The impact of Ownership structure on M&A premium

(1)

The impact of Ownership structure on

M&A premium

University of Amsterdam Master of International Finance Supervisor: Jens Martin

Name: Siqi Zhou (11353007) Academic year: 2016-2017

(2)

Abstract

This thesis focuses the impact on M&A premiums from ownership structure and tries to look into the internal mechanism of the impact. Samples are chosen from

companies listed on Shenzhen Stock Exchange, Shanghai Stock Exchange, Shenzhen SME Board Stock Exchange and Shenzhen Stock Exchange Chi Next Board and the M&A cases happened in 2007 to 2016 ten years duration. In order to conduct

regression models, ownership structure is deconstructed into five measureable factors- ownership concentration, ownership balance degree, nature of actual controller, stock liquidation and ratio of transferred stock. The gap between transfer price and stock price is used to measure M&A premium. The relationship between M&A premium and ownership structure is examined by combining multiple linear and nonlinear regression model. The study finds out: (1) Ownership concentration level and M&A premium have U Curve relationship; (2) Ownership balance degree and ratio of transfer shares are negatively related to M&A premium; (3) Ratio of tradable shares is positively correlated with M&A premium. Based on the empirical studies, some conclusions and suggestions are come up to give enlightenment to companies carry out M&A in the future and Chinese M&A market.

Key words: Ownership Structure; M&A premium; Ownership Concentration; Agency Cost; Chinese Listed Companies

(3)

1. Introduction ... 1

2. Literature Review ... 3

3. Hypotheses ... 5

4. Measures & Methodology ... 8

4.1 Measures ... 8

4.2 Methodology ... 11

5. Data Analysis and Results ... 12

6. Discussion and Conclusion ... 25

(4)

1. Introduction

Stock market has the functions of stock issuance and trading. The core of stock trading is the stock price. A reasonable stock pricing system is important for the development of capital market and economic. Stock price is actually the external expression of ownership price because different ownership structures lead to the difference in stock prices. Ownership structure can be represented by ownership concentration, ownership balancing degree, the nature of shareholders, mobility and transfer of equity stake. Theoretically, all these elements should have impact on stock price during trading of stocks.

When choosing a start-up project, venture funds and private equities not only care about the business model, start-up team, operating condition and development of future, but also pay attention to the exit methods. Exit methods include initial public offering, secondary buyout and M&A. Nowadays in China, the process of IPO on main boards is very long and complicated. Companies have to meet a long list of requirements which are difficult for most of them. Secondary buyout is not common because once funds decide to invest on a project they will keep the investment. Exiting on the next round offering would frustrate the confidence of other investors. Therefore, merger and acquisition is the most appropriate and common way to exit. For example, DiDi and KuaiDi, 51 job and GanJi, MeiTuan and DianPind and Ctrip and Qunar are influential M&A cases during the last few years.

In the M&A cases, prices of the deals especially the premiums are the focus of both the two sides. M&A premiums concern the relocation of rights and obligations of

stakeholders, the future operation of companies and financial management. So how ownership structure affects M&A premiums makes the topic of this thesis. Existing studies show that M & A does not necessarily create value for companies, which is at

(5)

least partly attributable to high acquisition premiums in M&A (Hunter & Jagtiani, 2003). So what factors affect the acquirer's premium? Answer to this question mainly focus on 3 levels: First on market level, such as the market auction hypothesis believes that the presence of one or more competitors in the process of mergers and acquisitions makes the competition for company control become fierce (Slusky & Caves, 1991). Acquiring companies will pay relatively high premium in order to gain control right. Second on enterprise level, synergy hypothesis argues that when both assets of sides join together in certain way, the outcome is better than when they are independent. When the synergy is higher, the acquirer is willing to pay the premium level is higher (Sirower, 1997). Third on ownership level, principal-agent hypothesis tells that because executives act for their own interests rather than the interests of shareholders, they participate in M&A in order to promote reputation, power and compensation (Jensen & Meckling, 1976). In the process of transformation of China economic environment, the ownership structure of companies has gone through large change. As there are already many studies focus on effect of M&A from market and enterprise aspects, I would like to dig into the internal relationship between Chinese company ownership structure and M&A premium and try to find out the mechanism.

Generally, this thesis will look into the mechanism of the impact of ownership structure on M&A premiums. Frist build the theoretical framework of the effect ownership structure has on cost of principal agent, company governance and profitability, and further on M&A premiums. On one hand, the segregation of managing and controlling can result in fraud of managers. On the other hand, differences exist in the synergy effect of M&A companies and liquidity of purchased shares. Both these issues can account for different degree of M&A premiums.

Based on the theoretical framework, this thesis will quantitatively measure ownership structure, company performance and M&A premiums and carry out descriptive

(6)

statistical analysis on the data. Then ownership structure will be constructed on the impact of M&A premiums mechanism regression model. In the empirical research, the unique features of Chinese stock market and ownership structure will be included. And constructive advices for M&A regulations will be raised based on the analysis

outcomes.

2. Literature Review

This chapter offers an overview of previous related studies and literature on main theories and terms of M&A and ownership structure, which are used to give a dynamic and hopefully comprehensive insight.

One company’s ownership structure depends on, first the characters of different shareholders; second the rate of shares that different shareholders own. The shareholders enjoy rights of voting, receiving dividends according to their shares portion.

Berle & Means (1969) finds that the distribution of corporate control right is decided by the rate of stocks on hand. Therefore, acquirers would like to pay a premium to gain the control of the company (Dimopoulos & Sacchetto, 2014). As long as the premium is smaller than the profit and the synergistic effect they can get with the right of control, the merger is going to process (Barclay, 1989). Other aspects also lead to higher premium. For example, the antitakeover compensation for managers (Sokolyk, 2011) and other antitakeover actions of acquired companies (Souther, 2016). From the perspective of corporate governance, John et al. (2010) argue that the target companies with better investor’s protection will gain higher premium. When acquirers and target companies are in oligopoly competition, or they are competitors in stock market, the premium of M&A would be higher (Spiegel & Tooles, 2013).

(7)

On the other hand, the of CEO or management of acquiring companies also have big influence on the buyout price. El-Khatib (2015) find that overconfidence and network centrality of CEOs can cause them to offer an excessive premium. The overconfident CEOs always overvalue their own abilities and positively estimate the value of

mergers (Roll,1986). Considering the ownership structure of acquiring companies, the more dispersive the stock equity distribution, the higher M&A premium (Ferris et al., 1997). Government or pulic companies would offer higher premium than private corporate (Bargeron et al. 2008). When the financial index of acquiring companies is better, for example solvency ability and liquidity, the premium will be higher

(Gondbalekar et al. 2004).

For the Chinese domestic research, Chen (2011) finds that the less ownership

concentration, the more widely distributed the stocks, the higher ownership balancing degree and the better company performance. Thus, the valuation of less ownership concentrated company would be higher and should be offer a purchase premium. With the exception, Chen & Lu (2013) think under the situation of top management tie, companies with greater ownership concentration earn higher premium. According to Zhang et al. (2005)’s research, Chinese government shareholding ratio and agency cost present positively correlational dependence. Considering the correlation between proportion of largest shareholder and agency cost, there are different opinions on it. As agency cost of target company and price premium are negatively related, the effect of ownership concentration is under debate. Xiao, Chen (2006) argue that as the proportion of the largest shareholder increase, the agent conflict between controlling shareholders and small shareholders get severe and thus the agency cost increase. Nevertheless, with one big shareholder, the supervise on operator is easier and the company management is more stable. Under Chinese laws, the protection of investors is relatively lower, and one big shareholder situation is more suitable. Thus, the agency cost should be less when the proportion of largest shareholder increases.

(8)

To Sum up, existing literatures is focusing on the connection between corporate financial index and M&A premium or ownership concentration and M&A premium. However, there is still conflict on this topic when considering Chinese characteristic policy.

3. Hypotheses

3.1 Ownership Concentration and M&A Premium

Different forms of ownership arrangement will have different forms of agency problems. When the stock right is highly dispersed, there is a "free rider" mentality among the shareholders, which cannot effectively supervise the managers. Managers may act in their own interests against the maximization of shareholders' interests. The conflicts between the shareholders and managers of the cause is known as the first agency problem; and when the concentration of ownership in the hands of the controlling shareholder, although the first agency problem properly relieved, but holding between shareholders and outside shareholders may have more serious conflicts of interest. Controlling shareholders may take advantage of their control position to engage in self-interest and cause damage to the interests of external shareholders. This kind of agency problem caused by the conflict of interest between shareholders is called Agency Problem of the Second Type.

According to Du and Gan (2012), the first large shareholder is in favor of corporate operation performance. The management level moral hazard caused related to Agency Problem of the First Type will decrease. Therefore, when the proportion of largest shareholder gets higher, Agency Cost of First Type gets lower and M&A premium increases. However, the largest shareholder has to share the profit that gained from the

(9)

strengthening of corporate governance and operation with small shareholders. As the control over the company grows, the largest shareholder tends to gain personal profit and plunder company equity through Tunneling or Related Transaction. This behavior impairs the interests of minority shareholders and leads to increase of Agency Problem of Second Type.

Hypothesis 1: The ownership concentration and M&A premium have U curve relationship

3.2 Ownership Balance and M&A Premium

Song & Han (2005) offered empirical evidences with data of Chinese listed companies showing that ownership balance degree is positively relative to agency costs. In a structure with low ownership balance degree, the big shareholder’ supervision on agent manager is hard to be intervened by other shareholders. But as the balance degree increasing, other shareholders would give more limitations on big shareholder’s behavior and decisions. The contradiction between shareholders would be solved at a cost of company benefits. The time and energy they spend on supervision of agent managers would be reduced, so agency costs would increase. And higher agency costs lead to higher premium.

Hypothesis 2: The ownership balance and M&A premium have negative relationship.

3.3 Nature of Actual Controller and M&A Premium

The nature of the largest shareholder is especially important. If the company is state-owned, then it will have different standards for corporate development and company performance than private enterprise. State-owned companies should consider

(10)

stabilizing employment, enhancing enterprise welfare and achieving national strategic objectives. When M&A happens, it doesn’t take obtaining higher price as the only purpose. Instead it should consider the social benefits, qualification of acquiring companies and employee benefit issues. Therefore, if the actual controller is the country or government, the level of price premium will reduce.

Hypothesis 3: If the actual controller is the country or government, the level of price premium will reduce.

3.4 Stock Liquidity and M&A Premium

Amihud and Mendelson (1986) come up with the liquidity premium theory, saying that equity with low liquidity should be priced lower and on the contrary equity with high liquidity should be given more value. An interesting fact is that if managers sell their 5-year restricted shares with 20% to 70% discount, they will get more profit than holding them (Kahl et al., 2003). This means price of restricted shares will be lower than unrestricted shares when they are sold. Therefore, when a company is being merged, the higher the ratio of unrestricted shares, the more liquid its stock is. The acquiring company can liquidate the stock easily and it will pay higher premium.

Hypothesis 4: Stock liquidity has a positive relation with M&A premium.

3.5 Transfer of Shares and M&A Premium

When acquiring company gets to buy more shares of target company, its ownership ratio is greater and its control over the operation and management of target company is stronger. But at the meantime, the shares it holds is harder to liquidate and the price would be given a discount. Thus, I can have another hypothesis.

(11)

Hypothesis 5: The greater transfer of shares, the lower M&A premium.

4. Measures & Methodology

4.1 Measures

When choosing the variables, I seek to cover various characteristics of a Merger & Acquisition and factors corresponding with ownership structure. All the variables captured in the analysis are listed and defined in Table 2.

Dependent Variables

In General, there are two mainstream ways to measure M&A premium: (1) The

indirect way is to calculate the excess returns that target company gets from the M&A. (2) The direct way is to calculate the difference between the purchasing price that acquiring company offers and the actual stock price.

For Chinese domestic researches, Han & Ye (2004), Cheng (2009), Dai (2011) and Ge et al. (2014) all use the first way to measure M&A premium. The formula is as following:

M&A Premium= (Transfer Price per Share – Total Equity per Share) / Total Equity per Share

Using the second way, Moeller (2005) take share price on six days before announcement of M&A into account. The formula is as following:

(12)

M&A Premium= (Transfer Price per Share – Closing price on six days before announcement of M&A) / Closing price on six days before announcement of M&A

Based on the reality of Chinese stock market, I choose closing price on the announcement day

of M&A. The formula is as following:

M&A Premium= (Transfer Price per Share – Closing price on announcement day of M&A) / Closing price on announcement day of M&A

Considering that stock price is affected by a lot factors and has high variability, I will use the

average closing price of 10, 20 and 30 days before and 10, 20, and 30 days after the

announcement date to run a Robustness Test. The formula is as following:

M&A Premium= (Transfer Price per Share – Average Closing price of 10(20/30) days before (after) announcement date) / Average Closing price of 10(20/30) days before (after)

announcement date

Independent Variables

Based on the analysis and logics, I break down ownership structure into five aspects including ownership concentration, ownership balancing degree, nature of actual control, stock liquidity and transfer of equity stake. In order to carry out the empirical research, I use five variables to measure each aspect. First, I use proportion of the largest shareholder to measure ownership concentration. Second, I use ratio of the second to tenth largest shareholders to measure ownership balance. Third, nature of actual control can be represented by whether the actual controller is state-owned (Li,

(13)

2009). Fourth, stock liquidity can be measured by the ratio of unrestricted shares. Fifth, I use the ratio of transferred shares in total shares to represent transfer of shares.

Control Variables

Following the prior researches in M&A, a number of control variables are taken into consideration as described below. Company Size, prior studies state that the less intense competition, greater equity value and sizeable complexity of post-merger integration associated with large targets can lead to lower M&A premium

(Alexandridis et al., 2011). In Chinese stock market, there is a domain of blue chips of big size companies which have stable profitability and small stock price volatility. I use market value of Target Company to measure target size as a control variable. Growth Ability, which can be measured by Book to Market, tells the prospect growth potential if target company and should be taken as a control variable.

Financial theory suggests that macroeconomic situations can affect stock market and thus stock prices and returns. Considering the M&A cases happened in a 10 year range time and stock price of listed companies can vary from year to year, I added 9 dummy variables (Y2008-2016) to exclude the influence of years. For example, if one sample happened in 2008, Y2008 equals 1 otherwise equals zero.

(14)

Table 1. Variables Definition

This table summarizes the definitions of all dependent variables, independent variables and control variables that are used in the research model.

Dependent Variables

Premium: M&A Premium

PB30: Average Premium of 30 days before announcement date

PA10: Average Premium of 10 days after announcement date

Independent Variables

Share1: Ratio of shares of the largest shareholder to total shares

Share1^2: Square of ratio of shares of the largest shareholder to total shares

Share9: Ratio of shares of the second to tenth shareholders to total shares

Controller Nature: Dummy variable, equals 1 if the actual controller is state-owned, otherwise

counts 0

URS: Ratio of unrestricted shares to total shares

TS: Ratio of transferred shares to total shares

Control Variables

MV: Market Value in present of company size

BTM: Book to Market rate in present of growth ability

Y2008: Dummy variable, equals 1 if M&A happened in 2008, otherwise counts 0

4.2 Methodology

The study takes IBM SPSS 24.0 as the data analysis tool to conduct a multiple linear regression analysis to test conceptual model, proposed hypotheses and finally to reach the conclusion. As mentioned above, the regression model of ownership structure and M&A premium is formulated as follows to be estimate.

(15)

Model 1:

PR = α+β1*Shrare1+β2*Share9 +β3*G + β4*URS+ β5*TS + β6*MV+ β7*BMR+ ε

In the hypothesis 1, I assume that the ownership concentration and M&A premium have U curve relationship. Thus, the square of Ratio of shares of the largest shareholder to total shares is added to the model which is as follows.

Model 2:

PR = α+β1*Shrare1+γ*Share1^2+ β2*Share9 +β3*G + β4*URS+ β5*TS + β6*MV+ β7*BMR+ ε

In this formula, α is the constant term. And β and γ are regression coefficients, which measure the effect of independent variables have on dependent variable. Ε is random error. I assume it is not auto-correlated and has zero average and equal variance.

5. Data and Empirical Research

5.1 Data collection and Descriptive Statistics

All the data used in this research is collected from CSMAR Database (China Stock Market & Accounting Research Database) which is jointly developed by GTA Information Technology Co. Ltd, the University of Hong Kong and the China Accounting and Finance Research Center of the Hong Kong Polytechnic University. It contains around 50 GTA major databases and constitutes several parts, including Macroeconomics, China's Listed Companies, Stock Market, Bond Market and Banking. Among the M&A cases that happened and completed in the last ten years (2007-2016), I choose the subjects whose transfer object is stock of listed companies which include targets listed on Shenzhen Stock Exchange, Shanghai Stock Exchange, Shenzhen SME Board Stock Exchange and Shenzen Stock Exchange ChiNext Board. In the end, I collected 1035 samples in my dataset.

(16)

Table 2 shows the descriptive statistics including frequency, mean, and standard deviation, minimum and maximum. In order to check missing data among the

variables, a frequency test for all variables is performed. Among 1109 samples in total, 1104 samples are valid. The mean of M&A premium on the announcement day is -0.212, which tells us that transfer price is lower than closing price in most of the cases. The ratio of largest shareholder has an average of 31.63%, minimum of 3.62% and maximum of 89.41%. And the ratio of second to tenth shareholders has an average of 22.94%, minimum of 1.55% and maximum of 60.99%. This means the ownership structures of the target companies are very diversified- from highly

centralized to well-balanced. The average of ratio of unrestricted shares to total shares is 72.84%, while the minimum is 10.03%, which means the stock of most target companies is liquidated except for some extremely liquidated ones. The average of transferred stock rate is 13.48%, with a minimum of 5% and maximum of 68.73%. And the average of state-owned controller is 0.26 shows us that unstate-owned companies are more than state-owned companies.

(17)

Table 2. Descriptive Statistics

This table provides frequency and other descriptive statistics variables between 2007 and 2016 with 1109 M&A samples.

Frequency

Mean Std. Deviation Minimum Maximum Valid Missing Premium 1109 0 -.2118 .35993 -.99 1.89 PB30 1109 0 -.1628 .38308 -.99 2.29 PB20 1109 0 -.1679 .38187 -.99 2.21 PB10 1109 0 -.1580 .38972 -.99 2.35 PA10 1105 4 -.2642 .37776 -1.00 3.60 PA20 1105 4 -.2691 .36370 -1.00 3.02 PA30 1105 4 -.2714 .37942 -1.00 3.90 Share1 1109 0 31.6343 14.90864 3.62 89.41 Share9 1109 0 22.9367 12.59596 1.55 60.99 Controller Nature: 1109 0 .2642 .44111 .00 1.00 URS 1109 0 72.8398 25.19440 10.03 100.00 TS 1109 0 13.4756 9.53789 5.00 68.73 MV 1109 0 133.5794 241.36113 .00 3504.54 BTM 1104 5 .1580 .17072 -.48 .92

Graph 1 shows the number of M&A cases happened each year during 2007 to 2016. We can see a stable increasing trend after 2011 and huge jump between 2014 and 2015. In February 10th, 2015, the CBRC issued “Commercial Bank M&A Loan Risk Management Guidelines”, raising the duration of merger loans to 7 years, and the limitation of M&A loans increased to 60% of total transaction value. In the context of both positive demand and external policy, the impact of China's M&A fund on the market is deepening. In August 31st same year, China Securities Regulatory Commission and Ministry of finance, the SASAC, China Banking Regulatory Commission four ministries jointly issued “Notice of Encouraging Mergers and Acquisitions of Listed Companies, Cash Dividend and Share Repurchase”. Regarding the M&A supervisions, the scope of administrative examination and approval was reduced, and procedures were simplified. It encouraged innovation of payment tools

(18)

and financing methods. And directional convertible bonds as a payment tool were launched. The notice also urged state-owned listed companies rely on capital market to strengthen the integration of resources, adjust and optimize the industrial layout. In addition, the further deepening of reformation of state-owned enterprises, accelerate the pace of Chinese M&A projects continue, to optimize the allocation of assets, expand the scale of enterprises, realize strategic transformation and structural adjustment of production capacity and so on.

Graph 1. Number of M&A Samples 2007-2016

5.2 Empirical Research

5.2.1 Multicollinearity Test

Multicollinearity is a statistical phenomenon in which two predictor (independent) variables in a multiple regression model are highly correlated. The VIF (Variance Inflation Factor) test is conducted to check whether there is the multicollinearity existing between the independent variables. Table 3 is the Multicollinearity Test results of the variables. The examining results show that each individual VIF score is

124 75 79 55 45 64 70 113 263 221 0 50 100 150 200 250 300 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

(19)

below 2. According to Sueles and Craney (2002), there would be present a

multicollinearity among independent variables if a VIF is larger than 5 or larger than 10. Thus, it is believed that there is no worry about multicollinearity of the measures.

Table 3. Results of Multicollinearity Test

This table provides tolerance and VIF (Variance Inflation Factor) which are indicators of Multicollinearity Test of the five independent variables and Market Value and Book to Market.

Collinearity Statistics Tolerance VIF Share1 .735 1.360 Share9 .695 1.439 Controller Nature .946 1.057 URS .763 1.310 TS .906 1.104 MV .984 1.017 BTM .921 1.086

a. Dependent Variable: Premium

5.2.2 Regression Analysis- Model 1

In the process of regression analysis, first I added independent variables one by one to check their correlation with dependent variables separately. Then I did multi-linear regression with all the dependent variables according to Model 1. Table 4 shows the results of beta and t-statistic of all the variables.

From the table, we can see that the standardized coefficient of Share1 is significantly negative under 1% level. This means M&A premium decrease as ratio of shares of largest shareholder increase, which is agreed with Chen & Lu (2013)’s conclusion.

(20)

The standardized coefficient of Share9 is negative under confidence level of 5%, which means that ratio of shares of the second to tenth shareholders is negatively related to premium. Therefore, Hypothesis 1 is significant under confidence level of 5%.

With stable coefficients (β= 0.064 & 0.068), the nature of property right of the number one shareholder is positively related to M&A premium. This is contrary to my

hypothesis that the more balanced ownership structure is the higher M&A premium. Unfortunately, Hypothesis 2 is rejected.

Variable URS has significant positive coefficient at 5% level. Namely, as the proportion tradable shares to total share go up, the M&A premium increase. We can accept Hypothesis 4 under 5% confidence level.

The last independent variable TS show a negative correlation with dependent variable under 10% level. If the stock transfer rate increase, the acquiring company tend to offer higher price. Thus, Hypothesis 5 can be accepted under 10% confidence level.

For the control variables, Market Value of target companies has significantly negative correlation with M&A premium. And the value of coefficient is stable, between -0.06 to -0.046. This means the higher the market value is, the less premium targets will receive. The coefficient of Book to Market ratio is also very stable- around 0.08. It is very reasonable to take these two factors as control variables.

To sum up, all the hypotheses have been checked except one should be rejected. Because adjusted R-square equals 0.255, Model 1 is proved to be effective, explaining 25.5% total variance of the dependent variable. It can also be observed that adjusted

(21)

R-square is higher when all the variables are added to the regression model. Thus, we can conclude the ownership structure as an entire part has effect on M&A premium.

(22)

Table 4. Coefficients - Model 1

This table provides estimates of coefficients on Premium performed by Multi-leaner regression analysis for independent variables and control variables. In column 1, I first test all the control variables and they are all significant. In column 2-6, I add ownership structure variables one by one to check each variable’s own effect. Column 7 shows the entire model with all the variables included.

Premium Premium Premium Premium Premium Premium Premium

Share1 -0.129*** -0.120*** (-4.82) (-3.82) Share9 -0.053** -0.069** (-1.95) (-2.18) G 0.064** 0.068** (2.31) (2.48) URS 0.174*** 0.114** (5.71) (3.25) TS -0.054** -0.052* (-2.00) (-1.88) Y2008 0.191*** 0.192*** 0.186*** 0.194*** 0.171*** 0.191*** 0.174*** (5.86) (5.94) (5.69) (5.96) (4.23) (5.86) (5.43) Y2009 0.150*** 0.145*** 0.145*** 0.151*** 0.113*** 0.153*** 0.119*** (4.55) (4.46) (4.38) (4.61) (3.40) (4.65) (3.62) Y2010 0.198*** 0.202*** 0.190*** 0.201*** 0.161*** 0.199*** 0.169*** (6.36) (6.53) (6.03) (6.44) (5.14) (6.39) (5.41) Y2011 0.181*** 0.19*** 0.173*** 0.185*** 0.140*** 0.180*** 0.158*** (5.93) (6.29) (5.64) (6.08) (4.54) (5.92) (5.10) Y2012 0.215*** 0.213*** 0.209*** 0.220*** 0.147*** 0.212*** 0.163*** (6.75) (6.75) (6.53) (6.91) (4.37) (6.64) (4.83) Y2013 0.282*** 0.288*** 0.278*** 0.292*** 0.227*** 0.280*** 0.252*** (8.64) (8.88) (8.48) (8.89) (6.73) (8.56) (7.40) Y2014 0.328*** 0.333*** 0.322*** 0.342*** 0.253*** 0.324*** 0.287*** (9.29) (9.54) (9.10) (9.57) (6.79) (9.19) (7.56) Y2015 0.408*** 0.42*** 0.399*** 0.426*** 0.301*** 0.399*** 0.346*** (9.74) (10.10) (9.46) (10.02) (6.63) (9.47) (7.36) Y2016 0.623*** 0.627*** 0.616*** 0.643*** 0.524*** 0.618*** 0.568*** (15.152) (15.39) (14.96) (15.34) (11.88) (15.01) (12.52) MV -0.058** -0.046* -0.057** -0.060** -0.057** -0.059** -0.046* (-2.16) (-1.71) (-2.10) (-2.24) (-2.14) (-2.18) (-1.74) BTM 0.081*** 0.093** 0.08*** 0.068** 0.077*** 0.080*** 0.072** (2.86) (3.29) (2.82) (2.33) (2.74) (2.80) (2.52) N 1109 1109 1109 1109 1109 1109 1109 adj R-sq 0.218 0.234 0.22 0.221 0.24 0.22 0.255 t statistics in parentheses *p<0.1, ** p<0.05, *** p<0.01

(23)

5.2.3 Robustness Test

In consideration of other effects on stock price, I used the average closing price of 10, 20 and 30 days before and 10, 20, and 30 days after the announcement date to run a Robustness Test. Table 5 summarize the results of the test.

It shows that Market Value (MV), Ratio of shares of the largest shareholder to total shares (Share1), Ratio of shares of the second to tenth shareholders to total shares (Share9) and Ratio of unrestricted shares to total shares (URS) are robust both before and after the announcement date. Nevertheless, Nature of actual controller (G) is only robust before or around the announcement date. And Ratio of transferred stock (TS) is only robust after the announcement date.

(24)

Table 5. Results of Robustness Test

This table shows the coefficients results of robustness test. I substitute the main dependent variable- Premium with average premium during a certain time before and after the M&A announcement day to check the robustness of my model.

PB30 PB20 PB10 PA10 PA20 PA30

Share1 -0.131** * -0.127*** -0.122*** -0.100*** -0.103** -0.107*** (-4.09) (-3.99) (-3.82) (-3.17) (-3.30) (-3.41) Share9 -0.070** -0.074** -0.078** -0.013 -0.008 -0.002 (-2.16) (-2.29) (-2.41) (-0.409) (-0.27) (-0.056) Controller 0.054* 0.059** 0.064** 0.047* 0.044 0.040 Nature (1.94) (2.13) (2.30) (1.71) (1.61) (1.47) URS 0.129*** 0.131*** 0.120*** 0.131*** 0.140*** 0.134*** (3.61) (3.70) (3.39) (3.74) (4.02) (3.84) TS -0.032 -0.032 -0.034 -0.089** -0.084*** -0.078*** (-1.12) (-1.15) (-1.20) (-3.20) (-3.04) (-2.81) MV -0.050* -0.050* -0.050* -0.064** -0.060** -0.060** (-1.85) (-1.85) (-1.88) (-2.41) (-2.29) (-2.26) BTM 0.054* 0.055* 0.042 0.135*** 0.145*** 0.015*** (1.85) (1.89) (1.46) (4.69) (5.11) (5.24) N adj R-sq 1109 1109 1109 1109 1109 1109 0.23 0.242 0.237 0.253 0.27 0.256 t statistics in parentheses *p<0.1, ** p<0.05, *** p<0.01

5.2.4 Regression Analysis- Model 2

To check the reality of Hypothesis 1-The ownership concentration and M&A premium have U curve relationship, the square of Ratio of shares of the largest shareholder to total shares is added in Model 2. Table 6 shows the results of beta and t-statistic of all the variables.

(25)

From the results we can see, the coefficients of both Share1 and Share1 Square are significant in all the scenarios. Share1 has negative correlation with Premium under 1% confidence level and Share1 Square has positive correlation with Premium under 10% confidence level. These outcomes all together prove that ratio of shares of the No.1 shareholder has U Curve relationship with M&A premium. Namely, As the

proportion of shares of largest shareholder gets bigger, The M&A premium will decrease to a degree in the beginning, touch the bottom and bounce back later on.

Then I take a further step to get a direct revelation of the U curve relationship between ratio of largest shareholder and M&A premium. From the descriptive statistics table, we know that Share1 is in a range from 3.62 to 89.41. Thus I divide it into 9 ranges: 0 to 10, 10 to 20, 20 to 30…80 to 90. Then Share1 is substituted with 8 dummy variables based on the range. For example, if value of Share 1 is between 0 and 10, Share1 Range 0-10 equals 1 otherwise equals 0. I dismiss range 50 to 60 because 9 ranges can be represented by 8 dummy variables. The regression of Model 2 is run again with 8 range dummy variables in place of Share1. Table 7 shows the coefficients of all the variables. We can see for the lower ranges, the coefficients are negative. And for the higher ranges, the coefficients are positive. Also, as the range gets higher, the coefficient has an increase trend which can be told directly from Graph 2. Each bin in Graph 2 stands for the coefficient of each range. This means when the largest shareholder owns little proportion of total shares, M&A premium decreases as the proportion increase. But the decrease gets smaller in this process. At some point the decreasing trend reverses. When the largest shareholder owns little proportion of total shares, M&A premium also increase as the proportion increase. In general, it shows a U curve correlation.

(26)

Table 6. Coefficients-Model 2

This table provides estimates of coefficients on Premium based on non-leaner regression analysis by including the square of ratio of share of largest shareholder.

P P P P P P Share1 -0.421** * -0.390** * -0.433** * -0.357** * -0.407** * -0.331** * (-3.92) (-3.65) (-4.05) (-3.33) (-3.74) (-3.05) Share1^2 0.279** 0.205* 0.290*** 0.244** 0.270** 0.193* (2.60) (1.90) (2.70) (2.28) (2.49) (1.77) Share9 -0.115** * -0.074** (-4.00) (-2.32) Controller Nature 0.074*** 0.072*** (2.69) (2.63) URS 0.144*** 0.109*** (4.63) (3.13) TS -0.020 -0.051* (-0.73) (-1.82) MV -0.046* -0.038 -0.048* -0.047* -0.046** -0.045* (-1.71) (-1.42) (-1.79) (-1.79) (-1.73) (-1.69) BTM 0.094*** 0.096*** 0.079*** 0.088*** 0.093*** 0.073** (3.34) (3.42) (2.75) (3.15) (3.30) (2.56) N adj R-sq 1109 1109 1109 1109 1109 1109 0.242 0.252 0.246 0.256 0.242 0.263 t statistics in parentheses *p<0.1, ** p<0.05, *** p<0.01

(27)

Table 7. Coefficients – With Share1 Ranges

This table provides estimates of coefficients on Premium performed by dividing Share1 into 9 ranges and use a dummy variable to represent each range. While the coefficients of other variables remain stable, the trend of the 8 range betas can give a direct implication of the non-linear

correlation between Share1 and Premium.

Premium Share1 Range 0-10 -0.022 (-0.73) Share1 Range 10-20 -0.140*** (-2.91) Share1 Range 20-30 -0.074 (-1.44) Share1 Range 30-40 -0.057 (-1.34) Share1 Range 40-50 0.014 -0.34 Share1 Range 60-70 0.059* -1.91 Share1 Range 70-80 0.003 -0.13 Share1 Range 80-90 0.023 -0.87 Share9 -0.080** (-2.45) Controller Nature 0.064** -2.33 URS 0.107*** -3.05 TS -0.051* (-1.80) MV -0.043 (-1.60) BTM 0.077*** -2.67 N adj R-sq 1109 0.258 t statistics in parentheses *p<0.1, ** p<0.05, *** p<0.01

(28)

Graph 2. Coefficients of Each Range of Largest Shareholder Ratio

This graph provides frequency of M&A cases happened during each year between 2007 and 2016.

6. Conclusion and Implications

Integrating the characteristics of Chinese acquired and acquiring companies, this research includes 1104 valid samples of M&A cases from 2007 to 2016. Multilinear and nonlinear models have been constructed to perform empirical analysis and demonstrate the hypotheses regarding correlation between ownership structure and M&A premium. Following conclusions can be drawn:

(1) As ownership gets more concentrated and the ratio of largest shareholders increases, M&A premium shows a U curve trend.

(2) Ownership balance degree has negative relationship with M&A premium. (3) State-owned targets tend to get more M&A premium than unstate-owned

targets.

(4) Acquiring companies tend to offer higher premium for stock with higher liquidation. -0.150 -0.100 -0.050 0.000 0.050 0.100 0-10 10-20 20-30 30-40 40-50 60-70 70-80 80-90

(29)

(5) Acquiring companies tend to offer higher premium when they get more shares of the targets.

Based on these five conclusions, several suggestions for companies going through mergers and acquisitions can be summarized. When choosing a proper acquire target, besides considering market and enterprise factors, companies can give more attention to the ownership structure. To make a good deal, acquiring companies should avoid targets with extreme ownership structure. For example, one big shareholder owns all the control or the control is too decentralized. When acquiring companies have

enough capital, they should seek to acquire more stock of targets to save the budget. If they don't care much about the liquidity of the equity, they can consider the targets with more restricted shares.

Unlike other indicators including EPS, Market Value or Free Cash Flow of corporate performance and valuation, ownership structure is a factor that is hard to measure. Thus, in previous M&A cases consultants would pay less attention to the impact of ownership structure. This research gives an insight in the effect of ownership structure on M&A premium with implications for practice in the future M&A cases. There exist some limitations in this study. I only include five factors to present ownership structure, other parts such as politic connection between management and ratio of management owned stock are not examined. Hopefully, in future studies this topic can be more profoundly discussed.

(30)

References

Amihud, Y. and Mendelson, H., 1986. Asset pricing and the bid-ask spread. Journal of

financial Economics, 17(2), pp.223-249.

Alexandridis, G., Fuller, K.P., Terhaar, L. and Travlos, N.G., 2013. Deal size,

acquisition premia and shareholder gains. Journal of Corporate Finance, 20, pp.1-13. Betton, S. and Eckbo, B.E., 2000. Toeholds, bid jumps, and expected payoffs in takeovers. The Review of Financial Studies, 13(4), pp.841-882.

Bargeron, L.L., Schlingemann, F.P., Stulz, R.M. and Zutter, C.J., 2008. Why do private acquirers pay so little compared to public acquirers?. Journal of Financial

Economics, 89(3), pp.375-390.

Berle, A.A. and Means, G.G.C., 1991. The modern corporation and private property. Transaction publishers.

Barclay, M.J. and Holderness, C.G., 1989. Private benefits from control of public corporations. Journal of financial Economics, 25(2), pp.371-395.

Chen D., and Chen Y., 2011. Relationship of Ownership Concentration, Ownership Balance and Corporate performance –Empirical research of companies listed on Chinese SME board during 2007 to 2009. Accounting Research, 1,pp. 38-43 . Chen S., & Lu C.. 2013. Inter firm executives' linkage and merger premium decision making: An Empirical Study Based on Inter Organizational imitation

theory. Management World(5), 144-156.

Carline, N.F., Linn, S.C. and Yadav, P.K., 2002. The influence of managerial ownership on the real gains in corporate mergers and market revaluation of merger partners: empirical evidence.

Demsetz, H. and Lehn, K., 1985. The structure of corporate ownership: Causes and consequences. Journal of political economy, 93(6), pp.1155-1177.

(31)

Du X., & Gan S.. 2012. Research on the relationship between capital structure, ownership structure and business performance -- Based on Chinese listed companies on GEM. Business Accounting(1), 73-75.

Dimopoulos, T. and Sacchetto, S., 2014. Preemptive bidding, target resistance, and takeover premiums. Journal of Financial Economics, 114(3), pp.444-470.

Eckbo, B.E., 1983. Horizontal mergers, collusion, and stockholder wealth. Journal of

financial Economics, 11(1-4), pp.241-273.

Eckbo, B.E. and Thorburn, K.S., 2000. Gains to bidder firms revisited: domestic and foreign acquisitions in Canada. Journal of Financial and Quantitative Analysis, 35(1), pp.1-25.

El-Khatib, R., Fogel, K. and Jandik, T., 2015. CEO network centrality and merger performance. Journal of Financial Economics, 116(2), pp.349-382.

Ferris, K., Melnik, A. and Rappaport, A., 1977. Cash tender offer pricing: an empirical analysis. Mergers and Acquisitions, 12, pp.9-14.

Gondhalekar, V.B., Raymond Sant, R. and Ferris*, S.P., 2004. The price of corporate acquisition: determinants of cash takeover premia. Applied Economics Letters, 11(12), pp.735-739.

Grinstein, Y. and Hribar, P., 2004. CEO compensation and incentives: Evidence from M&A bonuses. Journal of financial economics, 73(1), pp.119-143.

Hunter, W.C. and Jagtiani, J., 2003. An analysis of advisor choice, fees, and effort in mergers and acquisitions. Review of Financial Economics, 12(1), pp.65-81.

Jensen, M.C. and Meckling, W.H., 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of financial economics, 3(4),

(32)

John, K., Freund, S., Nguyen, D. and Vasudevan, G.K., 2010. Investor protection and cross-border acquisitions of private and public targets. Journal of Corporate

Finance, 16(3), pp.259-275.

Kahl, M., Liu, J. and Longstaff, F.A., 2003. Paper millionaires: How valuable is stock to a stockholder who is restricted from selling it?. Journal of Financial

Economics, 67(3), pp.385-410.

Lewellen, W., Loderer, C. and Rosenfeld, A., 1985. Merger decisions and executive stock ownership in acquiring firms. Journal of Accounting and Economics, 7(1-3), pp.209-231.

Lloyd, W.P., Modani, N.K. and Hand, J.H., 1987. The effect of the degree of ownership control on firm diversification, market value, and merger activity. Journal of Business

Research, 15(4), pp.303-312.

Mehran, H., 1995. Executive compensation structure, ownership, and firm performance. Journal of financial economics, 38(2), pp.163-184.

Moeller, T., 2005. Let's make a deal! How shareholder control impacts merger payoffs. Journal of Financial Economics, 76(1), pp.167-190.

Roll,R. 1986. The Hubris Hypothesis of Corporate Acquisitions［J］.Journal of Business ,（59）:197-216.

Sokolyk, T., 2011. The effects of antitakeover provisions on acquisition targets. Journal of Corporate Finance, 17(3), pp.612-627.

Souther, M.E., 2016. The effects of takeover defenses: Evidence from closed-end funds. Journal of Financial Economics, 119(2), pp.420-440.

Spiegel, M. and Tookes, H., 2013. Dynamic competition, valuation, and merger activity. The Journal of Finance, 68(1), pp.125-172.

Slusky, A.R. and Caves, R.E., 1991. Synergy, agency, and the determinants of premia paid in mergers. The Journal of Industrial Economics, pp.277-296.

(33)

Sirower, M.L., 1997. The synergy trap: How companies lose the acquisition game. Simon and Schuster.

Song L., Han L., 2005. Empirical analysis of the influence of the proportion of large

shareholders on agency cost. Nankai Business Review,8(1):30-34.

Xiao Z., & Chen D.. 2006. The influence of corporate governance structure on agency cost -- Empirical Evidence from Chinese Listed Companies. Finance & Trade

Economics (12), 29-35.

Zhang Z., Song L., & Zhang Q.. 2005. Empirical analysis of capital structure and equity agency cost of Listed Companies in China. Accounting Research(8), 44-49.