Revealing shorts: an examination of large short position disclosures. Charles M. Jones (Columbia) Adam V. Reed (UNC) William Waller (UNC)

Transcription

1 Revealing shorts: an examination of large short position disclosures Charles M. Jones (Columbia) Adam V. Reed (UNC) William Waller (UNC)

2 TIME Are Short Sellers to Blame for the Financial Crisis? By BILL SAPORITO Thursday, Sep. 18, 2008 "It was sad to see Merrill go down as well," said the voice from inside Lehman Brothers this week as he pondered his own future. "But at least they screwed the shorts. That was good to see." It was also, at least in the minds of many angry investment bank CEOs, a long time coming. In the months leading up to the current market chaos, the short sellers have been on the prowl. But now the witch hunt has begun. The shorts nailed Lehman and Bear Stearns by betting that their shares would continue to fall. And now they have Morgan Stanley and Goldman Sachs in their sights, sparking speculation that the last two remaining go-it-alone investment banking giants may have to find a deep-pocketed commercial bank to partner up with. "What's happening out there? It's very clear to me we're in the midst of a market controlled by fear and rumors, and short sellers are driving our stock down," fumed John Mack, CEO of Morgan Stanley, in a memo to employees. "You should know that the Management Committee and I are taking every step possible to stop this irresponsible action in the market. We have talked to Secretary [Hank] Paulson and the Treasury. We have talked to Chairman [Chris] Cox and the SEC." Cox is listening, and is reportedly proposing a temporary ban on short selling, subject to approval by the SEC's commissioners. If short sellers could be rounded up and roasted as heretics to the true bull market religion, there'd be a rush of people from Lehman and Merrill fighting to add wood to the fire. And Mack would bring the gasoline. Also on Thursday Britain said it will ban all short selling of financial stocks until at least next January, while New York Attorney General Andrew Cuomo announced that he was launching an investigation into complaints of short sellers spreading false rumors about targeted companies like Lehman Brothers, AIG, Goldman Sachs and Morgan Stanley. Some recent accusations It's very clear to me short sellers are driving our stock down -John Mack, CEO of Morgan Stanley New York Attorney General Andrew Cuomo launching an investigation into complaints of short sellers spreading false rumors about targeted companies like Lehman Brothers, AIG, Goldman Sachs and Morgan Stanley.

3 SEC: A Balancing Act Under normal market conditions, short- selling contributes to price efficiency and adds liquidity to the markets. At present, it appears that unbridled short- selling is contributing to the recent, sudden price declines in the securities of financial institutions unrelated to true price valuation. - Christopher Cox, chairman of the SEC

4 Short sellers: heroes or villains? Short sellers might be heroes if: They help markets get to the right price by identifying overvalued stocks. They help stabilize share prices.

5 Short sellers: heroes or villains? Short sellers might be heroes if: They help markets get to the right price by identifying overvalued stocks. They help stabilize share prices. Short sellers might be villains if: They drive prices below fundamental value. They destroy good companies. They destabilize stock prices. They make markets less liquid. This argument has been going on for a long time

6 Why adopt disclosure rules? Mainly if worried about abusive shorting If shorts are informed, could argue that prices are beter. Pros: Sunlight might discourage bad behavior Disclosure is a lighter touch than bans and restrictions Cons: Might also discourage fundamental shorts Loss of short seller s intellectual property Follow- on shorting might destabilize prices Disclosures might enable a short squeeze

7 Key question: What is the impact of disclosure regulation? What are stock returns around disclosures? q Is a strategy of mimicking disclosed positions profitable? Do we observe evidence of this type of follow- on behavior? Does the identity of the discloser mater? How do disclosures relate to corporate events? q Rights issues in particular

8 UK disclosure rules A short seller must disclose a short position in excess of 0.25% shares outstanding. Additional disclosures are required at 0.1% increments (0.35%, 0.45%,...). Final disclosure required if position drops below 0.25%. Apply to financial stocks and stocks with rights issues. Disclosures are required by 3:30 p.m. on the business day following the first day on which the position reaches, exceeds or falls below the disclosure threshold

9 Binning

10 Timeline of disclosure regulation

11 France and Spain France Disclosure rules apply French stocks listed on Euronext and Alternext. Sets threshold of 0.50% shares outstanding. Spain Extends to all Spanish exchanges (we focus on the Madrid exchange because of data availability). Sets threshold of 0.20% for filing with CNMV. Only positions above 0.50% are disclosed with identity of discloser. Small positions available only in an aggregated statistic.

12 Timeline of disclosure regulation

13 Timeline of disclosure regulation

14 Timeline of disclosure regulation

15 Sample and data Sample: Stocks listed on exchanges subject to disclosure regimes Analysis focuses on post- ban periods Disclosed firms and control firms Data: Disclosure database (from Data Explorers) Shorting statistics (from Data Explorers) Rights issue announcements (from Data Explorers) Return and volume data (from Yahoo! Finance) 13F data (from EDGAR)

16 A sample disclosure

17 A disclosure example

18 Summary stats

19 Most Active Short Sellers

20 Many Spanish have big shorts

21 Returns around disclosure Event Window CAR Abnormal Return Std Err Panel A: Full Sample (-30,-3) (-3,-1) (0,1) (0,2) (0,5) (0,10) (0,20) (0,30) (0,60) (0,90)

22 Simple Portfolio Strategy Long one share of the disclosed stock Short one share of the stock s one- digit Industry Classification Benchmark (ICB) sector index Hold for each date in the (0,30)- day event window Return set to zero if no stocks meet criteria on a given day

23 Monthly portfolio returns Percent Excess Returns

24 Portfolio returns in calendar time Panel B: Monthly Returns in Calendar Time 20% 10% Monthly Return 0% -10% -20% Start of UK Regime Start of Spanish Regime Start of French Regime -30% 01/31/ /31/ /31/ /31/ /31/ /31/2011

25 Returns around disclosure Event Window CAR Abnormal Return Std Err Panel A: Full Sample (-30,-3) (-3,-1) (0,1) (0,2) (0,5) (0,10) (0,20) (0,30) (0,60) (0,90) Panel B: Rights Issue Subsample (-30,-3) (-3,-1) (0,1) (0,2) (0,5) (0,10) (0,20) (0,30) (0,60) (0,90)

26 No rights issue, no action Panel C: Non-Rights Issue Subsample (-30,-3) (-3,-1) (0,1) (0,2) (0,5) (0,10) (0,20) (0,30) (0,60) (0,90)

27 Follow- ons, changes, & close- outs (0,1)-day Event Window N CAR Abnormal Return Std Err Panel A: Full Sample Initial Disclosed Positions All Disclosed Positions Closed-Outs Upticks Downticks Panel B: Rights Issue Subsample Initial Disclosed Positions All Disclosed Positions Closed-Outs Upticks Downticks Panel C: Non-Rights Issue Subsample Initial Disclosed Positions All Disclosed Positions Closed-Outs Upticks Downticks

28 Most profitable disclosers Discloser Avg. Daily Profits (0,2)-day Window Avg. Daily Profits (0,20)-day Window United Kingdom Disclosers 32 Cap Smaller Companies Europe Morton Holdings Inc BGI UK 32 Capital Fund RH Moore, LP Wellington Management Blue Ridge Capital Societe Generale SA Marshall Wace LLP Gartmore Investment Management Ltd Odey Asset Management LLP p-value for F-Test of Joint Significance of Fixed Effects

29 Matched sample Control firms are matched on: Country Market capitalization Turnover Short interest / shares outstanding Change in short interest over past month

30 Shorting activity around disclosure Window Short Interest Percent of Lenders Active Scaled Number of Open Loans Concentration of Loans Daily Cost of Borrowing Score Panel A: Full Sample (-3,-1) (-1,0) (-1,1) (0,2) (0,5) (0,10) (0,20) Panel B: Rights Issue Subsample (-3,-1) (-1,0) (-1,1) (0,2) (0,5) (0,10) (0,20) Panel C: Non-Rights Issue Subsample (-3,-1) (-1,0) (-1,1) (0,2) (0,5) (0,10) (0,20)

31 Rights issues Causality directions: Shorting could affect returns Returns could affect the level of shorting Third variable: corporate event causes returns and shorting SEOs and rights issues: Well documented negative announcement effect (e.g. Slovin, Sushka and Lai, 2000). Henry and Koski (2010) claim there is price pressure from shorting in U.S. SEO s.

32 Rights issue announcements Specification 1 Specification 2 Specification 3 Estimate Std Err Estimate Std Err Estimate Std Err Panel A: (0,5) CAR Intercept Disclosed # of Follow-on Disclosures Ratio of Rights to Total Shares Discount to Share Price Panel B: (0,20) CAR Intercept Disclosed # of Follow-on Disclosures Ratio of Rights to Total Shares Discount to Share Price

33 No reversal post- completion: Looks like info, not manipulation Disclosed Undisclosed Mean Std Err Mean Std Err Difference in Means (Completion, 5) CAR (Completion, 20) CAR (Completion, 60) CAR

34 Short interest and rights announcements

35 Diff- in- diff: rights issues with disclosures vs. without Window Short Interest Percent of Lenders Active Scaled Number of Open Loans Concentration of Open Loans Daily Cost of Borrowing Score (-3,-1) (-1,0) (-1,1) (0,2) (0,5) (0,10) (0,20)

36 What predicts disclosures? Logit model of the disclosure of a short position. The sample includes stock- day observations for all firms. Dependent variable q binary variable equal to one if a short position in the stock was disclosed on day t and equal to zero otherwise. Disclosure variable q binary variable equal to one if a short position in the stock was disclosed in the last week or month.

37 Predictive variables Natural candidates q q q Past firm returns Firm market capitalization Firm trading volume Prior short position disclosures Interaction of prior disclosure with discloser or position characteristics q 13F assets under management q Position size q Discloser headquartered in London or New York q Discloser centrality

38 Logit model results (Table 12) Specification 1 Specification 2 Specification 3 Specification 4 Specification 5 Marginal Marginal Marginal Marginal Marginal Estimate Estimate Estimate Estimate Estimate Effect Effect Effect Effect Effect Panel A: Full Sample (Baseline Probability = ) Disclosed {t 1,t 5} Disclosed {t 6,t 30} Disclosed {t 1,t 5} AUM Disclosed {t 6,t 30} AUM Disclosed {t 1,t 5} PositionSize Disclosed {t 6,t 30} PositionSize Disclosed {t 1,t 5} MoneyCtr Disclosed {t 6,t 30} MoneyCtr Disclosed {t 1,t 5} Centrality Disclosed {t 6,t 30} Centrality

39 Follow- on disclosers are nearby (Table 13)

40 Conclusions Empirical facts are clear: Strong evidence of follow- on shorting Discloser reputation and location mater But, no evidence of piling on in smaller short positions Portfolio returns show significant effect But, litle effect on returns outside of rights issues In rights issues: q q Evidence that disclosure is not driving the return, in itself No reversal post- completion, suggesting it s information, not price pressure Overall, no direct evidence of abusive shorting or big return effects