The Relationship Between Stock Repurchase and CEO
Compensation in Chinese A Share Market
Master Thesis
MSc Finance Asset Management
Student
: Zhang Bowen
Student ID : 11377143
Supervisor : Prof. dr. F. Lopez de Silanes
Date
: 2017.8.15
Statement of Originality
This document is written by Student Bowen Zhang who declares
to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is
original and that no sources other than those mentioned in the text
and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for
the supervision of completion of the work, not for the contents.
Abstract
Stock repurchase has many functions, it will release a positive signal to the market and then raise the stock price. In Chinese A share market, stock repurchase starts late due to economy and policy reasons. But in recent years, the behavior became more frequent. Since the market is not mature enough, whether the determinants of stock repurchase is in line with developed stock markets in the world is still questionable.
This study focuses on the relationship between stock repurchase and CEO compensation. I use the latest data from 2013.1 to 2017.6 in Chinese A share market and use the methodology of both Chinese and western literatures. The result shows that Stock repurchase has a tight connection with CEO compensation, especially in vega, CEO ownership, level of undervaluation and financial constraints.
Contents
1. Introduction ... 1
2. Background and Literature review... 2
2.1 Conception ... 2
2.2 Literature review ... 4
2.3 Background of Chinese A share market ... 8
3. Hypothesis and Methodology ... 12
3.1 Hypothesis... 12
3.2 Methodology ... 12
4. Data and descriptive statistics ... 15
4.1 Sample selection ... 15
4.2 Statistics ... 16
5. Results and robustness checks ... 18
5.1 Univariate results ... 18
5.2 Multivariate results ... 21
5.3 Actual repurchases ... 25
6. Conclusion ... 27
1
1. Introduction
Stock repurchase is becoming more common and important in global stock markets because of its varieties of benefits. When a company offers to buy back shares, they are releasing a signal to the market that they have healthy cash flows and their stock price are undervalued. They may also do stock repurchase to pay dividends for the employees instead of paying cash, adjust the capital structure to get rid of higher tax, or prevent their components from a hostile takeover. The investors will have more confidence on the company and finally the stock price will probably go up.
Managers also want to see stock repurchase. In recent years agency cost is becoming more important any many companies decide to pay their senior executives stock and options to keep them working in line with the company. As a result, when stock repurchase occurs, they will get considerable benefit with the increase stock price. People have realized the problem, and thus have made many research in academic and reality to minimum such cost. They design the contract in a tricky way to balance the evaluation of stock price and corporate finance. But still, stock repurchase is a profitable way to benefit the manager both in stock price and in evaluation, because the denominator in earning per share (EPS) will decrease and the manager seems to have better performance.
But it is questionable that this rule remains the same in the immature markets. Chinese A share market is an emerging market will a short history of 30 years, and stock repurchase just become common five years ago. Within the unstable market environment and the strong control of government, Chinese companies and scholars are
2
still exploring the road. Not many research production have been done, and most of are just qualitative analysis or concentrate on one company. With plenty of space for quantitative research and latest data, I want to find out whether Chinese A share market is in line with the most developed stock markets in the world, or they have much difference.
In this paper, I will follow the methodology of both Chinese and western research, using event study to figure out the relationship between stock repurchase and CEO compensation in Chinese A share market. I get the latest data from Wind and CSMAR database, and sample period is from 2013.1 to 2017.6, which covers around 90% of stock repurchase in China.
I will calculate the abnormal return related with stock repurchase in the past 5 years and test which characters of CEO compensation will have significant influence on stock repurchase, and I will do proper robustness check to support my results. I hope this paper will make contribution to corporate finance and stock market research of Chinese A share market.
2. Background and Literature review 2.1 Conception
Stock repurchase is becoming more common and important in stock markets. When a company offers to buy back shares or the shareholders want to increase tradable shares, the management and main shareholders must have confidence on the intrinsic value of the company, and their buyback will release a positive signal to the market. On the other
3
hand, buyback is an important part of the internal stability mechanism of market, it will play the role of stabilizing the market and restoring investors’ confidence. For the company itself, the ultimate purpose of stock repurchase is increase the value of the company with the following aspects:
First of all, release the signal to the market that the stock price is undervalued. Stock repurchase has an opposite effect to stock issuance, it is a positive signal to the market and the stock price will rise accordingly. Even if the stock is still undervalued after the repurchase, shareholders can still benefit from repurchasing at a low price.
Secondly, when the cash flow of the company obviously surpasses their need for investments, it could be used to repurchase shares, and this will result in the increase of earnings per share. Since the repurchase attempts to make the investors believe that the stock is attractive, the company is not wasting money on investments with bad benefits. Thirdly, stock repurchase can be used to pay dividends. When the company’s cash holding is not reliable enough for keeping a high dividend for a long period or they want to get rid of higher tax, they can pay the dividend by stock repurchase.
Moreover, stock repurchase can adjust the financial leverage. If the company thinks that the percentage of equity capital in the capital structure is relatively high, buyback shares will increase the ratio of liabilities, which will help to decrease the weighted average capital cost. Although we can also reduce shareholder’s equity and leverage up by paying cash dividend, the earning per share will be different when the benefit is the same. And we can quickly increase liability ratio by issue bonds to buy back shares.
4
Besides, the ownership structure will be adjusted. The shares bought back by the company can be used to exchange for those of merged companies, or to meet the demand of the warrant holders who want to subscribe the stock or convertible bonds. Stock repurchase can also be used when executing options to get rid of issuing new stocks and then dilute gains.
At last, it will reduce the hazard of hostile takeover. When the stock price is higher and the quantity of tradable stocks is decreased, their components need to spend more on the acquisition. As a result, there are less opportunities for potential takeovers.
However, stock repurchase is not only making profit for the company, because it is always accompanied by agency cost in practice. In corporate governance field in recent years, the contracts have improved a lot to keep the CEO work for the benefit of the company. Besides giving basic salary and bonus for short term motivation, offering stock and options are the most common methods to achieve the goal that CEO will represent the benefit of the company for long term. Managers will have some opportunities to maximize their own profits by buy back the stock, which will increase the stock price, and directly raise his own compensation. Moreover, their incentive will be read by the market to some extent, so the characters of CEO compensation will also influence the effect of the stock repurchase.
2.2 Literature review
Worldwide companies are spending a greater portion of their revenue on stock repurchase in recent years. From the data provided by Thrum (2013), we know that stock repurchase amount of the S&P 500 index companies in 2013 has reach $500
5
billion, and the growth rate is increasing steadily. As the stock repurchase becomes popular in real world, the topic is booming in academic field as well.
The most essential topic related to stock repurchase is signaling theory. Most stock repurchase announcements will result in abnormal returns, and the theory regard the behavior as a signal of stock undervaluation.
The research of Waqar and Ahmed (2015) indicates that CEO compensation arrangement will have an significant influence on the market reaction to the firm who announce stock repurchase programs. The paper shows that the market will get the signal from the repurchase, read the incentive and responds accordingly. This effect will be stronger if the firm is suffering from high information asymmetry, or the stock price is undervalued.
Signaling theory is originated from asymmetric information theory. In the viewpoint of Akerlof (1970), some participants in the stock market has an information asymmetry with the others, which means they know more about the business circumstance of the companies than the outsiders. Since the rest of market cannot get the correct information on time, including the expected value and risk, the asymmetry will lead to stock misevaluation (Grullon and Michaely, 2004). And the participants with information advantage can “earn” money before the market is able to get available information.
In the empirical study by Ikenberry et al. (1995), they use share repurchase announcement data from the Wall Street Journal between 1980 and 1990. The results show that repurchasing firms have an abnormal return of over 12% over the following
6
4 years. Compared with growth firms with low book-to-market ratio (B/M) that do not have any abnormal performance, undervalued firms with high B/M earn an abnormal return of over 45% in the 4-year period post repurchase announcement.
For further support for their findings, Ikenberry et al. conduct a similar study in 2000. The database changed to 1,060 stock repurchase announcements in the Canadian market but the evidence is similar to their former study. Undervalued firms still significantly outperform growth firms before the announcement, while this time growth firms also earn abnormal returns. consistent with this undervaluation hypothesis, Chan et al. (2004) also find similar results in the US stock market.
In an ideal circumstance, managers will work for the interest of long-term shareholders. When the stock price is undervalued, they will try to correct it by sending positive signal to the market. However, in the well-established literature of Spence (1973), the importance of signaling cost is brought into our view. The cost is important as they provide credibility. In the absence of signaling costs, all managers will have the incentive to mimic positive signals, which results in a “pooling equilibrium”. The market fails to distinguish good companies from bad ones, hence all the companies will have an average value.
However, stock repurchase announcements is costless, because it just represents the company’s authorization to repurchase shares, not the actual transaction, and there are no commitments. From the research of Stephens and Weisbach (1998) and Bhattacharya and Dittmar (2008), a great portion of companies choose not to execute their stock program. Chan et al. (2010) presents their worry that lack of commitment
7
and the flexibility of stock repurchase will also result in mimicking behavior. In US, the regulatory and disclosure requirements related to the execution of stock repurchase are at minimum. Reputational penalty for failing to realize the repurchase is also not significant as well.
In the research of Grullon and Michaely (2002), companies also start to do stock repurchase to distribute free cash flows since 1980s. In line with the free cash flow hypothesis by Jensen (1986), managers will be less likely to abuse excess cash because of stock repurchase. Even so, according to the research of Barclay and Smith (1988), Gong et al. (2008) and Brockman et al. (2008), equity stakes owned by the managers provide them with the incentive to depress stock price, then wealth transfer from shareholders who sell their stock to the remaining shareholders (including managers themselves) is accelerated. They find supportive evidence that managers will deflate their earnings and announce bad news before share repurchase. The higher equity stakes they have, the more they want to lower the repurchase price. This phenomenon is due to pre-repurchase downward earning management to some extent, and it reduces the effect of equity compensation, for its purpose is to align executives and management.
Taxation motive is another reason for stock repurchase in the early days. In the research of John and Williams (1985) and Allen et al. (2000), stock repurchase is treated as capital gain and the tax rate is lower than dividend income, so it is a more efficient way to transfer wealth to shareholders compared to dividends. Other papers by Fama and French (2001); DeAngelo et al. (2004) and Fatemi and Bildik (2012) are supporting this viewpoint.
8
Stock repurchase can also be used to adjust capital structure (Dittmar ,2000). In a survey of finance directors top 200 businesses by Dixon et al. (2008), reaching an optimal structure is the most cited reason for stock repurchase.
In addition, managers’ selfish incentives also induce them to do stock repurchase for their own benefit. Given the fact that market reacts positively to stock repurchase announcements, the managers will benefit directly due to the increase in the value of their stock and options. Moreover, their performance evaluation will be enhanced as well by improving earnings per share (EPS). In the research of Steven Young et al. (2011), they find out the EPS related incentives will induce managers to buy back stocks.
Another research by Sheng-Syan Chen et al. (2016) finds out if the CEO have high equity compensation and low outside monitor, like independent directors and analysts, he will be able to release more bad news to earning forecast and hold the good news before stock repurchase. This will result in positive abnormal stock returns in long term. In contrast, with low equity compensation and more monitor, his pre-repurchase management opportunism will be constrained.
2.3 Background of Chinese A share market
Chinese A share is the common shares released by Chinese domestic corporations and traded by China mainland institutions, organizations or individuals with RMB (Chinese yuan). Its official name is RMB ordinary shares. It has a “T+1” delivery system and a price limit of 10% per day. A share market is the symbol of Chinese stock market, which only has a short history of about 30 years.
9
repurchase compared with Americans, they are persuaded to do so in 1980s when there was no Chinese stock market. The development of economy, finance environment and policies result in the late start together.
Since 1990s, Chinese companies started stock repurchase in practice. But with the statistics result of Wind database, until 2014, the number of companies who has done stock repurchase is only 186 and the repurchase amount is about 70 billion Chinese yuan. The number is neglectable compared with the new coming tradable shares.
On the other end of the earth, listed companies in United States have consistently bought back their own stocks since 2009, and they can even thoroughly private the ownership. In the following 5 years, the repurchase fund is 1.6 times of the IPO and additional issues. Considering paying dividends, American those companies offered their shareholders 2.7 times than their demand for fund. This behavior is one of the essential reasons of the bull market in last years and shows the approve of the companies’ fundamental value.
Compared with release amount, the repurchase amount of listed companies in A share market is neglectable, but the number is growing. Since the end of 2012, the number of companies and amount of money is growing continuously.
Since 2008, A share market experienced a long period of bear market and the market is undervalued by less than 10 times. In theory, the listed companies have enough motivation to release positive signals to the investors, however, we did not see many repurchase occurs. From the Wind statistics results, 104 companies did stock repurchase 160 times in total during 2009-2013. Sixteen billion RMB was involved,
10
which is only 0.08% of market capitalization of that time.
Does Chinese companies lack those motivation? We can find the answer among the Chinese stocks launched in United States and Hong Kong. Even though they locate most of their business in mainland China, they did frequent stock repurchase, especially when the market is reaching the bottom. As a result, the reason why A share stocks have little interest to stock repurchase must be unique in China.
Chinese policy has the most influence power. Firstly, Chinese Company Law has a strict attitude to stock repurchase that it was forbidden in principle, only expectations are allowed, including reducing registered capital, M&A, paying staff, and shareholders’ meeting.
Since the law is published in 1993, only the last 2 parts are added, little things have been changed in principle.
Secondly, stock repurchased need to be officially approved by shareholders' meeting, so it is lack of flexibility. Stock repurchase is easily affected by the fast-changing market, but the meeting has a complicated process and takes a long period.
Thirdly, the execution process is limited strictly. There are detailed regulations for the listed company, information reveal, price, time, share amount and sources of fund. The trade is more complicated in practice.
Another essential reason is that Chinese companies’ attitude toward cash. Most companies have difficulty in financing, so they prefer sufficient cash. In a long period, they rarely pay dividends or pay it with shares, not in cash. Other companies just do not have enough money for stock repurchase.
11
A unique emphasis in China will back up this argument, that stock dividend is considered to increase value. This method is a better substitution than stock repurchase, because companies will not pay cash and the tax rate is 50% off in the secondary market.
Policy always plays an important role in Chinese stock market. Since 2009, cash dividends are more common in A share market, because China Securities Regulatory Commission(CSRC) published a series of policies to force the companies to pay it. As a result, companies have less and less motivation and cash to buyback stocks.
In the coming years, A share market experienced a huge shake. In 2014.6, Shanghai index starts at 2048 points, and goes upward insanely, surpassing 5000 points in a year and reaching the peak of 5178 points. Then the stock market crash happened, and the index quickly fell to 2638 in half a year.
To make the stock price stable in the shaking market, many listed companies did more stock repurchase. From Wind database, we can find that over 900 stock repurchase behaviors occurred since 2014. In general, there were a certain degree of premium in the price of repurchased shares. The companies who can buy back their own stocks when the price is falling will send a positive signal to the market, and the price will get support from investors, the assessed value will be fixed then.
During the same period, CSRC published more policies to encourage stock repurchase, which also accelerate its blowout. In 2015.8, they published a notice to facilitate the companies to timely buy back shares — When the stock price is lower than EPS, or P/E ratios is lower than the average level in the same industry, the company could set a threshold value to trigger stock repurchase and select varieties of tools to
12
raise capital, such as issuance of preferred stock and bonds.
3. Hypothesis and Methodology 3.1 Hypothesis
Chinese A share market is an emerging market. Within the unstable market environment and the strong control of government, Chinese companies and scholars are still exploring the road. Not many research production have been done, and most of are just qualitative analysis or concentrate on one company. With plenty of space for quantitative research and latest data.
I set up the hypothesis that stock repurchase behavior and effects in Chinese A share market is in line with the most developed stock markets in the world. It will cause excess return and thus raise CEO compensation. Furthermore, when the CEO wealth is less related with stock price and volatility, and CEO have more ownership of the company, stock repurchase will lead to a higher excess return.
3.2 Methodology
In this paper, I base my methodology on the research of Chinese and western papers. First, following the research of Huang and Liu (2007), I use event study in order to calculate the excess return of stock repurchase. I set up an event window of 11 trading days (-5, +5) around the date when stock repurchase finished. Then I will calculate the expected return 𝑅𝑖𝑡 in the clean period of 30 trading days (-35, -6). Using the market regression model, we can figure out the relationship between single stock and the market:
13
𝑅𝑖𝑡 = 𝛼𝑖 + 𝛽𝑖 ∗ 𝑅𝑚𝑡+ 𝜇𝑖𝑡
where 𝑅𝑖𝑡 is the return of stock i in time t, and it could be calculated by 𝑅𝑖𝑡 = 𝑃𝑖𝑡− 𝑃𝑖(𝑡−1)
𝑃𝑖(𝑡−1)
where 𝑃𝑖𝑡 is the closing price of stock i in time t. 𝑅𝑚𝑡 is the return of market portfolio. This paper will use the Shanghai composite index and Shenzhen composite index as the symbol of market portfolio, following
𝑅𝑚𝑡 = 𝐼𝑡− 𝐼𝑡−1 𝐼𝑡−1
Where 𝐼𝑡 is the closing price of the index in time t.
Besides, 𝛼𝑖 measures the average return that will not be explained by the market, and 𝛽𝑖 measures the sensitivity of the stock to the market. 𝜇𝑖𝑡 is the residual. From
the regression, we can find out the expected return of stock i at time t. Later, we will figure out the abnormal return following
𝜇𝑖𝑡 = 𝑅𝑖𝑡− 𝛼𝑖− 𝛽𝑖∗ 𝑅𝑚𝑡
Then we can calculate the cumulative average abnormal return 𝐶𝐴𝑅𝑖𝑡 with 𝐶𝐴𝑅(𝑡1, 𝑡2) = ∑ 𝜇𝑖𝑡
𝑡2
𝑡=𝑡1
Then we can get the mean of CAR by calculate 1
𝑛∑ 𝐶𝐴𝑅(𝑡1, 𝑡2)
𝑛 𝑖=1
Then, following Waqar and Ahmed (2015), I use delta and vega to measure the CEO’s wealth sensitivity to fluctuation of stock price and risk. With the definition of Core and Guay (2002) and Coles et al. (2006), delta equals to the change of executive’s wealth (in RMB value, originally in dollar value) for a 1 percentage point change in
14
stock price, and vega equals to the change of executive’s wealth (in RMB value) for a 1 percentage point change in annualized stock return volatility. We can calculate with Black-Sholes-Merton option valuation model, since delta and vega are first derivatives with respect to price and volatility respectively. Besides, I use percentage of shares owned by CEO to represent CEO ownership.
Then I create a dummy variable named “Compensation” to combine the three variables above. I give the value of 1 when the delta is above average, vega is below average and CEO ownership is higher than median.
Finally, the regression to estimate the impact of CEO wealth sensitivity on excess return of stock repurchase is:
𝑅𝑖,𝑡 = α + 𝛽1(delta𝑖,𝑡−1) + 𝛽2(vega𝑖,𝑡−1) + 𝛽1(CEO ownership𝑖,𝑡−1) + ∑ 𝛽𝑖
𝑛 𝑖=3
(𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠𝑖,𝑡−1) + 𝜖𝑖
where Ri is the excess return, and delta, vega and CEO ownership are as defined above. For the control variables, I include Financial constraints, which will have less ability to do stock repurchase and the effect will not be so ideal. The variable is measured by Kaplan and Zingales (1997) (KZ) index as follows:
𝐾𝑍 = −1.002 ( CF𝑇 TA𝑇−1) − 39.368 ( DIV𝑇 TA𝑇−1) − 1.315 ( CA𝑇 TA𝑇−1) + 3.139(LEV) + 0.283(Q)
where CF is cash flow, DIV is the cash dividend, CA is current assets of the company in the fiscal year, TA is total assets of the firm, LEV is the ratio of debt and book value of assets, while Q is the market-to-book value.
15
company in the fiscal year, “book-to-market ratio(B/M)” which is the ratio of the firm’s book value to its market value, “cash” and “cashflow” to measure if the company have sufficient fund storage, “but-and-hold return” which is the return in 30-day clean period prior to the announcement, “percent sought” which is the percentage of repurchased shares to outstanding shares.
4. Data and descriptive statistics 4.1 Sample selection
Stock repurchase have a very short history in Chinese A share market, the data is very limited. However, thanks to the prosperity of finance in recent years, stock repurchase behaviors are also booming, so I can get sufficient data for research.
The data of this research consists of three parts, stock repurchase data, stock data and CEO compensation. Stock repurchase data is from the Shanghai and Shenzhen stock research block of Wind database, while the daily stock data and annual CEO compensation data are from CSMAR database.
I choose the time spanning from 2013.1 to 2017.6, which is the latest data of all research. During the sample period, Chinese companies finished stock repurchase 935 times (by contrast, all the stock repurchase before 2013 is less than 100 times in total). This number has excluded the repurchases with less than 100 RMB to trim the extreme values, or they might have special reasons and not helpful for the research. For the stock data, since I use event study as methodology to calculate the stock return, the actual data is from December 2012 to July 2017 to cover the event window.
16
4.2 Statistics
Table 1 Statistics of stock repurchase amount
Year N Mean s.d. p5 p50 p95
2013 84 1.03E+08 5.54E+08 156834 4186994 3.34e+08
2014 157 5.75E+07 5.05E+08 62700 1043200 1.58e+08
2015 195 2.10E+07 1.06E+08 77050 1125300 4.58e+07
2016 313 2.82E+07 1.81E+08 100631 1688100 6.39e+07
2017(first half) 186 1.03E+07 2.97E+07 116467 1911563 4.50e+07
Total 935 3.48e+07 2.90e+08 96910 1570916 7.83e+07
Table 1 presents the summary statistics of stock repurchase amount. During the sample period of 4 and a half year, we can see the increasing trend of stock repurchase behaviors in Chinese A share market, especially in recent 2 years, the number of stock repurchase is surpassing 300 per year on average. Moreover, the standard deviation is quite big in the early years, showing that the market is not mature enough to cover varieties of companies. The median also illustrated that when stock repurchase was not so popular, only companies with sufficient cash could afford large amount of stocks.
Table 2 shows the announcement returns and CEO compensation variables and other characteristics of the companies. The announcement return in the (-5,5) event window is 1.52 percent, which is in line with former research that companies will benefit from stock repurchase announcement. The mean of delta and vega are 991.36 and 149.82 (thousand Chinese yuan) respectively, and the median of CEO ownership is 0.49. These numbers are used in the Compensation dummy to judge whether the executive is well aligned with shareholder.
17
Table 2: Descriptive statistics of returns, CEO wealth sensitivity measures and firm characteristic
CAR is the 11 day (-5 to +5) cumulative abnormal return around the announcement date (day 0) using value weighted market return as benchmark. Delta is the RMB change in the executive wealth for 1 percentage point change in stock price. Vega is the RMB change in the executive wealth for 1 percentage point change in annual volatility. CEO ownership is the CEO’s stock ownership of the firm expressed as a fraction of total shares outstanding. Size is the market value of the firm. B/M is the ratio of book value of firm to its market value. Prior BHR is the cumulative buy-and-hold return of the firm for 35 days prior to the announcement to 6 days before the announcement (-35 to -6). Leverage is the ratio of total debt to total assets of the firm. Cash is cash level of the firm scaled by total assets. Cash Flow is the operating cash flow of the company scaled by total assets. Financial constraints are measured by KZ index (discussed in methodology section). All variables are in the fiscal year of stock repurchase. Actual repurchases is the percentage of shares repurchase as a fraction of intended repurchase size (percent sought). Sales growth is the increase in revenues over previous year revenues.
Percentiles
Variable N Mean Std Dev 25th Median 75th
CAR 935 1.52 7.11 -1.68 1.12 5.03 Delta (thousand) 935 991.36 6285.23 105.12 399.03 1690.26 Vega (thousand) 935 149.82 491.08 15.44 79.19 163.95 CEO ownership (%) 935 1.80 5.10 0.05 0.49 2.73 Size 935 7923.05 19403.25 930.52 2408.49 5023.47 Book-to-market 935 38.05 33.51 27.44 37.52 54.48 Prior BHR 935 -0.03 0.19 -0.12 -0.01 0.06 Leverage 935 0.47 0.13 0.38 0.60 0.71 Cash 935 0.08 0.13 0.01 0.06 0.21 Cash Flow 935 0.09 0.17 0.02 0.09 0.15 Financial Constraints(KZ) 935 1.94 2.95 1.04 1.60 2.87 Actual Repurchases 935 0.61 0.69 0.19 0.54 1.22 Sales growth 935 1.14 0.79 1.02 1.10 1.24
18
From 35 to 5 days Before the announcement, the company will suffer a loss of 3%. There are two reasons to explain this phenomenon, one is that they release bad news on purpose to lower the price, so the stock repurchase will have better influence. The other is that when the stock price falls, the company decide to do stock repurchase to send a good signal to the market.
5. Results and robustness checks 5.1 Univariate results
Table 3 illustrates the correlation matrix of excess returns with five main variables. The coefficient of announcement return and vega is -0.039, and it is significant at 5% level. This means that when the manager is more like risk seeker, the market will be less willing to give positive reaction toward stock repurchase. The relationship between announcement return and delta is not significant enough.
CEO ownership is a more intuitive variable to understand the CEO’s asset compared to that of company. The data is in line with the theory that the coefficient of ownership is 0.059 with the highest significance level.
Cumulative buy-and-hold return is also negatively related to announcement return with 5% significance. This is consistent with former research as well, that when the company has more potential to be undervalued, the market will react stronger to the stock repurchase announcement.
19
Table 3: Correlation Analysis of announcement returns and the executive compensation sensitivity measures
The table report the correlation coefficients. CAR represents 11 day (-5 to +5) cumulative abnormal return around the event date (day 0). Vega is the RMB change in manager wealth for 1 percentage point change in firm annual volatility. Delta is the RMB change in manager wealth for 1 percentage point change in firm share price. Prior BHR is the cumulative buy and hold return of the firm for 35 days prior to the announcement to 6 days before the announcement date (-35 to -6). Ownership represents the executive percentage share ownership of the firm. N shows the number of observations.
Variable 1 2 3 4 5 1.CAR 1 2.Vega -0.039** 1 3.Delta -0.002 0.324** 1 4.Prior BHR -0.088** 0.028 0.040* 1 5.Ownership 0.059*** -0.041* 0.229*** -0.087*** 1 N 935 935 935 935 935
*, **, *** represent significance at 10%, 5% and 1% significance level respectively.
Table 4 shows the characteristics of CAR by delta and vega. I sort the firms into 5 quintiles based on the delta and vega values respectively. Quintile 1 has companies with lowest delta and vega values, while quintile 5 contains firms with highest ones.
Panel A shows the short term CAR and the companies are sorted into quintiles by vega. We can see the clear trend of mean announcement returns, when the companies are in the higher vega quantile, they will have lower abnormal return. The median of returns shows a similar trend. The result is in line with the hypothesis that companies with higher vega will basically have a lower return.
20
Table 4: Mean CAR by vega and delta quintiles
The table reports the mean, median, standard deviation, minimum and maximum values of CARby delta and vega quintiles. CAR represents 11 day (-5 to +5) cumulative abnormal return around the repurchase announcement event (day 0). Vega is the RMB change in manager wealth for 1 percentage point change in firm annual volatility. Delta is the RMB change in manager wealth for 1 percentage point change in firm stock price. In order to sort firms into quintiles, for each year of data I sort all the firms into quintiles based on their delta and vega values. I then assign the repurchase sample firms to these groups based on their delta and vega values for the fiscal year of repurchase announcement. Quintile 1 contains firms with lowest delta and vega values and quintile 5 has firms with highest delta and vega values. N shows the number of firms in each quintile.
Panel A: Mean short term return (CAR) by vega quintiles
Quintiles N MEAN MEDIAN STD MIN MAX
1 186 2.08*** 1.46 7.24 -30.83 59.94
2 199 1.77*** 1.78 6.58 -37.62 40.32
3 212 1.38*** 1.15 7.19 -50.01 42.18
4 188 1.21*** 1.09 6.03 -28.10 30.28
5 150 0.94*** 1.02 6.12 -46.52 31.95
Panel B: Mean short term return (CAR) by delta quintiles
1 145 1.49*** 1.46 7.99 -31.22 59.33
2 177 0.91*** 1.19 6.83 -31.84 46.21
3 201 1.63** 1.28 6.82 -27.59 39.03
4 213 1.29*** 1.04 6.01 -31.02 28.62
5 199 1.15*** 1.22 6.92 -48.04 34.04
*,**,*** represent significance at 10%, 5% and 1% significance level respectively.
Panel B shows the short term CAR and the companies are sorted into quintiles by delta. Compared with Panel A, there is no clear trend, but we still find out that the announcement returns of companies in the lower quintile have better chance to surpass that of companies in a higher quintile.
21
5.2 Multivariate results
In table 5, I regress I regress CAR on delta and vega in different regression models following Coles et al. (2006). I regress them together to avoid endogeneity and control the effect of each other, because the variables may vary from different companies. Then I cluster on year for robust standard errors.
In Model I, I regress CAR on vega and delta. The results are reasonable, for the coefficient of vega is negative and that of delta is positive. Both are significant at 5% level.
In model II, I only regress CAR on CEO ownership. The coefficient is 0.06 with a significance level of 1%. The result offers another strong support for my hypothesis of CEO ownership.
In model III, I regress CAR on delta, vega and CEO ownership. Here vega and CEO ownership do not show much difference in coefficient and significance level, but delta is not significant any more. The conclusion will remain the same.
In model IV, all the control variables are included. The coefficient of 3 major variables
are like in model III. We can find out that prior buy-and-hold returns and financial
constraints are at 1% significance level, and cash flow are also significant at 10%. All of the three control variables are negatively correlated with abnormal return.
These outcomes illustrate that stock repurchase announcement return has strong explanatory power with vega, CEO ownership, level of undervaluation and financial
characters. The result is quite in line with Waqar and Ahmed (2015) that these factors
22
Table 5: The effect of executive compensation on the initial market reaction to sharerepurchase announcement: Multivariate Analysis
The table presents the results of regressing initial buyback announcement abnormal returns (CAR) on executive compensation variables, CEO share ownership and control variables in different models. CAR represents three day (-1 to +1) cumulative abnormal return around the repurchase announcement date (day 0). Vega is the RMB change in manager wealth for 1 percentage point change in firm annual volatility. Delta is the RMB change in manager wealth for 1 percentage point change in firm stock price. Delta and vega values are in $000’s. CEO ownership is the CEO’s stock ownership of the firm expressed as a fraction of total shares outstanding. Percent sought is the percentage of outstanding shares that the management states that it intends to buyback. Prior BHR is the cumulative buy and hold return of the firm for 30 days prior to the announcement to 2 days before the announcement (-30 to -2). Financial constraints are measured by KZ index (discussed in methodology section). Size is the market value of the firm. B/M is the ratio of book value of firm to its market value. Cash Flow is the operating cash flow of the company scaled by total assets. Actual buyback dummy is a dummy variable equal to 1 if actual repurchase is higher than the sample median and zero otherwise. All variables are in the fiscal year prior to the announcement except actual buyback dummy.
Label Parameter Estimate
Model I Model II Model III Model IV
Intercept 1.38*** 1.22*** 1.29*** 1.41*** (8.57) (7.89) (10.64) (4.76) Vega -0.76*** -0.68*** -0.458*** (5.02) (3.49) (2.91) Delta 0.0113** -0.003 0.004 (1.98) (0.09) (0.44) CEO ownership 0.06*** 0.06*** 0.0724*** (4.45) (3.66) (4.12) Percent sought 0.155 (0.67) Prior BHR -4.01*** (3.29) Financial constraint -0.135*** (3.18) Size -0.001 (0.014) B/M 0.0002 (0.07) Cash Flow -1.25* (0.62)
Actual buyback dummy -0.133
(0.09)
Year cluster Yes Yes Yes
R-square 0.0019 0.0042 0.005 0.015
F-value 5.07*** 9.73*** 5.99*** 3.44***
23
Since I use the latest database (until 2017.6), I cannot do the long-term research, and here delta seems not a good variable in this part. If the vesting period for options (normally 3 years) is included in the data, delta can also be justified as it aligns the managers’ incentive over long term. Compared with delta, CEO ownership always reflects the importance of executives’ wealth on stock repurchase announcement immediately. And the investors will concentrate more on vega instead of delta, because according to behavior finance theory, people will concentrate more on losses than gains. For robustness test, I use the compensation dummy defined in methodology part. The dummy variable is a combination of vega, delta and CEO ownership, it will connect shareholders and managers. When the value is 1, it means that the incentives are coincident, otherwise it will be 0.
Table 6 tests the effect of CEO compensation on market’s reaction to stock repurchase announcement.
In Model I, the coefficient of compensation dummy is 0.88 with 1% significance level. The result shows that executives’ compensation will be read by the investors and the market will react accordingly
In model II, firm size dummy is introduced. In theory, when information asymmetry is high, stock repurchase will have more power of signaling. Firm size is a proper variable to measure information asymmetry because smaller firms have fewer analyst or media coverage, their information is less known by the public. I combine firm size dummy and the compensation dummy. The value will be 1 if the firm size is below average and 0 if the firm is a big one.
24
Table 6: The effect of executive compensation on the initial market reaction to stock repurchaseannouncement: Multivariate Analysis
The table presents the results of regressing initial buyback announcement abnormal returns (CAR) on compensation dummy and control variables in different models. CAR represents 11 day (-5 to +5) cumulative abnormal return around the repurchase announcement date (day 0). Compensation dummy represents the wealth alignment between the executive and shareholder. It takes the value of 1 when delta is high (above average), vega is low (below average) and CEO ownership is above median and 0 otherwise. Percent sought is the percentage of outstanding shares that the management intends to buyback on announcement date. Prior BHR is the cumulative buy and hold return of the firm for 35 days prior to the announcement to 6 days before the announcement (-35 to -6). Financial constraints are measured by KZ index (discussed in methodology section). Size is the market value of the firm. B/M is the ratio of book value of firm to its market value. Cash Flow is the operating cash flow of the company scaled by total assets. Actual buyback dummy is a dummy variable equal to 1 if actual repurchase is higher than the sample median and zero otherwise. All variables are in the fiscal year prior to the announcement except actual buyback dummy. Size and value dummy is 1 for small and value firms respectively and 0 otherwise.
*,**,*** represent significance at 10%, 5% and 1% significance level respectively. Parameter Estimates
Model I Model II Model III
Variables Coeff. t-stat Coeff. t-stat Coeff. t-stat
Intercept 1.51 4.91 1.49 4.81 1.68 4.32 Compensation dummy 0.88*** 2.88 -0.47** -2.19 -0.23 -0.61 Percent sought 0.03 0.72 0.03 0.63 0.02 0.35 Prior BHR -5.01*** -3.90 -4.71*** -3.77 Financial constraint -0.20*** -3.67 -0.22*** 3.23 -0.16*** -3.83 Size (x10-5) -0.37 -0.73 -0.42 -0.86 B/M (x10-2) 0.12 0.33 0.11 0.91 0.11 0.38 Cash Flow -1.47 -0.71 -1.26 0.49 -1.40 -0.71
Actual buyback dummy -0.25*** -3.85 -0.28 0.60 -0.31
Size dummy -0.05 -0.09 0.04 0.08
Compensation x size dummy 1.73*** 3.94
Value dummy 0.42 1.12
Compensation x value dummy 1.82*** 2.05
25
The outcome is almost doubled than that of only compensation dummy, and it is still significant at 1% level. As a result, when the information is highly asymmetric and CEO’s wealth is aligned with that of shareholders, the market will have a stronger reaction to stock repurchase.
In model III, I substitute firm size dummy with value dummy. I use the company’s buy and hold return to measure undervaluation. The coefficient turns out to be 1.82 with the best significance level. The result is similar with that of model II, when the company is more undervalued and CEO’s wealth is aligned with that of shareholders, the market will have a stronger reaction to stock repurchase.
5.3 Actual repurchases
In this part I will use actual repurchase behaviors instead of announcements. Since stock repurchase announcement has already released the signal, or the signal happens during post announcement period, the market will react accordingly and raise the stock price earlier than the implement of repurchase. Since the purpose is already achieved, actual repurchases start to become costly. As a result, the compensation variables should have an opposite relationship between returns and actual repurchases.
The hypothesis is tested in table 7, where actual repurchase is dependent variable. The results show that vega has a significantly positive relation with actual repurchase rate, while the other 2 compensation variables are not significant enough. For the control variables, book-to-market ratio, cash flow and prior return positively related with actual repurchase and percent sought is negatively related with it.
26
Table 7 The effect of executive compensation on actual share repurchases
The table regresses actual repurchases on CEO compensation sensitivity measures and other control variables. Actual repurchases are firms’ actual share re-acquisitions in the first four quarters from the quarter of stock repurchase announcement. Delta is the RMB change in manager’s wealth for a 1 percent change in firm’s share price. Vega is the RMB change in manager’s wealth for a 1 percent change in firm’s annual volatility. Delta and vega values are in thousand RMBs. CEO ownership is the CEO’s stock ownership of the firm expressed as a fraction of total shares outstanding. ln Size is the natural log of the market value of the firm. B/M is the ratio of book value of firm to its market value. Leverage is the ratio of total debt to total assets of the firm. Cash is cash level of the firm scaled by total assets. Cash Flow is the operating cash flow of the company scaled by total assets. Prior return is the cumulative buy-and-hold return of the firm for 35 days prior to the announcement to 6 days before the announcement (-35 to -6). Financial constraints are measured by KZ index (discussed in methodology section). Percent sought is the percentage of outstanding shares that the management states that it intends to buyback.
Model I Model II
Variables Coeff. t-stat Coeff. t-stat
Intercept/sigma 0.776 21.6 0.955 13.6 Vega 0.301*** 4.94 0.322*** 4.08 Delta -0.003 -1.78 -0.003 -1.92 CEO ownership -0.002 -0.47 -0.002 -0.47 ln Size -0.003 -0.26 B/M 0.001** 2.33 Leverage 0.051 1.28 Cash -0.022 -0.45 Cash flow 0.386*** 9.01 Prior return 0.279** 2.44 Financial constraint 0.003 0.28 Percent sought -0.041*** -8.46 pseudo/R-squared 0.0190 0.0812
27
6. Conclusion
In this paper, I study the relationship between stock repurchase and CEO compensation in Chinese A share market.
The research of relationship between stock repurchase and CEO compensation is relatively mature in global market. From background knowledge and former literature, the relationship is reasonable and significant. When a company do stock repurchase, it will release a positive signal to the market, and the market will have a higher expectation the company and then raise the stock price. Since stock and options are major part of CEO compensation especially in recent years, the stock repurchase by the company will directly raise the income of CEO.
Like most of the undeveloped stock markets in the world, stock repurchase in Chinese A share market has a late start. Until 2014, China Securities Regulatory Commission(CSRC) started to lift the control over stock repurchase. Hundreds of listed companies start to do more stock repurchase, but the stock market soon experienced a huge shock in the coming year. Such an emerging and immature stock market need more empirical researches to test whether it is in line with the global rules.
Based on the former literature, I set up the hypothesis that stock repurchase will cause excess return and thus raise CEO compensation. When CEOs have more ownership of the company, they will get higher benefit.
I tested the hypothesis using a sample of 935 times stock repurchase in Chinese A share market from 2013.1 to 2016.6. The outcome is in line with the hypothesis, the excess return exists in A share market and will affect the managers’ compensation in a
28
positive way. Moreover, vega and CEO ownership, level of undervaluation and financial characters will significantly influence the effect of stock repurchase. The paper shows that in Chinese stock market, stock repurchase can release a positive signal and thus get reasonable reaction from the market. The result is in line with western stock market, and hopefully it will make contribution to the development of Chinese stock market research.
29
References
A Edmans et al.(2013).“Equity Vesting and Investment”.Social Science Electronic Publishing .
Ahmed and Waqar. (2016). “Agency issues in share repurchase programmes”. Ethos.bl.uk.
A Srivastav et al.(2014).“CEO inside debt holdings and risk-shifting: Evidence from bank payout policies”.Journal of Banking & Finance , 47 (1) :41-53.
Bonaimé. A. A. (2012). "Repurchases, Reputation and Returns". Journal of Financial & Quantitative Analysis, 47(2): 469-491.
CC Liu and NY Chen.(2015).“”The Signaling Effect of Open-Market Block Share Repurchase Announcements”. Chinese Journal of Accounting ,1 11 (1) :1-28.
Chava. S. and A. Purnanandam. (2010). "CEOs versus CFOs: Incentives and corporate policies". Journal of Financial Economics, 97(2): 263-278.
Chen et al. (2012). "Financial constraints and share repurchases". Journal of Financial Economics, 105(2): 311-331.
D Andriosopoulos et al.(2016).“ Prediction of open market share repurchases and portfolio returns: evidence from France, Germany and the UK”. Review of Quantitative Finance & Accounting , 46 (2) :387-416.
E Liljeblom et al.(2013).“Do dividends signal future earnings in the Nordic stock markets”.Review of Quantitative Finance & Accounting , 44 (3) :493-511.
F Leng and K.M Zhao.(2012).“Insider trading around open-market share repurchases”.Journal of Economics & Finance , 38 (3) :461-491.
30
GL Caton et al. (2016). “Governance and post-repurchase performance”. Journal of Corporate Finance, 39:155-173.
G Lin et al.(2015).“Bank executive compensation structure, risk taking and the financial crisis”.Review of Quantitative Finance & Accounting , 45 (3) :609-639. HC Chen et al.(2014).“Insider Trading and Firm Performance Following Open Market Share Repurchase Announcements”.Journal of Business Finance & Accounting ,41 (1-2) :156–184.
Henrik Cronqvist et al. (2012). “Behavioral consistency in corporate finance: CEO personal and corporate leverage”. Journal of Financial Economics, 103(1):20-40. J Delisle et al.(2014).“Share Repurchases and Wealth Transfer Among Shareholders”.Social Science Electronic Publishing.
JY Suh.(2017).“Managerial Incentives for Risk-Taking and Internal Capital Allocation”.Australian Journal of Management , 42.
K Chan et al. (2010). "Share repurchases as a potential tool to mislead investors". Journal of Corporate Finance, 16(2): 137-158.
L Fei and G Noronha.(2013).“Information and Long-Term Stock Performance Following Open-Market Share Repurchases”.Financial Review,, 48 (3) :461–487. MJ Alderson et al.(2014).“Secured debt and managerial incentives”.Review of Quantitative Finance & Accounting , 43 (3) :423-440.
N Wesson et al. (2014).“Market underreaction to open market share repurchases on the JSE”.South African Journal of Business Managemen, 45 (4) :59-69.
31
Shareholders and Financial Reporting Quality”.Chinese Journal of Accounting, 9 (1) :1-33.
U Bhattacharya and S E. Jacobsen. (2016) .“The Share Repurchase Announcement Puzzle: Theory and Evidence”. Review of Finance , 20 (2).
W Ahmed，et al. (2015). “Executive Compensation and Open Market Share Repurchases”. fma.org
