Transcription
Can Short Sellers Detect the Risk of Corporate Governance Failure?ABSTRACT: Using the first-time disclosure of material internal control weaknesses (ICW)under Section 404 of the Sarbanes-Oxley Act, as a negative shock to a firm’s governanceenvironment, we establish that short sellers trade on negative governance information. We findnegative abnormal returns, both in the long- and short-term, around severe ICW revelationevents, and correspondingly, a short intensity build-up prior to the events. We do not find asimilar build-up when ICW are less severe. We also show that the short intensity build-up is notexplained by parameters known to be associated with internal control failure. These results areconsistent with short sellers who develop unique expertise in predicting ICW. We further showthat this expertise is most pronounced when the information environment quality of the firm ispoor. This is the first study to demonstrate that short sellers take positions in anticipation ofeventual governance deficiency disclosure.Keywords: short selling; abnormal short intensity; internal control weaknesses; severe internalcontrol problems; section 404 of the Sarbanes-Oxley Act.1
I. IntroductionThe heated debate on the role of short sellers in capital markets, among practitioners,regulators, and academics is ongoing (Jones and Lamont, 2002; Lamont and Stein, 2004).Opponents of short selling claim that short sellers are noise traders who manipulate prices forpersonal gain (Wilchins, 2008). Proponents however, argue that those investors serve asinformation intermediaries who contribute to the price discovery process, and thus, enhancemarket efficiency (Einhorn, 2008). Despite academic evidence consistent with the latter view,that short sellers identify overvalued stocks and help incorporate negative information into stockprices (e.g., Christophe et al., 2004; Diether et al., 2008; Christophe et al., 2010; Karpoff andLou, 2010; Liu et al., 2012), following the financial crisis of 2008, and accusations that shortsellers accelerated the market price downward spiral (Wilchins, 2008; Zarroli, 2008), 1 inSeptember 2008, the US Securities and Exchange Commission (SEC) temporarily banned certainshort-selling activities in financial stocks, and placed new requirements on share borrowing.2 TheSEC is further considering the adoption of a new disclosure policy that will require short sellersto publicly disclose their positions, which can thus potentially limit their impact on securityprices.3,4While evidence on the ability of short sellers to anticipate certain important eventsalready exists (e.g., Daske et al., 2005; Boehmer et al., 2008; Boehmer and Wu, 2013), most ofthese studies focus on short selling prior to events with immediate adverse effects on earnings,1See the article in the Wall Street Journal on July 24 2008, titled, “What the SEC really did on short selling.”See the article in the Wall Street Journal on September 19 2008, titled, “SEC is set to issue temporary ban againstshort selling.”3See the article in the Securities Technology Monitor on July 20, 2011, titled, “SEC hears good, bad on short-sellingdisclosure.”4The UK, France and Spain already require short sellers to publicly disclose large short positions in all stock listedin those countries (Jones et al., 2011).22
such as restatements and accounting fraud (Dechow et al., 1996; Efendi et al., 2005; Desai et al.,2006; Karpoff and Lou, 2010), negative earnings surprises (Christophe et al., 2004), and assetwrite-downs (Liu et al., 2012). While these studies show that short interest contains viableinformation on firms’ returns and earnings, whether or not short sellers are also interested in,and capable of identifying firms with corporate governance deficiencies is largely unknown.Corporate governance deficiency is of great importance to investors in financial markets. Theliterature has already established that governance characteristics, such as the strength andindependence of the board of directors and of the audit committee, ownership structure, and antitakeover provisions, have an impact on the firm’s performance (Agrawal and Koneber, 1996;Gompers et al., 2003; Dey, 2008), excess CEO pay (see Bebchuk et al., 2002 for acomprehensive review), earnings quality (Becker et al., 1998; Klein, 2002; Xie et al., 2003;Vafeas, 2005), and overinvestment of excess cash (Richardson, 2006). Studies have also shownthat negative corporate governance information is incorporated into firms’ financing cost ex-post(Ashbaugh-Skaife et al., 2009; Kim et al., 2011; Gordon and Wilford, 2012; Cassell et al., 2013).Yet, it is still to be determined whether some sophisticated investors (e.g., short sellers) wouldreact to negative governance information ex-ante and play a beneficial role to other marketparticipants.In this study, we use the first-time reporting of material internal control weaknesses overfinancial reporting (ICW) under Section 404 (SOX 404) of the Sarbanes-Oxley Act (SOX) tocapture a negative shock to a firm’s governance environment, and we study the role of shortsellers in detecting the firm-level risk of corporate governance failure. We are motivated to usethis corporate governance disclosure, given that prior studies (e.g., Hammersley et al., 2008;Ashbaugh-Skaife et al., 2009) show ICW reporting under SOX 404 to be a negative corporate3
governance event. 5 We find that, as sophisticated investors, short sellers trade in anticipation ofthis negative governance event.To carry out our analysis, we first identify firms that disclose ICW under SOX 404 forthe first time. We focus only on first-time ICW reporting, because market participants react morestrongly to unexpected information, and the first-time ICW disclosure is likely to be moreunexpected than a repeated one. 6 We focus on triggering events, rather than on the ICWdisclosure itself, since prior studies (Beneish et al., 2008; Hammersley et al., 2008; AshbaughSkaife et al., 2009) find only a weak market reaction in the short window around the disclosuretime, while Ghosh and Lee (2013) show that most of the information on internal control (IC)problems is incorporated into the stock price even prior to the IC disclosure, with limited newinformation generated around that date. We search for significant corporate events preceding theactual ICW disclosure that may be indicative of the upcoming adverse disclosure, which include:1) CEO and CFO departures for reasons other than retirement; 2) auditors switching from a Big4 to a non-Big-4 audit firm; 3) large restructuring or asset write-downs; 4) a delay in thereporting of either quarterly or annual financial statements (10-K/10-Q filings); and 5) ICWdisclosure under SOX Section 302 (SOX 302).7,8 We find that the likelihood of having at least5SOX is considered the most far-reaching reform in the financial reporting and corporate governance requirementsfor public companies since the 1930s, and SOX 404 is one of the most visible and tangible changes introduced(Donaldson, 2005; Li et al., 2008).6We confirm this assertion, as we find a slightly positive and significant 0.19 percent, 3-day stock returnsurrounding repeated ICW disclosure (untabulated).7SOX 302 also requires an IC disclosure. There are three main differences between SOX 302 and SOX 404. First,the amount of work, and level of testing, are much more comprehensive under SOX 404 (Iliev, 2010), which led tocomplaints by firms regarding compliance costs (AeA, 2005), and to numerous compliance delays. The morerigorous SOX 404 procedure is also evident by the facts that: 1) only 26 percent of first-time ICW reporting firmsunder SOX 404 previously reported control deficiencies under SOX 302; and 2) when in 2007, non-acceleratedfilers moved from reporting only under SOX 302 to reporting also under SOX 404, the percentage of disclosure ofineffective controls jumped from 9.1 percent to 23.7 percent (Kinney and Shepardson, 2011). Second, SOX 404disclosure by accelerated filers (but not by non-accelerated filers) requires auditor attestation, whereas SOX 302does not require an auditor report. The majority of our sample, 55.5 percent, is subject to auditor attestation. Third,disclosure under SOX 302 is done quarterly, while disclosure under SOX 404 is done annually.4
one of the above triggering events during the disclosure year is substantially higher for firms thatdisclose ICW under SOX 404 than for firms reporting adequate controls by the end of the year(69 percent vs. 28 percent, untabulated).9 If none of these events take place during the year, weconsider the actual ICW disclosure under SOX 404 as the triggering event. If multiple eventsoccur, we choose the earliest one as our triggering event. We then examine the short intensity (SI)in the 18-month period leading to the triggering event. Since firm fundamentals are known toaffect short-selling behavior (e.g. Dechow et al., 2001; D’Avolio, 2002; Asquith et al., 2005;Duarte et al., 2006), we also examine the abnormal short intensity (ABSI), which controls forthose firm fundamentals. We study the trading patterns separately for firms with more and lesssevere IC problems, in that it may be more difficult for investors to anticipate the ICW disclosurewhen IC problems are less severe. We distinguish more versus less severe ICW by the count ofweaknesses reported, with a higher count implying more severe problems (Gordon and Wilford,2012). We find a progressive short intensity build-up in the months leading to the triggeringevent, but only for firms with more severe ICW. Specifically, for firms reporting at least threeweaknesses (more severe ICW), we find that SI (ABSI) increases by 0.97 (0.82) percent from3.63 (1.65) percent to 4.56 (2.67) percent, over the 18 months leading to the disclosure, while forfirms reporting one or two weaknesses (less severe ICW), the increase over the same period is0.38 (-0.13) percent from 3.66 (0.91) percent to 4.04 (0.78).10 Multivariate tests that control for8In a similar vein, Karpoff and Lou (2010) search for potential triggering events that lead to the disclosure offinancial misconduct by the SEC.9CEO/CFO turnover can be for good reasons, such as a promotion, or for bad reasons, such as firing. The fact thatturnover is 4.5 times higher for ICW firms (11.7 percent) than for the control sample (2.6 percent) suggests thatmost of the turnover is for negative reasons. Auditor change is 3.9 times higher for our sample (11.7 percent) thanfor the control sample (3.0 percent). A change from a Big-4 to a Non-Big-4 auditor might be due to firm downsizing.We find that average sales change in the year prior to the switch for switching firms is 8.1 percent, and 4.0 percentfor non-switching firms, with the difference being statistically insignificant. The lack of evidence on sales drop rulesout downsizing as the reason for the switch. Overall, we believe that our triggering events are valid constructs.10Anecdotal evidence supports this finding. In an article published in the Financial Times on June 22nd, 2008, titled,“Governance flaws a ‘red flag’ for short selling,” Bill Hwang, chief executive and founder of Tiger Asia Asset5
factors known to be associated with short intensity show a significant increase in SI and ABSIprior to the triggering event, only for firms with more severe ICW. 11 Consistent with thetriggering event conveying negative news to investors, we document average negative andsignificant abnormal returns in the 3 days surrounding those events, and in the 12 months leadingto it.In the second part of our analysis, we apply a two-stage regression approach to studywhether short sellers simply trade on parameters that are known to be associated with IC failure(Ge and McVay, 2005; Ashbaugh-Skaife et al., 2007; Doyle et al., 2007; Ghosh and Lee,2013),12 or trade on unique information. In the first stage, we regress the number of weaknesseson those firm parameters, and generate predicted and unpredicted ICW components. In thesecond stage, we regress the short selling on the two components and find that it is theunpredicted ICW component that is associated with short selling. This finding suggests that shortsellers have superior information and unique capabilities to identify upcoming corporategovernance failure. In the cross-section analysis, we predict that the association between shortselling and the unpredicted ICW component would be more pronounced when the informationenvironment of a firm has a higher degree of asymmetry. Using firm size and monthly returnvolatility to proxy for the degree of information asymmetry (Drake et al., 2011), we find resultsconsistent with this prediction. This finding underscores the valuable information involving firms’corporate governance, which is conveyed in short sellers’ trades.13Management, which has approximately 12 billion invested in short selling positions in Korean, Japanese andChinese equities, explains that companies with serious corporate governance problems are like a big red flag forshort-selling in his company. He further explains: ”Short-sellers look for weakness and when they find it, they act.”11Similarly, Desai et al. (2006) conclude that short sellers identify and target the more egregious restatements in thesample, and Karpoff and Lou (2010) find that short selling increases in the severity of financial misconduct.12In the same spirit, Drake et al. (2011) show that short sellers provide value-relevant information about futurereturns beyond what is provided by analysts’ recommendations, and by eleven other predictive variables.13We caution against concluding that short sellers are able to gather private information about these companies.While this is possible, it is also possible that short sellers develop superior expertise to process publicly available6
We conduct several robustness tests to increase the confidence in our findings. First,throughout the paper, we employ measures of both short selling and abnormal short selling, andwe examine both the level of, and the change in short intensity, to ensure that our results are notmeasure-specific dependent. Second, we construct a placebo test in which we randomly assignan ICW disclosure “event date” within our sample period, and then test whether there is anassociation between short intensity and the severity measure on the pseudo “event date.” We donot find such an association using the random date.This study contributes to the extant literature in three ways. First, our study contributes tothe literature on short sellers’ abilities. While prior studies have shown short sellers’ ability toanticipate adverse earnings-related events, such as upcoming restatements and negative earningsannouncements, the prediction of these events may be due to a mere in-depth financial statementanalysis. 14 , 15 The ability to anticipate a major corporate governance failure in the form ofmaterial weaknesses over internal controls is likely to require more than a thorough examinationof accounting figures. In other words, short sellers may trade on intangible information (e.g.,managerial abilities or reputations) that are not embedded in the accounting and/or firmcharacteristics related to governance deficiency. Second, we contribute to the corporategovernance literature by demonstrating that short sellers, a group of sophisticated investors, aresensitive to firms’ IC effectiveness, and trade on governance failure information in advance ofthe IC report. This finding is new to the literature, and reinforces the importance of governancequality to investors (Ashbaugh-Skaife et al., 2009; Kim et al., 2011; Singer and You, 2011;information (Engelberg et al., 2012) that is yet unknown to be associated with ICW, in order to identify companieswith deterioration in their IC environment.14For example, Jim Chanos, a well-known short seller, was able to anticipate the upcoming troubles at Enron bynoticing huge broadband investments Enron made at the time many broadband companies were failing. In addition,he was unable to understand how the company was very profitable in the years prior to its failure (Tauli, 2004).15We further isolate the corporate governance failure effect by removing from the sample firms that restated theirfinancial statements within one year surrounding the triggering events.7
Gordon and Wilford, 2012; Cassell et al., 2013), to the point that they are willing to spend thetime to detect governance flaws. Moreover, our evidence also suggests that short selling can beused as a predictor for firms’ upcoming IC problems, echoing Jensen (2005)’s suggestion thatboard members could better monitor firms’ corporate governance systems by consulting withshort sellers. Third, we identify a setting in which short sellers play a beneficial role in securityprice setting, by trading on negative governance information, thereby pointing out theimportance of short selling as an integral and active part of an efficient capital market. Byhighlighting short sellers’ abilities and their contribution to price informativeness, we inform theongoing debate on the role of short sellers in the capital market. The remainder of the paper isorganized as follows. The following section surveys the related literature and develops testablehypotheses. Section three describes the sample construction, the measures of short intensity, andprovides descriptive statistics. Section four reports the main empirical results, and Section fivepresents robustness tests. Section six concludes the study.II. Literature Review and Hypothesis DevelopmentLiterature on Short SellingWhether short-selling activities are beneficial or detrimental to financial markets hasbeen an ongoing controversy (e.g., Jones and Lamont, 2002; Lamont, 2004; Wilchins, 2008).The recent financial crisis directed heavy criticism toward short sellers and accusations ofcreating a cascade of selling activities, subsequently creating market overreaction to bad newsand deviation of firm prices from fundamental values (Wilchins, 2008). These accusations wereechoed by the SEC Chairman Christopher Cox, who claimed that short selling contributed tolarge share price declines at financial institutions, such as Lehman Brothers, Bear Stearns, Fannie8
Mae and Freddie Mac.16 Consequently, in September 2008, the SEC issued a temporary ban on“naked” short sales in 799 financial stocks, in order to prevent traders from profiting from furtherprice declines.17 Similarly, many countries around the globe, such as the UK and Japan, imposeda ban on short selling for "as long as it takes" to stabilize the financial markets (Acharya andRichardson, 2009). Beber and Pagano (2013) report that out of 30 European and developed nonEuropean countries, only ten countries did not impose a ban on short selling during the periodspanning from January 1, 2008 to June 23, 2009. More recently, in May 2011, the SEC startedconsidering the implementation of a more transparent short-selling disclosure policy.18In contrast to the regulatory move toward limiting the scope of short selling, academicstudies in general take a positive view of short-selling activities, finding that short sellers identifyovervalued stocks, and thus, can effectively signal to market participants upcoming unfavourablenews in a timely manner (e.g., Desai et al. 2002; Christophe et al., 2004; Diether et al., 2008;Christophe et al., 2010; Karpoff and Lou, 2010; Liu et al., 2012).19,20 For example, Dechow et al.(2001) find that short sellers target firms that are overpriced relative to their fundamentals.21Drake et al. (2011) find that short sellers exploit a wide range of information from financialstatements, press releases, and other public available channels to establish their short positions.Engelberg et al. (2012) find that a substantial portion of short sellers’ trading advantage is16See the article on Bloomberg.com, July 15, 2008, titled, "SEC to limit short sales of Fannie, Freddie, brokers."Naked short selling is the practice of short selling a tradable asset without first borrowing the security or ensuringthat the security can be borrowed.18A recent paper by Beber and Pagano (2013) finds that bans were detrimental for liquidity, slowed the pricediscovery process, especially in bear markets, and failed to support prices, except possibly for U.S. financial stocks.In addition, Saffi and Sigurdsson (2011) find that relaxing short-sales constraints is not associated with an increasein either price instability or the occurrence of extreme negative returns.19Based on Fama (1991)’s definition of an efficient capital market, it is impossible to predict future performancefrom trading, based on public data. Therefore, in an efficient market, rational investors who are willing to short sell,despite the high cost of short selling, must possess information to trade on.20One exception is Henry and Koski (2010), who find manipulative short selling, immediately prior to SecondaryEquity Offering (SEO) issue dates. However, in non-SEO periods, they still find an informative role for short selling.21As anecdotal evidence, Cassell et al. (2011) show that Enron’s short intensity ratio increased from 0.5 percent inAugust 1999 to 2.35 percent in October 2001.179
derived from their ability to more effectively analyze publicly available information. Combiningdata on short sales and news releases, the authors show that the well-documented negativerelation between short sales and future returns is twice as large on news days and four times aslarge on negative news days, as on non-news days. Studies also focus on short sellers’ abilities toidentify firms with accounting manipulation or corporate fraud. It has been shown that shortsellers accumulate positions prior to earnings restatements (Efendi et al., 2005; Desai et al., 2006)that lead to allegations of securities fraud (Griffin, 2003; Efendi and Swanson, 2009), or to SECenforcement actions for financial misrepresentation (Karpoff and Lou, 2010). Furthermore, shortsellers can convey benefits to a wide range of market participants. Pownall and Simko (2005)conclude that short sellers can serve as information intermediaries when analyst coverage of afirm is low. In a study on agency costs of overvalued equity, Jensen (2005, p.16) suggests:“Compensation and audit committees might well discover important information about failingsin their company’s strategy and/or management team by communicating with short sellers whohave bet on future declines in the price of the company’s stock.” Indeed, a recent study byCassell et al. (2011) finds evidence that auditors take information from short sellers’ activitiesinto account when setting audit fees. This may suggest that short sellers provide incrementalinformation about misreporting risk, which is orthogonal to information known to auditors. Moredirectly related to corporate governance and informed trading, Ferreira and Laux (2007) showthat openness to the market for corporate control leads to more informative stock prices, as itencourages the collection of, and trading based on private information. However, little is knownas to whether investors trade on the effectiveness of corporate governance in general, and if shortsellers target ex-ante firms with ICW, in particular.Literature on Internal Control over Financial Reporting10
Internal controls over financial reporting (ICFR) is defined by the SEC as “a processdesigned by, or under the supervision of, the registrant's principal executive and principalfinancial officers, or persons performing similar functions, and effected by the registrant's boardof directors, management and other personnel, to provide reasonable assurance regarding thereliability of financial reporting and the preparation of financial statements for external purposesin accordance with generally accepted accounting principles.” 22 This process includes themaintenance of records that accurately and fairly reflect the transactions and dispositions of thecompany’s assets. It should assure the prevention or timely detection of unauthorizedacquisitions and the use or deposition of assets that could have a material effect on the financialstatements. An internal control deficiency exists when the design of the control, or its operation,does not allow the prevention or detection of misstatements on a timely basis (AS2, Paragraph 8).A significant deficiency in ICFR can lead to unintentional estimation errors due to a lack ofadequate policies, training of personnel, or inconsistent application of these policies, which cancause transactions to be recorded incorrectly. ICW can also lead to intentional misstatementsthrough misrepresentations and omissions by employees or managers, which are not prevented ordetected due to inappropriate controls. Both intentional and unintentional misstatements maycause financial statements to be less accurate, more biased (in the case of intentionalmisrepresentations), less reliable, and less relevant (see Ashbaugh-Skaife et al., 2009, Section 2for a comprehensive discussion of ICFR and internal control deficiencies). Tighter internalcontrols can significantly lower the likelihood of such misstatements (AS2; Donaldson, 2005).Prior to the massive accounting scandals of the early 2000s (Enron, WorldCom andGlobal Crossing, among others), limited attention was paid to ICFR. This has dramatically22SEC release No. 33-8238 (Jun. 5, 2003).11
changed with the enactment of SOX, which places great importance on ICFR; in particular,Section 404 of the Act. SOX 404 requires that both the management and the external auditorannually assess the IC structure and procedures, and then issue a report of their conclusions onthe IC effectiveness, which is included in the annual financial statements.23 Since ICW increasesthe risk of financial misreporting, ICW disclosure is viewed negatively by the market (asconfirmed in our tests), and as such, provides short sellers with incentives to gather informationthat will help them detect upcoming failures. This, in turn, may help convey this negativeinformation to the capital market in a timely fashion and may lower subsequent losses borne byinvestors when this information becomes publicly known.Several studies have investigated the determinants and outcomes of ICW disclosureunder SOX. Ge and McVay (2005) find that firms with ICW tend to be more complex, smaller,and less profitable than firms with effective IC. Doyle et al. (2007) further find that ICW firmsare younger, rapidly growing, and undergo restructuring, whereas Ashbaugh-Skaife et al. (2007)find that those firms also have fewer available resources to invest in IC, and thus have a greateraccounting risk. In terms of the consequences of a weak IC system, ICW firms pay higher auditfees (Hogan and Wilkins, 2006), experience negative market returns in the short windowsurrounding the disclosure, though somewhat mute (Hammersley et al. 2008), suffer a higher23SOX 404 initially became effective only for accelerated filers, for fiscal years ending after November 15, 2004.An accelerated filer is a company with a worldwide market value of common equity held by non-affiliated members(“float”) of at least 75 million, which has filed at least one annual report under Section 13(a) or 15(d) of theExchange Act, and is not eligible to file quarterly or annual reports as a “small business” (Forms 10-QSB or 10KSB). After numerous delays, non-accelerated filers (companies with a float of less than 75 million) have alsostarted complying with SOX 404 for the fiscal years ending after December 15, 2007 for management assessmentonly. The requirement for auditor attestation with respect to non-accelerated filers was further delayed numeroustimes, until it was finally waived by the Dodd–Frank Wall Street Reform and Consumer Protection Act (Section 989G) on July 29, 2010. Foreign filers were also subject to a gradual size-based implementation schedule.12
yield spread in the bond market (Kim et al., 2011), and have a higher cost of equity capital(Ashbaugh-Skaife et al., 2009; Gordon and Wilford, 2012).24Hypothesis DevelopmentPrior studies have shown that corporate governance failure in the form of ICW disclosureis a negative news event associated with both short- (a negative event return) and long-term (anincrease in the cost of equity) consequences (e.g., Beneish et al., 2008; Hammersley et al., 2008;Ashbaugh-Skaife et al., 2009). Thus, short sellers can profit from trading on firms with ICW.Prior research has also provided evidence that short sellers are sophisticated investors (e.g.,Drake et al., 2011; Engelberg et al., 2012) who are capable of predicting and trading in advanceof certain news events (e.g., Dechow et al., 2001; Efendi et al., 2005; Desai et al., 2006). Thisfinding suggests that it could be of interest for short sellers to collect information on firms’ ICquality, prior to IC disclosure under SOX 404, and trade on it accordingly. However, to the bestof our knowledge, no study to-date has examined whether or not short sellers have the interest orthe ability to ex-ante detect governance deficiencies. We predict that short sellers will collectinformation on the IC quality of firms in order to trade on this information. It is worth noting thatnot all ICW disclosures under SOX 404 are equally informative, as it is very likely that first-timeICW disclosure would be informative and perceived as a negative disclosure event, whilerepeated ICW disclosure would contain less new information (as we indeed confirm for oursample and discuss in the introduction section). Thus, short sellers are likely to be interested onlyin identifying the first-time ICW. This leads us to hypothesize that there will be a progressiveaccumulation of short positions in the period leading to first-time ICW disclosure under SOX404. Our first hypothesis, thus, stated in the alternative form, is as follows:24Ogneva et al. (2
In this study, we use the first-time reporting of material internal control weaknesses over financial reporting (ICW) under Section 404 (SOX 404) of the Sarbanes-Oxley Act (SOX) to capture a negative shock to a firm's governance environment, and we study the role of short sellers in detecting the firm-level risk of corporate governance failure.
