Selling Short and Short Interest (How Smart Are the Bears?)
AEM 428 Manny Dong
Materials from Charles Lee (2003) & Bhaskaran Swaminathan (2005)
SEC's Action on Financial Crisis
On September 18, 2008, the SEC imposed a ban until Sep. 30, 2008, on short sales of hundreds of financial stocks. On Oct 1, 2008 SEC Extended ban to Oct. 17, 2008. Supporter: Prevent the market from declining Opponent: Cause market inefficiency How do you conduct a research to address the concerns? (797 banned stocks)
09/22/09
A ban on short sales of financial stocks comes at a high price. Working paper by Zhang et al. 2008
Did the ban accomplish the SEC's goals? What were the unintended consequences? Targeted firms (797) Control firms Short Reduced Price 12.5%(d0) 10.9%(d1) 5%(d0&d1) Spread 77bp 32bp Volatility 10.6 7.2
Ban did achieve basic goals Ban distorted markets for targeted firms, degrading the market quality of stocks and reducing liquidity.
09/22/09
Review: Short Selling
What are long sellers and short sellers in common?
09/22/09
Review: Short Selling
Benefits (Seller):
09/22/09
Review: Short Selling
Risks and restrictions (seller):
09/22/09
Review: Short Selling
Benefits (Lender):
09/22/09
Review: Short Selling
Risks and restrictions (lender):
09/22/09
Part II. Research Roundup
09/22/09
Research Questions
Market Statistics What do we know about short-interest in aggregate? How does the market for borrowing work? How Smart are the Bears? The information content of short interest How much shorting is "normal"? Does short interest predict future returns? If so, what measures seem to work best. => Dechow et al. (2001); Desai (2002); Asquith-Meulbroek (1995): Execution Cost Issues What do we know about short trades? Frequency and execution costs. Short-horizon profitability of short sales. => Angel (2002) Is short interest information really exploitable? Are profits from Dechow et al. (2001) and Desai (2002) illusionary? The role of short-sell constraints 09/22/09 => Boehme et al. (2002)
Market Statistics
Where can we find information about shortCaveats: Only interest?
Prime
as of 15 of
Brokers each month; But difficult to get overall picture on every reported only if more than stock 250,000 shares WSJ and Barron's are shorted Has up-to-minute information on availability NYSE Stocks (reported around 20th to 23rd of each month) NASDAQ Stocks (reported around 25th to 28th of each month) NASDAQ Trading Data Each trade initiated by a short-seller is marked Jim Angel (FAJ, 2002) NYSE/AMEX/NASDAQ http://Shortsqueez.com 09/22/09
Market Statistics
What type of information is reported?
Short
Positions: Number of Shares Shorted Short Interest: As a Percentage of Shares Outstanding Short Ratio: As a Percentage of Average Daily Volume (typically over past month)
(NYSE Short Interest Data Sample)
3/28/2008
Symbol A AA AAI AAP AAR AAV AB ABA ABB ABC ABD St ock_Descr ipt ion AGI L ENT TECH I NC AL COA I NC. AI RTRAN H OL DI NGS,I NC ADVANCE AUTO PARTS AM R CRP 7.875% Advant age Ener gy I ncome Fund AllianceBer nst ein H olding L .P. Alabama Power Company SR-J ABB L TD. AM ERI SOURCEBERGEN CO ACCO Br ands Cor por at ion St ock_Split Pr ev_Rev Avg_Daily_Vol 2,745,227 15,706,023 2,450,384 953,466 21,285 609,798 617,604 11,653 5,295,839 1,806,169 380,348 Cur r _Shor t _Pos 5,918,400 24,840,945 19,042,513 2,113,804 13,664 371,544 544,205 17,220 3,388,435 3,148,715 6,549,662 Pr ev_Shor t _Pos 4,311,309 21,273,462 19,286,638 1,925,883 13,445 319,411 466,290 21,996 1,694,800 5,447,106 6,366,144
Market Statistics
Short Interest on the rise
Source: Dechow et al. (2001)
09/22/09
Market Statistics
On the other hand, the typical firm has very little short interest
Source: Dechow et al. (2001)
09/22/09
The Market for Borrowing
D'Avolio (JFE 2002) Obtained data from a large financial institution Detailed information on loan supply, variation in fees, and the incidence and nature of lender recall April 2000 to September 2001 (18 months) Key Findings: Most stocks are easy to borrow. The value-weighted cost to borrow a sample portfolio is 25 bp per year. 91% of the stocks are "general collateral" and have a VW-mean fee of 17 bp per year. S&P500 stocks are almost always in this group. Only 7% of the loan capacity for U.S. equity is typically utilized. Approximately 9% of the stocks (about 206 per day) have fees of over 1% per year. These "specials" have a mean fee of 4.3% per year. The vast majority of these are in the bottom size decile 09/22/09 and under $5.
The Market for Borrowing
Key Findings (cont'd): Fewer than 1% of the stocks on loan are "extremely special" (roughly seven per month). These demand negative rebate rates (e.g., KKD and PALM), and have loan fees as high as 30 to 50%. "Special" stocks display significant investor disagreement: they have high turnover, high dispersion in analyst forecasts, increased message board activity, and low cash flows. Recalls are rare. In an average month, 2% (61) of the stocks on loan are recalled. Having been recalled, the mean (median) time before the short can be reestablished with the lender is 23 (nine) trading days.
How smart are bears?
In theory, the bears might be smarter because they face greater transaction costs Diamond and Verrecchia (JFE 1987) argues that fewer liquidity traders will sell short because of the costs higher Predict that stocks which have heavy short-interest will experience lower subsequent returns On the other hand, short interest might be the result of other activities that have nothing to do with bearish sentiments Event arbitrage during merger situations
09/22/09
How smart are bears?
Also, short sellers face a myriad trading constraints
Some
institutions prohibited from taking short positions. Availability of shares for borrowing; Access to proceeds; Exposure to dividends; Higher trading costs (NYSE Uptick rule).
How smart are bears?
What do we know?
Earlier evidence suggests that the average bear is pretty much average
Figlewski (1981), Conrad (1986), Vu and Caster (1987), Brent, Morse and Stice (1990), Choie and Hwang (1994), Senchak and Starks (1993), Woolridge and Dickinson (1994) Asquith and Meulbroek (1995): Focus on stocks with at least 2 % of their outstanding shares shorted. Show that, for NYSE/AMEX firms, high short interest is bad news for the stock (incremental to a myriad of risk factors). Desai et al. (2002): Using more recent data and more powerful tests, show the same for Nasdaq stocks. 09/22/09
But more recent evidence suggest D-V might be right after all
=> The Bears look to be pretty smart!
What else do we know?
Dechow et al. (2001): Bears are Value Investors
They target low fundamental-to-price stocks
09/22/09
What else do we know?
Dechow et al. (2001): they are pretty good at it
They seem to be able to pick out the ones that are likely to do worse.
09/22/09
Desai et al. (JF 2002)
Out-of-sample tests of Asquith-Meulbroek (1995):
Nasdaq
stocks Obtain monthly short interest data from Nasdaq June 1988 to December 1994 Approximately twice as many observations
Desai et al. (JF 2002)
Calendar-time portfolio approach
Require
firms in the portfolio to meet the threshold level in the preceding month Thresholds: 2.5, 5.0, 7.5 & 10% EN is the entry month (when shortinterest first exceeds cutoff) EX is the exit month (when shortinterest drops below cutoff) One pair of EN-EX results in one observation
Desai et al. (JF 2002)
Key
Findings
Table
2 (and page 7): Characteristics of heavily shorted firms
These firms are typically larger (8th Nasdaq size decile), more liquid (9th dollar volume decile), high turnover (9th decile), low value-to-price (3rd B/M decile; 4th CF-to-price decile).
09/22/09
Desai et al. (JF 2002)
Tables
3 and 4: Heavily shorted firms experience negative subsequent returns
Approximately 1% per month negative excess returns The effect increases monotonically with the amount shorted These firms load positively on SMB and negatively on PR1YR
(they earn positive abnormal returns when small firms do well relative to large firms; they earn negative abnormal returns when momentum stocks do well).
Desai et al. (JF 2002)
Calendar
abnormal returns for heavily shorted firms from month EN=1 to Ex+1
Execution Costs
A Close Look at Short Selling on the Nasdaq Market (Angel et al.,2003, FAJ )
What's New?
All
prior studies have used short interest data (monthly snapshots at a point in time) This study uses individual transaction-level data from the Nasdaq exchange Isolates each trade initiated by a short-seller to examine frequency and market impact issues Sample period: Sept 13, 2000 to December 12, 2000 Sample firms: 1,393 Nasdaq stocks
Execution Costs
Key Findings
Short
sale activities have increased further since the mid1990's (Table 1) On average, one out of every 42 trades is a short sale Shorts tend to focus on Stocks that had the highest contemporaneous returns (Table 2)
Execution Costs
Key Findings
Stocks that had the highest contemporaneous returns (Table 2)
Execution Costs
Key Findings (Continued)
Stocks with higher volume (Table 3) and volatility (Table 4) No day-of-the-week effect (Table 5) High short-selling days precede days of unusually low returns Shorts seem to be able to time the market (Table 6)
But is it exploitable?
Boehme et al. (2002): Short-sale Constraints and Overvaluation Research Focus: Which of the heavily shorted stocks really tank? The effect is focused on stocks with
Shares that are difficult to borrow (smaller size, higher relative short interest percentage, and no exchange-traded options), and Dispersion of investor opinion is high (high recent turnover and high volatility)
Key
Results:
For stocks that are easy to borrow and has low dispersion of opinion, short interest is uncorrelated with future returns
Boehme et al. (2002)
Sample:
NYSE
(Jan. 1988 to July 1999) Nasdaq (Jan. 1993 to July 1999)
=> A larger and more recent data set than prior studies Commentary:
But
none of the results are all that new Basically, the question of exploitability is still an open one
Summary
How smart are the bears?
Large
short interest is bad news for a firm's future returns (incremental to 3-factor model) Heavily shorted stocks look like late-stage glamour stocks (high V/P ratios, high turnover) But profitability of this strategy is mainly from heavily shorted stocks with high volume and volatility
Summary
Is this signal exploitable?
Not
100% clear from the evidence Much depends on what you assume about availability of shares and execution costs As a simple heuristic, if you can short an already heavily shorted stock (and you can stomach the short-run volatility), you probably should
Good luck to your investment!