CHINA JOURNAL OF ACCOUNTING STUDIES 2019, VOL. 7, NO. 3, 293–316 https://doi.org/10.1080/21697213.2019.1676044 ARTICLE Short-selling and corporate innovation: evidence from the Chinese market a b c d Chuntao Li , Hongmei Xu , Liwei Wang and Peng Zhou Department of Finance, School of Finance, Zhongnan University of Economics and Law, Wuhan, China; Department of Finance, International Business College, South China Normal University, Guangzhou, China; c d Department of Clearing, Zhengzhou Commodity Exchange, Zhengzhou, China; Department of Finance, Economics and Management School, Wuhan University, Wuhan, China ABSTRACT KEYWORDS Short-selling; corporate Short selling has been demonstrated as an important external innovation; information corporate governance mechanism, which can discipline managerial asymmetry behaviours and mitigate principle-agent conﬂicts in developed markets. With a data set of Chinese public companies, we examine the governance eﬀect of short selling on corporate innovation in China. We ﬁnd that short selling has a signiﬁcantly positive eﬀect on corporate innovation regarding both innovation quantity and qual- ity. Our cross-sectional tests show that the positive eﬀect of short- selling is more pronounced for ﬁrms with weaker internal and external corporate governance. Lastly, we document that short- selling improves corporate innovation through lowering ﬁrms’ information asymmetry and improving the eﬃciency of managerial contract. Our results indicate that short-selling is a necessary com- plementary mechanism of ﬁrms’ corporate governance system in China. 1. Introduction As an important trading mechanism in ﬁnancial markets, short-selling has attracted great attention from governments, academia and investors. After the ﬁnancial crisis in 2008, western developed markets (e.g. the US) have imposed more stringent constraints on short-selling; however, China has launched a long-awaited pilot scheme on 31 March 2010, allowing pilot ﬁrms on a designated list to be sold short (Chang, Luo, & Ren, 2014).Theeventprovidesuswith a rare opportunity to study the impact of short-selling on an underdeveloped ﬁnancial market, such as China. This paper aims to investigate whether short-selling aﬀect corporate innovation in China. Prior studies have investigated the impact of short-selling on security returns and ﬁrm policies from various perspectives. On the one hand, traditional literature has examined the eﬀects of short-selling on asset prices and returns (Bris, Goetzmann, & Zhu, 2007; Chang, Cheng, & Yu, 2007;Miller, 1977;Saﬃ &Sigurdsson, 2010). For example, major literature has found that short-selling could improve price eﬃciency by compounding negative CONTACT Hongmei Xu firstname.lastname@example.org Department of Finance, International Business College, South China Normal University, Guangzhou 510631, China Paper accepted by Guliang Tang. © 2019 Accounting Society of China 294 C. LI ET AL. information into stock prices (Bris et al., 2007; Diamond & Verrecchia, 1987;Saﬃ & Sigurdsson, 2010). On the other hand, empirical studies focusing on the eﬀects of short- selling on corporate decisions are burgeoning in recent years. The intuition underlying is that short sellers may provide external governance forces to discipline managers’ opportu- nistic behaviour by providing short-selling threat. For example, Massa, Zhang, and Zhang (2015) and Fang, Huang, and Karpoﬀ (2016) ﬁnd short-selling has a signiﬁcant impact on earnings management. Chang, Lin, and Ma (2018) and Chu (2015)investigate theeﬀects of short-selling on managers’ mergers & acquisitions and product market development deci- sions, respectively. On the Chinese side, prior literature also ﬁnds that short-selling reduces ﬁrms’ earnings management as well as the probability of ﬁnancial restatements (Chen & Liu, 2014; Zhang, Zhou, & Li, 2016). Few studies have examined the impact of short-selling on corporate innovation. For example, He and Tian (2016) have found evidence that short-selling could boost ﬁrms’ innovation output by increasing ﬁrms’ exposure to patenting-related litigation risk. Massa, Wu, Zhang, and Zhang (2016) have explored the eﬀect of short-selling on innovation input in a cross-country setting. However, all these studies have based their analysis on the market conditions of developed countries, such as the US and the European markets. For under- developed markets (e.g. China) that are experiencing economic transactions, the underlying mechanisms that short-selling aﬀect corporate innovation may diﬀerent due to the under- developed law systems and market supervisory systems. Therefore, in China, how short- selling aﬀects ﬁrms’ decision of corporate innovation and the underlying mechanisms through which short-selling aﬀect corporate innovation are still unclear. Based on existing theories of short-selling, we promote two competing hypotheses on how short-selling can aﬀect corporate innovation: ﬁrst, short-selling may improve corporate innovation by holding-up managers’ sub-optimal investments in innovative projects. The principal-agent theory predicts that managers not properly monitored have intentions to under invest long-term and risky innovative projects to enjoy private beneﬁts such as ‘quiet life’ or to eliminate career concerns (Bertrand and Mullainathan, 2003;Narayanan, 1985). Even if they invest in innovative projects, they have incentives to push up innovation quantity than quality due to grandstanding or entrenchment concerns (Gompers, 1996). Short sellers play a disciplining role in the ﬁnancial markets by detecting managers’ value- destroying behaviours and short selling ﬁrms’ stocks. Consequently, when short sellers ﬁnd managers’ sub-optimal investment in innovation, they may choose to short sell ﬁrms’ stocks. The treat of depressing stock price will mitigate managers’ myopia on innovation project since their compensation (e.g. stock option and bonuses) is contingent to stock prices. Therefore, managers may discipline themselves ex ante when making decisions on innova- tive projects and focus more on value-enhancing innovative projects. Second, short-selling may also deter corporate innovation by creating pressure on stock prices. Short sellers’ main objective is to identify underperformed ﬁrms that are overvalued and short sell these stocks. Thus, they have lower tolerance on short-term failures of a ﬁrm. Due to this, managers may have incentives to focus on short-term activities to avoid being shorted. Graham, Harvey, and Rajgopal (2005) point out that about 78% of executives tend to sacriﬁce long-term value to meet short-term targets. Yet, tolerance for failure is critical for the long-term and risky innovative projects (Manso, 2011). Consequently, managers who care more about short-term performance may cut oﬀ innovation input or long-term and risky innovative projects. CHINA JOURNAL OF ACCOUNTING STUDIES 295 Since the above hypotheses have contradicted predictions regarding the eﬀect of short- selling on corporate innovation, it is necessary to conduct an empirical study to explore the association between short-selling and corporate innovation. However, in practice, it is hard to detect the real eﬀect of short-selling due to the endogenous nature of short sales. First, Corporate innovation may adversely aﬀect short sellers’ short actions. Due to the long-term, risky, idiosyncratic and unpredictable nature of innovation projects, ﬁrms with innovative projects are more likely to be targeted by short sellers. Therefore, corporate innovation may result in a higher level of short-selling. Second, on the Chinese markets, the 2010 pilot scheme only allows ﬁrms on a designated list to be shorted. These ﬁrms are not randomly selected and tend to be big and blue-chip companies that have higher innovative level. As a result, empirical studies that use the Chinese setting to study the real eﬀect of short-selling may face self-selection problem. Third, some unobservable factors, such as culture and diﬀerent levels of government regulations may aﬀect short-selling and corporate innova- tion simultaneously, thus raise endogeneity problems caused by omitted variables. To tackle the above endogeneity problems, we ﬁrst exploit a quasi-experiment, the deregulation of short-selling constraints starting in 2010, to investigate the impact of short- selling on corporate innovation. We adopt a diﬀerence-in-diﬀerence (DiD) method by com- paring the change in ﬁrms’ innovation output of pilot ﬁrms with that of non-pilot ﬁrms surrounding the deregulation. We ﬁnd that pilot ﬁrms’ innovation output (in terms of both innovation quantity and quality) increased signiﬁcantly after the removing of short-selling constraints compared to the non-pilot ﬁrms. To further eliminate the endogeneity concern, we use the value of shares available for lending (SSP) as a proxy for short-selling and the ETF funds share-holdings as an instrument variable to further investigate the association between short-selling and corporate innovation. We still ﬁnd a signiﬁcantly positive relationship between short-selling and corporate innovation. Besides, our basic results also hold when we conduct a series of other robust checks. Next, we conduct a series of cross-sectional tests to investigate whether the eﬀect of short-selling on ﬁrm innovation vary in ﬁrms with diﬀerent levels of internal and external corporate governance. We ﬁnd that in ﬁrms with less developed ﬁnancial markets and lower internal corporate governance level, the positive eﬀect of short- selling on corporate innovation is more signiﬁcant. Lastly, we explore the potential mechan- isms through which short-selling aﬀect corporate innovation. We present that short-selling lowers the information asymmetry and improves the eﬃciency of managerial incentive contracts, which contribute to the improvement of corporate innovation in Chinese ﬁrms. This study contributes the literature in several aspects. First, it provides additional evidence on theimpactofshort-selling.Speciﬁcally,weemphasise thegovernanceroleplayed byshort sales and ﬁnd it can improve corporate innovation by lowering information asymmetry and improving the eﬃciency of managerial incentive contracts. Although prior literature has investigated the impact of short-selling on corporate innovation in developed markets (He &Tian, 2016;Massa et al., 2016), our paper is the ﬁrst to comprehensively examine the relationship between short-selling and corporate innovation in an underdeveloped market. Besides, we demonstrate that short-selling aﬀect corporate innovation through diﬀerent mechanisms as argued by He and Tian (2016), since the law systems and ﬁnancial regulations in China are not as developed as in other developed markets like the US. Second, we provide evidence that government regulations on ﬁnancial markets could aﬀect corporate behaviour, for example, ﬁrms’ investment in long-term and risky innovative projects. Thus, our study deepens the understanding of the real eﬀects of macro-policies on ﬁrm behaviours. Since 296 C. LI ET AL. China is under the economy transitional stage, our result indicates that government’s policies on ﬁnancial markets could beneﬁtenterprises’ upgrading by improving external monitoring. The rest of the paper proceeds as follows: Section 2 describes related literature and the institutional background of removing the constraints of short-selling in China. Section 3 presents the research design. Section 4 documents the main ﬁndings of this paper. Section 5 deals with the endogeneity concerns and presents robust checks. Finally, the conclusions and implications are presented in Section 6. 2. Related literature and institutional background 2.1. Related literature Our paper related to three strands of studies. First, it is related to the traditional literature on ﬁnancial markets’ feedback mechanism. Although individual investors have no information advantages compared to the ﬁrm managers, the whole investor group has signiﬁcant power to transfer information to the market through trading behaviours (Subrahmanyam & Titman, 1999). For example, the information on the innovation level of the current industry and the perception of the investors and consumers for the ﬁrm (Januszewski, Köke, & Winter, 2002; McGrath, 2001). Such information aggregation function of the ﬁnancial market reduces information asymmetry (Grossman, 1976;Hellwig, 1980). Edmans et al. (2015) ﬁnds that the current decision makers (e.g. the management, the funds providers, the board, the customers, the regulators, and the employees) are able to learn information from the stock prices and make use of it to make decisions. Short sellers are speculated investors that good at gathering private information and conducting shorting-selling activities to make proﬁts. On the one hand, it has been argued that short-selling promotes price discovering, thus enhances price eﬃciency in the capital market (Saﬃ &Sigurdsson, 2010). On the other hand, short-selling has also been criticised that short-selling deters market liquidity, ampliﬁes market risks and increase market volatility (Henry & McKenzie, 2006;Karpoﬀ &Lou, 2010; Keim & Madhavan, 1995). In general, although traditional literature has largely studied the eﬀects of short-selling on market eﬃciency, there are limited studies have investigated the impacts of short-selling on corporate decisions. Our paper also relates to the literature investigating the factors that aﬀect corporate innovation. Prior studies have investigated the various factors that aﬀect managers’ innova- tion decisions from the micro-ﬁrm level and the macro-economic level. For example, they ﬁnd that the institutional investors, corporate venture capital, ownership structure, unions, bank competitions, stock liquidity and ﬁnancial development could aﬀect corporate inno- vation decisions (Aggarwal, Erel, Ferreira, & Matos, 2011;Bradley,Kim,&Tian, 2013; Chemmanur, Loutskina, & Tian, 2014; Cornaggia, Mao, Tian, & Wolfe, 2015; Fang, Tian, & Tice, 2014; Hsu, Tian, & Xu, 2014; Megginson & Netter, 2001). However, there are few studies have examined the eﬀect of short-selling on corporate innovation. He and Tian (2016) regard corporate innovation output as a proxy for managerial myopia and ﬁnd asigniﬁcantly positive relationship between short-selling and corporate innovation; Massa et al. (2016) document that short sellers could aﬀect ﬁrms’ innovation input in a cross-country setting using data from 33 countries. So for, we have found no other study has investigated the eﬀect of short-selling on corporate innovation in the Chinese setting. Since China is becoming one of the biggest economies around the world, it is CHINA JOURNAL OF ACCOUNTING STUDIES 297 meaningful to examine whether short-selling as an eﬃcient market mechanism, also aﬀect corporates’ innovation decisions. Further, it is also interesting to investigate whether the mechanisms that short-selling aﬀect corporate innovation in western countries also work in a transitional economy like China. Finally, our paper is related to literature on the role of short-selling on corporate behaviour. Prior literature argue that short sellers are speculated investors that act as information intermediaries in the ﬁnancial markets, thus play an external corporate govern- ance mechanism for corporations. For example, Desai, Krishnamurthy, and Venkataraman (2006) have found that short-selling forces increase signiﬁcantly before listed ﬁrms’ ﬁnancial restatement. Christophe, Ferri, and Hsieh (2010) argue that the short positions increase signiﬁcantly before the analysts’ downgrade, and the change of the short position is signiﬁcantly related to the negative post-event return. Li and Zhang (2015) further verify the ‘price pressure hypothesis’ and ﬁnd that the price pressure created by short-selling would increase managers’ willingness of voluntary of disclosure. Massa et al. (2015)and Fang et al. (2016) found that short-selling has a negative impact on earnings management. Chang et al. (2018) found that short-selling could promote managers’ decisions on M&A. Chu (2015) provides evidence that short-selling could also aﬀect ﬁrms’ product markets, speciﬁcally, short-selling increases ﬁrms’ market shares. In China, short-selling has been prohibited until 2010, a pilot programme has been conducted by the CSRC. The programme aims to remove the constraints on marginal trading and short-selling in the Chinese ﬁnancial markets. Since then, pilot ﬁrms that are included in the designated short-sale list have been allowed in China. Few studies have investigated the real eﬀects of short-selling on the Chinese markets. Recently, Chinese studies that examining the real eﬀects of short-selling have found that removing the short-selling constraint has a positive eﬀect for lowering market volatility, improving market eﬃciency as well as price eﬃciency (Gu, Hao, & Zhang, 2011;Li, Chen,&Lin, 2015;Xiao& Kong, 2014). Besides, some studies have also investigated the impacts of short-selling on ﬁrm behaviours. For example, Chen and Liu (2014) document that short-selling has anegative eﬀect on earnings management in Chinese listed ﬁrms; Zhang et al. (2016) ﬁnd that short-selling reduces listed ﬁrms’ probabilities of ﬁnancial restatements. However, few studies have examined the real eﬀect of short-selling on corporate innovation. In this paper, we try to ﬁll this gap by investigating how short-selling aﬀect Chinese listed ﬁrms’ innova- tion output, regarding both innovation quantity and quality. 2.2. Institutional background In China, due to the immature ﬁnancial legal system and ﬁnancial markets, regulators have prohibited short-selling for a long period to prevent uncontrolled market volatility and crash. Before 2010, there were no markets in China for stock options and futures that investors can use to create synthetic shorts (Chen, Lin, Lu, & Ma, 2018). However, in the last decade, with the signiﬁcant expansion of the Chinese economy and the improvement of the ﬁnancial legal system, Chinese regulators are searching to remove the short-selling constraints gradually. Apart from the prohibition stages of short-selling, the Chinese stock markets have experienced three stages to remove the short-selling constraint. First, the preliminary preparation of related law and regulations. In October 2005, the new Chinese Security Law has taken the ﬁrst step to bring the marginal trading and short- 298 C. LI ET AL. selling into the legal fold. The law has regulated that security companies planning to carry out marginal trading and short-selling business should be endorsed by the securities regulatory body under the state council. On July 2006, the CSRC has released the Measures for Marginal Trading and Short-selling for Security Companies, which regulated the implementing details of the marginal trading and short-selling business for the security companies. The above law and regulation have laid a good foundation for the pilot programme of the marginal trading and short-selling. Second, the In-house tests stage. On October and November 2008, the CSRC has organised the In-house tests of the marginal trading and short-selling system in 11 security companies to make sure the system has worked smoothly. In 2010, the state council has approved that the pilot programme of marginal trading and short-selling can be implemented. Third, the implementation of the pilot programme. On March 2010, the pilot programme has been implemented under the lead of the CSRC. In the ﬁrst designated list of the pilot programme, 90 listed ﬁrms have been chosen as pilot ﬁrms that are allowed to be short- sold. Since then, the list is expanding with time, new ﬁrms are continuing added into the list while old ﬁrms that no longer accord with the selection criteria of pilot ﬁrms are removed from the list. By March 2015, there are in total 900 listed ﬁrm on the designated list. In the following Table 1, we present the details of the changes of the pilot ﬁrms each year. According to Chen et al. (2018), to avoid short-selling bringing too much market volatility and even market crash in the ﬁnancial market, the CSRC has regulated ﬁrms that are eligible for short-selling should meet the following criteria: ﬁrst, a minimum of 200 million tradable shares; second, public ﬂoat should be more than 800 million; third, more than 4000 share- holders; fourth, daily turnover rate should be higher than 15% of index turnover rate; ﬁfth, daily trading volume that is more than 50 million RMB. The pilot programme of removing short-selling constraints in China has provide us a quasi-experiment to investigate the real eﬀect of short-selling by using a DiD identiﬁcation strategy. Since the pilot ﬁrms on list are changing over time, which creating both time-series and cross-sectional variations in short-selling restrictions for ﬁrms. Therefore, this staggered setting is helpful to eliminate the potential concern of omitted variables. However, with the Table 1. The changes of the pilot ﬁrms on the designated list. Event dates Nr. of ﬁrms added Event date Nr. of ﬁrms deleted 2010/3/31 90 2010/7/1 5 2010/7/1 5 2010/7/29 1 2010/7/29 1 2011/12/5 1 2011/12/5 189 2013/1/31 54 2013/1/31 276 2013/3/6 1 2013/4/10 1 2013/3/7 1 2013/9/16 206 2013/3/26 1 2014/9/22 218 2013/3/29 2 2013/5/2 1 2013/5/3 1 2014/3/28 1 2014/4/1 1 2014/4/29 1 2014/5/5 2 2014/9/22 13 Added ﬁrms in total 986 Deleted ﬁrms in total 86 Total ﬁrms in total (2015) 900 The data are extracted from the Shanghai and Shenzhen stock exchanges. CHINA JOURNAL OF ACCOUNTING STUDIES 299 above selection criteria, the chosen ﬁrms tend to be big and blue-chip ﬁrms. This may bring selection bias when we use the pilot ﬁrms as treatment ﬁrms and other listed ﬁrms as control ﬁrms in a DiD identiﬁcation strategy to test the relationship between short-selling and corporate innovation. To eliminate this endogeneity concern, we also use IV strategy and PSM method to make robust checks in latter parts of this paper. 3. Research design 3.1. Research data and sample The sample construction starts with a comprehensive list of Chinese common stocks on Shanghai and Shenzhen stock exchanges between 2005 and 2015. The accounting data, patent data and the data of ﬁrm board and management are retrieved from the China Stock Market Trading Research (CSMAR) database provided by GuoTaiAn (GTA) Company, a major provider of Chinese data. Besides, the designated stock list for short-selling is collected from the Shanghai and Shenzhen stock exchange websites. We clean the data as follows: ﬁrst, we delete observations with missing values of the major variables in this paper; second, considering the pilot programme was began in 2010, we exclude ﬁrms that were listed after 2009; third, we also delete ﬁrms that are in the insolvent status; fourth, we delete ﬁrms that are in the ST or *ST status; ﬁfths, we winsorsize all continues variables at the 1% and 99% levels to mitigate the eﬀects of outliers. Finally, our sample includes 1671 A-share listed ﬁrms on the Shanghai and Shenzhen stock exchanges and 11,469 ﬁrm-year observations between 2005 and 2015. 3.2. Variable deﬁnitions 3.2.1. Measures of corporate innovation Following He and Tian (2016), we use a ﬁrm’s patenting activity to measure corporate innovation. In China, patents can be classiﬁed into three categories, the invention patent, utility model patent and design patent. According to the originality of the patents, Tan, Tian, Zhang, and Zhao (2015) argue that the invention patents are the most original ones among the three types, then follows by the utility model patents, and the design patents rank the third place. In this paper, we use three proxies to measure a ﬁrm’s innovation level. First, we use the total number of the three types of patents granted (Patents)to measure the total quantity of a ﬁrm’s innovation output. Second, we use the number of the invention patents granted (Invention) to measure a ﬁrm’s invention quality since invention patents represents a ﬁrm’s most original inventions. Third, since the design patents involve limited technological advancements, we construct another measure, the total number of invention patents and utility patents granted (Inv&Des) to measure a ﬁrm’s innovation output. Due to the time-lag of the patent application, we use the 1 year lagged patent number to construct the above three measures. Besides, to address The existing literature uses the number of citation a patent receives as a measure for patent quality because it assumes that more inﬂuential and higher quality patents have a larger number of subsequent citations. However, it is practically diﬃcult to extract the patent citation data since the Chinese SIPO database has not provide this data. Therefore, in this paper, we follow Tan et al. (2015) to use the number of invention patents as a proxy for innovation quality. 300 C. LI ET AL. the concern that patent number are right skewed, we use the natural logarithm of one plus the above three proxies to make analysis. 3.2.2. Measure of short-selling We construct a dummy variable Treat to donate pilot ﬁrms in the designated short-sale list. Speciﬁcally, Treat equals to 1 if a ﬁrm is eventually included in the short-sale list by the end of our sample period, and 0 otherwise. In this paper, the pilot ﬁrms constitute the treatment sample and non-pilot ﬁrm serve as control sample in our DiD models. Moreover, we build adummyvariable, Post to indicate the sub-periods before and after pilot ﬁrms are included in theshort-salelist. Post equals to 1 if a ﬁrm’s stock has been included in the list in a given year, and 0 otherwise. In the robust checks of this paper, we follow Chang et al. (2018), to use the value of shares available for lending (SSP) as a proxy for short-selling. We deﬁne it as the daily lending volume minus the daily reimbursed volume then divided by the ﬁrm’s market capitalisation 1 day before. We use the average SSP during a year (365 days) as the main explanatory variable in our regression models. 3.2.3. Other control variables Following prior literature, we control a vector of variables at the ﬁrm level and industry level that are supposed to aﬀect corporate innovation in our regression models. These variables include: ﬁrm size (Size), leverage ratio (LEV), ﬁrm growth rate (Growth), ﬁrm age (Age), tangible asset ratio (Tangibility), capital expenditure ratio (CapEx), board size (BoardSize), board independence (Indep), the duality of chairman and general manager (Duality), institutional shareholding (institution). Besides, considering that a ﬁrm’s patent- ing may grow with time, we also control the patent growth (PatGrow) in our models. We provide the deﬁnitions of the main variables in the following Table 2. 3.3. Summary statistics In Table 3, we present the summary statistics of the main variables in this study. The means of Patent, Invention and Inv&Des are 2.714, 1.871 and 2.535, respectively. These descriptive statistics are similar to He and Tian (2016). However, in general, the Chinese ﬁrms’ innova- tion outputs are relatively smaller than the US ﬁrms. The mean (median) of Treat is 0.494 (0.500), which indicating about 50% of the observations are pilot ﬁrms. Regarding other control variables, our descriptive statistics are also consistent with prior literature (He & Tian, 2016). Thus, we do not report the details of the summary statistics of these control variables. 3.4. Empirical strategies Our empirical strategy mainly relies on the quasi-exogenous shocks created by the pilot programme of short-selling launched in 2010. Since the pilot ﬁrms on the short-sale list are changing overtime, we get a staggered setting to conduct the DiD identiﬁcation strategy. Speciﬁcally, we use the following speciﬁcation to investigate the relation between short-selling and corporate innovation: Innovation ¼ α þ βTreat Post þ γZ þ δ þ φ þ ε (1) i;tþ1 i;t i;t i i;t i CHINA JOURNAL OF ACCOUNTING STUDIES 301 Table 2. Variable deﬁnitions. Variable name Variable deﬁnitions Dependent Patent The total innovation output, measured as the natural logarithm of 1 plus the number of variable patents granted, including invention patent, utility model patents and design patents. Invention The innovation quality, measured as the natural logarithm of 1 plus the number of intention patents granted. Inv&Des The innovation output, measured as the natural logarithm of the 1 plus the number of invention patents and the utility model patents. Independent Treat A dummy variable, equals to 1 if the ﬁrm has been included in the short-sale list, and 0 variables otherwise Post A dummy variable, equals to 1 for the years after the treated ﬁrm is included in the short-sale list, and 0 otherwise. SSP The value of shares available for lending, measured as the diﬀerences between the daily lending volume and daily reimbursed volume then divided by the market capitalisation of the prior day. We use the average SSP during a year (365 days) as the main explanatory variable in our regression models. Control Size Firm size, measured as the nature logarithm of total annual sales. variables LEV Leverage ratio, measured as the ratio of total liabilities to total assets. Growth Firm growth rate, measured as the total assets divided by the one-year lagged total assets. Tangibility Tangible asset ratio, measured as the total ﬁxed assets divided by total assets. CapEx Capital expenditure ratio, measured as the capital expenditure divided by total assets. PatGrow Average patent growth rate for the past three consecutive years. R&D R&D expenditure ratio, measured as the annual R&D expenditure divided by total assets. KZ The KZ index following Kaplan and Zingales (1997) ROA Return on assets, measured as net proﬁt divided by total assets. Institution Institutional shareholding, measured as the percentage of shares held by institutional investors. Duality The duality of chairman and general manager, a dummy variable, equals to 1 if the chairman and general manager is the same person, and 0 otherwise. BoardSize Board size, measured as the nature logarithm of the number of board of directors. Indep Board independence, measured as the number of independent directors divided by the number of the board of directors. Age Firm age, measured as the nature logarithm of the number of years since the ﬁrm is established. Table 3. Summary statistics. variable N mean sd min p25 p50 p75 max Patent 11,469 2.714 3.812 0 0 0 2.197 5.771 Invention 11,469 1.871 2.935 0 0 0 1.386 4.860 Inv&Des 11,469 2.535 3.625 0 0 0 2.079 5.561 Treat 11,469 0.494 0.500 0 0 0 1 1 Post 11,469 0.147 0.354 0 0 0 0 1 Size 11,469 21.41 1.469 16.86 20.50 21.32 22.25 25.28 LEV 11,469 0.496 0.197 0.068 0.351 0.506 0.646 0.925 ROA 11,469 0.040 0.055 −0.205 0.014 0.036 0.066 0.207 Growth 11,469 0.164 0.273 −0.347 0.016 0.107 0.236 1.767 Tangibility 11,469 0.265 0.184 0.002 0.120 0.231 0.383 0.764 CapEx 11,469 0.059 0.055 0 0.017 0.043 0.083 0.257 R&D 11,469 0.003 0.012 0 0 0 0 0.074 KZ 11,469 1.791 0.667 0.220 1.360 1.815 2.214 4.273 PatGrow 11,469 0.061 0.469 −1.522 0 0 0.134 1.522 Institution 11,469 0.079 0.107 0.002 0.015 0.044 0.099 0.606 Duality 11,469 0.155 0.362 0 0 0 0 1 BoardSize 11,469 9.220 1.878 5 9 9 10 15 Indep 11,469 0.365 0.052 0.273 0.333 0.333 0.375 0.571 Age 11,469 13.75 4.965 3 10 14 17 27 302 C. LI ET AL. where Innovation is the dependent variable of interest, and Treat is a dummy variable, which indicates whether a ﬁrm belongs to the pilot ﬁrms. Post is a dummy variable, indicating whether a ﬁrm is allowed to be short-sold in a given year. Our main variable of interest in Equation (1) is the interaction term Treat*Post. Z is a vector of control variables, which we have introduced in Table 2. Besides, we also include time and year ﬁxed eﬀects in our model to control the macro factors that change with the time and the time- invariant unobserved heterogeneity at the ﬁrm level. Lastly, the standard errors are clustered at the ﬁrm level. 4. Empirical results 4.1. Baseline results Our ﬁrst set of tests examines how short-selling aﬀect corporate innovation under the staggered deregulation of short-selling constraint began in 2010. Table 4 reports the base- line results. In column (1)–(3), we include the year and industry ﬁxed eﬀects in the baseline regressions. We ﬁnd that the coeﬃcients of Treat*Post are 0.193, 0.219 and 0.211 and all signiﬁcant at 1% level. The results indicate that short-selling leads to a signiﬁcant increase in corporate innovation regarding to both innovation quantity and innovation quality. In column (4)–(6), we include year and ﬁrm ﬁxed eﬀects and found short-selling still have signiﬁcantly positive eﬀects on the innovation quantity and quality. In general, our empirical tests document a positive association between short-selling and corporate innovation. 4.2. The eﬀects of internal governance environment In this section, we investigate how ﬁrms’ internal governance aﬀect the association between short-selling and corporate innovation. Since the Chinese authorities have not published a comprehensive corporate governance index, we use two proxies to measure listed ﬁrms’ internal corporate governance. First, we use ownership structure to present ﬁrms’ internal governance environment. In China, the state-owned ﬁrms occupy most resources compared with private ﬁrms; how- ever, the innovation production eﬃciency and innovation eﬃciency of the state-owned ﬁrms are much lower (Dong, Zhao, & Yuan, 2014; Wu, 2012). Prior literature argues that the severe agency problems have led to relatively lower corporate governance in state- owned ﬁrms, which deters ﬁrms’ innovation eﬃciency (Wu, 2012). Yet, short-selling is a compensation to ﬁrms’ governance mechanisms (Chang et al., 2018), therefore, we predict that in state-owned ﬁrms, short-selling has a signiﬁcantly positive eﬀect on corporate innovation. To test the above assumption, we divide our sample into state-owned ﬁrms and private ﬁrms and re-estimate our basic speciﬁcations in the two sub-samples. If the state holds more than 50% shareholdings of a ﬁrms, we deﬁne it as a state-owned ﬁrm, otherwise, we deﬁne it as a private ﬁrm. We present the regression results in the following Table 5.We ﬁnd that in state-owned ﬁrms, short-selling has signiﬁcantly positive eﬀects on both innovation quantity and quality. However, we did not ﬁnd signiﬁcant eﬀects in the private ﬁrm sample. The above results indicating that short-selling is an eﬀective replenishment mechanism of ﬁrms’ internal corporate governance. CHINA JOURNAL OF ACCOUNTING STUDIES 303 Table 4. Short-selling and enterprise innovation. (1) (2) (3) (4) (5) (6) F.Patent F.Invention F.Inv&Des F.Patent F.Invention F.Inv&Des Treat*Post 0.193*** 0.219*** 0.211*** 0.0942** 0.0977*** 0.119*** (4.14) (5.47) (4.66) (2.25) (2.70) (2.95) Sale 0.214*** 0.172*** 0.192*** 0.0362 0.0279 0.0306 (16.88) (15.77) (15.68) (1.64) (1.53) (1.42) LEV −0.373*** −0.311*** −0.198* 0.341** 0.236** 0.345** (−3.11) (−3.16) (−1.74) (2.40) (1.98) (2.53) ROA 0.917*** 0.753*** 0.878*** 0.235 0.337* 0.310 (3.70) (3.70) (3.75) (0.96) (1.80) (1.40) Growth 0.0330 0.0301 0.0366 −0.0154 0.00272 −0.0168 (0.72) (0.84) (0.85) (−0.46) (0.10) (−0.52) Tangibility 0.00715 −0.0868 0.0476 0.258** 0.234** 0.263** (0.09) (−1.36) (0.63) (2.07) (2.26) (2.17) CapEx 0.865*** 0.881*** 1.051*** −0.0757 0.0228 0.000111 (3.75) (4.59) (4.77) (−0.38) (0.13) (0.00) RD 7.512*** 8.222*** 7.769*** 1.280 2.563 1.518 (5.47) (6.78) (5.76) (0.80) (1.51) (0.92) KZ −0.0630** −0.0114 −0.0910*** −0.0987*** −0.0630** −0.101*** (−2.08) (−0.47) (−3.20) (−3.07) (−2.56) (−3.34) PatGrow 0.659*** 0.492*** 0.622*** 0.224*** 0.170*** 0.207*** (23.18) (20.79) (23.06) (8.59) (7.99) (8.28) Institution 0.373*** 0.271*** 0.409*** 0.145 0.164 0.172 (3.04) (2.75) (3.52) (0.86) (1.05) (1.06) Duality 0.201*** 0.187*** 0.187*** −0.0195 −0.00906 −0.0254 (5.90) (6.50) (5.81) (−0.52) (−0.25) (−0.69) BoardSize 0.119* 0.257*** 0.174*** 0.214* 0.235** 0.219** (1.75) (4.58) (2.73) (1.92) (2.52) (2.01) Indep −0.0360 0.284 0.0914 0.384 0.310 0.387 (−0.14) (1.29) (0.37) (1.28) (1.25) (1.32) Age −0.484*** −0.306*** −0.471*** 0.496*** 0.516*** 0.528*** (−13.61) (−10.47) (−13.78) (3.16) (3.81) (3.45) Constant −3.400*** −3.462*** −3.215*** −1.555** −1.867*** −1.629** (−10.24) (−12.05) (−10.08) (−2.17) (−3.12) (−2.34) Firm No No No Yes Yes Yes Year Yes Yes Yes Yes Yes Yes Industry Yes Yes Yes No No No N 11,469 11,469 11,469 11,469 11,469 11,469 R 0.400 0.346 0.409 0.099 0.105 0.114 * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; t value is reported in the parentheses. Table 5. Impact of ownership structure. State-owned enterprises Private enterprise F.Patent F.Invention F.Inv&Des F.Patent F.Invention F.Inv&Des Treat*Post 0.127** 0.132*** 0.150*** −0.00301 −0.00979 0.0122 (2.29) (2.69) (2.76) (−0.05) (−0.20) (0.22) Constant −3.154*** −2.972*** −3.137*** −0.995 −1.796** −1.068 (−2.87) (−3.24) (−2.89) (−1.08) (−2.34) (−1.21) Control Yes Yes Yes Yes Yes Yes Firm Yes Yes Yes Yes Yes Yes Year Yes Yes Yes Yes Yes Yes N 7039 7039 7039 4430 4430 4430 R 0.129 0.135 0.141 0.071 0.079 0.089 * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. Second, we use the ownership concentration as a proxy for ﬁrms’ internal governance. Prior literature ﬁnds that the higher the ownership concentration, the lower the ﬁrms’ corporate governance. Therefore, we predict that in ﬁrms with higher ownership 304 C. LI ET AL. concentration, the short-seller would play as an eﬀective outside governance mechanism and improve corporate innovation. We use the median of ﬁrms’ largest shareholder’s shareholdings to divide our sample into to sub-samples. If a ﬁrm’s largest shareholder holds more 50% shares, it belongs to the higher ownership concentration group, otherwise belongs to the lower ownership concentration group. We test our assumption by re-estimating our basic speciﬁcation in the two sub-samples. Table 6 reports the regression results. We ﬁnd that when ﬁrms have higher ownership concentration, short-selling improves corporate innovation signiﬁ- cantly regarding to both innovation quantity and quality. These results are consistent with our predictions. 4.3. The eﬀect of external governance environment The external ﬁnancial environment may also aﬀect the real eﬀect of short-selling. If a region has higher ﬁnancial marketisation, the listed ﬁrms will face higher supervision from the regulators. Consequently, the eﬀective external governance environment will substitute the external governance role-played by short-selling (Massa et al., 2015). Thus, we predict that in regions that have lower ﬁnancial marketisation, short-selling will improve corporate innovation signiﬁcantly. To test the above prediction, we use the ‘Financial Marketisation Index of 2009’ developed by Fan, Wang, and Zhu (2010) as an indicator variable for a region’s levels of ﬁnancial marketisation. We use the sample median of the index to divide our sample into two sub-samples with higher and lower ﬁnancial marketisation, respectively. Then, we estimate our basic speciﬁcation in the two samples and report the results in Table 7. It is documented that in ﬁrms that located in regions with lower ﬁnancial marketisation, short-selling improve corporate innovation signiﬁcant, which is consistent with our prediction. 4.4. The underlying mechanism In this section, we explore the underlying mechanisms that short-selling aﬀect corporate innovation under the Chinese setting. We propose two potential mechanisms that short- selling could improve corporate innovation. Table 6. Impact of ownership concentration. High ownership concentration Low ownership concentration F.Patent F.Invention F.Inv&Des F.Patent F.Invention F.Inv&Des Treat*Post 0.141*** 0.110*** 0.165*** 0.00571 0.0592* 0.0328 (3.29) (3.07) (3.98) (0.15) (1.75) (0.87) Constant −1.808** −1.833*** −1.620** −3.527*** −3.311*** −3.749*** (−2.28) (−2.76) (−2.12) (−4.96) (−5.39) (−5.49) Control Yes Yes Yes Yes Yes Yes Year Yes Yes Yes Yes Yes Yes Firm Yes Yes Yes Yes Yes Yes N 5947 5947 5947 5522 5522 5522 R 0.146 0.134 0.124 0.158 0.157 0.145 * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. CHINA JOURNAL OF ACCOUNTING STUDIES 305 Table 7. Impact of external market environment. Low ﬁnancial marketisation High ﬁnancial marketisation F.Patent F.Invention F.Inv&Des F.Patent F.Invention F.Inv&Des Treat*Post 0.120* 0.155*** 0.176*** 0.0744 0.0419 0.0682 (1.82) (2.71) (2.77) (1.46) (0.93) (1.38) Constant −1.667* −1.424* −1.308 −1.624 −2.565*** −2.249** (−1.68) (−1.69) (−1.33) (−1.54) (−2.90) (−2.22) Control Yes Yes Yes Yes Yes Yes Year Yes Yes Yes Yes Yes Yes Firm Yes Yes Yes Yes Yes Yes N 5808 5808 5808 5661 5661 5661 R 0.128 0.127 0.141 0.076 0.092 0.093 * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. First, short-selling may improve corporate innovation by eliminating the information asymmetry in the targeted ﬁrms. Short sellers are sophisticated investors that have strong incentives to dig ﬁrms’ private information and make proﬁt by shorting the targeted stocks. Therefore, when ﬁrm managers invest in sub-optimal innovation projects, short sellers tend to discover these activities and compound the negative information into stock prices by shorting the stocks. Consequently, managers may discipline themselves ex ante when making decisions on innovative projects and focus more on value-enhancing innovative projects. With the above argument, we predict that when a ﬁrm face more severe information asymmetry, short sellers’ disciplinary role is stronger, and as a result, ﬁrm managers may tend to invest in innovative projects with brighter prospect and better quality. We use the Sobel mediation tests to examine the above underlying mechanism. We use the standard error of the analyst forecast (F_SD) to measure a ﬁrm’s information asymmetry and conduct the following steps to make the mediation tests. First, we regress corporate innovation on short-selling (Treat*Post); second, we regress the mediating variable, ﬁrm information asymmetry (F_SD) on short-selling (Treat*Post); third, if we ﬁnd short-selling have signiﬁcant impacts on corporate innovation and ﬁrm information asymmetry in the above regressions, we put the mediating variable into the regression of corporate innovation on short-selling. Then, we observe the changes of the coeﬃcients of short-selling compared with that in the ﬁrst regression in step one. If we ﬁnd that the coeﬃcient of short-selling become smaller or the signiﬁcance of the coeﬃcient decreases, we can prove that a part of the eﬀect of short-selling on corporate innovation comes from the mediating variable. In Table 8, we report the results of the Sobel mediation tests. In column (1), we ﬁnd that short-selling has a signiﬁcant negative eﬀect on ﬁrms’ information asymmetry. In column (2), short-selling improves corporate innovation output signiﬁcantly at the 5% level. In column (3), we add the variable Treat*Post and F_SD together into the regres- sion and ﬁnd that the magnitude and the signiﬁcance of the coeﬃcient of Treat*Post decreases. In column (4)–(7), we repeat the above steps using another to dependent variables, the Invention and the Inv&Des,wealso ﬁnd similar results. Thus, our results prove that short-selling improves corporate innovation by partly reducing ﬁrm informa- tion asymmetry. 306 C. LI ET AL. Table 8. Short selling, information asymmetry and corporate innovation. (1) (2) (3) (4) (5) (6) (7) F_SD FPatent FPatent FInvention FInvention FInv&Des FInv&Des Treat*Post −0.00267*** 0.0942** 0.0594 0.0977*** 0.0693* 0.119*** 0.0869** (−5.31) (2.25) (1.28) (2.70) (1.73) (2.95) (1.96) F_SD −2.182** −1.451* −1.997** (−2.22) (−1.81) (−2.11) Constant 0.00936 −1.555** −2.033** −1.867*** −2.180** −1.629** −1.894* (0.78) (−2.17) (−2.00) (−3.12) (−2.46) (−2.34) (−1.90) Year Yes Yes Yes Yes Yes Yes Yes N 8630 11,469 8630 11,469 8630 11,469 8630 R 0.141 0.099 0.099 0.105 0.103 0.114 0.114 ind2ratio 6.6% 4.1% 4.7% * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. Second, short-selling may also improve corporate innovation by improving the eﬃ- ciency of managerial incentive contracts. Eﬃciency managerial incentive contracts could discipline managers from enjoying private beneﬁts such as ‘quite life’ (Bertrand and Mullainathan, 2003; Narayanan, 1985). If managers’ incentive contracts include proper percentages of stock options, the value of the stock options will ﬂuctuate with the stock prices. Short-selling put treat on depressing a ﬁrm’s stock prices, which will mitigate managers’ myopia on innovation project since their compensation (e.g. stock option and bonuses) are contingent to stock prices. Therefore, managers may discipline themselves ex ante when making decisions on innovative projects and focus more on value-enhan- cing innovative projects. Therefore, we propose that short-selling may aﬀect corporate innovation by improving the eﬃciency of managerial incentive contracts. To test the above prediction, we use the managerial stock options (Incentive) as a proxy for the eﬃciency of a ﬁrm’s managerial incentive contracts and conduct the above Sobel mediation tests. Here, we use the variable Incentive as the mediating variable. Table 9 reports the results of the Sobel mediation tests. In column (1), we ﬁnd that short-selling improves the eﬃciency of managerial incentive contracts signiﬁcantly. In column (2), the results indicate that short-selling improves ﬁrms’ innovation output signiﬁcantly. In column (3), we include both Treat*Post and Incentive into the regression. We ﬁnd that both the magnitude and the signiﬁcance of the coeﬃcient of the variable Treat*Post reduced compared to column (2). In column (4)–(7), we repeat the above regression steps using another two dependent variables, the Invention and the Inv&Des,we ﬁnd similar results. In general, our Sobel mediation tests show that short-selling also impact corporate innovation through improving the eﬃciency of managerial incentive contracts. 5. The endogeneity concerns and robust checks 5.1. The parallel trend assumption An important identiﬁcation assumption for DiD analysis is the parallel trend assumption, which argues that there should be no diﬀerent development of the dependent variable in treated and control ﬁrms before the exogenous staggered shocks. To investigate whether the assumption is fulﬁled, we investigate the time dynamics around the shocks in the following tests. Speciﬁcally, we ﬁrst construct time variables Before3-After4 to indicate the CHINA JOURNAL OF ACCOUNTING STUDIES 307 Table 9. Short-selling, managerial incentive and corporate innovation. (1) (2) (3) (4) (5) (6) (7) Incentive FPatent FPatent FInvention FInvention FInv&Des FInv&Des Treat*Post 0.537*** 0.0942** 0.0899** 0.0977*** 0.0944*** 0.119*** 0.115*** (3.24) (2.25) (2.15) (2.70) (2.59) (2.95) (2.84) Incentive 0.00788* 0.00632* 0.00777** (1.91) (1.87) (2.00) Constant −17.45*** −1.555** −1.660** −1.867*** −1.951*** −1.629** −1.723** (−6.83) (−2.17) (−2.27) (−3.12) (−3.22) (−2.34) (−2.43) Year Yes Yes Yes Yes Yes Yes Yes N 11,469 11,469 11,469 11,469 11,469 11,469 11,469 R 0.062 0.099 0.100 0.105 0.105 0.114 0.115 ind2ratio 27.6% 14.6% 23% * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. time trend. Before3 (After4) equals to 1 if a ﬁrm is 3 (2/1/0/4) year(s) before (after) including the short-sale list, and 0 otherwise. We then substitute these dummy variables into our basic speciﬁcation and make regressions. Table 10 reports the estimated results. In column (1)–(3) we only add Before3-Before1 into our models. We ﬁnd that before the exogenous shocks, the treatment ﬁrms (pilot ﬁrms) and control ﬁrms have no signiﬁcant diﬀerences regarding to corporate innovation. In column (4)–(6), we only add Current- After4 into the models. The results show that after the exogenous shocks, the pilot ﬁrms experience signiﬁcantly increase in corporate innovation relative to the control ﬁrms. In column (7)–(9), we add Before3-After4 together into our speciﬁcations. The results still show that the diﬀerences between the pilot ﬁrms and control ﬁrms regard to corporate innovation are only signiﬁcant after the shocks. In general, the above results conﬁrm that our results satisfy the parallel trend assumption of the DiD identiﬁcation strategy. 5.2. The propensity score matching (PSM) method In China, the CSRC has regulated certain criteria to select the pilot ﬁrms on the short-sale list. Therefore, the ﬁrms on and oﬀ the short-sale list maybe fundamentally diﬀerent in ﬁrm characteristics. We follow Rosenbaum and Rubin (1983), use propensity score matching to further construct a treatment group and a control group that have no signiﬁcant diﬀerences in ﬁrm characteristics and re-estimate our main speciﬁcation. We conduct the following matching procedure: ﬁrst, we use pilot ﬁrms’ characteristic 1 year before the listing inclusion event as the treatment dataset and the ﬁrms that were not included in the short-sell list as the matching sample; second, we estimate the propensity score using a logit model in which dependent variable is Treat, and performing a nearest neighbour matching strategy, using the closest propensity score and with a propensity score match within 0.01 to match each treatment ﬁrm with one control ﬁrm. We retain all pairs in the case of multiple matching. The logit model includes all control variables in Equation (1). Besides, we also include the year and the industry ﬁxed eﬀects in the model. After the matching, we get in total 654 pairs of treatment and control groups, 1308 ﬁrm-year observations. In Table 11,we present the summary statistics and the mean diﬀerences of the after-matching treatment and control ﬁrms. It is shown that after PSM matching, the two groups of ﬁrms have no signiﬁcant diﬀerences regarding ﬁrm characteristics. The PSM method helps to dampen the 308 C. LI ET AL. Table 10. The parallel assumption. (1) (2) (3) (4) (5) (6) (7) (8) (9) Patent Invention Inv&Des Patent Invention Inv&Des Patent Invention Inv&Des Before3 0.0104 0.0125 −0.00057 0.0453 0.0503 0.0314 (0.19) (0.29) (−0.01) (0.84) (1.16) (0.62) Before2 0.0256 0.0408 0.0236 0.0698 0.0894** 0.0640 (0.50) (0.99) (0.49) (1.32) (2.11) (1.29) Before1 −0.0136 0.00233 −0.0292 0.0465 0.0681 0.0265 (−0.27) (0.06) (−0.61) (0.87) (1.59) (0.53) Current 0.135** 0.136*** 0.128** 0.157*** 0.165*** 0.144*** (2.54) (3.18) (2.55) (2.84) (3.72) (2.77) After1 0.0542 0.0772 0.0469 0.0764 0.107** 0.0634 (0.91) (1.62) (0.84) (1.25) (2.17) (1.10) After2 0.205** 0.216*** 0.211** 0.233*** 0.254*** 0.231*** (2.33) (3.06) (2.55) (2.58) (3.51) (2.72) After3 0.315*** 0.288*** 0.295*** 0.338*** 0.319*** 0.312*** (3.52) (4.01) (3.50) (3.73) (4.37) (3.64) After4 0.266* 0.287** 0.227 0.292* 0.322*** 0.246* (1.76) (2.38) (1.60) (1.93) (2.65) (1.73) Constant −3.309*** −3.928*** −3.333*** −2.889*** −3.520*** −2.910*** −2.779*** −3.372*** −2.824*** (−11.72) (−17.31) (−12.52) (−9.74) (−14.77) (−10.40) (−9.09) (−13.73) (−9.79) Control Yes Yes Yes Yes Yes Yes Yes Yes Yes Industry Yes Yes Yes Yes Yes Yes Yes Yes Yes Year Yes Yes Yes Yes Yes Yes Yes Yes Yes N 11,469 11,469 11,469 11,469 11,469 11,469 11,469 11,469 11,469 R 0.407 0.356 0.420 0.408 0.357 0.421 0.408 0.357 0.421 * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. CHINA JOURNAL OF ACCOUNTING STUDIES 309 Table 11. Summary statistics of the after-matching sample. Control Group(pilot = 0) Experimental group(pilot = 1) Variable N mean Sd p50 N mean sd p50 Mean Diﬀerence Patent 654 1.261 1.516 0 654 1.413 1.761 0 −0.151 Invention 654 0.884 1.191 0 654 1.017 1.428 0 −0.132 Inv&Des 654 1.199 1.461 0 654 1.326 1.699 0 −0.127 Size 654 21.71 1.19 21.64 654 21.99 1.311 21.92 −0.28 LEV 654 0.505 0.209 0.514 654 0.499 0.197 0.503 0.006 ROA 654 0.047 0.044 0.039 654 0.055 0.051 0.044 −0.008 Growth 654 0.172 0.282 0.115 654 0.196 0.24 0.156 −0.024 Tangibility 654 0.232 0.177 0.194 654 0.228 0.175 0.191 0.004 CapEx 654 0.0541 0.050 0.040 654 0.059 0.052 0.044 −0.004 RD 654 0.005 0.016 0 654 0.006 0.016 0 −0.001 KZ 654 1.798 0.758 1.806 654 1.779 0.62 1.827 0.018 PatGrow 654 0.105 0.471 0 654 0.099 0.448 0 0.006 Institution 654 0.073 0.114 0.033 654 0.087 0.103 0.057 −0.013 Duality 654 0.154 0.362 0 654 0.15 0.357 0 0.005 BoardSize 654 2.19 0.204 2.197 654 2.211 0.204 2.197 −0.022 Indep 654 0.371 0.055 0.333 654 0.371 0.054 0.333 −0.001 Age 654 2.674 0.359 2.708 654 2.653 0.356 2.639 0.021 potentially confounding ﬁrm characteristics diﬀerences known to aﬀect corporate innova- tion, helping alleviate concerns that the results are driven by general time trends. Next, we estimate our baseline DiD speciﬁcation in the PSM sample and report the results in Table 12.We ﬁnd that in the PSM sample, we still ﬁnd that the coeﬃcients of Treat*Post are signiﬁcantly positive, indicating short-selling improves corporate innova- tion in both the innovation quantity and quality. 5.3. The instrument variable method Hennessy and Strebulaev (2015) argued that conducting a DiD identiﬁcation strategy in a quasi-experiment setting may not eliminate all endogeneity concerns. To further mitigate the endogeneity concern in this paper, we use the value of shares available for lending (SSP) as a proxy for short-selling and the ETF funds share-holdings as its instru- ment variable to further investigate the association between short-selling and corporate innovation. On the one hand, the ETF fund pursue proﬁt maximisation and risk minimisa- tion, thus ETF funds tend to use short-selling to mitigate systematic risk. Consequently, the ETF funds shareholding has a positive relation with the value of shares available for lending (Massa et al., 2015). On the other hand, the ETF funds only care about the proﬁtability rather than corporate operation, thus has no direct relation with ﬁrms’ innovation output. Therefore, it is suitable to use ETF funds shareholdings as instrument variable in our research setting. We perform the following 2SLS regressions: Step1 : SSP ¼ α þ βETF þ γZ þ δ þ φ þ ε (2) i;t i;t i;t i i;t Step2 : Innovation ¼ α þ βSSP þ γZ þ δ þ φ þ ε (3) i;tþ1 i;t i;t i i;t In Equation (2), we provide the ﬁrst-stage regression, in which SSP is the dependent variable and ETF is the independent variable. In this model, we predict the value of shares available for lending (SSP). In Equation (3), we present the second-stage regression. We regress corporate innovation on the predicted SSP from the ﬁrst-stage regression and the 310 C. LI ET AL. Table 12. DiD estimation in the PSM sample. (1) (2) (3) F.Patent FInvention FInv&Des Treat 0.0198 0.0405 0.0180 (0.67) (1.63) (0.63) Post 0.0485 0.0674 0.0539 (0.92) (1.53) (1.06) Treat*Post 0.0951* 0.0968** 0.105** (1.73) (2.11) (1.99) Constant −2.298*** −2.557*** −2.182*** (−6.72) (−8.97) (−6.67) Control Yes Yes Yes Industry Yes Yes Yes Year Yes Yes Yes N 10,169 10,169 10,169 R 0.4148 0.3655 0.4228 * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. other control variables used in Equation (1). Thus with this approach, the predicted value from the ﬁrst-stage regression is no longer correlated with the error term of the second- stage regression, and the estimated coeﬃcient is consistent. Table 13 reports the results. In column (1)–(3), we provide the results of OLS regres- sions, in which the dependent variable is innovation output and in the independent variable is SSP. We found a signiﬁcantly positive relationship between SSP and ﬁrms’ innovation output, which is consistent with our baseline results in the last section. In column (4)–(6), we document the results of the 2SLS regressions. It is shown that even when we use the instrument variable to mitigate the endogeneity concerns, we still ﬁnd signiﬁcantly positive relationship between the short-selling and corporate innovation, regarding both innovation quantity and quality. 5.4. Other robust checks 5.4.1. The pilot programme of transfer securities In the Pilot programme of marginal trading and short-selling started in 2010, the security ﬁrms are only allowed to lend their own funds and securities, which constrains the number of securities that are available for short-selling. In 2013, the CSRC started another pilot scheme, which allows banks, insurances companies and funds companies to transfer their securities to the security companies for lending. This pilot programme of transfer securities includes 11 security companies and approves 90 pilot ﬁrms’ securities to be transferred. The pilot programme is another exogenous shock for increasing the securities that are available for short-selling, thus provides another quasi-experiment setting to examine the real eﬀects of short-selling. Based on this setting, we use the DiD identiﬁca- tion strategy to investigate the relation between short-selling and corporate innovation again. Speciﬁcally, we deﬁne a dummy variable Trans_Post, which equals to 1 if ﬁrm i is included in the transfer security list in a given year t, otherwise equals to 0. Then, we use this dummy variable as our main independent variable in the DiD speciﬁcation. We also control the same control variables in Equation (1), the year ﬁxed eﬀect as well as ﬁrm ﬁxed eﬀect in the speciﬁcation. CHINA JOURNAL OF ACCOUNTING STUDIES 311 Table 13. The 2SLS regressions. OLS 2SLS (1) (2) (3) (4) (5) (6) F.Patent F.Invention F.Inv&Des F.Patent F.Invention F.Inv&Des SSP 0.617*** 0.592*** 0.749*** 1.347*** 1.535*** 1.620*** (3.22) (3.16) (3.84) (4.52) (6.06) (5.65) Sale 0.0325 0.0250 0.0265 0.0187 0.00726 0.0101 (1.48) (1.38) (1.24) (1.05) (0.48) (0.59) LEV 0.335** 0.229* 0.337** 0.339*** 0.234** 0.342*** (2.36) (1.94) (2.49) (2.92) (2.37) (3.06) ROA 0.268 0.366* 0.348 0.345* 0.466*** 0.441** (1.10) (1.96) (1.58) (1.70) (2.69) (2.25) Growth −0.0101 0.00751 −0.0106 0.000442 0.0212 0.00206 (−0.30) (0.27) (−0.33) (0.01) (0.80) (0.07) Tangibility 0.261** 0.237** 0.267** 0.268*** 0.246*** 0.275*** (2.10) (2.30) (2.21) (2.91) (3.14) (3.11) CapEx −0.0960 0.00337 −0.0245 −0.122 −0.0307 −0.0559 (−0.48) (0.02) (−0.12) (−0.68) (−0.20) (−0.33) RD 1.248 2.539 1.484 1.104 2.354*** 1.313 (0.78) (1.50) (0.90) (1.23) (3.07) (1.52) KZ −0.0948*** −0.0593** −0.0964*** −0.0906*** −0.0539** −0.0914*** (−2.98) (−2.43) (−3.21) (−3.55) (−2.48) (−3.72) PatGrow 0.224*** 0.169*** 0.206*** 0.223*** 0.168*** 0.205*** (8.58) (7.99) (8.27) (14.09) (12.52) (13.48) Institution 0.156 0.173 0.184 0.204 0.234** 0.241* (0.94) (1.11) (1.16) (1.56) (2.11) (1.92) Duality −0.0216 −0.0112 −0.0280 −0.0223 −0.0121 −0.0288 (−0.58) (−0.31) (−0.77) (−0.73) (−0.47) (−0.98) BoardSize 0.215* 0.236** 0.220** 0.207*** 0.226*** 0.211*** (1.93) (2.53) (2.02) (2.70) (3.47) (2.86) Indep 0.376 0.304 0.378 0.336 0.252 0.330 (1.25) (1.23) (1.29) (1.44) (1.27) (1.47) Age 0.515*** 0.535*** 0.551*** 0.544*** 0.572*** 0.585*** (3.31) (3.98) (3.65) (5.91) (7.30) (6.61) Constant −1.576** −1.875*** −1.646** −1.595*** −1.818*** −1.623*** (−2.36) (−3.35) (−2.52) (−3.36) (−4.51) (−3.56) Firm Yes Yes Yes Yes Yes Yes Year Yes Yes Yes Yes Yes Yes N 11,469 11,469 11,469 11,469 11,469 11,469 R 0.102 0.108 0.117 * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. In Table 14, we report the estimated results. In column (1)–(3), we ﬁnd that the coeﬃcients of Trans_Post are all signiﬁcantly positive, indicating short-selling could improve corporate innovation. In column (4)–(6), we further control the eﬀect of the pilot programme started in 2010 by adding the interaction term Treat*Post. The results document that both Trans_Post and Treat*Post have signiﬁcantly positive coeﬃcients. These results further prove that short- selling could improve corporate innovation under the Chinese setting. 5.4.2. Other proxies for corporate innovation In this section, we use other proxies for corporate innovation to make robust checks. Since ﬁrms’ patent applications sometimes take more than 1 year, we use the two and 3 years lagged innovation output as our dependent variables to test the eﬀect of short-selling on corporate innovation. Speciﬁcally, we use the lagged 2 years, 3 years and 4-years Invention and Inv&Des as dependent variables to re-estimate Equation (1). Table 15 reports the 312 C. LI ET AL. Table 14. The eﬀect of the pilot programme of transfer securities. (1) (2) (3) (4) (5) (6) F.Patent F.Invention F.Inv&Des F.Patent F.Invention F.Inv&Des Trans_Post 0.165*** 0.132*** 0.158*** 0.126** 0.0933** 0.117** (3.00) (2.87) (2.99) (2.26) (2.00) (2.18) Treat*Post 0.115*** 0.114*** 0.122*** (4.24) (5.03) (4.67) Size 0.0445*** 0.0410*** 0.0405*** 0.0354** 0.0321** 0.0309* (2.72) (3.01) (2.58) (2.15) (2.34) (1.96) LEV 0.265** 0.168* 0.248** 0.275** 0.179* 0.260** (2.37) (1.80) (2.31) (2.46) (1.92) (2.42) ROA −0.365* −0.200 −0.315* −0.326* −0.163 −0.275 (−1.88) (−1.24) (−1.69) (−1.68) (−1.00) (−1.48) Growth −0.0575* −0.0435* −0.0490* −0.0537* −0.0397 −0.0449 (−1.94) (−1.76) (−1.72) (−1.81) (−1.61) (−1.58) Tangibility 0.164* 0.194*** 0.184** 0.167* 0.197*** 0.186** (1.85) (2.63) (2.16) (1.88) (2.66) (2.19) CapEx 0.0437 0.127 0.0715 0.0415 0.125 0.0691 (0.25) (0.88) (0.43) (0.24) (0.87) (0.42) RD 5.372*** 5.705*** 5.448*** 5.264*** 5.597*** 5.333*** (6.20) (7.89) (6.55) (6.07) (7.75) (6.41) KZ −0.0594** −0.0397* −0.0616*** −0.0604** −0.0407** −0.0626*** (−2.42) (−1.94) (−2.61) (−2.46) (−1.99) (−2.66) PatGrow 0.324*** 0.206*** 0.299*** 0.324*** 0.207*** 0.300*** (21.24) (16.22) (20.45) (21.27) (16.25) (20.48) Institution 0.0571 0.0189 0.0687 0.0886 0.0500 0.102 (0.46) (0.18) (0.57) (0.71) (0.48) (0.85) Duality −0.0489* −0.0192 −0.0429 −0.0479 −0.0182 −0.0418 (−1.66) (−0.78) (−1.52) (−1.63) (−0.74) (−1.48) BoardSize 0.255*** 0.239*** 0.285*** 0.247*** 0.230*** 0.276*** (3.45) (3.87) (4.02) (3.34) (3.74) (3.90) Indep 0.441** 0.416** 0.563*** 0.408* 0.384** 0.529** (1.96) (2.22) (2.62) (1.82) (2.05) (2.46) Age 0.778*** 0.811*** 0.825*** 0.786*** 0.818*** 0.833*** (8.83) (11.02) (9.75) (8.92) (11.14) (9.85) Constant −2.655*** −2.953*** −2.904*** −2.462*** −2.812*** −2.669*** (−6.24) (−8.33) (−7.12) (−5.46) (−7.48) (−6.17) Control Yes Yes Yes Yes Yes Yes Firm Yes Yes Yes Yes Yes Yes Year Yes Yes Yes Yes Yes Yes N 11,469 11,469 11,469 11,469 11,469 11,469 R 0.022 0.023 0.041 0.023 0.025 0.043 * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. results. It is shown that the coeﬃcients of Treat*Post are signiﬁcantly positive in all the following regression, indication short-selling improves corporate innovation in a long term. In sum, our results are robust when we use other proxies of corporate innovation. 5.4.3. The tobit model and tests in a smaller sample In our sample, the median of all the three proxies of innovation output is zero, indicating half of the ﬁrms has no patents applications. Therefore, we use the Tobit models to re- estimate our basic speciﬁcations. We ﬁnd consistent results with our basic models. We did not report these results due to space limitations. Besides, in general industries, patenting activities are relatively fewer. However, in high-technology industries, patenting activities are more active. Speciﬁcally, we deﬁne manufacturing industry, information transmission industry, software development CHINA JOURNAL OF ACCOUNTING STUDIES 313 Table 15. Other proxies for innovation output. (1) (2) (3) (4) (5) (6) F2.Invention F2.Inv&Des F3.Invention F3.Inv&Des F4.Invention F4.Inv&Des Treat*Post 0.0552** 0.0869*** 0.0823** 0.144*** 0.0689 0.144*** (2.03) (2.83) (2.31) (3.61) (1.59) (2.98) Constant −1.584*** −1.669*** −1.104** −1.110** 0.231 0.335 (−4.02) (−3.74) (−2.53) (−2.27) (0.47) (0.62) Control Yes Yes Yes Yes Yes Yes Year Yes Yes Yes Yes Yes Yes Firm Yes Yes Yes Yes Yes Yes N 10,134 10,134 8773 8773 7344 7344 R 0.0666 0.0767 0.0495 0.0565 0.0391 0.0404 * * *, * *, * represent the signiﬁcance level of 1%, 5%, 10%, respectively; the output of t value is reported in the parentheses. industry and information technology service industry as high-technology industries. In this section, we extract high-technology industries according to the CSRC industry index and estimate our basic speciﬁcation only in this sample. We also ﬁnd consistent results with our basic models. We also did not report these regressions due space limitations. 6. Conclusions and implications In this paper, we use the deregulation of short-selling constraint started in 2010 as a quasi- experiment to investigate the real eﬀect of short-selling on corporate innovation. We ﬁnd that in China short-selling improves corporate innovation signiﬁcantly regarding both innovation quantity and quality. Besides, our cross-sectional tests show that short-selling could improve corporate innovation in companies with worse internal and external governance environment. The results indicate that short-selling is a necessary comple- mentary mechanism of ﬁrms’ corporate governance system. Besides, we also examine the possible underlying mechanisms through which short-selling aﬀect corporate innovation. We document that short-selling improves corporate innovation through lowering ﬁrms’ information asymmetry and improving the eﬃciency of managerial contract eﬃciency. Our results are robust after eliminating multiple endogeneity concerns. For example, our results satisfy the parallel trend; our results are robust when we use the PSM to mitigate the fundamental diﬀerences between the pilot ﬁrms and control ﬁrms; our results are also robust when we use the instrument variable strategy. Lastly, we also conduct other robust checks, such as use other proxies for corporate innovation, use the Tobit models, etc. In general, the robust checks are consistent with our basic results. Our empirical results conﬁrm that positive role that short-selling plays in the Chinese ﬁnancial markets. Prior literature has argued that short-selling is critical for improving market eﬃciency and ﬁrm performance in developed countries, such as, in the US. Our paper provides evidence that short-selling also improves market eﬃciency in under- developed markets like China. Besides, short-selling also disciplines managerial beha- viour and mitigates agency problems in Chinese companies; thus, it is an important complementary mechanism of ﬁrms’ corporate governance system. Due to the positive eﬀects of short-selling in China, we suggest regulators could expand the scales of the pilot programme step by step and encourage ﬁnancial innovation in Chinese ﬁnancial markets. 314 C. LI ET AL. Acknowledgments The authors acknowledge the ﬁnancial support from the Humanities and Social Sciences of Ministry of Education of China (No. 19YJA790038) and the ﬁnancial support from the National Natural Science Foundation of China (No. 71802113). Chuntao Li also acknowledges the support from Collaborative Innovation Center of Industrial Upgrading and Regional Finance (Hubei). Hongmei Xu also acknowledges the support from the West Bank of the Pearl River Research Center in South China Normal University. All errors are our own. Disclosure statement No potential conﬂict of interest was reported by the authors. Funding This work was supported by the Humanities and Social Sciences of Ministry of Education of China [19YJA790038]; National Natural Science Foundation of China . References Aggarwal, R., Erel, I., Ferreira, M., & Matos, P. (2011). Does governance travel around the world? Evidence from institutional investors. Journal of Financial Economics, 100(1), 154–181. Bertrand, M., & Mullainathan, S. (2003). Enjoying the quiet life? Corporate governance and manage- rial preferences. Journal of Political Economy, 111(5), 1043–1075. Bradley, D., Kim, I., & Tian, X. (2013). The causal eﬀect of labor unions on innovation, SSRN paper. Available at SSRN 2232351. Bris, A., Goetzmann, W.N., & Zhu, N. (2007). Eﬃciency and the bear: Short sales and markets around the world. The Journal of Finance, 62(3), 1029–1079. Chang, E.C., Cheng, J.W., & Yu, Y. (2007). Short-sales constraints and price discovery: Evidence from the Hong Kong market. The Journal of Finance, 62(5), 2097–2121. Chang, E.C., Lin, T.C., & Ma, X. (2018). Does short-selling threat discipline managers in mergers and acquisitions decisions? Journal of Accounting and Economics, forthcoming. Chang, E.C., Luo, Y., & Ren, J. (2014). Short-selling, margin-trading, and price eﬃciency: Evidence from the Chinese market. Journal of Banking & Finance, 48, 411–424. Chemmanur, T.J., Loutskina, E., & Tian, X. (2014). Corporate venture capital, value creation, and innovation. The Review of Financial Studies, 27(8), 2434–2473. Chen, H., & Liu, F. (2014). The governance roles of margin trading: A perspective of earnings management. Accounting Research, 9,45–52. Chen, S., Lin, B., Lu, R., & Ma, H. (2018). The disciplinary eﬀects of short sales on controlling shareholders. Journal of Empirical Finance, 46,56–76. Christophe, S.E., Ferri, M.G., & Hsieh, J. (2010). Informed trading before analyst downgrades: Evidence from short sellers. Journal of Financial Economics, 95(1), 85–106. Chu, Y. (2015). Short selling and the product market: Evidence from SHO, SSRN paper. Available at SSRN Cornaggia, J., Mao, Y., Tian, X., & Wolfe, B. (2015). Does banking competition aﬀect innovation? Journal of Financial Economics, 115(1), 189–209. Desai, H., Krishnamurthy, S., & Venkataraman, K. (2006). Do short sellers target ﬁrms with poor earnings quality? Evidence from earnings restatements. Review of Accounting Studies, 11(1), 71–90. Diamond, D.W., & Verrecchia, R.E. (1987). Constraints on short-selling and asset price adjustment to private information. Journal of Financial Economics, 18(2), 277–311. CHINA JOURNAL OF ACCOUNTING STUDIES 315 Dong, X., Zhao, J., & Yuan, P. (2014). Research on innovation eﬃciency loss of state-owned enterprises. China Industrial Economic, 2,97–108. Edmans, A., Goldstein, I., & Jiang, W. (2015). Feedback eﬀects, asymmetric trading, and the limits to arbitrage. American Economic Review, 105(12), 3766–3797. Fan, G., Wang, X., & Zhu, H. (2010). NERI INDEX of marketization of China’s provinces 2009 report. Beijing: Economic Science Press. Fang, V.W., Huang, A.H., & Karpoﬀ, J.M. (2016). Short selling and earnings management: A controlled experiment. The Journal of Finance, 71(3), 1251–1294. Fang, V.W., Tian, X., & Tice, S. (2014). Does stock liquidity enhance or impede ﬁrm innovation? The Journal of Finance, 69(5), 2085–2125. Gompers, P.A. (1996). Grandstanding in the venture capital industry. Journal of Financial Economics, 42(1), 133–156. Graham, J.R., Harvey, C.R., & Rajgopal, S. (2005). The economic implications of corporate ﬁnancial reporting. Journal of Accounting and Economics, 40(1–3), 3–73. Grossman, S. (1976). On the eﬃciency of competitive stock markets where trades have diverse information. The Journal of Finance, 31(2), 573–585. Gu, Z., Hao, X., & Zhang, Y. (2011). Short - selling constraints, investor behavior and A - share market pricing bubble. Journal of Financial Research, 2, 129–148. He, J., & Tian, X. (2016). Do short sellers exacerbate or mitigate managerial myopia? Evidence from patenting activities. Social Science Electronic Publishing. Hellwig, M.F. (1980). On the aggregation of information in competitive markets. Journal of Economic Theory, 22(3), 477–498. Hennessy, C., & Strebulaev, I.A. (2015). Natural experiment policy evaluation: A critique. CEPR Discussion Paper No. DP10455. Retrieved from https://ssrn.com/abstract=2572453 Henry, Ó.T., & McKenzie, M. (2006). The impact of short selling on the price-volume relationship: Evidence from Hong Kong. The Journal of Business, 79(2), 671–691. Hsu, P.H., Tian, X., & Xu, Y. (2014). Financial development and innovation: Cross-country evidence. Journal of Financial Economics, 112(1), 116–135. Januszewski, S.I., Köke, J., & Winter, J.K. (2002). Product market competition, corporate governance and ﬁrm performance: An empirical analysis for Germany. Research in Economics, 56(3), 299–332. Kaplan, S.N., & Zingales, L. (1997). Do investment-cash ﬂow sensitivities provide useful measures of ﬁnancing constraints? The Quarterly Journal of Economics, 112(1), 169–215. Karpoﬀ, J.M., & Lou, X. (2010). Short sellers and ﬁnancial misconduct. The Journal of Finance, 65(5), 1879–1913. Keim, D.B., & Madhavan, A. (1995). Anatomy of the trading process empirical evidence on the behavior of institutional traders. Journal of Financial Economics, 37(3), 371–398. Li, Y., & Zhang, L. (2015). Short selling pressure, stock price behavior, and management forecast precision: Evidence from a natural experiment. Journal of Accounting Research, 53(1), 79–117. Li, Z., Chen, C., & Lin, B. (2015). Does short-selling improve the pricing eﬃciency of China’s stock market? – Evidence from natural experiments. Economic Research Journal, 50(4), 165–177. Manso, G. (2011). Motivating innovation. The Journal of Finance, 66(5), 1823–1860. Massa, M., Wu, F., Zhang, B., & Zhang, H. (2016). Saving long-term investment from short-termism: The surprising role of short selling. INSEAD Working Paper No. 2015/11/FIN. Retrieved from SSRN: https://ssrn.com/abstract=2558876 Massa, M., Zhang, B., & Zhang, H. (2015). The invisible hand of short selling: Does short selling discipline earnings management? The Review of Financial Studies, 28(6), 1701–1736. McGrath, R.G. (2001). Exploratory learning, innovative capacity, and managerial oversight. Academy of Management Journal, 44(1), 118–131. Megginson, W.L., & Netter, J.M. (2001). From state to market: A survey of empirical studies on privatization. Journal of Economic Literature, 39(2), 321–389. Miller, E.M. (1977). Risk, uncertainty, and divergence of opinion. The Journal of Finance, 32(4), 1151–1168. 316 C. LI ET AL. Narayanan, M. (1985). Managerial incentives for short-term results. The Journal of Finance, 40(5), 1469–1484. Rosenbaum, P.R., & Rubin, D.B. (1983). The central role of the propensity score in observational studies for causal eﬀects. Biometrika, 70(1), 41–55. Saﬃ, P.A., & Sigurdsson, K. (2010). Price eﬃciency and short selling. The Review of Financial Studies, 24(3), 821–852. Subrahmanyam, A., & Titman, S. (1999). The going-public decision and the development of ﬁnancial markets. The Journal of Finance, 54(3), 1045–1082. Tan, Y., Tian, X., Zhang, X., & Zhao, H. (2015). The real eﬀects of privatization: Evidence from China’s split share structure reform. Kelley School of Business Research Paper, (2014-33). Wu, Y. (2012). The dual eﬃciency losses in Chinese state-owned enterprises. Economic Research Journal, 3,15–27. Xiao, H., & Kong, A. (2014). A study on the mechanism of the impact of the securities margin trading on the ﬂuctuations of the special nature of the stock price: A test based on the diﬀerence-in- diﬀerence model. Management World, 8,30–43. Zhang, X., Zhou, P., & Li, C. (2016). Short selling and earnings quality: Evidence from corporate ﬁnancial restatement. Journal of Financial Research, 8, 175–190.
China Journal of Accounting Studies
– Taylor & Francis
Published: Jul 3, 2019
Keywords: Short-selling; corporate innovation; information asymmetry