Short Selling: A Review of the Literature and Implications for Future Research

Abstract

This systematic literature review critically analyzes studies on the determinants of short selling and the implications for information distribution, real economic decisions, financial reporting, and external auditing. We select and review studies within a research framework, identifying two of the most important areas in the literature: short sellers as important information intermediaries and short sellers’ influence on accounting, auditing, and other corporate decisions as the ‘spillover effect’ of the information distribution. Of the two, the former has a strong emphasis on financial markets, whereas the latter extends this traditional topic in finance to financial economics, corporate governance, and accounting. Our review highlights that, although short sellers use both private information and public information in selecting stocks for shorting, we know little about how they use private and non-financial information to influence managerial economic decisions and firms’ financial reporting decisions. In discussing potential future research, we emphasize that penetrating the information ‘black box’ and positioning research regarding the information that short sellers use and how they use it are necessary to advance the short-selling literature.

Keywords:

• Short selling
• Information intermediaries
• Financial reporting
• Corporate decisions

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Supplemental Data and Research Materials

Supplemental data for this article can be accessed at https://doi.org/10.1080/09638180.2020.1788406.

TABLE A1. Summary of research on short sellers’ information production and the securities that short sellers target 2

TABLE A2. Short selling and real economic decisions 22

TABLE A3. Short selling, financial reporting, and auditing outcomes 23

TABLE A4. Miscellaneous 26

Notes

1 The advantage of systematic reviews lies in a ‘replicable, scientific, and transparent process that enables the researcher to provide an audit trail, justifying his/her conclusions’ (Tranfield et al., 2003, p. 218).

2 Some examples of the costs and risks of short selling include the uptick rule, restrictions on access to proceeds from short sales, unlimited loss with increases in the stock price, legal constraints on short selling by certain institutions, and negative rebate rates (Diamond & Verrecchia, 1987). Engelberg et al. (2018) use variance of lending fees as a proxy for short-selling risk and find that higher short-selling risk is associated with lower future returns, decreased price efficiency, and less short-selling activity by arbitrageurs.

3 Zhao (2020) finds that activist short sellers tend to target opaque firms that experience about three times as negative abnormal returns in both the short and the long term as non-opaque targets. Although this evidence is suggestive of informative shorts, Zhao (2020) also finds that opaque targets experience more dramatic price reversals, particularly when there is a substantial initial drop in targets’ prices immediately after the attacks: evidence of manipulative short selling.

4 We find a high degree of overlap in the journal rankings in the accounting and finance field in the ABDC and the Association of Business Schools (ABS) journal rankings. The ABDC tends to be a more inclusive list, and we decide to follow the ABDC ranking. For a complete list of journals of the ABDC and ABS ranking systems, please see ABDC 2019 Journal Rankings, retrieved from https://abdc.edu.au/research/abdc-journal-list/2019-review/; Last accessed: June, 18, 2020, and ABS 2018 Journal Rankings, retrieved from https://charteredabs.org/academic-journal-guide-2018/; Last accessed: June, 18, 2020.

5 For example, former short-selling expert Kathryn Staley (1996) notes in her book The art of short selling that ‘Short sellers accumulate volumes of disparate facts and observations then they make an intuitive leap based on the information at hand. Frequently, the signs point to large problems that will not be revealed in total until after the collapse.’ In a similar manner, James Chanos, President of Kynikos Associates, in his testimony before the Securities and Exchange Commission (SEC) Roundtable on Hedge Funds, mentions that, before choosing stocks for short sale, his firm (Kynikos Associates) pursues a rigorous financial analysis and focuses on securities issued by companies that seem to have (i) materially overstated earnings; (ii) an unsustainable or operationally flawed business plan; and/or (iii) engaged in outright fraud. https://www.sec.gov/spotlight/hedgefunds/hedge-chanos.htm; Last accessed: June, 18, 2020.

6 Brunnermeier and Oehmke (2014) provide a model that shows that temporarily depressed stock prices from short selling force a financial institution to fire sell long-term assets to repay debt and satisfy the leverage constraint. Such losses allow predatory short sellers to make a profit on their short positions.

7 Short sales were restricted in mainland China until March 31, 2010, when a pilot scheme permitted 90 designated stocks to engage in short selling in China’s A-share markets to improve information efficiency. Short-selling activities have been on the rise since the Chinese Government relaxed the short-sale restriction on stocks in 2010 (Jin et al., 2018). Hong Kong lifted restrictions on short sales after January 1994 and has been identifying eligible stocks for shorting every year through a selection mechanism ever since (Chang et al., 2007).

8 Although Miller (1977) hypothesizes that dispersion of investor opinion in the presence of short-sale constraints leads to stock price overvaluation, empirical tests of Miller’s hypothesis examine the valuation effects of only one of these two necessary conditions. For example, Figlewski (1981) finds some weak evidence that more heavily shorted firms underperform less heavily shorted firms using a limited sample from 1973 to 1979. Although his least shorted firms produce positive abnormal returns with statistical significance, his most shorted deciles do not produce significant negative abnormal returns. However, Boehme et al. (2006) find strong evidence of significant overvaluation for stocks that are subject to both conditions simultaneously. However, Mashruwala and Mashruwala (2018) provide evidence supporting the beneficial effect of accounting conservatism in reducing overvaluation for firms with the highest short-selling constraints and investor disagreement.

9 We acknowledge one of the reviewers for this insightful comment.

10 Using short interest as a proxy for shorting demand, however, is problematic, because the quantity of shorting represents the intersection of supply and demand. The quantity of shorting should respond to both the cost and the benefit of shorting the stock, so stocks that are very costly to short will have low short interest. Impossible-to-short stocks have an infinite shorting cost, yet the level of short interest is zero. Thus, short interest can be negatively correlated with the shorting demand, overpricing, and shorting costs (Lamont, 2012).

11 The benchmark period consists of all the trading days in the short volume dataset that do not fall within an 11-day earnings announcement period window (Christophe et al., 2004).

12 Beneish et al. (2015) consider ‘limits to the supply of lendable shares’ as a key impediment to pricing efficiency. However, Reed (2015) argues that this may not be the case, as the supply of available shares is closely related to other variables, like the demand and price for borrowing, quantity borrowed, and short interest. Therefore, Reed (2015) suggests that ‘supply should not be considered as a completely independent view of the equilibrium relationship among quantity, price, and demand’ (p. 97).

13 Table A4 (see online appendices) summarizes the papers reviewed in Section 2 that are mainly in the areas of short selling regulations and are not included in Table A1-A3.

14 The four information categories include accounting (earnings surprise, accruals, and capital expenditures), valuation (market value of equity, earnings-to-price ratio, book-to-market ratio, and average daily stock turnover), growth (sales growth and forecasted long-term growth), and momentum (earnings forecast revision and price momentum).

15 Drake, Myers, Scholz, et al. (2015) use daily short volume data over short windows around restatement announcements as opposed to the monthly short interest data that Desai et al. (2006) use. The use of short window data ‘increases the likelihood that short-seller activities are directly related to the restatement characteristics rather than to other information about earnings quality’ (pp. 220–221).

16 For example, Moody’s 2002 Special Comment (Fons et al., 2002, p. 5) states: ‘Moody’s will use confidential non-public information that issuers provide to Moody’s only for the purpose of assigning ratings. Moody’s will not, without the permission of the issuer, disclose the information in the press release or other research reports published in connection with the rating, or in discussions between Moody’s analysts and investors, or other issuers … Moody’s believes that the efficiency of capital markets is best served by permitting issuers to disclose to rating agencies material non-public information for use solely in rating decisions.’

17 Firms often engage in confidential conversations with potential investors (e.g. hedge funds) before the public announcement of the private placement of their shares. The Securities Act in the US prohibits insiders who have received private information during these conversations from trading.

18 Geczy et al. (2002) also find that, while average borrowing costs are initially elevated for IPOs, they appear to be insufficient to explain the long-run underperformance of IPOs. Edwards and Hanley (2010) find evidence of active short selling in the first trading week of IPOs, a finding that is inconsistent with the notion that short-sales constraints are binding in the immediate aftermarket. However, Patatoukas et al. (2020) predict and find that the IPOs that are the most susceptible to overpricing (using accounting measures of valuation uncertainty as a proxy) have first-day sizable returns of +44 percent but lockup expiration returns of -10 percent. They also find that these IPOs experience severe short-sales constraints that peak around the lockup expiration. With respect to the informational efficiency of short sellers around seasoned equity offerings (SEOs) announcements, Deshmukh et al. (2017) find that firms with large increases in short interest prior to the SEO announcement exhibit more negative announcement period returns and inferior long-term operating and stock price performance following the equity issue. However, using daily short-selling data around seasoned equity offerings (SEOs), Henry and Koski (2010) find no evidence of informed short selling, which is consistent with manipulative trading.

19 Au et al. (2009) also reveal that heavily shorted firms underperform their lightly shorted counterparts, although the magnitude of this underperformance using equal and value-weighted returns is only 0.59 and 0.19 percent per month, respectively, which is much lower than that reported in the US.

20 However, short selling also provides an external governance mechanism to monitor managers and to facilitate information transparency, a view that Jin et al. (2018) test and confirm using data from China.

21 The choice of selecting post-announcement period short-selling activity is premised on the empirical findings that short sellers react strongly to pro forma earnings announcements (Christensen et al., 2014). Such evidence implies that short sellers could be more reactive than proactive and hence might exhibit abnormal shorting activity during the post-announcement period, particularly for unfavorable announcements.

22 In a Wall Street Journal article, Zweig (2018) reports that, over the nine years before 2018, short funds lost more than 90 percent of their money.

23 We are grateful to an anonymous reviewer for drawing our attention to this strand of the short-selling literature.

24 Although most empirical studies on short selling suggest that short sellers short overvalued stocks because they are well informed, a few studies find that arbitrage needs motivate short selling rather than only exploiting the gain by shorting overvalued stocks. For instance, Liu and Wu (2014) find no abnormal short selling in stock acquirers ahead of a business merger announcement, indicating that short sellers are not informed about upcoming mergers. Their evidence also shows that merger arbitrage needs mainly motivate the increased short selling occurring on announcements of stock mergers and that the arbitrage price pressure that short selling generates contributes to the observed negative returns to stock acquirers around merger announcements.

25 However, Bergsma and Tayal (2019) provide evidence on short interest-related mispricing—overpricing (underpricing) in high (low) short-interest stocks. They find that the lottery stocks with a greater proportion of unsophisticated retail traders have a higher arbitrage risk. The high risk results in costly arbitrage, so overvalued (undervalued) stocks exhibit more negative (positive) alphas.

26 However, Kolasinksi et al. (2013) report that, during the 2008 short-selling ban period in the US, short-selling activity became more informative. Beber and Pagano (2013,, pp. 347–348) nevertheless reconciled this contradictory evidence by noting the following, ‘in the presence of a partial short-selling ban, banned stocks may feature slower price discovery … yet their price may become more sensitive to the short sales that investors are allowed to carry out on other stocks—especially if the ban is accompanied by increased disclosure of short sales, as indeed was the case in the United States during the crisis.’

27 Clinch et al. (2019) find that the inconsistent results are attributable to the inclusion of prior forecasting behavior as a control variable (p. 13), a variable that Li and Zhang (2015) do not include.

28 It is possible to submit exchange orders for immediate execution even if the execution prices fall outside the prevailing National Best Bid and Offer (NBBO). In contrast, dark pools are necessary to match orders at prices within the NBBO. Although this matching can reduce the bid–ask spread costs, if no match is available within the NBBO, then no trade takes place (Reed et al., 2019).