Article information

Abstract

Trans Abstract

1. Introduction

“Banning short-sellinginterferes with price formation, thereby increasing uncertainty. That can only artificially amplify volatility and probability of default, the opposite effect to that claimed, and hampers the ability of markets to serve the real economy. It is not – and never has been – true that bans have any other, positive effect on market activity or price levels.”

- 30 March 2020, The World Federation of Exchanges (WFE)¹⁾ —

WFE criticizes recent bans on short-selling against the stock market crash due to the COVID-19 pandemic.^2), The European Union and a number of emerging countries, such as South Korea, Spain, and France decided to engage in short-selling to protect their capital market. In contrast, the U.S., U.K., and Japan had refused to engage in short-selling in their stock market, as they believe that “there is no evidence that it is a driver of market routs.”³⁾

Short-selling restrictions havebeen extensively studied in the finance literature, typically since the 2008 financial crisis, however, academicand practical evidence on the effect of short-selling is still controversial (e.g., Jiang et al., 2022). On the one hand, short-sellers who might be sophisticated investors canplay an important and positive role in price efficiency. On the other hand, short-selling activities may manipulate stock prices and destabilize the market, and thus, regulators react to the crisis period by restricting short-selling.

Nevertheless, empirical specification for examiningthe impact of short-selling is challenging because of endogeneity concerns, such as selection bias and reverse causality. Addressing such concerns is difficult without exogenous change in short-selling activities. Thus, the literature has attempted to constructa quasi-natural experiment using the regulatory change on short-selling. For instance, Beber and Pagano (2013) focus on the regulatory interventions imposing bans on short-selling around the world during the 2008 financial crisis. The U.S. short-selling Regulation SHO also can be viewed as an exogenous shock, and thus, allows conducting empirical design in prior studies (Deng et al., 2020; Fang et al., 2016; Grullon et al., 2015; Li and Zhang, 2015).

With this regard, the current Korean stock market provides an ideal setting to overcome endogeneity issues because of its unique regulations on short-selling from 2020 to 2021. In March 2020, short-selling activities for all the listed stocks on the Korea Exchange (KRX)^4), were suspended due to the COVID-19 shock. After about 14-months later, since May 2021, short-selling activities have been allowed but only for stocks that are components of the following two representative indices in the KRX: KOSPI200 and KOSDAQ150.^5), Therefore, when the constituents of these indices are updated, newly indexed or excluded stocks in 2021 have experienced exogenous shocks in their short-selling availability. By comparing impacts of the index addition or out both in 2020 and 2021, we attempt to examine the relatively pure impacts of short-selling activities in Korea. More specifically, in order to assume that entering in or removing from KOSPI200 and KOSDAQ150 constituents is randomly determined, which is a key condition of a quasi-experimental setting, we restrict our sample into similar stocks in market size and trading volume.In other words, we set the control group as such similar stocks but do not experience inclusion or exclusion at the date of KOSPI200 and KOSDAQ150 constituents change, while the treatment group as the stocks that have entered in or removed from the indices. As a result, the treatment and control groups are basically similar, but after the date of indices’constituents change, only the treated stocks are suddenly available/forbidden for short-selling in the market because of their affiliation changes. Note that although a number of prior studies have investigated short-selling in Korea (e.g., Jung et al., 2013; Lee and Wang, 2022; Wang, 2023), our study is the first attempt to employ this quasi-experimental regulatoryto understand impacts of short-selling permission or ban.

Using this quasi-experimental setting due to the unique short-selling regime change in the Korean stock market, we first confirm that stocks belonging to KOSPI200 or KOSDAQ150 index in 2021 have experienced a significant increase in their short-selling volume, supporting the validity of our empirical design. For instance, if short-sellers in the Korean stock market are not interested at all to short newly indexed stocks, our empirical strategy could be meaningless; however, we find that such newly indexed stocks have significant short-selling volume immediately after their index inclusion.In addition, short-selling volume for stocks that are excluded from KOSPI200 or KOSDAQ150 index after the index update date immediately drop.

Most importantly, we find that short-selling permission in the Korean stock market enhances the affected stocks’ price efficiencies, andthus, it can contribute to improving market efficiency. This finding is also consistent with the view in recent studies that suggest a prominent role of short-selling in price discovery (e.g., Bushman and Pinto, 2024; Luu et al., 2023). Our results further imply that short-selling permission appears to be related to stock price appreciation. However, we find no evidence short-selling permission/ban strongly affects stock volatility, which is contrary to the general argument on short-selling in the Korean stock market.

We ensure that this paper provides practical and policy implications. According to our empirical evidence, short-selling permission appears to improve price efficiency, thereby contributing to the KRX. It suggests that short-sellers (using their information) play an important role in price discovery to some extent. Moreover, our paper shows no evidence that short-selling activities are strictly detrimental to stock returns or volatilities. Although our finding (at least partially) supports the positive side of short-selling in the Korean stock market, we note that its generalization to smaller stocks should be carefully studied further as we fail to find the consistency of evidence with the KOSDAQ-only sample. This suggest that stocks with low liquidity might have different outcomes with short-selling permission/ban.

The remainder of this paper is organized as follows. Section 2 reviews therelated literature on short-selling. Section 3 describes the unique institutional background about short-selling regime in the Korean stock market and our empirical specification. Section 4 presents the empirical results. Section 5 concludes the paper.

2. Related literature

Short-selling is a well-known trading strategy, that is “selling high and buying low.” Starting from early studies that theoretically compare the cost and risks ofshort-selling (e.g. Diamond and Verrecchia, 1987; Miller, 1977), the related literature generallyposits that short sellers are sophisticated or informed traders in the market as they will short stock only if they believe that their targeting stocks will compensate for additional costs and risks. Such bearing costs and risks in short-selling are, for example, unlimited loss with stock price incline, which is also recently highlighted by the GameStop episode (Atmaz et al., 2024), and restrictions on access to proceeds and some legal constraints (Diamond and Verrecchia, 1987).

Miller (1977) argues that overpriced stocks exist in the market if there is disagreement about these values among investors. In addition, if the market restricts pessimistic investors who might want to short such overpriced stocks (which could be viewed as short-selling bans), this overpricing phenomenon is more likely to occur.For instance, if short sellers are allowed to short an overpriced stock before its bad earnings announcement, and they are rational and informed investors, they can cool down such an overpriced stock’s price even before its earnings announcement, implying a positive role in price discovery of short sellers.

However, a stream of research argues that short-selling is detrimental in society, as short sellers are indeed uninformed but just predatory traders (in general, speculative institutional investors). As a result, short-selling per se is likely to be associated with stock price manipulation, increase in market volatility, and signaling of selling pressure that makes profits for predatory short sellers (Allen and Gale, 1992; Brunnermeier and Oehmke, 2014).

Given the nature of the short-selling strategy, the endogeneity threat is critical in the short-selling literature, as highlighted by Jiang et al. (2022), which is a review paper on short-selling. For instance, it is difficult to argue whether a stock’s short-selling pressure causes its subsequent price drop or short-sellers indeed target such a stock that will drop in the near future. Therefore, a large body of empirical studies attempts to address endogeneity concerns by using exogenous change in short-selling activities.

As a result, regulations on short-selling have been extensively studied in the finance literature, particularly exploiting the 2008 financial crisis in which many financial regulators imposed a ban onshort-selling to reduce volatility and to stem the market crash. However, the empirical evidence is mixed or even contradicts the reason for banning short-selling.

Boulton and Braga-Alves (2010) find that short-selling bans, in fact, do not reduce market volatility on average exploiting the US Securities and Exchange Commission’s (SEC) temporary restrictions on naked short sales of the stocks of 19 financial firms in July 2008. Saffi and Sigurdsson (2011)use the global sample from 2005 to 2008, and they find that short-selling restrictions are harmful to market efficiency, and the subsequent relaxing of such restrictions does not induce volatility or the extreme stock price drop. Bris et al. (2007) also highlight the positive side of removing short-sale restrictions, using regulatory intervention on whether short-selling is prohibited or practiced in 46 countries. Chang et al. (2007) who examine the Hong Kong sample find that short-selling bans on individual stocks tend to cause overvaluation, in line with the argument of Miller (1977). More recently, Luu et al. (2023) examine the effect of short-selling in the US market after the COVID-19 pandemic. The authors find that, during the COVID-19 shock in the stock market, short-selling activities are only concentrated on overpriced stocks, suggesting that short-selling plays a role in improving price discoveries insteadof triggering stock market crashes.

In terms of the Korean stock market, several papers have examined the effect of short-selling in the stock market (Jung et al., 2013; Chung and Wang, 2020; Lee and Wang, 2019, 2022; Wang, 2023; Wang et al., 2017; Wang and Lee, 2015). Notably, most research focusing on the Korean sample also exploits the Korean government’s intervention in October 2008, which was temporary bans for all forms of short-selling due to the financial crisis. Jung et al. (2013) suggest that short-selling (by individual investors) contributes to price efficiency in the stock market. Wang and Lee (2015) find that short-selling in the Korean stock market is mainly driven by foreign investors, and their short-selling does not induce market volatility. More recently, Wang (2023) examines the predictability of short-selling on future returns in the Korean stocks. Further, the author finds that short-selling is more concentrated for larger firms with better credit grades, in contrast to the US market, and short-selling has a predictivepower on stock returns only for such larger firms.

It is worth to note that, to the best of our knowledge, there is no research yet exploiting short-selling regime changes in the Korean stock market against the COVID-19 pandemic, although there are several prior studies on short-selling using the Korean sample. We will discuss our empirical strategy that exploits the unique institutional setting in Korea around the COVID-19 pandemic in the next section.

3. Institutional background and empirical identification

In March 2020, the Korean government agency, the Financial Services Commission (FSC) decides to engage in short-selling activities in the Korean market. We collect related announcements from press releases of the FSC, summarized in Panel A of Table 1. Specifically, the FSC warns about the worldwide spread of COVID-19 (also, in Korea) and its negative shock on the stock market as well as significant rocketing volatility and short-selling volume. Therefore, they decide to impose a ban on short-selling for the 6-months period as a stock market stabilization instrument. This short-selling ban, however, has continued exceptionally long for about 14-months from March 2020 in Korea. After that, the FSC decided to resume short-selling activities in the KRX, but more importantly, allowing only for KOSPI200 and KOSDAQ150-indexed stocks.

<Table 1>

Key information for empirical specification

Note. This table illustrates key information for our empirical identification. Panel A summarizes the Korean government’s intervention on short-selling in KRX against the COVID-19 pandemic shock from early 2020 to 2021. Panel B summarizes the regular updates in 2020 and 2021 for KOSPI200 and KOSDAQ150. For these two representative indices, KRX regularly announces update twice in a year (every June and December). Panel C shows the treatment and control groups in our empirical specification around the four regular updates of the indices in 2020 and 2021. In Panel C, Index_in (Index_out) equals to one for stocks that are newly indexed (excluded) in terms of KOSPI200 and KOSDAQ150, and zero for the control group. We define the control group as ten (five)nearest neighbors with each of the treated stocks by performing the propensity score matching technique among stocks that have not experienced changesin their affiliations for KOSPI200 or KOSDAQ150 index at the dates of the four regular updates of the indices, denoted by Larger (Smaller) control group.

Panel A: Short-selling related announcements in 2020 and 2021 by the Korean government

This paper focuses on the date of KOSPI200 and KOSDAQ150 constituents change in 2021 as the index-included/excluded firms suddenly experience short-selling permission/ban. Note that constituents of the index are updated in June and December every year in a regular manner (hereafter, the regular update), as summarized in Panel B of Table 1. Nevertheless, still there is a potential concern because the index’s regular update might be related to money flow from passive funds in financial institutions. For instance, some passive funds may include only stocks belonging to the KOSPI200 or KOSDAQ150 indices, thus stocks that are indexed or excluded may experience exceptional changes in demand from the passive funds. To address this concern, our empirical strategy is to use the same regular update events that occurred in 2020 as a base.

The only difference inwhether a stock belongs to KOSPI200 or KOSDAQ150 index between 2020 and 2021 is short-selling availability. Thus, most importantly, we can estimate the impact of short-selling permission/ban by comparing the index’s regular update (i.e., the index addition/out) in 2020 and 2021.⁶⁾ Assuming that entering in or removing from KOSPI200 and KOSDAQ150 constituents among similar stocks around the threshold of such indices is (almost) random, our empirical strategy can be viewed as a quasi-natural experiment. Overall, we implement the following difference-in- difference (DID) specification:

where the dependent variable is a target variable of interest, such as short-selling volume and price efficiency measures. Index_change_it is an indicator that equals one for the treatment group: stocks changing their affiliation at the date of the index’s regular update, and more specifically, Index_in and Index_out, and zero for the control group: stocks that are around KOSPI200 and KOSDAQ150 indices’ threshold but are unchanged in terms of their affiliations at the index’s regular update. We describe available observations for the treatment and control groups in Panel C of Table 1. More specifically, to construct the control group, we first identify stocks that have not experienced changes in their affiliations for KOSPI200 or KOSDAQ150 index. Among these stocks, we then perform the propensity score matching technique to find ten or five nearest neighbors with each of the treated stocks (hereafter, denoted by Larger or Smaller control group, respectively).⁷⁾ The indicator Y2021 is the post variable of the unique short-selling regime in the Korean stock market, which equals to one for observations in 2021 and zero for ones in 2020.

It is noteworthy that, since we include observations at the regular update events in 2020, β₁ can represent the impact of the index addition/out in 2020, such as money flow from passive funds. The coefficient of Y2021 (i.e., β₂) is econometrically identical to year fixed effect, thus, our empirical tables do not mention year fixed effect. Most importantly, our main DID estimator β₃ indicates difference in the effect of the index addition/out between 2020 and 2021, thereby referring to the pure impact of short-selling permission/ban.

Furthermore, we control for the stock’s market capitalization and trading volume. To control for market-specific and industry-specific effects, we also include an indicator that equals one for stocks listed on KOSPI and zero for KOSDAQ, and the industry fixed effects (δ_j) based on 17 sections of the Korean Standard Industrial Classification (KSIC).

We manually collect data on KOSPI200 and KOSDAQ150 constituent changes from the KRX announcement,^8), and provide the list of stocks that experienced index-in and index-out (i.e., the treatment group) from 2020 to 2021 in Appendix A. We also obtain stock price and trading data from FN Dataguide. Then, we construct a cross-sectional sample around the regular update in June and December 2020 and 2021 (i.e., four events). All the continuous variables are winsorized at the 1st and 99th levels to reduce the effect of outliers. Panels A and B of Table 2 show the detailed definitions of variables and the descriptive statistics, respectively.

<Table 2>

Sample summary

Note. This table illustrates definitions of the variables and summary statistics in Panels A and B, respectively. We winsorize all continuous variables at 1st and 99th percentiles to remove the influence of outliers. Panel C reports the mean difference in stocks’ market capitalization and trading volume between the treatment (Index_in = 1 or Index_out = 1) and the corresponding control groups (Larger and Smaller control group).

Panel A: Definitions of variables

4. Empirical analysis

4.1. Validity test

Before we investigate the impact of short-selling permission/ban, it is necessary to check the validity of our empirical specification. In Panel C of Table 2, we first examine whether our two main control variables (i.e., the stock’s market capitalization and trading volume) significantly differ by the treatment and control groups.⁹⁾ The first and second rows in Panel C focus on the samples of Index_in and Index_out, and the last two columns report the t-statistics for the mean difference between the treatment group versus the Larger or Smaller control group, respectively. Both columns show that the t-statistics for the mean difference are sufficiently small. Thus, the results of Panel C suggest that the treatment and control groups are basically similar in our sample construction, further supporting the underlying assumption of our empirical strategy.

Next, we examine whether short-selling activities of stocks that have entered in (Index_in) or removed from (Index_out) KOSPI200 and KOSDAQ150 indices are indeed affected, compared to the control group. For instance, if short-selling volume per se of the treated stocks (i.e., newly-entered stocks in KOSPI200 and KOSDAQ150 indices) is unaffected, it may indicate that short-sellers are, in fact, not interested in the treated stocks that are close to around those indices’ threshold; if so, our empirical specification could be invalid.

Table 3 presents the results of this validity test.Odd-numbered columns indicate Larger control group, whereas even-numbered columns use a more restrictive criterion, Smaller control group, for selecting the control group as described in the previous section. Following the literature on short-selling (Diether et al., 2009; Lee and Wang, 2019), we calculate relative short-selling as the trading volume of short-selling divided by the total trading volume, and we take its difference between those for1-, 2-, and 3-months (i.e., 20-, 40-, 60-trading-days) before and after the index’s regular update. As we expect, the interaction terms between Index_in (Index_out) and Y2021 show positive (negative) estimates, whereas both the standalone estimates of Index_in and Index_out are insignificant in Table 3. These results imply that belonging to KOSPI200 or KOSDAQ150 index in 2021 is significantly associated with short-selling activities, supporting that our empirical design can be a quasi-natural experiment.

<Table 3>

Effects of short-selling permission/ban on short-selling volume

Note. This table reports the results of validity tests where the dependent variable is short-selling volume changes, the difference in short-selling relative volume between before and after the index updates (based on 20, 40, and 60 trading days). Models (1)-(6) and (7)-(12) examine the impact of short-selling permission and ban, respectively. Definitions of the variables are described in Panel A of Table 2. Parentheses report t-statistics based on standard errors adjusted for heteroscedasticity. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively.

4.2. Main results: Impact of short-selling permission/ban

Table 4 analyzes the impact of short-selling permission/ban on price efficiency. In particular, we adopt two measures of price efficiency proposed by Hou and Moskowitz (2005), also referred to as price delay in the literature (e.g., Saffi and Sigurdsson, 2011). Note that a lower value of price delay indicates a higher price efficiency, and thus, a more efficient price discovery in the stock market. In Table 4, the dependent variable is the difference in between price delay calculated by 3-months (i.e., 60-trading-days) before and after the index’s regular update. In columns (1)–(4), β₃ (i.e., the interaction terms between Index_in and Y2021) is negatively estimated on price delay measures at least at the 10% significance level, suggesting that short-selling permission appears to improve the affected stocks’ price efficiencies. We find no evidence that short-selling ban significantly influences price efficiency in our empirical design (columns (5)–(8)). The results in Table 4 are generally consistent with the prior literature revealing the positive impact of short-selling on price efficiency (Beber and Pagano, 2013; Saffi and Sigurdsson, 2011).

<Table 4>

Effects of short-selling permission/ban on price efficiency

Note. This table reports the main results where the dependent variable is changes in price efficiency measures. We calculate Hou and Moskowitz’s (2005) measures, Price delay 1 or 2, for each stock and take its difference between 60 trading days before and after the index update. Models (1)-(4)and (5)-(8) examine the impact of short-selling permission and ban, respectively. Definitions of the variables are described in Panel A of Table 2. Parentheses report t-statistics based on standard errors adjusted for heteroscedasticity. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively.

Next, we examine whether the affected stocks by short-selling permission/ban exhibit different patterns of stock returns or volatilities. Market participants in the Korean stock market (typically, retail investors) often argue that short-selling should be banned because, for example, foreign investors intensively short-sell Korean stocks and take earnings out of the Korean stock market. It is not difficult to see news suggesting that short-selling is one significant factor of market meltdown.

Table 5 presents the results where we focus on stock returns for 1-, 2-, and 3-months (i.e., 20-, 40-, 60-trading-days) after the index’s regular update. In columns (1)–(6), the treatment group in 2020 (i.e., Index_in) is likely to have negative returns compared to the control group. However, β₃ (i.e., Index_in × Y2021) in these columns generally have significantly positive coefficients; we cautiously interpret that this could be due to the improved price discovery (as shown in Table 4), but it needs to be studied further. In addition, β₃ for Index_out shows the opposite direction in columns (7)–(12), although the statistical significance levels are relatively weak.

<Table 5>

Effects of short-selling permission/ban on stock returns

Note. This table reports the main results where the dependent variable is stock returns for the next 20, 40, and 60 trading days after the index updates. Models (1)-(6) and (7)-(12) examine the impact of short-selling permission and ban, respectively. Definitions of the variables are described in Panel A of Table 2. Parentheses report t-statistics based on standard errors adjusted for heteroscedasticity. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively.

We conclude that the results of Table 5 contradict the general belief about short-selling in the Korean stock market. On the one hand, at the very least, short-selling permission is not destroying stock returns on average; instead, we find that it could be positively related to stock returns. On the other hand, there is no evidence that short-selling ban could be helpful to protect stock returns. Our interpretations are consistent with Saffi and Sigurdsson’s (2011) argument that relaxing short-selling constraints does not lead to negative stock returns.

Table 6 examines the impact of short-selling permission/ban on stock volatility. To measure the change in stock volatility around the index’s regular update, we calculate a variance of daily stock returns during 1-, 2-, and 3-months (i.e., 20-, 40-, 60-trading-days) before and after the index’s regular update. We take its difference (between before and after value) and standardize by its before value. Here, we find that the treated stock’s volatility by short-selling permission does not show significant changes on average during all the three periods (see, β₃ in columns (1)–(6)). Index_out × Y2021 also shows insignificant coefficients in columns (7)-(12), concluding that short-selling ban also has no significant impact on stock volatility.

<Table 6>

Effects of short-selling permission/ban on stock volatility

Note. This table reports the main results where the dependent variable is changes in stock volatility. We calculate a variance of daily stock returns for each stock and take its ratio of the difference between before and after the index updates (based on 20, 40, and 60 trading days) to the before value. Models (1)-(6) and (7)-(12) examine the impact of short-selling permission and ban, respectively. Definitions of the variables are described in Panel A of Table 2. Parentheses report t-statistics based on standard errorsadjusted for heteroscedasticity. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively.

Taken together for Tables 5 and 6, we repeatedly note that our empirical evidence is against the general belief about short-selling in the Korean stock market. For instance, if short-selling activities are associated with (inappropriate) selling pressure, and thus, negatively influence stock price, short-selling permission/ban could be bad/good news for shareholders. However, we rather find the opposite results which are weakly positive (negative or insignificant) stock returns even after short-selling permission (ban). The results on stock volatility, regardless of whether the event is Index_in or Index_out, are also not strong enough to claim the detrimental role of short-sellers.Note that our results are also consistent with prior studies that suggest the absence of evidence for speculative short-sellers in the Korean stock market (Jung et al., 2013; Wang and Lee, 2015).

Notably, our empirical evidence suggests that short-selling permission in the Korean stock market improvesprice efficiency. In addition, we argue that stock returns and volatility of the treatment groups are not strongly affected compared to the control groups in our empirical setup. Thus, we argue that there is no strongevidence supporting the evil side of short-selling in the Korean stock market, such as dropping stock price, limiting market efficiency, and destabilizing market by predatory short sellers; rather, our study supports the bright side by suggesting an important role of short-selling for price efficiency and discovery (Beber and Pagano, 2013; Saffi and Sigurdsson, 2011).

5. Conclusion

In November 2023, the Korean FSC again decided to suspend short-selling activities in the Korean stock market. This suspension will continue to March 2025, and until that time, the Korean government is aiming to build the system to monitor speculative short sellers. With this regard, ourpaper can provide both academic and practical implications, by attempting to employ the recent regulation changes on short-selling in the Korean stock market.

First, we suggest that the current regime in the Korean stock market can be an ideal setting to construct academic research on the impact of short-selling. More specifically, we adopt a quasi-natural experiment design using the constituents change in KOSPI200 and KOSDAQ150 indices in 2020 and 2021, aiming to overcome potential endogeneity concerns. Since endogeneity threat is critical in the short-selling literature, this novelapproach alleviates such concerns and enables us to examine thepure impact of short-selling permission/ban.

Second, our empirical analysis shows evidencethat at least some short-sellers play an important role in price discovery, rather than its detrimental role in society. In particular, short-selling permission in the Korean stock market appears to enhance the affected stocks’ price efficiencies, and thus, it can contribute to improving market efficiency. Notably, in line with our results, recent studies on short-selling also suggest the positive role of short-selling (e.g., Bushman and Pinto, 2024; Luu et al., 2023).

Third, there is no clear evidence that short-sellers are predatory traders on average in Korea. In our empirical results, we never findthat short-selling activities are harmful to the stock market, neither from the perspective of stock returns nor volatilities. This result is also consistent with prior studies, which strongly suggestthe absence of evidence for speculative short-sellers in the Korean stock market (Jung et al., 2013; Wang and Lee, 2015).

Recently, in June 2024, Morgan Stanley Capital International (MSCI) announced that South Korea failed to include in its index of the developed market.¹⁰⁾ In particular, MSCI pointed out that the Korean government’s intervention in short-selling ban in November last year is not desirable. This MSCI view is in linewith one from the WFE statement in March 2020 (as illustrated in the Introduction). Consistent with these views, our paper concludes that banning short-selling might not be helpful for stabilizing the market, further casting doubts on the reason for banning short-selling. Yet, some future studies are required to check whether short-selling rules towards smaller stocks also show the same pattern of empirical evidence in this paper.

Article information Continued