The Disciplinary Effects of Short Sales on the Controlling Shareholders

Transcription

1 The Disciplinary Effects of Short Sales on the Controlling Shareholders Shenglan Chen School of Economics and Management Inner Mongolia University Hohhot , China Bingxuan Lin College of Business Administration University of Rhode Island Kingston, RI Rui Lu* Lingnan (University) College Sun Yat-Sen University Guangzhou , China Hui Ma School of Accountancy Shanghai University of Financial and Economics Shanghai , China * Corresponding author We acknowledge funding for project xxxxxxx from the National Natural Science Foundation of China. 1

2 The Disciplinary Effects of Short Sales on the Controlling Shareholders Abstract: Although the literature (Massa et al., 2015; Hirshleifer, Teoh, and Yu, 2011; Karpoff and Lu, 2010) demonstrates the disciplinary effects of short sales on managers, no study has looked at how short sales can also mitigate the agency costs of the controlling large shareholders. Using the introduction of short sales in China as an exogenous event, we show that short sales also serve as an effective mechanism in curbing expropriation by the controlling large shareholder. Such a disciplinary effect is more pronounced for firms with higher ownership concentration and bankruptcy risk. Meanwhile, the disciplinary effects of short sales are muted when other alternative governance mechanisms are in place. Our paper provides new insights into the effects of short sales in the corporate governance domain, specifically in mitigating the agency cost between controlling and minority shareholders. Key Words: Short Sales; Controlling Shareholder; Expropriation; Agency Cost; Corporate Governance; Chinese Market 2

3 1. Introduction Beyond the obvious effects of short sales on price efficiency (Miller, 1977; Hong and Stein, 2003; Danielsen and Sorescu, 2001; Boehmer and Wu, 2013), some studies have investigated how short sales may serve as a corporate governance mechanism and play a disciplinary role. The intuition is that short sellers who specialize in information gathering and processing and are eager to profit from negative corporate news may provide governance forces to discipline managers opportunistic behavior (Jensen, 2005; Karpoff and Lou, 2010). Focusing on the U.S. equity market, studies have shown that the downside price pressure resulting from informed short selling helps to constrain managers incentives to manipulate earnings (Massa, Zhang, and Zhang, 2015; Fang, Huang, and Karpoff, 2016) and engage in value-destroying acquisitions ex ante (Chang, Lin, and Ma, 2015). He and Tian (2015) find that value-destroying sub-optimal investments in innovative projects due to agency problems can be mitigated by the threat of depressing stock prices from short sellers. Taken together, this evidence suggests that managers are deterred from pursuing their self-interests under the pressure of short selling. However, the literature does not adequately address how short sales can also mitigate the agency cost of large shareholders. It is important to note that well-dispersed ownership is relatively rare outside of the U.S., and that the presence of large shareholders with 3

4 substantial blocks of shares is more common for most European and Asian companies (La Porta, Lopez-de-Silanes, and Shleifer, 1999; La Porta et al., 2002; Faccio and Lang, 2002). In this broader setting, the central agency problem is the conflict of interest between the controlling and minority shareholders. Specifically, controlling shareholders pursue private benefits of control, which is referred to as tunneling (Johnson et al., 2000). Empirical studies have found support for such tunneling behavior of large shareholders. For example, Baek, Kang, and Lee (2006) find evidence for tunneling in Korean business groups. Claessens et al. (2002) argue that the expropriation of minority shareholders by controlling large shareholders is a major concern in East Asian economies. Liu and Tian (2012) show that controlling shareholders in China can directly intervene in corporate decisions and keep corporate resources strictly under their control. Under the weak legal institutions and imperfect investor protections in China, tunneling is prevalent. Controlling shareholders in China obtain private benefits through various forms of tunneling activities such as fund occupation (Jiang, Lee, and Yue, 2010; Wang and Xiao, 2011) and related party transactions (Cheung, Rau, and Stouraitis, 2006; Peng, Wei, and Yang, 2011). The intuition that short sales can serve as a governance mechanism against controlling shareholders is considered as follows. The power of controlling shareholders to expropriate outside investors is moderated by 4

5 their financial incentives not to do so (La Porta et al., 2002). An important source of disincentive to exploit company wealth is the cost of expropriation. Short sellers are incentivized to uncover corporate misbehavior and unfavorable information that is not incorporated into the current stock price. When controlling shareholders engage in the extraction of private benefits, short sellers can trade against the controlling shareholders misconducts, leading to a stock price decline and damage to large shareholders wealth. Due to liquidity constraints, controlling shareholders cannot sell their entire stakes upon a negative signal, and suffer significant losses once the stocks are highly shorted (Bond, Edmans, and Goldstein, 2012; Wang, 2014). Under this circumstance, as long as the controlling shareholders are concerned about the expropriation costs, the threat of short selling can serve as an external governance mechanism and thus prevent controlling shareholders from engaging in tunneling activities ex ante. The recent introduction of short sales in China and the widespread problem of tunneling by controlling large shareholders create an ideal setting in which to explore the disciplinary role of short sales on controlling shareholders expropriation behavior. Using a sample of Chinese listed firms between 2007 and 2014, we find that an exogenous increase in short-selling pressure can restrain controlling shareholders extraction of company resources. The magnitude of this effect is economically significant. 5

6 Specifically, after controlling for a set of firm characteristics, we find that the removal of short sale constraints leads to a 7.35% decrease in total related party transactions, a 19% decrease in industry-adjusted related party transactions, and a 20% decrease in exploitive related party transactions (i.e., excluding specific terms from total annual related party transactions) for the treatment (pilot) group compared with the control group. To establish the channel through which short sales affect controlling shareholders behavior, we also find evidence that a greater downside risk in stock price is associated with higher disciplinary effects. Our additional tests also examine how alternative governance mechanisms may be effective substitutes in curbing controlling shareholders misbehavior. We include several robustness checks to make sure our results are free of bias or other alternative explanations. First, we use a propensity score matching approach in our test to make sure our results are not driven by potential endogeneity concerns. Second, we subject our results to a placebo test, where firms are fictionally included in a short sales list in a different time. Our findings no longer appear in the placebo test. We also perform two diagnostic tests and verify that the parallel trend assumption holds for our difference-in-differences analysis. Finally, we rule out the possibility that our results are driven by the ex post learning effect rather than the ex ante disciplining effect from short-selling pressure by controlling for the 6

7 short interest ratio in our model. Our paper emphasizes the governance role played by short sales and contributes to the literature in the following ways. First, this is the first analysis of how short sales affect controlling shareholders behavior in general and related party transactions in particular. With a special focus on the agency conflicts between the controlling and minority shareholders in emerging markets, this paper provides new evidence that short-selling activity can serve as a governance mechanism to mitigate the agency cost of controlling large shareholders. Second, our paper is related to the growing studies arguing for the real effects of financial markets. Starting from Hayek (1945), who argues that prices are a useful source of information, researchers have found that managers use the new information learned from financial market prices to guide their real decisions (Goldstein and Guembel, 2008; Bond, Edmans, and Goldstein, 2012). Our findings show that the short-selling market can also affect controlling shareholders decisions. Third, our paper adds to the corporate governance literature, which studies the trade-off between voice and exit (Maug, 1998). This stream of literature focuses on voice as the primary disciplining device, even though recent studies have shown that exit is also a governance mechanism in itself (Admati and Pfleiderer, 2009; Edmans, Fang, and Zur, 2013). Our evidence shows that a preventative exit such as a short sale can be a powerful force used to discipline 7

8 controlling shareholders. Finally, our findings contribute to the policy debate over the benefits and costs of short selling. Previous research has demonstrated that short sellers can not only enhance price informativeness, but also reduce principal-agent problems. Our results indicate that the prospect of short selling conveys additional external benefits to investors by disciplining controlling shareholders expropriation. This paper is organized as follows. Section 2 gives the institutional background of short sales in China. Section 3 describes the sample and data. Section 4 reports the empirical results. Section 5 examines the cross-sectional heterogeneity in the disciplinary effects of short sales. Section 6 reports the results for a battery of robustness checks. Section 7 studies the economic consequences of such disciplining effects. Section 8 concludes the paper. 2. Institutional Background The Chinese stock market began exploring the possibility of making margin trading and short selling available around In 2008, the China Securities Regulatory Commission (CSRC) picked 11 top brokerages to run a trial program, which lasted for about 2 years. A more structured pilot program was introduced in March 2010, when 90 constituent stocks from the Shanghai Stock Exchange (SSE) 50 Index and the Shenzhen Stock Exchange (SZSE) Component Index were made 8

9 eligible for margin trading and short selling. Since then, the list of firms eligible for short selling has changed over time, with the CSRC revising the qualification list in November 2011, January 2013, September 2013, and September The number of pilot stocks increased from 90 to 900 during , accounting for nearly 16% of our total sample. Of the 900 pilot stocks, 55.56% (500) are listed on the SSE and 44.44% (400) on the SZSE. The removal of short-selling constraints put pressure on pilot stocks relative to non-pilot stocks. The pilot program thus eliminated the need to directly estimate changes in short-selling pressure, a difficult task (Chang, Lin, and Ma, 2015). Using an exogenous change that causes an increase in short-selling pressure for pilot firms can reduce the omitted variable bias and allow us to establish a causal relationship between short-selling pressure and expropriation by the controlling shareholder. The most interesting feature of this experiment is that the list of firms eligible for short selling has changed over time, which creates both time-series and cross-sectional variations in the short-selling restrictions for firms. In other words, different from the single event of Regulation SHO, the events we use here are staggered over time. A key advantage of the staggered deregulation is that there are multiple exogenous shocks that affect different firms at different times. Our study is therefore free of a common identification difficulty faced by studies with a single shock, that is, the 9

10 existence of potential omitted variables coinciding with the shock that directly affect tunneling. Finally, in developed economies, short sellers have alternative ways to create short positions, such as options and futures, which makes it harder to isolate the effect of certain short-selling mechanism from others (Boehmer and Wu, 2013; Stratmannn and Welborn, 2013). However, there is no clear substitute for short sales in China, and we can directly estimate the real effects of short sales on controlling shareholders expropriation. 3. Sample Descriptions, Variable Definitions, and Descriptive Statistics 3.1 Sample Descriptions We start our sample with all of the companies listed on the SSE and SZSE in March 2010, when the short sales program was initially introduced. We then obtain data from the Chinese Stock Market and Accounting Research (CSMAR) database to construct proxies for controlling shareholders expropriation and control variables. To construct our final sample, we follow several procedures. We begin by deleting firms in the financial services and utilities industries, as disclosure requirements and accounting rules are significantly different for these regulated industries. To ensure that controlling shareholders have meaningful stakes in their companies, we follow prior studies (La Porta, 10

11 Lopez-de-Silanes, and Shleifer, 1999; Faccio and Lang, 2002; Lin et al., 2011) and require the ownership of a controlling shareholder to be at least 10%. Finally, we also remove observations with missing values when we construct the control variables. To mitigate the effects of outliers, all of the continuous variables are Winsorized at the 1% level. Our final sample includes 10,969 firm-year observations between 2007 and Variable Definitions Short-selling constraints were removed gradually beginning in 2010, and stocks were added to the designated list based on a set of criteria such as market capitalization, turnover, trading age, and the number of individual shareholders (see Appendix A1). To identify the effects of short sales on controlling shareholders expropriation, we first create an indicator variable, Treat, to denote firms in the pilot program as the treatment group. Specifically, Treat equals one if a firm s stock is designated as a pilot stock in year t under pilot program and zero otherwise. A large body of empirical evidence has shown that controlling shareholder usually use related party transactions as a means of tunneling to extract private benefits from minority shareholders, especially in emerging markets (La Porta et al., 1999, 2000; Jian and Wong, 2010; Claessens, Fan, and Lang, 2006). Following these studies, we construct four measures to proxy for the controlling shareholders expropriation of 11

12 minority shareholders. Related party transactions are defined as transactions with related entities such as shareholders, members of the board of directors, and affiliated companies. The first measure is total annual related party transactions (RPTa), defined as total annual related party transactions scaled by total assets. To mitigate the influence of industry, the second measure is industry-adjusted related party transactions (RPTb), defined as the difference between the RPTa and its industry median. In addition, given that some types of related party transactions such as cooperation projects, license agreements, R&D outcomes, and managerial compensations are more likely driven by regular company operations, we exclude these types of related party transactions from RPTa and construct our third measure: industry-adjusted exploitive related party transactions (RPTc). The final measure (RPTd) is abnormal related party transactions, as proposed by Jian and Wong (2010). This measure is simply the residual term of an OLS regression model capturing any normal components of related party transactions that are associated with industry classifications and firm characteristics. Following the tunneling literature (Maury and Pajuste, 2005; Bharath, Jayaraman, and Nagar, 2013), we control for a vector of firm characteristics that may affect controlling shareholders expropriation in our analysis. Our control variables include firm size (Size), leverage (Lev), firm value (Tobin s Q), firm performance (ROA), ownership of the 12

13 controlling shareholder (SH1), divergence between the cash-flow rights and voting rights of the controlling shareholder (CV), ownership of the second largest shareholder (SH2), board size (BSize), proportion of independent directors (IndepR), managerial compensation (Salary), and ownership propriety (SOE). We provide detailed variable definitions in Appendix A Descriptive Statistics To minimize the effect of outliers, we Winsorize all control variables at the 1 st and 99 th percentiles. Table 1 presents the statistics for the main variables. Specifically, Table 1 reports the number of observations and the mean, median, standard deviation, and deciles (99% and 1%) and quartile (75% and 25%) distributions of the variables. The results show that the means of RPTa, RPTb, RPTc, and RPTd are 0.367, 0.157, 0.156, and 0.003, respectively. The mean and median of Treat are 15.9% and 0.0%, respectively, and the 75% quartile is also 0.0%, indicating that the treatment group is rather small compared with the control group. Regarding the control variables, an average firm has a book value asset of 3.62 billion, a leverage of 0.482, Tobin s Q of 1.935, controllers shareholding of 0.419, a divergence between cash-flow rights and voting rights of 0.831, a proportion of independent directors of 0.207, and managerial compensation of 0.39 million. {Insert Table 1 here} 13

14 4. Empirical Results 4.1 Baseline Difference-in-Differences Results When stocks are added to the designated short-selling list, they start to experience short-selling pressure. Meanwhile, the CSRC s decision is unlikely to be related to controlling shareholders odds of tunneling. As such, the removal of short-selling constraints satisfies both the relevance and exclusion conditions for our empirical tests. Moreover, the list of firms eligible for short selling changes over time, which creates both time-series and cross-sectional variations in terms of the short-selling restrictions for firms. Thus, we use the removal of short-selling constraints as a quasi-natural experiment to identify the causal effect of short-selling pressure on controlling shareholders expropriation. Specifically, following Bertrand and Mullainathan (2003), we adopt the difference-in-differences approach as follows: RPT Treat X (1) it t t i t it Where RPTit is four different measures of controlling shareholders expropriation activities for firm i in year t; Treatt is an indicator variable equal to one if the stock is specified as pilot stock in year t and zero otherwise; and Xt is a vector of control variables as discussed in Section 3. αi is the firm fixed effect used to control for time-invariant omitted firm characteristics and ensure that estimates of β represent average, within-firm changes in controlling shareholders expropriation over time 14

15 rather than simple cross-sectional correlations. at is the year fixed effect used to account for nationwide factors such as macroeconomic conditions that may simultaneously affect controlling shareholders expropriation as well as the likelihood of a pilot firm s stock being chosen. An advantage of our experiment is that different firms are specified as pilot firms at different times, which avoids a common identification difficulty faced by studies with a single shock, that is, the existence of potential omitted variables coinciding with the shock that directly affect expropriation. Thus, the staggered deregulation of short-selling constraints implies that the control group is not restricted to only those firms that are never chosen as pilot firms. The regression results estimating equation (1) are reported in Table 2. The dependent variables are total annual related party transactions (RPTa), industry-adjusted total annual related party transactions (RPTb), industry-adjusted exploitative related party transactions (RPTc), and abnormal related party transaction (RPTd). The coefficient of interest is the DiD estimator, β. In column (1), the DiD estimator, which is the coefficient estimate on Treat, is and significant at the 10% level, suggesting that pilot firms experience a decline of RPTa that is lower than that of control firms. This difference is economically sizeable, as it represents approximately 7.35% (0.027/0.367) of the average change of RPTa for all of the firms in our sample (0.367). The results also show a 15

16 negative and statistically significant relationship between controlling shareholders expropriation and the removal of short-selling constraints in columns (2) and (3). In terms of economic significance, the coefficient estimates imply that pilot firms experience 19% (0.03/0.157) and 20% (0.031/0.156) decreases in related party transactions. Overall, the preceding results 1 indicate that short-selling pressure has a disciplining effect on controlling shareholders expropriation behavior. {Insert Table 2 here} 4.2 The Role of Price Downside Risk In theory, after removing the constraints on short sales, stock prices display a greater sensitivity to negative information than to positive information, suggesting that a pilot program affects risk asymmetrically by increasing the downside risk only (Goldstein and Guembel, 2008). We can anticipate a corresponding change in managerial behavior when a firm s risk changes. Angelis, Grullon, and Michenaud (2016) find that a pilot group must modify its CEO incentive contracts to respond to the change in increased downside risk. Consistent with this line of argument, we suggest that downside risk is a potential channel through which short-selling pressure disciplines controlling large shareholders. If a firm s stock is more sensitive to negative news, then the cost of tunneling becomes greater, as 1 As the measure in column (4) is the residual of the regression, it has different meanings when compared with Models (1) (3). The negative coefficient suggests that the presence of short sales lowers the likelihood of expropriation by controlling shareholders; however, we cannot measure the decline of related party transactions like we do in the other columns. 16

17 the holdings held by controlling shareholders are subject to more downside risks. In this section, we first replicate Grullon, Michenaud, and Weston s (2015) study for our sample, and confirm that the stock prices for the pilot firms become more sensitive to bad news. We examine the behavior of daily returns of both the pilot and control firms during bearish and bullish stock market days. The objective of these tests is to provide evidence that pilot stocks experience an asymmetric shock to stock price risk. We test the firms stock price reactions to bad market-wide news. Specifically, we resort to difference-in-differences analyses in which we sort daily market-wide returns into five quintiles to test whether the returns of firms in the pilot group become more negative on very bad market days (first quintile of market returns) after initiating a pilot program than before. Panel A of Table 3 reports the results of this analysis. The two groups of firms do not display different patterns of returns on bad market days before the removal of the short-selling constraints. However, the firms in the pilot group have more negative returns than the control firms after the adoption of a short sales program during the worst market days (lowest quintile). The difference-in-differences coefficient is statistically significant at the 5% level. The results confirm that pilot firms are exposed to more downside risk. Next, we use a triple differences technique to examine whether the 17

18 disciplining effects are more pronounced for pilot firms that experience more downside risk. Following Angelis, Grullon, and Michenaud (2016), we create a dummy variable that equals one if the firm is in the top quintile of changes in stock price return sensitivity to negative market returns around the announcement date (DS_Risk). We measure the changes in stock price return sensitivity to negative market returns as changes in the firms stock returns when the daily stock market returns fall into the lowest quintile of stock market return days. The results are reported in Panel B of Table 3. The coefficients on DS_Risk Treat are significantly negative in all of the columns, indicating that controlling large shareholders in firms with higher downside risk are less likely to engage in tunneling activities. The results provide support for our conjecture that short sales result in greater price downside risks and curb controlling shareholders self-serving behavior that may destroy firm value. 5. Cross-sectional Analyses To better understand how short sales may serve as a governance mechanism and mitigate the agency problem of the controlling shareholder, we examine how our results may vary for different levels of controlling shareholders ownership and bankruptcy risk. 5.1 Controlling Shareholders Ownership Large shareholders may use their control rights to maximize their own utility, which may come at the expense of other shareholders (Shleifer and 18

19 Vishny, 1997). However, it is not rational for controlling shareholders to pursue such private benefits of control when there is an increasing level of short-selling pressure. As short sellers target and attack the misconduct of firms, stock prices go down and lower the value of controlling shareholders ownership. The presence of short sales thus increases the expropriation cost and reduces controlling shareholders incentive to tunnel. As controlling shareholders face a higher cost if they have high ownership stakes in their firms, we expect that the disciplining effect is more pronounced for firms with higher ownership concentrations. Table 4 provides the results. We find that the coefficients on SH1 Treat are negative and significant in all of the models, suggesting that firms with higher levels of controlling ownership experience less tunneling after the introduction of short sales. {Insert Table 4 here} 5.2 Bankruptcy Risk Even if short-selling constraints are removed, an investor must locate the shares for borrowing to subsequently deliver them to the buyer when he or she intends to short a particular firm, and thus bears higher searching costs to locate shares and loan fees (Duffie, Garleanu, and Pedersen, 2002). Thus, short sellers target certain firms to lower the shorting cost and maximize their trading profits. Desai, Krishnamurthy, and Venkataraman. (2006) show that heavily shorted firms experience 19

20 poor subsequent performance and a higher rate of delisting. Maffett, Owens, and Srinivasan (2016) also find that short selling increases the informativeness of equity market indicators of financial distress. It appears that short sellers tend to target firms experiencing financial distress. Therefore, we argue that controlling shareholders in distressed firms face greater disciplinary action, as these firms tend to attract more short sellers. We hypothesize that the disciplining effect is greater for high-risk firms than for financially healthy firms. In particular, we use the Altman Z-score to measure the bankruptcy risk. Following Altman (1968), we compute the Z-score as follows: Z -score 1.2 X 1.4 X 3.3 X 0.6 X 1.0 X (2) where X1 is the ratio of working capital to total assets, X2 is the ratio of retained earnings to total assets, X3 is the ratio of earnings before interest and taxes to total assets, X4 is the ratio of market value of equity to the book value of total debt, and X5 is the ratio of sales to total assets. We report the results in Table 5. We find that the coefficients on the interaction term Z-score Treat are negative and significant at the 5% level. These results confirm our expectation that the disciplining effect of short selling on controlling shareholders expropriation is more pronounced for firms with higher bankruptcy risk. {Insert Table 5 here} 20

21 6. Additional Tests Here, we consider how alternative governance mechanisms, including monitoring by other non-controlling large shareholders (NCLSs) and the legal environment, may interact with short-selling pressure to mitigate the agency cost of large shareholders. 6.1 Non-Controlling Large Shareholders (NCLSs) Theoretically, NCLSs play important roles in corporate governance (Bennedsen and Wolfenzon, 2000) because they have power as well as incentives to monitor the controlling shareholder. Empirical evidence also shows that NCLSs can enhance a firm s value by monitoring its controlling shareholder s expropriation of private benefits (Laeven and Levine, 2008; Attig, Ghoul, and Guedhami, 2009). As such, we hypothesize that the disciplinary role of short sales varies with the monitoring of NCLSs. Specifically, the role of short sales is most significant when NCLSs are unable to perform their monitoring duty. However, when NCLSs can effectively monitor a controlling shareholder and prevent him or her from looting the company, then the presence of additional market forces such as short sales do not result in significant improvements in lessening the agency cost of large shareholders. We use the ratio of the sum of the top 10 independent NCLS voting rights to controlling shareholder voting rights to measure the governance effect of NCLSs (Balance). The results are presented in Table 6. We find 21

22 that the coefficients on Balance Treat are significantly positive, suggesting that the disciplining effect of short selling on controlling shareholders expropriation is less pronounced when NCLSs play a greater monitoring role. {Insert Table 6 here} 6.2 Legal Environment The importance of the legal system in preventing expropriation by controlling shareholders has been well emphasized by La Porta et al. (1998), La Porta et al. (2002), and Dyck and Zingales (2004). Intuitively, controlling shareholders are less likely to engage in tunneling in markets that better protect minority shareholders. If the removal of short-selling constraints has a disciplining effect on controllers expropriation, it should be more effective for firms located in places with weaker legal systems. We use the sub-index of marketization (legal environment) developed by Fan, Wang, and Zhu (2011) to gauge the extent to which the legal environment influences the disciplinary role of short sales. In Table 7, we find that the disciplining effect is less pronounced for firms located in high quality legal environments. {Insert Table 7 here} 6.3 Addressing the Endogeneity Issue Using the pilot program as a quasi-natural experiment, our results 22

23 show that short-selling pressure has a disciplining effect on controlling shareholders expropriation. However, our findings are likely to be endogenous, as stocks must meet certain criteria (e.g., turnover, market capitalization) to be included in the list (Chang, Cheng, and Yu, 2007). Unobservable firm heterogeneity correlated with both inclusion in the list and controlling shareholders expropriation is likely to bias the results. To mitigate the endogeneity problem, we conduct the tests using propensity score matching and then a placebo test. Our matching technique follows a study by Rosenbaum and Rubin (1983). The purpose of matching is to pair each pilot firm with a non-pilot firm in year t on the basis of fundamentals. To eliminate the order effect, observations are randomly ordered before matching (Dehejia and Wahba, 2002). Specifically, we estimate the propensity score using a logistic regression model to regress Treat on the 10 baseline covariates in Model 1, and then use nearest-neighbor matching to specify a new controlled group. Finally, we use the difference-in-differences matching estimator. Standard matching estimators in combination with the difference-in-differences methodology can not only significantly improve the quality of non-experimental evaluation results (Blundell and Costa Dias, 2000), but also eliminate unobserved time-invariant differences in controlling shareholders expropriation between pilot and non-pilot firms that standard matching estimators fail to eliminate (Smith and Todd, 2005). 23

24 The results of Table 8 show that coefficients on Treat are significantly negative, which is consistent with the baseline findings reported in Table 2. {Insert Table 8 here} The second identification strategy is to conduct a placebo test using the fictional deregulation of short-selling constraints. We conduct a placebo test as follows. For the first-time inclusion of pilot firms in a share-sales list, we create a fictional inclusion that occurs four years before the actual inclusion. 2 At the same time, we maintain the assignment of the treated and control firms, that is, the treated and control firms in the placebo test are the same treated and control firms as those in the baseline tests. If baseline results are driven by the inherent and unobservable differences between treated and control firms, we can obtain similar results even using the fictional deregulation. If the baseline results do not hold, then we can conclude that our findings are not driven by an omitted variable bias. The results are presented in Table 9. We find that the coefficients on Treat are insignificant and become positive, which is different from our baseline results. Overall, these two identification tests suggest that short-selling pressure has a negative causal effect on controllers expropriation. {Insert Table 9 here} 6.4 Assumptions for DiD Tests 2 We choose four years as the fictional time of inclusion to ensure we can assign all of the sample firms outside of our initial sample period between 2010 and

25 The validity of the difference-in-differences estimates depends on the parallel trend assumption, i.e., in the absence of the treatment, the average change in controlling shareholders expropriation exhibits parallel trends. We perform two diagnostic tests 3 and find corresponding evidence to show that the parallel trend assumption is not violated. The first test is conducted to estimate the pre-deregulation difference in controlling shareholders expropriation between pilot and non-pilot firms. The second test is conducted to test the timing of changes in expropriation around the deregulation of short-selling constraints (Roberts and Whited, 2012; Fairhurst and Serfling, 2016). We find that a small and statistically insignificant decrease in expropriation before short-selling constraints are removed, but a large statistically significant decline in deregulation of the constraints. Overall, these findings confirm that the parallel trend assumption is satisfied and that it is appropriate to use a difference-in-differences approach. 6.5 Alternative Explanation Bond, Edmans, and Goldstein (2012) point out that the learning effect is an important channel through which financial markets affect real corporate decisions. Therefore, our results may be driven by an ex post learning effect from short-selling activities rather than an ex ante disciplining effect from short-selling pressure. One may argue that 3 To save space, we do not report the tables here. They are available upon request. 25

26 controlling shareholders can learn information from stock prices and use it to guide their expropriation decisions. This is known as the feedback effect explanation. To differentiate the disciplining effects of short sales from the feedback effects stemming from short interest, we follow Chang, Lin, and Ma (2015) to incorporate short interest into our model and re-estimate our main results. The results are reported in Table 9. We find results consistent with those reported in earlier tables after controlling for short interest, thus ruling out this alternative feedback effect explanation. {Insert Table 10 here} 7. Market Response to the Disciplinary Role of Short Sales We show that expropriations by controlling shareholders are less severe when the firm is faced with short-selling pressure. Here, we explore whether market investors interpret the disciplinary role of short sales in the same way. We estimate the market reaction toward announcements of related party transactions and determine whether cumulative abnormal returns (CAR) differ between firms in the pilot program and firms not in the pilot program. In general, we expect CAR to be less negative if short sales play a disciplinary role in curbing controlling shareholders from exploiting company wealth. To examine the effect on market reactions, we conduct the following regression model: 26

27 CAR (- n, n) it Treatt Controlst -1 i t it (3) where CAR (-n, n)it refers to the cumulative abnormal returns of the related party transactions for firm i in year t and is computed as the residuals of the market model based on 5-, 11-, and 21-day event windows (-2, 2), (-5, 5), and (-10, 10), respectively, where day 0 is the announcement date. Specifically, we use the daily stock return during the period of (-210, -11) to estimate the market model, with the CSMAR value-weight return as the market return. We then use the coefficients derived from this stage to compute the residual from the market model during the event windows. Controls is a set of control variables. Table 11 reports the results. We find that the coefficients on Treat are significantly positive, suggesting that firms under short-selling pressure experience better (less negative) market responses to their related party transaction announcements. {Insert Table 11 here} 8. Conclusion Using a sample of Chinese listed firms between 2007 and 2014, we document that an exogenous increase in short-selling pressure can restrain controlling shareholders expropriation of company wealth. We also find supporting evidence that the disciplinary role of short sales come from the price downside risks brought forth by active shorts. The price downside risk provides a disincentive for controlling shareholders to seek private 27

28 benefits due to higher expropriation cost. Additionally, we find that the disciplinary effects of short sales are greater when other governance mechanisms are less effective. The benefits of short sales in improving market price and liquidity have been well documented. Recently, many studies have shown how short sales can function as a governance mechanism in mitigating the agency problem between managers and shareholders. Our paper extends this line of research and shows that it can also help to lower the agency cost of controlling large shareholders. Our findings add to the debate over the pros and cons of short sales in the corporate governance domain. They also enrich the literature on the market mechanism of corporate governance and its interaction with internal governance systems. 28

29 References: [1] Admati, A. R., P. Pfleiderer The Wall Street walk and shareholder activism: Exit as a form of voice. Review of Financial Studies. 22(7): [2] Altman, E. I Financial ratios, discriminant analysis and the prediction of corporate bankruptcy. The Journal of Finance. 23(4): [3] Angelis, D. D., G. Grullon, S. Michenaud The effects of short-selling threats on incentive contracts: Evidence from a natural experiment. Working Paper. [4] Attig, N., S. E. Ghoul, O. Guedhami Do multiple large shareholders play a corporate governance role? Evidence from East Asia. Journal of Financial Research. 32(4): [5] Baek, J.-S., J.-K. Kang, I. Lee Business groups and tunneling: Evidence from private securities offerings by Korean chaebols. Journal of Finance. 61: [6] Bennedsen, M., D. Wolfenzon The balance of power in closely held corporations. Journal of Financial Economics. 58(1-2): [7] Bertrand, M., S. Mullainathan Enjoying the quiet life? Corporate governance and managerial preferences. Journal of Political Economy. 111(5): [8] Bharath, S. T., S. Jayaraman, V. Nagar Exit as governance: An empirical analysis. The Journal of Finance. 68(6): [9] Blundell, R., M. Costa Dias Evaluation methods for non-experimental data. Fiscal Studies. 21(4): [10] Boehmer, E., J. Wu Short selling and the price discovery process. Review of Financial Studies. 26(2): [11] Bond, P. E., A. Edmans, I. Goldstein The real effects of financial markets. Annual Review of Financial Economics. 4(1): [12] Chang, E. C., J. W. Cheng, Y. Yu Short-sales constraints and price discovery: Evidence from the Hong Kong market. Journal of Finance. 62(5): [13] Chang, E. C., T. Lin, X. Ma Does short-selling threat discipline managers in mergers and acquisitions decisions? Working Paper. [14] Chen, Z., G. N. Dong, M. Gu The causal effects of margin trading and short selling on earnings management: A natural experiment from China. Working Paper. [15] Cheung, Y., L. Rau, R. Stoutaitis Tunneling, propping, and expropriation: Evidence from connected party transactions in Hong Kong. Journal of Financial and Economics. 82(2): [16] Claessens, S., J. P. H. Fan, L. H. P. Lang The benefits and costs of group affiliation: Evidence from East Asia. Emerging Markets Review. 7(1): [17] Claessens, S., S. Djankov, J. P. H. Fan, and L. H. P. Lang Disentangling the Incentive and Entrenchment Effects of Large Shareholders. Journal of Finance. 57(6): [18] Danielsen, B. R., S. M. Sorescu, Why do option introductions depress stock 29

30 prices? A study of diminishing short sale constraints. Journal of Financial and Quantitative Analysis. 36: [19] Dehejia, R. H., S. Wahba Propensity score-matching methods for non-experimental causal studies. Review Economics Statistics. 84(1): [20] Desai, H., S. Krishnamurthy, K. Venkataraman, Do short sellers target firms with poor earnings quality? Evidence from earnings restatements. Review of Accounting Studies. 11: [21] Duffie, D., N. Gârleanu, L. H. Pedersen Securities lending, shorting, and pricing. Journal of Financial Economics. 66(2 3): [22] Dyck, A., L. Zingales Private benefits of control: An international comparison. Journal of Finance. 59(2): [23] Edmans, A., V. W. Fang, E. Zur The effect of liquidity on governance. Review of Financial Studies. 26: [24] Faccio, M., L. H.P Lang The ultimate ownership of Western European corporations. Journal of Financial Economics. 65(3): [25] Fairhurst, D. J., M. Serling Employment protection, investment, and firm growth. Working Paper. [26] Fan, G., X. Wang, H. Zhu NERI Index of Marketization of China s Provinces 2011 Report. Beijing: Economic Science Press. (In Chinese) [27] Fang, V., A. Huang, J. M. Karpoff Short selling and earnings management: A controlled experiment. Journal of Finance. 71(3): [28] Goldstein, I., A. Guembel Manipulation and the allocational role of prices. Review of Economic Studies. 75(1): [29] Grullon, G., S. Michenaud, J. Weston The real effects of short-selling constraints. Review of Financial Studies. 28(6): [30] Hayek, F. A The use of knowledge in society. American Economic Review. 35(4): [31] He, J., X. Tian SHO time for innovation: The real effects of short sellers. Working Paper. [32] Hirshleifer, D., S. H. Teoh, J. J. Yu Short Arbitrage, Return Asymmetry, and the Accrual Anomaly. Review of Financial Studies. 24(7): [33] Hong, H., J. C. Stein Differences of opinion, short-sales constraints, and market crashes. Review of Financial Studies. 16(2): [34] Jensen, M Agency costs of overvalued equity. Financial Management. 34(1): [35] Jian, M., T. J. Wong Propping and tunneling through related party transactions. Review of Accounting Studies. 15: [36] Jiang, G., C. Lee, H. Yue Tunneling through inter-corporate loans: The China experience. Journal of Financial Economics. 98: [37] Johnson, S., R. La Porta, A. Shleifer, F. Lopez-de-Silanes Tunneling. American Economic Review. 90(2): [38] Karpoff, J. M., X. Lou Short sellers and financial misconduct. Journal of Finance, 65(5): [39] Laeven, L., R. Levine Complex ownership structures and corporate 30

31 valuations. Review of Financial Studies. 21(2): [40] La Porta, R., F. Lopez-de-Silane, A. Shleifer, R. W. Vishny Law and finance. Journal of Political Economy. 106(6): [41] La Porta, R., F. Lopez-de-Silane, A. Shleifer Corporate ownership around the world. Journal of Finance. 54: [42] La Porta, R., F. Lopez-de-Silane, A. Shleifer Investor protection and corporate governance. Journal of Financial Economics. 59: [43] La Porta, R., F. Lopez-de-Silane, A. Shleifer, R. W. Vishny Investor protection and corporate valuation. Journal of Finance. 57: [44] Lin, C., Y. Ma, P. Malatesta, Y. Xuan Ownership structure and the cost of corporate borrowing. Journal of Financial Economics. 100(1):1-21. [45] Liu, Q., G. Tian Controlling shareholder, expropriations and firm s leverage decision: Evidence from Chinese non-tradable share reform. Journal of Corporate Finance. 18(4): [46] Maffett, M. G., E. L. Owens, A. Srinivasan, Pessimistic-trading restrictions and the information environment: Evidence from short-sale constraints and default prediction around the world. Working Paper. [47] Massa, M., B. Zhang, H. Zhang Invisible hand of short selling: Does short-selling discipline earnings management? Review of Financial Studies. 28(6): [48] Maug, E Large shareholders as monitors: Is there a tradeoff between liquidity and control? Journal of Finance. 53(1): [49] Maury, B. A. Pajuste Multiple large shareholders and firm value. Journal of Banking and Finance. 29(7): [50] Miller, E Risk, uncertainty, and divergence of opinion. Journal of Finance. 32(4): [51] Peng, W., K. Wei, Z. Yang Tunneling or propping: Evidence form connected transactions in China. Journal of Corporate Finance. 17(2): [52] Petersen, M. A Estimating Standard Errors in Finance Panel Data Sets: Comparing Approaches. Review of Financial Studies. 22(1): [53] Roberts, M., T. Whited Endogeneity in empirical corporate finance. Handbook of the Economics of Finance. [54] Rosenbaum, P. R., D. B. Rubin The central role of the propensity score in observational studies for causal effects. Biometrika. 70(1): [55] Shleifer, A., R. W. Vishny A survey of corporate governance. Journal of Finance. 52(2): [56] Smith, J., P. Todd Does matching overcome Ladondes critique of nonexperimental estimators. Journal of Econometrics. 125(1 2): [57] Stratmann, T., J. Welborn The options market maker exception to SEC Regulation SHO Journal of Financial Markets. 16(2): [58] Wang. K, X. Xiao Controlling shareholders tunneling and executive compensation: Evidence from China. Journal of Accounting and Public Policy. 30(1): [59] Wang, Z Short sellers, institutional investors, and corporate cash holdings. 31

32 Working Paper. 32

33 Appendix A1: Short Requirements issued by China Security Regulatory Commission Requirements 1 Qualified stocks have been traded on an exchange for more than three months. 2 Qualified stocks should have more than 200 million tradable shares outstanding (or more than RMB800 million in market value). 3 Qualified stocks should have no fewer than 4,000 individual shareholders. 4 In any given day during the past three months, the average daily turnover should be no lower than 15% of the market index turnover (or the average daily trading value should be no lower than RMB50 million), the average daily return should not deviate more than 4% from the market return, and the return volatility should not reach five times the market index volatility. 5 Qualified stocks have completed non-tradable shares reform, so that all shares are tradable. 6 Qualified stocks are not investigated currently by the CSRC, i.e., the firm is not under a merger, buyout, reorganization, or investigation of possible illegal activities. Summarized by Chen, Dong, and Gu (2015). 33

34 Appendix A2: Variables Definitions Variables Definitions RPTa Total annual related party transactions, the ratio of total annual related party transactions to the total assets. RPTb Industry-adjusted total annual related party transactions, the difference between RPTa and the median RPTa in the same year and in the same industry classification. RPTc Exploitative related party transactions (industry-adjusted), the ratio of exploitative related party transactions (excluding some types of related party transaction from the total annual related party transactions) to total assets. RPTd Abnormal related party transactions, the residual of the forecast related party transaction model. See Jian and Wong s (2010) study. Treat A dummy variable equal to one if the firm is chosen as a pilot firm in year t and zero otherwise. Size The natural logarithm of total assets. Lev The ratio of total debt to total assets. Tobin s Q The ratio of market value to book value of total assets. ROA The ratio of net income to total assets. SH1 The ownership of the controlling shareholder. CV The ratio of cash-flow rights to voting rights of the controlling shareholder. SH2 The ownership of the second largest shareholder. BSize The natural logarithm of the number of directors on the board. IndepR The proportion of independent directors on the board. Salary The natural logarithm of the top three highest executive compensations SOE A dummy variable equal to one if the firm is state-owned and zero otherwise. DS_Risk A dummy variable equal to 1 if the firm is in the top quintile of changes in stock price return sensitivity to negative market returns around the announcement date. We measure changes in stock price return sensitivity to negative market returns as changes in the firms stock returns when the daily stock market returns fall into the lowest quintile of stock market return days (as shown in Table III). Z-score Bankruptcy risk, computed based on the methodology used in Altman (1968). Balance The ratio of the ownership of non-controlling shareholders to the ownership of controlling shareholders. MKT The Legal Environment Index developed by Fan Wang, and Zhu. (2011). Short Interest The ratio of shares in short position to total shares outstanding. 34