms and for different classes of proportions of voting rights. This table reveals that the means of both dividend ratios increase as the voting rights of the controlling shareholders increase; the means of both dividend ratios for the state-controlled firms are higher for the firms controlled by non-state entities. These results are in line with the hypotheses in this study.

Table 5: Means of dividend-to-earning ratio and dividend-to-sales ratio by the voting rights and type of controlling shareholders

Means of dividend-to-earning ratio Means of dividend-to-sales ratio

3.3.3 Summary statistics of the dataset

Table 6 presents the summary statistics of the dataset by showing the numbers of observations, mean, standard deviation, minimum, maximum, skewness, kurtosis of the variables and the results of Kolmogorov-Smirnov test.

Table 6: Summary statistics of the dataset

Variable N Min. Max. Mean Standard

Deviation Skewness Kurtosis

Kolmogorov -Smirnov Sig. K-S Voting 3797 6.21 92.00 38.50 15.377 0.361 -0.307 0.053 0.000 Cash-flow -to-voting 3797 0.02 1 0.822 0.251 -0.18 0.137 0.202 0.000 Dividend-to -earning 3797 0 1 0.192 0.232 0.766 0.237 0.210 0.000 Dividend-to -sales 3797 0 0.423 0.022 0.044 0.88 1.122 0.208 0.000 Growth of sales 3797 -48.42 883.20 28.22 59.061 7.933 85.324 0.256 0.000 Size 3797 7.13 11.99 9.13 0.574 0.125 0.101 0.018 0.000 Return on equity 3797 -78.91 51.28 3.48 10.169 19.954 356.672 0.363 0.000 Leverage 3797 0.91 98.01 48.81 17.868 7.933 85.324 0.029 0.003

The results of Kolmogorov-Smirnov tests indicate that all the variables are normally distributed. Skewness measures the degree and direction of asymmetry; kurtosis is a measure of the heaviness of the tails of a distribution (UCLA Academic Technology Service). High positive values of skewness and kurtosis occur for growth of sales,

return on equity and leverage. However, nearly normal distributions have skewness

Figure 1: Normal P-P Plot of Leverage

Normal P-P Plot of LEV

Observed Cum Prob

1.00 .75 .50 .25 0.00 E xp e ct e d C u m P ro b 1.00 .75 .50 .25 0.00

As the modeling approach used in this study requires normality, appropriate data transformations are sought to improve normality. Following La Porta et al. (2000) and Wei et al. (2003), firms are partitioned into ten equal-size groups in ascending order of average growth of sales and ranked 1-10. The ranks are the transformed data for growth of sales. Similar transformation is applied to return on equity as well. Growth

of sales decile and return on equity decile stand for the variables with the transformed

data. Both of them are normally distributed after the transformation. The empirical results presented in the next section are based on transformed data.

presence of non-linear relations (UCLA Academic Technology Service). Figures in Appendix A present the outcome of an example of the residual diagnosis for one of the estimated linear regression models with dividend-to-earning as the dependent variable and voting as the independent variable. Such residual diagnosis is applied on all models estimated in this study. The results indicate that the assumptions are acceptable.

Another assumption for OLS is the absence of collinearity. Collinearity can be investigated before the estimation of the regression by checking bivariate correlations of independent variables and control variables, the results of which are reported in table 7. The high correlations among independent variables present a potential case of collinearity. A possible solution is not to apply the independent variables in one model, but to run the regressions using one of them at a time.

Table 7: Bivariate correlations of the independent variables and the control variables

V CTV State GS decile ROE decile Size LEV

V 1 CTV 0.166** 1 State 0.161** 0.345* 1 GS decile 0.125* 0.065 -0.052 1 ROE decile 0.159 -0.011 -0.093 0.36* 1 Size 0.182 0.028 0.121** 0.235 0.022 1 LEV -0.046 -0.041 0.068* 0.1 0.065 0.153 1

4. Results

Table 8 presents the regression results of dividend-to-earning and dividend-to-sales on the variables of ownership structure for all sample firms. As argued before, different types of controlling shareholders have different incentives to influence the dividend payment. Hence, I rerun the regressions for the firms controlled by the state and for the firms controlled by non-state entities. The results are shown in table 9. In all the regressions, model (1), model (2) and model (3) are applied to test hypothesis 1, hypothesis 2 and hypothesis 3 respectively.

Table 8: Regression results of dividend policy on ownership structure for all firms

Table 9: Regression results of dividend policy on ownership structure for the firms controlled by the state and for the firms controlled by non-state entities

State-controlled firms Non-state-entities-controlled firms

DTE DTS DTE DTS Model (1) (2) (1) (2) (1) (2) (1) (2) Intercept 0.123*** (3.689) 0.191*** (3.314) 0.024*** (3.827) 0.028*** (3.423) 0.066** (1.916) 0.107*** (2.914) 0.003* (0.470) 0.001 (0.068) Voting 0.138*** (3.989) 0.101** (3.037) 0.099** (2.114) 0.111** (2.426) Cash-flow-to -voting 0.029 (2.251) 0.033 (1.003) 0.001** (0.022) 0.070** (1.539) State 0.077** (2.081) 0.089** (2.389) 0.042 (1.081) 0.034 (0.961) 0.056* (1.616) 0.063* (1.258) 0.029 (0.617) 0.033 (0.688) Growth of sales decile 0.024** (1.637) 0.032** (1.918) 0.025* (1.121) 0.024** (1.160) 0.024** (1.637) 0.032* (1.918) 0.025 (1.121) 0.024* (1.160) Size 0.113*** (2.974) 0.131*** (3.540) 0.241*** (6.777) 0.257*** (7.262) 0.070* (1.426) 0.085* (1.737) 0.205*** (4.256) 0.220*** (4.596) Return on equity decile -0.154*** (-4.528) -1.162*** (-4.731) -0.275*** (-8.447) -0.281*** (-8.593) -0.197** (-2.098) -0.193** (-2.217) -0.084** (-1.843) -0.091** (-2.007) R2 0.079 0.071 0.155 0.147 0.057 0.058 0.087 0.085 Adjusted R2 0.074 0.065 0.151 0.142 0.049 0.049 0.082 0.077 F 17.265*** 15.046*** 37.029*** 34.625*** 14.444*** 14.425*** 18.945*** 18.107***

*, **, *** Significant at the 10%, 5%, and 1% levels, respectively.

4.1 Voting rights and dividend payout

entities pay higher dividends to build reputations for good performance and not expropriating minority shareholders. Controlling higher voting rights enables them to do so.

The results are in line with the discussion in previous section, which predicts that the unique institutional background in China makes the dividend payout resulted from the voting rights of controlling shareholders different from that of Europe (Maury and Paujuste, 2002; Gugler and Yurtoglu, 2003) and other Asian countries (Harada and Nguyen, 2006). Several literature in China, e.g. Yuan (2001), Lee and Xiao (2003) and Lv and Zhou (2005), find similar results to this study, even though the voting rights measured in these studies are based on the direct control.

4.2 Cash-flow-to-voting ratio and dividends

Hypothesis 2 expects a positive relationship between cash-flow-to-voting ratio and dividends. As shown in table 8 model (2), cash-flow-to-voting ratio is significantly positively related to dividend-to-sales ratio at the 10% level; however, no significant relationship is found between cash-flow-to-voting ratio and dividend-to-earning ratio. Moreover, the results of model (2) in table 9 show that, for firms controlled by the state, dividend-to-earning ratio and dividend-to-sales ratio are positively but insignificantly related to cash-flow-to-voting ratio. For the firms controlled by non-state entities, the relationship between cash-flow-to-voting ratio and both dividend payout ratios is significantly positive at the 5% level, which indicates that if a non-state entity controls the voting rights with relative lower cash flow rights, she intends to keep the dividends at a lower level. Therefore, hypothesis 2 is supported for only the firms controlled by non-state entities.

4.3 Type of controlling shareholder and dividends

Model (3) of table 8 tests the effect of the type of controlling shareholder and dividend policy. I hypothesize that state-controlled firms have higher payout ratios than firms controlled by non-state entities. Since on average the proportion of voting rights of the controlling shareholder is higher in state-controlled firms than in non-state-controlled firms, control for the proportion of voting rights when testing hypothesis 3. Table 8 presents a highly significant and positive relationship between state and both dividend payout ratios, which indicates that state control results in higher dividends. Hence, hypothesis 3 is supported. This result is consistent with Wei et al. (2003) and Bradford et al. (2006).

4.4 Other findings

Growth of sales decile is found to have significant and positive effect on dividend-to-earning ratio in all the regression models, which means that dividend-to-earning ratio is higher in firms with better investment opportunities. However, there is no significant relationship between growth of sales decile and dividend-to-sales ratio. Furthermore, size and return on equity decile are positively related to both dividends ratios in most of the models, and the relationship between leverage and dividend ratios is significantly negative. The interpretation is that larger, more profitable firms and firms with lower debts to assets pay higher level of dividends. These findings are consistent with previous discussion.

4.5 Robustness check

re-estimate the models using industry-adjusted dividends ratios. Following La Porta et al. (2000), all companies in the sample are grouped into twelve industry categories identified by CSRC after dropping the financial industry.4 For each company in a given industry, the industry-adjusted dividends variable is measured by the difference between the company’s dividends ratio and the median dividends ratio in this industry.

Moreover, to smooth out noise and transitory factors, for each firm of each year, I apply the average dividend ratios for the last three years as the dependent variables to re-examine the models. For example, the average dividend-to-earning ratio of firm in 2007 is calculated as the means of dividend-to-earning ratios from 2005 to 2007. For the firms with incomplete data for three years, I calculate the average value over the years with available data.

Regression results of robustness check are presented in appendix B. It seems that adjusting the effect of industries and considering the effect of time in the regressions do not alter the results a lot except that cash-flow-to-voting ratio is found to be significantly and positively related to dividends ratios in all models. Hypothesis 2 is supported, which suggests that firms whose controlling shareholder owns the voting rights with relatively lower cash flow rights pays lower dividends. In addition, the relationship between growth of sales decile and both dividend ratios is significant. Therefore, consistent with the prediction in previous section, dividends are higher in firms with better investment opportunities.

4

The industries include: (1) Agriculture, forestry, herd and fishery; (2) Mining; (3) Manufacturing; (4) Electricity, gas and water supply; (5) Construction; (6) Traffic, storage and post; (7) Electron city; (8) Wholesale and retail; (9) Estate; (10) Social service; (11) Culture, sports and entertainment; and (12) General.

5. Conclusion

The purpose of this study is to investigate how dividend policy is related to ownership structure in Chinese listed firms from the agency perspective. Specifically, I examine the effects of the voting rights of the controlling shareholder, the divergence between cash flow rights and voting rights and the type of controlling shareholder on dividend payout. The findings of this paper are consistent to the agency model stressed in Jensen and Meckling (1976) and La Porta et al. (2000), according to which, due to the weak legal protection of minority shareholders, dividend policy is a vehicle used by the controlling shareholder to expropriate the minority shareholders.

firms tend to pay higher dividends than the firms controlled by non-state entities.

Based on the empirical analysis of 3797 Chinese listed firms issuing only A-share over the period 2005-2007, I find that dividend payout is significantly positively related to the voting rights of the controlling shareholders. I also find a significant and positive relationship between dividend payout and cash-flow-to-voting ratio for all the sample firms after adjusting the effect of either industry or time period, which means that more divergence of cash flow rights and voting rights of the controlling shareholders results in higher dividend payout. Furthermore, the firms controlled by the state pay higher dividends than those controlled by non-state entities. Moreover, dividend payout is positively related to firm’s growth opportunities, profitability and size as well, and leverage ratio has negative impact on dividend policy.

This study is by far not giving a full picture of ownership structure and incentives to pay dividends in Chinese listed firms. There are some limitations in this study; nevertheless, it opens an interesting discussion on how different ownership structures influence firm behavior and dividend policy in China.

As an important aspect of ownership structure, the form of control chain is a hot topic in recent years. However, the measurement of the voting rights and cash flow rights in this study is simplified for convenience and to some extent inaccurate, so the influence of the divergence between cash flow rights and voting rights due to the control chain may be controversy. How to accurately proxy the efficiency and power of control chain and thus the impact on dividend policy should attract more attention in the future.

References

Berle, A., & Means, G. 1932. The Modern Corporation and private property. New York, Macmillan.

Black, F. 1976. The dividend puzzle. Journal of Portfolio Management, 2: 5-8.

Bradford, W., Chen, C., & Zhu, S. 2006. Ownership structure, control chains, and cash dividend policy: Evidence from China. Sun Yat Sen University's Securities and Financial Markets International Conference, Kaohsong, December 2006.

Brandt, L., & Li, H. 2003. Bank discrimination in transition economies: Ideology, information or incentives? Journal of Comparative Economics, 31: 387-413.

Claessens, S., Djankov, S., & Lang, L. H. P. 2000. Expropriation of Minority Shareholders in East Asia. Journal of Financial Economics, 58: 81–112.

Claessens, S., Fan, J. P. H., & Lang, L. H. P. 2002. The benefits and costs of group affiliation: Evidence from East Asia. CEPR Discussion Paper, 3364.

De Angelo, H., De Angelo, L., & Skinner, D. 1992. Dividends and losses. Journal of

Finance, 47: 1837-1863.

Easterbrook, F. H., 1984. Two agency-cost explanations of dividends. American

Economic Review, 74: 650-659.

Faccio, M., Lang, L. H. P., & Young, L. 2001. Dividends and expropriation.

Fama, E. F., & French, K. R. 2001. Disappearing dividends: Changing firm characteristics or lower propensity to pay. Journal of Financial Economics, 60: 3-43.

Fan, J. P. H., Wong, T. J., & Zhang, T. 2007. Organizational structure as a decentralization device: Evidence from corporate pyramids. Working Paper.

Gomes, A. 2000. Going public without governance: Managerial reputation effects.

Journal of Finance, 52 (2): 615–646.

Gugler, K., & Yurtoglu, B. 2003. Corporate governance and dividend pay-out policy in Germany. European Economic Review, 47: 731-758.

Holz, C. A. 2007. Have China scholars all been bought? Far Eastern Economic

Review, 170(3): 36–40.

Jensen, M., & Meckling, M. 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3: 305-360.

La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. 1999. Corporate ownership around the world. Journal of Finance, Vol. 54, No. 2: 471-517.

La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. 2000. Agency problems and dividend policies around the world. Journal of Finance, 55: 1-33.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R.W. 2002. Investor protection and corporate valuation. Journal of Finance, 57: 1147–1170.

Lintner, J. 1956. Distribution of incomes of corporations among dividends, retained earnings, and taxes. American Economic Review, 46: 97-113.

Liu, Q. 2006. Corporate governance in China: current practices, economic effects and institutional determinants. Economic Studies, 52(2): 415-45.

Liu, G. S., & Sun, P. 2003. Identifying ultimate controlling shareholders in Chinese public corporations: An empirical survey. Asia Program Working Paper, No. 2, June 2003.

Lv, C., & Wang, K. 1999. Empirical analysis on the dividend policy of listed companies in China. Economic Research Journal, 12: 31-39.

Maury, C., & Pajuste, A. 2002. Controlling shareholders, agency problems, and dividend policy in Finland. Working Paper, Stockholm School of Economics.

Miller, M. H., & Modigliani, F. 1961. Dividend policy, growth, and the valuation of shares. Journal of Business, 34: 411-433.

Qiao, L. 2006. Corporate governance in China: current practices, economic effects and institutional determinants. CESifo Economic Studies, 52(2): 415-453.

Shleifer, A., & Vishny, R. 1986. Large shareholders and corporate control. Journal of

Political Economy, 94: 461–488.

Shleifer, A., & Vishny, R. 1997. A survey of corporate governance. Journal of

Finance, 52: 737–783.

the Chinese public corporation. Management and Organization Review, 4(1): 17-38.

Tang, Y., Xie, R., & Zhang, X. 2007. Counterbalance mechanism of blockholders and tunneling of cash dividend: Evidences from Chinese listed companies from 1999 to 2003. Frontiers of Business Research in China, 1 (4): 521-543.

Tian, L. 2001. Government shareholding and the value of China’s modern firms.

William Davidson Institute Working Paper, No. 395.

Wei, J. G., & Xiao, J. Z. 2005. Shareholder Preferences and Dividend Payment in China. Working Paper. Cardiff Business School, Cardiff University.

Xu, X., & Wang, Y. 1997. Ownership structure, corporate governance and corporate performance: The case of Chinese stock companies. World Bank Policy Research

Working Paper, 1794, June.

Xu, X., & Wang, Y. 1999. Ownership structure and corporate governance in Chinese stock companies. China Economic Review, Vol. 10: 75-98.

Yuan, H. 2001. Analysis on the dividend policy of listed companies in China. The

Study of Finance and Economics, 27: 33-41.

Zou, H., & Xiao, J. Z. 2006. The financing behavior of listed Chinese firms. The

British Accounting Review, 38: 239–258.

Internet sources:

Cninfo Database: http://www.cninfo.com.cn/

Scatterplot

Dependent Variable: DTE

Regression Standardized Predicted Value

5 4 3 2 1 0 -1 -2 -3 R e g re ss io n S tu d e n ti ze d D e le te d ( P re ss ) R e si d u a l 10 8 6 4 2 0 -2 -4

Appendix

Appendix A

Normal P-P Plot of Regression Standardized Residual Dependent Variable: DTE

Observed Cum Prob

Appendix B

Table B-1: Regression results of dividend policy on ownership structure for all firms adjusting industry effects

Table B-2: Regression results of dividend policy on ownership structure for all firms with three-year average data