Authors: Hui Zhou
Categories: Research Article, Business performance, Employee stock ownership plan, Initial announcement, Learning effect, Market expectations, Stock price reaction, Termination announcement
Source: Heliyon
Authors: Hui Zhou
Based on a sample of Employee Stock Ownership Plans announced by Chinese listed companies, this paper explores whether managers learn from the market in the decision relating to the termination of the Employee Stock Ownership Plan. Results show Managers listen to the market and terminate Employee Stock Ownership Plans following market expectations. The more negative the market reaction to the initial announcement, the more likely management is to terminate it, and this effect is more pronounced for firms with non-overconfident managers and in lower market share rank. Moreover, the initial announcement return is reversed by the termination announcement return. When managers follow market expectations to terminate the Employee Stock Ownership Plan with a negative market reaction to the initial announcement, the market reaction to the termination is more positive. Further research finds that compared with the terminations of Employee Stock Ownership Plans that are not consistent with market expectations, the terminations that follow market expectations improve business performance more significantly.
Although Employee Stock Ownership Plan (hereafter ESOP) has gained wide usage in other economies, it has yet to become a common practice for Chinese listed companies until the China Securities Regulatory Commission (hereafter CSRC) released the guidance on ESOP in 2014. Recent studies have primarily focused on the shareholders’ wealth effect of ESOP and its impact on corporate performance. However, in practice, ESOPs have not always been implemented Some companies made initial announcements regarding proposed ESOPs but announced their termination subsequently without implementing them. According to Wind Financial Terminal data, 10% of Chinese listed companies that made initial announcements regarding ESOPs from 2014 to 2021 eventually terminated the plans. For example, a Chinese listed company, Infotmic, announced that “given the significant changes in the securities market, we have decided to terminate the ESOP.” Some companies experienced a significant drop in stock prices following the announcement and implementation of ESOPs, resulting in losses for their employees. Chen et al. [1] found that, in general, the implementation of ESOPs failed to facilitate employee benefit, thus having limited effectiveness in terms of boosting business performance. This implies that in certain circumstances, the termination of ESOPs may not necessarily be detrimental to firm value. The ESOPs implemented at the wrong time, designed in improper ways, or adopted with purposes other than incentive plans may have negative impacts on employees and corporate performance.
It is essential that management is equipped with the necessary information for planning the timing and implementation of ESOPs in order to ensure their effectiveness. The capital market brings together the information about macroeconomic and industry environment, as well as company-specific information, from a large number of market participants including investors and analysts. This information is incorporated into stock prices through market trading [2]. Prior studies have established the presence of a market-learning effect, which suggests that managers rely on the information conveyed by stock prices in decisions relating to investment, mergers and acquisitions (hereafter M&A), and dividend change [[3], [4], [5], [6], [7]]. In this paper, I seek to explore whether managers learn from the market when making decisions regarding the termination of ESOPs. I found that when the market reaction to the initial announcement of ESOP is relatively negative, indicating that the market does not support the ESOP, managers will listen to the market and terminate the ESOP.
This paper's possible contributions Firstly, it adds to the literature on ESOP. Academic studies have discussed the motivations and consequences of implementing ESOPs. They provided evidence for not only the bright sides [[8], [9], [10], [11], [12], [13], [14], [15]] but also the dark sides of ESOP [[16], [17], [18], [19], [20]]. Meanwhile, a number of problems relating to the planning and implementation of ESOP have surfaced in practice [1]. Thus, termination of an ESOP may be deemed necessary, under certain circumstances. However, scant attention has been devoted to such a decision. This study is the first to directly test reasons for the termination of ESOP and examine the consequences of terminating an ESOP. No prior studies have been conducted to explain why firms choose to terminate an incentive scheme such as Employee Stock Ownership Plans, except Liu and Feng [21], Yu [22], and Yu and Shao [23]. They investigated the termination of Equity Incentive Plans, which mainly include stock options and restricted stock plans. My study is related to, but different from these studies because (1) my paper focuses on the termination of the Employee Stock Ownership Plan that is used to motivate employees at all levels, while Liu and Feng [21], Yu [22], and Yu and Shao [23] focused on the Equity Incentive Plans, which have a pre-specified exercise price and are awarded to senior management and core employees; (2) my paper provides an explanation for why firms decide to terminate ESOPs from the perspective of the learning effect, while Liu and Feng [21], Yu [22], and Yu and Shao [23] contended that the probability of exercising the stock option, corporate governance structure, corporate financial performance, and long-term market performance could be used to explain why firms terminate the Equity Incentive Plans; (3) aside from explaining why firms terminate ESOPs, my paper also explores the consequence of ESOP terminations and demonstrates that the learning behavior could help improve business performance, thus telling a more complete story.
Secondly, it enriches the literature on the market-learning effect. Studies by Foucault and Frésard [3], Luo [4], Liu and Mcconnell [5], Tan et al. [6], and Amedeo and Winifred [7] reported that managers rely on the information incorporated in the stock prices when making decisions relating to investment, M&A, and dividend change. This study investigates whether managers learn from the market when deciding to follow through or terminate ESOP, thus providing new evidence for the learning theory. Moreover, it sheds light on the impact of learning behavior on the short-term and long-term consequences of ESOP terminations by examining how the learning behavior influences the market reactions at termination and business performance after termination.
Prior studies have focused on the motivation for carrying out ESOPs. First, Firms may adopt ESOP to increase worker motivation and satisfaction [8]. Employee owners show more positive attitudes, and their interests are more aligned with that of corporations. In line with this, Sun et al. [24] and Song and Liu [25] found that firms in high-tech industries and with a higher proportion of high-tech personnel are more likely to adopt ESOPs, as these firms have the greatest need to motivate and retain their employees. Besides, granting employees shares help send positive signals to the financial market. In line with the signaling theory, Sun et al. [24] and Song and Liu [25] documented that the probability of adopting ESOP is higher for firms with higher profit growth rates and undervalued share prices. Another strand of literature pointed out that ESOPs can also be implemented for purposes other than motivating employees. The entrenchment theory stated that firms adopt ESOP to lower the probability of a takeover. Consistent with the entrenchment theory, Zheng et al. [26] revealed that companies with lower controller shareholding ratios and higher risks of being acquired are more likely to implement ESOP, suggesting that ESOP has become a tool for large shareholders to strengthen corporate control right. Other studies posited that ESOPs are implemented for large shareholders to liquidate their shares or for short-term market value management objectives when the controlling shareholder's shares are pledged [1,19,20].
As for its economic consequences, there was empirical evidence for both the bright and dark sides of ESOP. On the bright side, the literature emphasizes that the positive effects of employee ownership on corporate performance are primarily due to enhanced work attitudes [8,27]. In the long run, ESOP may improve corporate governance, promote innovation and productivity, decrease corporate financialization, and reduce the cost of equity capital [[9], [10], [11], [12], [13], [14]]. Thus, the capital market reacts positively to the adoption of ESOP, and the announcement return is affected by the characteristics of ESOP, corporate financial performance, and corporate governance structure [[28], [29], [30]]. Another body of literature on ESOP focuses on its dark side. As ESOPs can be used as an entrenchment tool [26], they may exert a significant negative impact on corporate financial performance. Faleye et al. [18] documented that the implementation of ESOP negatively affects market valuation, long-term investment, research and development expenses, operating risk, sales growth, employment growth, total factor productivity, and labor productivity. Chen et al. [1] and Sun and Liu [20] found that if the ESOP is adopted for short-term market value management intentions, employees may suffer from losses due to the ESOP. If this is the case, ESOP cannot motivate employees and improve business performance. In line with these studies, Chang and Mayers [16] and Conte et al. [17] proved negative market reactions when ESOPs are implemented. Hao et al. [19] used a sample of ESOPs announced by Chinese listed companies to assess the effects of ESOP, demonstrating that if the ESOP is employed for large shareholders to cash out their shares, investors will detect this, leading to unfavorable market reactions.
In short, past literature has discussed the motivations and consequences of implemented ESOPs. Given that both the bright and dark sides of ESOP have been documented by prior studies, the consequence of ESOP remains a controversial issue. Just as it is found in the academic literature on the dark side of ESOP, a number of problems associated with the design and implementation of ESOP have appeared in practice [1]. Thus, it seems unavoidable or even necessary to terminate an ESOP under certain circumstances. In other words, the termination of an improperly designed ESOP may not necessarily be detrimental to firm value. However, prior literature paid little attention to the termination of ESOP. When studying the market reaction to the initial statement, prior research did not differentiate the implemented and terminated ESOPs. In terms of the long-term impact of ESOP, existing literature only studied the impact of successfully implemented ESOPs on corporate performance and ignored those terminated. Moreover, past literature did not discuss the important questions of why firms decide to terminate ESOP and what the consequences of terminating an ESOP are.
While past literature paid little attention to the termination of ESOP, other studies discussed the termination of another type of incentive scheme-Equity Incentive Plans. Liu and Feng [21] found that long-term market performance is an essential factor to be considered by Chinese listed firms in decisions regarding the termination of Equity Incentive Plans. Yu [22] established that the lower the likelihood that the equity incentive unlocking condition can be met, and the lower the premium of the stock price relative to the exercise price, the more likely the company will terminate the Equity Incentive Plans. Yu and Shao [23] documented that a number of corporate financial performance and corporate governance indicators affect the decision to terminate Equity Incentive Plans. Another strand of the literature focused on the consequence of the termination of Equity Incentive Plans. Yu and Shao [31] found that terminating Equity Incentive Plans generally does not improve the business performance of Chinese listed companies. In contrast, Carter and Lynch [32] revealed that the repricing of stock options (canceling and reissuing new options with a different strike price) enhances firms’ abilities to retain employees. For a sample of employee stock options repricing announcements made by Canadian firms, Grein et al. [33] found a positive three-day announcement-date mean excess return of 4.9%, suggesting that the market anticipates the termination of executive stock options to help retain key employees.
At the initial announcement of ESOP, investors utilize their own information to evaluate it and trade accordingly, leading to stock price reactions in the market. Stock prices will rise if the market supports the adoption of ESOP. Contrarily, if the market perceives that it cannot add value, for example, investors may dispute the timing of implementation and how ESOP is designed, or they believe that it is adopted for purposes other than motivating employees, then stock prices will fall [19]. By observing the market reactions, managers could obtain feedback from the market and use it for decision-making. The reasons First, the stock price is an important source of information [2]. Managers can obtain information on the macroeconomic environment and industry prospects owned by market participants, which is incorporated in stock prices. Previous studies have proved that managers listen to the market and rely on the information conveyed by stock prices in corporate decisions [[3], [4], [5], [6], [7]]. Although managers have detailed information regarding their own companies, the planning and implementation of ESOPs are still affected by external factors including macroeconomic policies, regulatory changes, the stock market environment, industry prospects, and competitors' strategies [21]. And market participants may be more familiar with this external information. Allen [34] pointed out that when business operations become more complex, external information will be increasingly crucial for decision-making. In line with this, Liu and Feng [21] have reported that long-term market performances affect the possibility of Chinese firms terminating their Equity Incentive Plans. This suggests that managers do consider external factors when making decisions relating to the termination of an incentive scheme. Secondly, compared with managers, a single investor may not have an informational advantage. However, market participants, such as individual investors, institutional investors, and analysts, may have a comparative informational advantage as a whole [35]. As long as market participants possess valuable information not known to managers, the latter can utilize it when making decisions. Finally, the stock market generates, aggregates, and disseminates costly and serendipitous information collected from outside investors [35,36]. Given the dispersed nature of information, the stock market provides a venue for investors who are adept at information production to supply their talents to the firm. Investors’ profit-seeking trading motivations save the firm extra search costs typically associated with other information sources, thus reducing the cost of acquiring information for managers [37].
In short, in order to improve decision-making, managers have the motivations to learn from the market, i.e., to obtain and use the information conveyed by stock price signals. When the initial announcement of ESOP triggers a relatively negative stock price reaction, managers may adhere to market expectations, reevaluate their original strategies, and finally decide to terminate the ESOP. The learning hypothesis is summarized as follows.Hypothesis 1The probability that management will terminate ESOP is higher when the market reaction to the initial announcement is more negative.Previous literature has not explored the market reaction to the termination of ESOP; however, the termination of M&A has been studied extensively. Bradley et al. [38] documented a negative shareholder wealth effect for acquirers when the transaction is terminated. Based on the M&A deals in the US, Cole et al. [39] and Liu and McConnell [5] found that the market reaction to the initial announcement and the termination of M&A are negatively related. Using a sample from the Chinese stock market, Tan et al. [6] also revealed a negative relationship between the initial announcement return and the termination announcement return of M&A. Based on this, I expect that market reaction to the initial announcement and termination announcement of ESOP are The negative market reaction to the initial announcement can be interpreted as market investors' disapproval regarding the implementation of the proposed ESOP. If managers take the market's initial expectations into account and then proceed to terminate the ESOP with a relatively negative market reaction to the initial announcement, the market is likely to reward such action with a favorable reaction to the termination announcement. When the market reacts favorably to the initial announcement, indicating market investors' approval of the proposed ESOP, if managers still choose to terminate it afterwards, then the market may react unfavorably to the termination. Hypotheses are summarized as follows.Hypothesis 2AThe market reaction to the initial and termination announcement of ESOP are negatively related.Hypothesis 2BWhen managers listen to the market and terminate the ESOP with a negative market reaction to the initial announcement, the market will react more favorably to the termination than when managers do not listen to the market.
I begin with the (proposed) Employee Stock Ownership Plans announced by Chinese listed companies from 2014 to 2021. To be included in the final sample, each observation must satisfy the following (1) To be included in the sample, the ESOP has to be announced by non-financial companies. I delete the financial companies because they operate under different regulations and have different natures and classifications of financial statement items [40]. (2) To be included in the sample, the ESOP has to be either completed or terminated. For research purposes, I delete the ESOP samples that are still in progress, with the results unknown by December 20, 2022. (3) The company that announced ESOP must have stock price data, corporate accounting data, and corporate governance data available in the China Stock Market & Accounting Research (hereafter CSMAR) Database. These criteria result in a final sample of 1309 ESOP samples used for estimating the baseline regression. Appendix A presents the sample selection procedure and the sample distribution by industry. The sample covers a broad range of industries.
The data of ESOP are from www.cninfo.com.cn, the Choice data platform under East Money, and the Wind Financial Terminal. First, I collect information on the initial announcement date, the termination announcement date, and the characteristics of (proposed) ESOP from www.cninfo.com.cn, which is the official information disclosure website for Chinese listed companies. I search the listed companies' announcements section of www.cninfo.com.cn with “Employee Stock Ownership Plan” in the title of the announcement to find listed companies that issued drafts of ESOP between 2014 and 2021. According to the
Corporate finance indicators, corporate governance indicators, CSRC industry code, stock price data, controlling shareholders' stock pledge data, data on the release of restricted shares, and management earnings forecasts data are from the CSMAR Database. Data on the educational background of employees of listed companies are collected from the Wind Financial Terminal. Firms’ high technology industry status is determined based on Shenyin & Wanguo industry code, which is from the CSMAR database.
To test hypothesis 1, following Liu and McConnell [5], I construct the probit regression model as (1)TERMINATEi = a+b1CAR-Ii + b2CONTROLi+ε
In equation (1), the dummy variable TERMINATE equals one if the company has announced to terminate the ESOP, and otherwise zero. CAR-I is the cumulative abnormal return calculated using the market-adjusted model, i.e., the sum of the differences between the firm's daily stock return and the market return over the three-day interval (−1, 1) around the initial announcement (t = 0) of ESOP. It measures the stock price reaction to the initial announcement and the market's perception of ESOP.
CONTROL is the vector of control variables. Firstly, prior studies pointed out that firms' financial condition and corporate governance structure can affect their decision to terminate an equity scheme [21,22]. Thus, I control for a series of corporate finance indicators and corporate governance indicators, including ①firm's total assets, FIRMSIZE, measured by the natural logarithm of total assets of the firm; ②debt ratio, FIRMLEV, measured by the ratio of total debt to total assets; ③return on total assets, ROA, measured by the ratio of net income to total assets; ④sales growth, GROWTH, measured by the growth rate of sales revenue; ⑤board size, BOARD, measured by the natural logarithm of the number of directors on the board; ⑥top 10 shareholders' shareholding, DISP, measured by the fraction of shares held by the top 10 shareholders; and ⑦management shareholding, MANHOLD, measured by the fraction of shares held by management. The corporate finance indicators and corporate governance indicators are lagged by one year relative to the year in which the initial announcement of ESOP is made.
Secondly, I control for the characteristics of the proposed ESOP, because they may also affect firms' decision to terminate the ESOP. These variables ⑧Non-public offering dummy, ESOPSHARE, equals one if the (proposed) ESOP shares are from non-public offerings, and zero otherwise. ESOP shares can be obtained from non-public offerings, secondary-market purchases, and others. ESOPs with shares from non-public offerings are subject to more regulations than others1; thus, they are more likely to be terminated. ⑨Source of fund, ESOPLEV, equals one if the ESOP is financed by employees' own wealth and salaries, and equals zero if the ESOP is financed by loans from controlling shareholders and other ways. The fact that employees subscribe to ESOP shares with their own wealth and salaries is an indication of their confidence in the company's future performance. Terminating an ESOP financed by employees' own wealth and salaries may demotivate employees to a greater extent than ESOP financed by loans from controlling shareholders. Thus, the probability of termination is lower for ESOP financed by employees' own wealth and salaries than for ESOP financed by loans from controlling shareholders. ⑩Size of ESOP, ESOPSIZE, is measured by the natural logarithm of the (proposed) size of ESOP fund. ⑪Number of employees, ESOPNUM, is measured by the natural logarithm of the (proposed) number of employees participating in ESOP. Terminating an ESOP with a larger number of participating employees may have a more detrimental effect on employee morale than an ESOP with fewer employees involved. Hence, the probability of terminating an ESOP decreases with the number of employees involved in the ESOP. ⑫Duration of ESOP, ESOPDUR, is measured by the natural logarithm of the (proposed) duration of ESOP (in months). The longer the duration of an ESOP, the longer the employees will work for the company to receive the benefits of the ESOP, i.e., the longer the employees' interests are aligned with the company's interests, and the better the expected performance of the ESOP implementation will be [30]. As a result, firms are less likely to terminate ESOPs with longer duration than those with shorter duration.
Thirdly, Liu and Feng [21] pointed out that long-term market performance is a key factor influencing the decision of Chinese firms to terminate an incentive scheme--the Equity Incentive Plans. Therefore I control for ⑬the long-term market performance of the sample firms’ stocks, FIRMRET, proposed by Liu and Feng [21]. For firms that have terminated their ESOPs, FIRMRET is defined as the monthly average return of the stock over the 12 months prior to the termination announcement date of the ESOP; while for firms that have completed their ESOPs, FIRMRET is defined as the monthly average return of the stock over the period starting from the initial announcement date to the end of 2021.
Finally, the industry fixed effect and year fixed effect are also controlled. Appendix B reports the definitions of the main variables.
Equations (2), (3) are constructed to test hypotheses 2A and 2B using the sample of terminated ESOPs. CAR-T is the cumulative abnormal return calculated using the market-adjusted model over the event window of (−1, 1) surrounding the announcement day (t = 0) of ESOP terminations. LISTEN is the dummy variable of listening to the market. LISTEN equals one if management terminates the ESOP with a negative market reaction to the initial announcement (CAR-I<0), and zero otherwise.
The determinants of market reaction to the termination announcement of ESOP have not been examined by prior studies. Nevertheless, a few studies [29,30] have explored the determinants of the market reaction to the initial announcement of ESOP, which may also affect the market reaction to the termination of ESOP. Following Zhang et al. [29] and Xiao et al. [30], I control for a series of corporate finance and governance indicators, ①firm's total assets (FIRMSIZE), ②debt ratio (FIRMLEV), ③return on total assets (ROA), ④sales growth (GROWTH), and ⑤board size (BOARD). The corporate finance indicators and corporate governance indicators are lagged by one year relative to the year in which the termination announcement of ESOP is made. Equations (2), (3) also control for the characteristics of the proposed ESOP, including ⑥non-public offering dummy (ESOPSHARE), and ⑦source of fund (ESOPLEV).(2)CAR-Ti = a+b1CAR-Ii + b2CONTROLi+ε(3)CAR-Ti = a+b1LISTENi + b2CONTROLi+ε
Table 1 presents the summary statistics of the main variables. According to the mean of TERMINATE, 10.0% of the samples have terminated the ESOP. The standard deviation of CAR-I and CAR-T are much greater than their median and mean, respectively, which implies that both the initial announcement and the termination announcement of ESOPs may elicit different market reactions depending on various factors. Since the market response to the initial announcements can vary with particular characteristics and implementation purposes and timing of the proposed ESOP, so may the market response to the termination of ESOP.Table 1Summary statistics of the main variables.Table 1VariablesObservationsMedianMeanSt. Dev.Min.Max.TERMINATE130900.1000.26001CAR-I13090.0190.0290.106−0.4020.553CAR-T131−0.005−0.0010.072−0.2450.324LISTEN13100.4430.46501FIRMSIZE130922.02022.2121.19919.95126.039FIRMLEV13090.3880.4020.1900.0610.859ROA13090.0570.0620.052−0.1460.231GROWTH13090.1660.3860.756−0.5484.765BOARD13092.1972.1050.1791.3862.944DISP13090.6010.5890.1370.1090.918MANHOLD13090.1060.1950.20800.830ESOPSHARE130900.1110.31301ESOPLEV130910.7920.40601ESOPSIZE13098.9238.9051.1764.18012.429ESOPNUM13097.7587.8561.2212.55212.384ESOPDUR13093.5863.5570.4102.4854.970FIRMRET13090.0050.0150.071−0.1300.151Note: The definitions of the variables are provided in Appendix B.Source: Processed data.
Table 2 presents the correlation matrices of the variables in equation (1) (Panel A) and equations (2), (3) (Panel B). It shows that the correlation matrix between the two variables in the correlation matrix has a value of less than 0.90. Hence, there is no issue with multicollinearity. Moreover, the results of the multicollinearity test show that the mean Variance Inflation Factor (VIF) value of equations (1), (2), (3) is 1.66, 1.59, and 1.61, respectively. This indicates that the regression models do not suffer from multicollinearity issues.Table 2Correlation matrix of the main variables.Table 2TERMINATECAR-IFIRMSIZEFIRMLEVROAGROWTHBOARDDISPPanel A Correlation matrix of the variables in equation (1)TERMINATE1.0000CAR-I−0.06261.0000FIRMSIZE−0.0262−0.04281.0000FIRMLEV0.03000.03580.57531.0000ROA−0.0667−0.0691−0.0014−0.23181.0000GROWTH0.04600.0357−0.02080.0909−0.06161.0000BOARD−0.0362−0.07520.26250.15630.0205−0.10101.0000DISP−0.0431−0.0598−0.0284−0.07290.18750.0410−0.08221.0000MANHOLD−0.0133−0.0205−0.3748−0.28680.1122−0.0346−0.13100.2429ESOPSHARE0.14570.1173−0.05950.0229−0.0005−0.00860.0309−0.0522ESOPLEV−0.0592−0.01650.11680.0965−0.0699−0.10290.0771−0.1062ESOPSIZE0.0347−0.00280.41120.17820.12230.01420.07220.0011ESOPNUM−0.0528−0.05440.60980.39210.1149−0.20040.2370−0.0315ESOPDUR−0.04550.02500.14780.00940.0800−0.03870.08370.0002FIRMRET−0.16130.1275−0.1422−0.09380.0606−0.04400.02520.0786MANHOLDESOPSHAREESOPLEVESOPSIZEESOPNUMESOPDURFIRMRET**MANHOLD1.0000ESOPSHARE−0.01591.0000ESOPLEV−0.12560.07441.0000ESOPSIZE−0.15130.18000.09351.0000ESOPNUM−0.2285−0.0053−0.06450.35311.0000ESOPDUR−0.01440.3344−0.03950.09900.16261.0000FIRMRET0.11090.08410.0337−0.0617−0.01650.06111.0000CAR-TCAR-ILISTENFIRMSIZEFIRMLEVROAGROWTHBOARDESOPLEVESOPSHAREPanel B Correlation matrix of the variables in equations (2), (3)CAR-T1.0000CAR-I−0.10811.0000LISTEN0.1564−0.70791.0000FIRMSIZE−0.0608−0.0676−0.01121.0000FIRMLEV−0.08020.1169−0.24500.50461.0000ROA0.1295−0.18000.09780.0306−0.34981.0000GROWTH0.16990.0035−0.0178−0.2004−0.14480.09751.0000BOARD−0.0259−0.1045−0.00240.27080.1124−0.0048−0.04671.0000ESOPLEV0.0565−0.01040.11750.18720.1040−0.0455−0.19500.12661.0000ESOPSHARE0.1556−0.15700.0971−0.10130.02660.13140.0181−0.0744−0.18591.0000Note: The definitions of the variables are provided in Appendix B.Source: Processed data.
According to Table 3, the mean of CAR-I is significantly positive, indicating that the adoption of ESOP is viewed favorably by the market in most cases, which is consistent with previous studies [[28], [29], [30]]. Yet, the frequency of CAR-I being negative is 38.9%, suggesting that the market could still react unfavorably to the initial announcement of ESOP under certain circumstances. For example, the market may perceive that the ESOP implemented with improper timing, designed in improper ways, or adopted with objectives beyond incentive schemes, may lead to losses for the firm. Particularly, evidence from the Chinese capital market shows that when ESOPs are adopted for short-term market value management purposes or for large shareholders to cash out their shares, employees may suffer from losses due to the ESOP and thus are not motivated by ESOPs [1,20], in which case investors will react unfavorably to the initial announcement of ESOP [19]. The average value of CAR-T is negative but not significant, while the frequency of CAR-T being positive is 40.4% (=100%–59.6%). This implies that although the majority of investors react unfavorably to the termination of ESOP, some may still view the termination of ESOP as good news. For example, the market may support the termination of ESOP implemented with improper timing, designed in improper ways, or adopted with purposes other than incentive plans, as such ESOP may be detrimental to firm value. This finding is consistent with Carter and Lynch [32] and Grein et al. [33], who contended that the repricing (canceling and reissuing new options with a different strike price) of improperly designed stock options enhances firms’ abilities to retain employees.Table 3Market reaction to the initial announcement and termination announcement of ESOP.Table 3VariableMeanFrequency of <0t-statisticsCAR-I0.02938.9%11.212CAR-T*−0.00159.6%−0.191Note: ,* and* indicate statistical significance at 10%, 5% and 1%, respectively.Source: Processed data.
Next, as previous studies provided evidence from the Chinese capital market that ESOPs may be adopted for purposes other than motivating employees [1,19,20], I classify the ESOP samples into different groups based on the purpose of carrying out ESOPs and compare the market reactions between these groups. Firstly, I compared the market reaction to ESOP announcements of companies whose controlling shareholdings were pledged in the previous year with the market reaction to ESOPs of companies whose controlling shareholdings were not pledged in the previous year, as ESOPs could be adopted for short-term market value management when controlling shareholders' stocks are pledged [1,20]. Secondly, I compare the market reaction to ESOP announcements of companies with restricted shares released in the previous year with the market reaction to ESOP announcements of companies with no restricted shares released in the previous year, as ESOPs could be adopted for shareholders to cash out their shares after the release of restricted shares2 [19]. Table 4 shows a lower mean and median of CAR-I for ESOPs announced by companies with controlling shareholders' stocks pledged or restricted shares released in comparison with the mean and median of CAR-I for ESOPs announced by companies with controlling shareholders' stocks not pledged or no restricted shares released, while CAR-T exhibits the opposite pattern. In addition, compared to companies with controlling shareholders' stocks not pledged or no restricted shares released, the frequency of CAR-I being negative and CAR-T being positive (=100%-Fre. of <0) is higher for companies with controlling shareholders' stocks pledged or restricted shares released. These results are consistent with the conjecture that in comparison to ESOPs adopted for motivating employees, investors tend to react less favorably (more favorably) to the initial announcement (termination announcement) of ESOPs adopted for large shareholders’ self-interest.Table 4Comparison of market reactions.Table 4Controlling shareholders' stocks pledgedControlling shareholders' stocks not pledgedRestricted shares releasedNo restricted shares releasedVariableMeanMedianFre. of <0MeanMedianFre. of <0MeanMedianFre. of <0MeanMedianFre. of <0CAR-I0.0270.01741.9%0.0340.02335.1%0.0290.01840.8%0.0300.02036.2%CAR-T0.0030.00149.4%−0.011−0.01374.4%0.0050.01145.7%−0.009−0.01167.6%Source: Processed data.
Table 5 shows that the CAR-I of terminated samples is significantly lower than that of completed ones, which is consistent with the hypothesis that the probability that management will terminate ESOP is higher when the market reaction to the initial announcement is more negative.Table 5Market reaction to the initial announcement--completed ESOP and terminated ESOP.Table 5VariableMeanDifference in mean (①-②)t-statistics①CAR-I (completed ESOPs)0.0320.0211.97②CAR-I (terminated ESOPs)0.011Note: ,* and* indicate statistical significance at 10%, 5% and 1%, respectively.Source: Processed data.
Table 6 reports the results of equation (1). Column 1 shows a significantly negative coefficient of CAR-I, indicating that the more negative the market reaction to the initial announcement, the more likely management is to terminate it, which confirms hypothesis 1. This result implies that managers extract information from investors through stock price movements in the market when making decisions relating to the termination of ESOP. My finding supports the market learning hypothesis, which suggests that managers consider the market's perception when deciding whether to terminate the ESOP or follow through with the ESOP.Table 6Market-learning effect and the termination of ESOP.Table 6Variable(1)(2)(3)(4)(5)TERMINATETERMINATETERMINATETERMINATETERMINATEAll ESOPsHigher market share rankLower market share rankOverconfident managersNon-overconfident managersCAR-I−0.150** (−2.45)−0.065 (−0.62)−0.175** (−2.22)−0.093 (−1.06)−0.203*** (−2.89)FIRMSIZE0.001 (0.08)−0.021 (−0.81)0.004 (0.35)−0.017 (−0.95)−0.013 (−0.72)FIRMLEV0.067 (1.49)0.091 (1.12)0.061 (1.05)0.168** (1.99)−0.038 (−0.66)ROA−0.277** (−2.10)−0.378 (−1.64)−0.185 (−1.05)−0.464** (−2.22)−0.499*** (−2.61)GROWTH0.011 (1.18)0.039*** (2.63)−0.002 (−0.17)−0.014 (−0.78)0.015 (1.28)BOARD−0.050 (−1.38)−0.092 (−1.40)−0.023 (−0.49)−0.113* (−1.87)−0.062 (−1.01)DISP−0.048 (−0.97)0.046 (0.48)−0.056 (−0.93)0.024 (0.78)−0.109*** (−2.91)MANHOLD0.034 (1.00)−0.092 (−1.54)0.124*** (2.84)−0.047 (−0.52)−0.070 (−1.20)ESOPSOURCE0.203*** (8.47)0.218*** (5.20)0.213*** (7.02)0.050** (2.11)0.063*** (2.88)ESOPLEV−0.043*** (−2.71)−0.091*** (−3.33)−0.013 (−0.66)−0.122** (−2.44)−0.243** (−2.39)ESOPSIZE0.002 (0.26)0.007 (0.48)0.002 (0.24)−0.007 (−0.93)−0.012 (−1.49)ESOPNUM−0.005 (−0.51)−0.007 (−0.37)−0.008 (−0.68)−0.020* (−1.88)−0.014 (−1.47)ESOPDUR−0.067*** (−3.79)−0.058* (−1.75)−0.074*** (−3.50)−0.058*** (−3.11)−0.051*** (−2.99)FIRMRET−0.396** (−2.01)−0.127* (−1.93)−0.351** (−2.22)−0.232* (−1.90)−0.369** (−2.08)YEARYESYESYESYESYESINDYESYESYESYESYESN1309684625600688Pseudo R^2^0.1850.1560.1540.1880.186Note: ,* and*** indicate statistical significance at 10%, 5% and 1%, respectively. Z-statistics are reported in parentheses. The constant is not reported.Source: Processed data.
As for the control variables, ESOPs with shares from non-public offerings are more likely to be terminated, as such ESOPs are subject to more regulations than others. ESOPs with longer duration or financed by employees' own wealth and salaries are less likely to be terminated, as terminating such ESOPs may demotivate employees and reduce the expected benefit of ESOPs to a larger extent [30]. Besides, the probability of terminating an ESOP decreases with the stock market performance. Employees holding their firm's stocks will suffer from losses due to poor stock market performance, in which case ESOPs are more likely to be terminated. This finding is consistent with Liu and Feng [21].
To further test the learning hypothesis, I then examine what factors help explain managers' propensity to learn from the market. Firstly, managers of firms in lower market share ranks may not have comparative informational advantages, thus having the greatest need and learning motive to collect information from the market to help with decision-making. Secondly, overconfident managers may put more weight on their own knowledge, information, and ability [41], thus ignoring the importance of external information conveyed by market prices. To test the above conjecture, I classify the firm-year samples with sales revenue lower than the industry-year median into the group with the lower market share rank; the rest are classified into the group with the higher market share rank. Following Lin et al. [41], I define management as overconfident if the forecasted earnings are higher than the actual earnings. As shown in Columns 2–5 of Table 6, the coefficients on CAR-I are significantly negative in groups with lower market share rank within the industry and non-overconfident managers but not significant in groups with higher market share and overconfident managers. This suggests that the learning effect is more pronounced for firms with non-overconfident managers and in lower market share ranks, as these firms have a higher propensity to learn from the market.
According to Column 1 of Table 7, the relationship between CAR-I and CAR-T is significantly negative, suggesting that when the market reaction to the initial announcement of ESOP is more positive (negative), the reaction to its termination is more negative (positive). This shows that the market reactions to the initial announcement and termination announcement are reversed, which confirms hypothesis 2A. In Column 2 of Table 7, the coefficient of CAR-I remains significantly negative when the samples with positive CAR-I are excluded, still supporting hypothesis 2A. The findings in Columns 1 and 2 of Table 7 are consistent with Cole et al. [39], Liu and McConnell [5], and Tan et al. [6], who found that the valuation effects in response to withdrawn mergers are worse when the initial stock price reaction at the time of the announced merger bid was higher. The termination effect appears to be a reversal of the initially anticipated benefits or losses that were impounded in the stock price at the time of the initial announcement. Since the termination of ESOP can reverse the perceived benefits (losses) that were initially expected, the favorable (unfavorable) valuation effects at the time of the initial announcement may be eliminated at the time of the termination announcement.Table 7Market expectations and stock price reaction to the termination of ESOP.Table 7Variable(1)(2)(3)CAR-TCAR-TCAR-TTerminated ESOPsTerminated ESOPs with CAR-I<0Terminated ESOPsCAR-I−0.092* (−1.76)−0.272** (−2.28)LISTEN0.041* (1.94)FIRMSIZE−0.001 (−0.16)0.002 (0.16)−0.001 (−0.21)FIRMLEV−0.020 (−0.50)−0.087 (−1.26)−0.010 (−0.25)ROA0.180* (1.77)−0.127 (−0.56)0.181* (1.82)GROWTH0.013 (1.49)0.019* (1.79)0.014 (1.52)BOARD−0.024 (−0.55)0.019 (0.38)−0.002 (−0.09)ESOPLEV0.004 (0.30)−0.003 (−0.10)0.003 (0.19)ESOPSHARE0.020 (1.33)0.028 (1.25)0.021 (1.37)Observations13158131Adj. R^2^0.0510.0690.059Note: ,* and*** indicate statistical significance at 10%, 5% and 1%, respectively. t-statistics are reported in parentheses. The constant is not reported.Source: Processed data.
Column 3 of Table 7 shows a significantly positive relationship between LISTEN and CAR-T, indicating that the accumulative abnormal return at the termination increases with the probability that managers listen to the market. When managers follow market expectations and terminate the ESOP that has a negative market reaction at the initial announcement, the market reaction at termination is more favorable than when managers do not follow market expectations. This confirms hypothesis 2B. Table 1 shows that the mean and median of CAR-T are negative. Table 3 shows that the frequency of CAR-T being negative is 59.6%. Thus, the statistics from Table 1, Table 3 suggest that most of the CAR-T are negative, i.e., the market investors generally consider the termination of ESOP bad news and react negatively to the terminations. However, when managers listen to the market and terminate the ESOP that has a negative initial announcement return, the market reaction at termination is more favorable or at least less negative than when managers do not listen to the market. Thus, when firms terminate an ESOP disapproved by investors in the first place (indicated by the negative market reaction to the initial announcement), this kind of termination may not necessarily be considered detrimental by the investors. Instead, investors may even perceive that it can add value to the firm.
Firstly, as I only include firms that made the initial announcement of ESOP in the sample, there might be a sample selection bias in the empirical tests. If a firm's decision to announce an ESOP is systematic, the pool of ESOP samples is non-random. Hence, the sample selection bias needs to be addressed, i.e., selection into firms announcing ESOP. Failure to account for the sample selection bias could lead to biased estimates.
To correct for sample selection bias, I follow Chen et al. [42] and apply the sample selection approach proposed by Heckman [43] as follows.
In the first step, the probit model is used to estimate the probability that a listed firm makes an initial announcement of ESOP. The sample selection correction term—the inverse Mill's ratio—is calculated accordingly and is denoted by IMR. This sample selection correction term is used as an additional regressor in the baseline regression estimated in the second step to correct for the sample selection bias. The dependent variable of the probit regression is ESOPDUM (a dummy variable that equals one if the firm makes an initial announcement of ESOP, and zero otherwise). Prior studies [[24], [25], [26]] provided explanations of why firms decide to carry out ESOP and make an initial announcement of ESOP. Following Sun et al. [24], Song and Liu [25], and Zheng et al. [26], I include several variables that are deemed to be potential predictors of firms making an initial announcement of ESOP, ①high-tech industry dummy, HIGHIND, equals one if the company is in one of the four high-tech communication, electronics, computer, and biomedicine, and zero otherwise3; ②PB ratio, PB, measured by the ratio of market share price to book value of equity; ③proportion of technical personnel, HIGHEM, measured by the ratio of the number of employees with a bachelor's degree or above to the total number of employees at the end of the year; ④profit growth rate, PROFITG, measured by the average growth rate of net income over the past two years; ⑤controller shareholding, SHRCR, measured by the fraction of shares held by the controlling shareholder. Corporate financial indicators such as ⑥firm's total assets (FIRMSIZE) and ⑦debt ratio (FIRMLEV) are also included.
Panel A of Table 8 reports the results of the probit regression estimated in the first step. It shows that firms in high-tech industries tend to announce ESOPs. Firms with higher profit growth rates, lower controller shareholding, larger sizes, and lower debt ratios are also more likely to propose to carry out ESOPs. These findings are consistent with prior studies [[24], [25], [26]].Table 8Endogeneity Heckman two-step models.Table 8Variable(1)ESOPDUMPanel A First Probit regressionHIGHIND0.174*** (4.63)PB−0.009 (−1.31)HIGHEM0.002 (1.51)PROFITG0.118*** (3.11)SHRCR−0.008*** (−6.74)FIRMSIZE0.077*** (4.82)FIRMLEV−0.266*** (−2.78)N13,628Pseudo R^2^0.168Variable(1)TERMINATEAll ESOPsPanel B Second Controlling for the sample selection correction termCAR-I−0.187** (−2.56)IMR0.066 (1.11)CONTROLYESYEARYESINDYESN1136Pseudo R^2^0.189Note: ,* and*** indicate statistical significance at 10%, 5% and 1%, respectively. t-statistics are reported in parentheses. The constant is not reported.Source: Processed data.
In the second step, the inverse Mill's ratio obtained from the results of the probit regression estimated in the first step, IMR, is added as an additional control variable in the baseline regression equation (1). Panel B of Table 8 presents the estimation results. The coefficient of CAR-I remains negative after I correct for the sample selection bias.
Secondly, I measure CAR-I and CAR-T over the five-day (instead of three-day) interval of (−2,2) surrounding the announcement day (t = 0). Results in Panel A of Table 9 show that CAR-I is still significantly and negatively related to TERMINATE and CAR-T, and the coefficient on LISTEN remains significantly positive.Table 9Robustness tests.Table 9Variable(1)(2)(3)TERMINATECAR-TCAR-TAll ESOPsTerminated ESOPsTerminated ESOPsPanel A Measure CAR over five-day intervalCAR-I−0.157*** (−2.62)−0.108* (−1.81)LISTEN0.041* (1.92)CONTROLYESYESYESYEARYESNONOINDYESNONON1309131131Pseudo R^2^ or Adj. R^2^0.1880.0480.059Variable(1)(2)(3)TERMINATECAR2-TCAR2-TAll ESOPsTerminated ESOPsTerminated ESOPsPanel B Use market model to calculate CARCAR2-I−0.166** (−2.38)−0.091* (−1.69)LISTEN0.040* (1.90)CONTROLYESYESYESYEARYESNONOINDYESNONON1278128128Pseudo R^2^ or Adj. R^2^0.1840.0460.054Variable(1)(2)(3)TERMINATECAR-TCAR-TAll ESOPsTerminated ESOPsTerminated ESOPsPanel C Include financial companiesCAR-I−0.149** (−2.44)−0.092* (−1.76)LISTEN0.039* (1.88)CONTROLYESYESYESYEARYESNONOINDYESNONON1327133133Pseudo R^2^ or Adj. R^2^0.1910.0490.062Variable(1)TERMINATEAll ESOPsPanel D Change the way of measuring FIRMRET**CAR-I−0.146** (−2.41)CONTROLYESYEARYESINDYESN1309Pseudo R^2^0.182Variable(1)CAR-TTerminated ESOPsPanel E Change the threshold of LISTEN**LISTEN0.045** (1.99)CONTROLYESYEARNOINDNON131Adj. R^2^0.057Note: ,* and*** indicate statistical significance at 10%, 5% and 1%, respectively. z-statistics or t-statistics are reported in parentheses. The constant is not reported.Source: Processed data.
Thirdly, CAR-I and CAR-T are calculated using the market model instead of the market-adjusted model. Following Shi et al. [44], I estimate the market model for each stock.(4)Ri,t = αi + βiRm,t + ε~i,t~
In equation (4), Ri,t and Rm,t, denote the daily return of stock i and market on trading day t, respectively. To estimate the intercept and slope, I set a subsample containing 250 trading days (300–50 days prior to the initial announcement date) as the estimation window and then calculate the value through OLS, displaying as αiˆ and βiˆ. Thus, the normal return is expressed by equation (5), as (5)E(Ri,t)=αi^+βi^Rm,t
Abnormal returns (ARi,t) can be defined as in equation (6).(6)ARi,t=Ri,t-E(R~i,t~)
Next, the initial announcement return based on the market model, CAR2-I, is calculated as the sum of the abnormal returns over the three-day interval (−1, 1) surrounding the initial announcement day (t = 0) of ESOP. Similarly, the termination announcement return based on the market model, CAR2-T, is calculated as the sum of the abnormal returns over the three-day interval (−1, 1) surrounding the termination announcement day (t = 0) of ESOP. Results in Panel B of Table 9 still favor hypotheses 1, 2A, and 2B when the cumulative abnormal returns are calculated using the market model.
Fourthly, the regression results reported in Table 6, Table 7 are based on a sample that excludes financial companies. As a robustness check, I include the financial companies in the sample and estimate the regressions again. The results reported in Panel C of Table 9 are quantitatively similar to the results reported in Table 6, Table 7 if I keep the observations of financial firms in the sample.
Fifthly, as described in section 3, I follow Liu and Feng [21] to define the long-term market performance, FIRMRET, as the monthly average return of the stock over the period starting from the initial announcement date to the end of 2021 for firms that have completed their ESOPs, and include FIRMRET as one of the control variables in equation (1). However, it is likely that some firms completed their ESOPs before 2021. As a robustness check, for firms that have completed their ESOPs, I measure the long-term market performance as the monthly average return of the stock over the period starting from the initial announcement date to the date of completion4 and estimate equation (1) again. Results in Panel D of Table 9 show that the conclusion still holds after I change the way of measuring the long-term market performance for firms that have completed their ESOPs.
Sixthly, as described in section 3, I define LISTEN = 1 as cases where the initial announcement return is less than the threshold of 0% and the ESOP is terminated. For robustness, I change the threshold and redefine LISTEN = 1 as cases where the initial announcement return is less than 0.96% (the median value of CAR-I for the terminated samples) and the ESOP is terminated. As shown in Panel E of Table 9, the coefficient on LISTEN remains positive after this robustness check.
Further, if managers terminate the ESOP with an unfavorable market reaction to the initial announcement following market expectations, will this market-learning behavior improve business performance? Equation (7) is constructed to test this conjecture, based on the terminated ESOP samples. ΔROA measures the percentage change in ROA after the termination of ESOP, compared with the previous year. The explanatory variable of interest is LISTEN, the dummy variable of listening to the market, as defined in Section 3. Control variables include FIRMSIZE, FIRMLEV, GROWTH, BOARD, DISP, and MANHOLD, defined in Section 3.(7)ΔROAi = a+b1LISTENi + b2CONTROL~i+~ε
Table 10 shows significantly positive coefficients of LISTEN, which means that the business performance (as measured by the percentage change in ROA) increases with the probability that managers listen to the market in decisions regarding the termination of ESOP. Compared to the terminations of ESOP that were not consistent with market expectations, those that followed market expectations improve business performance more significantly. Using a sample of terminated Equity Incentive Plans (another type of incentive scheme), Yu and Shao [31] found that the termination of Equity Incentive Plans generally does not improve the business performance of Chinese listed companies. However, when managers learn from the market and terminate the ESOP that was not supported by investors in the first place (as indicated by negative market price reactions to the initial announcement), this kind of termination will benefit the firm. This indicates that the management's market-learning helps improve their decision-making and business performance because stock price conveys valuable information to managers.Table 10Long-term economic consequences of ESOP termination.Table 10Variable(1)(2)ΔROAΔROATerminated ESOPsTerminated ESOPsLISTEN0.148** (2.44)0.130** (2.28)FIRMSIZE0.017 (0.33)0.027 (0.60)FIRMLEV−0.220** (−2.48)−0.170* (−1.91)GROWTH0.016** (2.44)0.013** (2.16)BOARD−0.080* (−1.88)−0.036 (−0.75)DISP0.006 (0.54)0.008 (0.61)MANHOLD−0.031 (−0.24)−0.029 (−0.16)YEARNOYESINDNOYESObservations123123Adj. R^2^0.1290.148Note: ,* and*** indicate statistical significance at 10%, 5% and 1%, respectively. t-statistics are reported in parentheses. The constant is not reported.Source: Processed data.
Based on a sample of Employee Stock Ownership Plans (ESOP) announced by Chinese listed companies, this paper discusses the termination of ESOP from the perspective of the market-learning effect. Results show Managers listen to the market and terminate ESOP following market expectations--the more negative the market reaction to the initial announcement, the more likely management is to terminate it. Moreover, the market reaction to the initial announcement and termination announcement are reversed. When managers follow market expectations to terminate the ESOP with a negative market reaction to the initial announcement, the market reaction to the termination is more positive. Further research shows that the termination of ESOP that followed market expectations improves business performance more significantly than the termination that did not follow market expectations.
This paper shows that stock prices convey useful information for managers and help improve decision-making. Therefore, managers from listed companies could consider their firm's stock price as a valuable information source. In addition, investor protection should be enhanced to encourage investors to collect information and make full use of it in their trade, thus increasing the information content of stock prices, which will then serve as better guidance for managers' decision-making.
This study has several limitations. Firstly, since ESOPs are still relatively new in the Chinese securities market, I was only able to obtain a relatively small sample for the empirical research. As more and more listed companies propose to carry out the ESOP, it will become more common in the Chinese securities market. Future research could examine the termination of ESOP using a larger sample and a longer sample period to check if the conclusion of my study still holds and whether the results vary with the macroeconomic environment. Secondly, while my research provides evidence from an emerging market that managers learn from the market when making decisions related to the termination of ESOP, there are bound to be differences in the learning behavior due to the differences in the regulatory and institutional environment between developed and emerging markets. Future research could explore this topic in the developed market to see if it will reach the same conclusion.
Hui Zhou: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.
Dr. Hui Zhou work was supported by 10.13039/501100010226Colleges and Universities Innovation Project of Guangdong Province of China [2022WTSCX179]; 10.13039/501100002338Humanities and Social Sciences Youth Foundation, Ministry of Education of the People's Republic of China [22YJC630221].
The authors do not have permission to share data.
The authors declare no conflict of interest.