Short Selling in Initial Public Offerings
ByAmy K. Edwards
andKathleen Weiss Hanley
U.S. Securities and Exchange Commission
For Presentation at theUniversity of Waterloo
2
Disclaimer
The U.S. Securities and Exchange Commission disclaims responsibility for any private publication or statement of any SEC employee or Commissioner. This study expresses the authors’ views and does not necessarily reflect those of the Commission, the Commissioners, or other members of the staff.
3
“.. short-selling is impossible during… the first days of
trading.”
-Hanley, Lee and Seguin (1996)
4
Short sale constraints and IPO pricing
Previous studies suggest that short sale constraints can lead to overvaluation and/or underpricing– Miller (1977), Derrien (2006) and Ljungqvist,
Nanda, and Singh
Constraints include:– Limits on underwriter lending shares during
first month of trading: Houge, Loughran, Suchanek and Yan (2001)
– Restricted supply of shares due to lock-ups: Ofek and Richardson (2000)
– Level of rebate rates: Ljundqvist, Nanda and Singh (2006), Geczy, Musto and Reed (2002)
5
MotivationRecent research questions the importance of short sale constraints in IPOs.– Rebate rates: D’Avolio (2002) and Geczy,
Musto and Reed (2002)– Grey markets: Dorn (2003), Ausseness,
Pichler and Stomper (2003), Cornelli Goldreich and Ljungqvist (2006)
Short sale transactions data publicly unavailable prior to January 2005 Examination of short selling at IPO provides natural experiment to examine effect of short sales (or lack thereof)
6
Empirical findings
Short selling occurs early in the aftermarketShort sales are positively related to the change in offer price and level of first day returns– Consistent with overvaluation due to either investor
sentiment or divergence of opinion
No evidence that investors are systematically circumventing constraints or rules on borrowing shares by engaging in “naked” short selling Short selling only marginally related to subsequent price returns Findings not due to market maker activity
7
OutlineShort selling institutional detailsData and summary statisticsEvidence on short sellingDeterminants of short sellingConstraints on short sellingEffect of and profitability of short sellingMarket makersConclusion
8
Timing
T=0– Offer date– First possible short selling date
T+3– First settlement date– IPO closing date – First possible fail to deliver date
9
Mechanics of a short saleShort sellers must borrow stock for delivery on T+3Brokers must locate shares before executing a short saleLocate occurs when the broker, not the investor, determines whether the stock can be borrowed– Market makers exempt from locate if
shorting for their own account
10
Cost of borrowingTo borrow stock, the short seller must post (proceeds of sale) collateral of about 102-105% of the stock valueA short seller is usually also required to post margin– Margin calls can be very costly if the short
seller has limited capital
The lender rebates interest on the collateral to the short seller– Rebate rate= Fed funds rate – stock loan fee– Depends on lending difficulty– Can be negative
11
Data
IPOs from SDC issued from January 1, 2005 through December 31, 2006 .– Final sample is 388 IPOs– No unusual characteristics except one IPO
Estimate market variables using TAQ and CRSPShort selling data– Publicly available transaction-level data from:
Amex, Aracex, Boston, Chicago, NASD, NASDAQ, National, NYSE, Phlx
– Aggregated to daily short selling volume
12
Additional variablesFirst day returns (CRSP)– From offer to open – From open to close– From offer to close
Change in Offer Price = (Pipo-Pmid)/Pmid (SDC)VolumeT+0/shares offered (CRSP)Price supported IPO dummy– IR=0 OR IPO is in the bottom quartile of % OAO
exercised (Bloomberg) OR top quartile for % percent trades, using TAQ, executed at the Pipo on offer day
Float= shares offer/shares outstanding (CRSP)Ability to execute– Percentage of the trading day when the rule allows
short sales to execute (TAQ)NASDAQ dummy
13
Offering statistics
Table 1 Mean Median
Panel A: Offering Statistics
Offer Price $14.82$14.5
0
Offer Amount (in mils) $188.53$114.
23
Change in Offer Price -4.18% 0.00%
Panel B: Offer Day Trading Statistics
First Day Return from Offer Price to Open 9.07% 2.84%
First Day Return from Open to Close 0.62% 0.00%
First Day Return 9.58% 4.17%
14
Does short selling exist?
Table 1 Mean Median
Short SalesT+0 /Shares Offered 7.26% 5.56%
Short SalesT+0/ Trading Volume T+0
12.02% 10.36%
Trading Volume/Shares Offered
58.94% 53.80%
Short SalesT+0/Shares Outstanding
3.02% 1.94%
15
Distribution of first day short sales
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
0% >0-5% >5-10% >10-15% >15-20% >20-25% >25%
First Day Short Selling/Shares Offered
Per
cent of Sam
ple
16
Daily short sales
-1%
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
11%
1 3 5 7 9 11 13 15 17 19 21 23 25 27
Trading Day
Shor
t Sa
les
/Shar
es O
ffer
ed a
nd R
eturn
s
-1%
4%
9%
14%
19%
24%
29%
34%
39%
44%
49%
54%
Vol
um
e/Sh
ares
Off
ered
Short SalesVolumeReturns
17
Intraday short sales
0%
10%
20%
30%
40%
50%
60%
Time Relative to Open in 15 Minute Increments
Per
cent of Tota
l on T
+0
Nasdaq Short Sales/Total Nasdaq Short Sales T+0
Nasdaq Trades/Total Nasdaq Trades T+0
NYSE & Amex Short Sales/Total NYSE & Amex Short Sales T+0
NYSE & Amex Trades/Total NYSE & Amex Trades T+0
18
Short selling in other studies
Diether, Lee and Werner (2006): Short sales comprise– 24% of daily trading volume in NYSE– 32% of daily trading volume in Nasdaq– Daily short selling volume is much
lower than short interest.
Trading volume initially much higher in IPOs than in individual stocks
19
Comparison to other studies
0%
5%
10%
15%
20%
25%
30%
1 3 5 7 9 11 13 15 17 19 21 23 25 27
Day
Shor
t Sa
les/
Vol
um
e
0%
10%
20%
30%
40%
50%
60%
70%
Vol
um
e/Sh
ares
Off
ered
All IPOs
Nasdaq IPOs
Nyse IPOs
Volume/Offer
20
Is short selling related to divergence of opinion?
Miller (1977) argues that
“the prices of new issues… are set not by the appraisal of the typical investor, but by the small minority who think highly enough of the investment merits of the new issue to include it in their portfolio. The divergence of opinion about a new issue [is] greatest when the stock is issued.”
21
Statistics by quartiles of first day return
Table 2 Lowest Q1 Q2 Highest
Number of IPOs 85 109 97 97
First Day Return from Offer Price to Close
-4.78% 1.11% 9.72% 31.54%
Volume T+0/Shares Offered 48.49% 46.56% 60.41% 80.53%
Change in Offer Price -12.55% -13.51% 0.08% 9.38%
Short SalesT+0 /Shares Offered
5.48% 5.20% 7.37% 11.00%
Short SalesT+0/ Trading Volume T+0
11.47% 10.17% 13.50% 13.12%
Cumulative Short SalesT+0 to
T+21
/Shares Offered12.83% 13.27% 18.94% 33.38%
22
Determinants of short salesTable 3: Dep Var=Short sales/Shares offered Model 1 Model 2
Intercept0.056
(4.70)***0.073
(6.41)***
Return from Offer to Open0.188
(8.88)***
Change in Offer Price0.173
(9.90)***
Price Supported IPO0.008(1.17)
0.004(0.55)
Float-0.005(-1.37)
-0.008(-2.30)**
Ability to Execute T+0-0.029
(-1.84)*-0.022(-1.43)
NASDAQ0.022
(3.57)***0.033
(5.41)***
Adj. R2 0.23 0.25
23
Short sale constraints
Ability to execute Uptick Rule and Nasdaq Bid TestGreater ability to execute, lower (not higher) is short selling
Ability to borrow shares– Float
Proxy for supply of lendable shares>Lower the float, higher (not lower) is short
selling
– “Naked” short selling
Cost of borrowing shares– Rebate rates
24
Ability to borrow shares
Settlement and clearing process– From trade date (T) to settlement date (T+3),
the clearing house (NSCC) aggregates buys and sells at clearing broker (not investor or level
Called “Continuous Net Settlement” or CNS– If a net seller does not have enough shares on
account DTC records a “failure to deliver” (FTD) in that clearing broker’s account and assigns a “failure to receive” (FTR) to account of clearing broker who was a net buyer
Fails to deliver as a proxy for “naked” short selling– Daily fails collected from CNS
Only aggregate fails of 10,000 shares or moreBalance variable not flow
25
Summary statistics on fails to deliver
Table 4 Mean Median
Panel A: All IPOs First Settlement Day (T+3)
Fails to Deliver/Shares Offered 4.23% 2.29%
Fails to Deliver/Short SalesT+0 1,083.37% 30.32%
Panel B: 210 IPOs with First Fail on First Settlement Day (T+3)
First Day Return 9.14% 3.67%
Fails to Deliver/Shares Offered 6.92% 5.73%
Fails to Deliver/Short SalesT+0 1,779.50% 99.70%
Panel C: 77 IPOs with First Fail Later Than First Settlement Day (T+4 to T+29)
First Day Return 10.88% 5.44%
First Fails to Deliver/Shares Offered 0.86% 0.46%
First Fails to Deliver/Short SalesT+0 31.47% 7.99%
Day of First Fails to Deliver 6.14 5.00
26
Daily fails to deliver
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
8.00%
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Trading Day
Fails
and S
hort
Sale
s A
s Per
cent of Share
s O
ffer
ed
Fails to Deliver
Short Sales
27
Determinants of fails to deliver (T+3)
Tobit analysisTable 5: Dep Var=FTD/Shares offered Model 1 Model 2
Intercept0.056
(3.40)***0.056
(3.44)***
Short SalesT+0/Shares Offered 0.39
(0.62)
First Day Return 0.029(1.06)
Price Supported IPO0.038
(4.53)***0.040
(4.55)***
Float-0.008(-0.81)
-0.008(-0.85)
Ability to Execute T+0-0.043
(-2.05)** -0.045
(-2.11)**
NASDAQ-0.038
(-4.40)*** -0.037
(-4.41)***
Log Likelihood 169.58 169.95
28
Threshold listLonger lived fails– Max of 10,000 shares or 0.5% of shares
outstanding for five consecutive days
T+7 is first day for inclusion on threshold list– Threshold list information from NYSE,
Amex, and Nasdaq
155 IPOs are on threshold list at some point during the first 30 days with 113 or 29% on list at T+7– Only 2% of non-IPO stocks on list in May
2006
29
Threshold listProbit analysis
Table 5: Dep Var=Dummy Threshold
Model 1 Model 2
Intercept-0. 405(2.16)
-0.322(1.43)
Short SalesT+0/Shares Offered 0.430(0.15)
First Day Return -0.495(0.92)
Price Supported IPO0.571
(15.39)***0.483
(8.58)***
Float0.038(0.20)
0.035(0.17)
Ability to Execute T+0-0.240(0.43)
-0.214(0.34)
NASDAQ-0.483
(10.65)***-0.463
(10.11)***
Log Likelihood 221.11 220.73
30
Could fails to deliver be related to underwriter price support
activities?Only offering amount plus exercise of the overallotment option (OAO) settles at DTCAny shares allocated in excess of the number of shares offered but not covered by the exercise of OAO at time of closing are considered “uncovered”– Must either be purchased in open market or
through OAO– Will result in fail to deliver if investor sells
shares prior to covering by underwriter
31
Cost of borrowing shares
Rebate rate data from anonymous vender– Includes 259 IPOs or 67% of sample– More likely to be in data
Greater is the short selling on first day and over first month IPO listed on NYSE/Amex
Loan fee=Fed funds rate-rebate rateAverage loan fee over first month of trading– Monthly: 0.15%– Annual: 1.88%– Compare to Geczy, et.al. (2002)
First day: 2.95%End of first month: 1.47%
32
Determinants of the cost of borrowing
Table 6: Dep Var=Avg weighted loan fee (T+3 to T+24) Model 3 Model 4 Model 5
Intercept0.014
(2.49)***0.013
(2.73)***0.019
(3.13)***
Short SalesT+0/Shares Offered 0.120(5.59)***
Cumulative Short SalesT+0 to T+21 /
Shares Offered
0.049(10.65)**
*
Fails to Deliver T+3/Shares Offered 0.068(2.14)**
R2 0.10 0.30 .01
Other independent variables not shown and are insignificant (N=259)
33
Short selling and returns
Short selling negatively related to subsequent returns– Diether, Lee and Werner (2007a)– Boehmer, Jones and Zhang (2008)
Use buy-and-hold returns adjusted for Nasdaq Composite IndexAdjust standard errors for clustering by month of IPO
34
Return predictability
Table 8
Panel BOne Month
Return From First Day
Close
Panel CThree Month
Return From First Day
Close
Panel DThree Month
Return With Loan Fees
(N=259)
Model 2Model
3Model 4
Model 5
Model 6 Model 7
Intercept0.083
(3.43)***
0.057(2.99)**
*
0.119(2.51)**
0.087(2.21)**
0.108(2.33)**
0.076(1.87)*
Short SalesT+0 /Shares Offered
-0.215(-1.35)
-0. 500(-1.67)
-0.491-(1.67)
Cumulative Short SalesT+0 to
T+21/
Shares Offered
0.050(1.33)
-0.031(-0.60)
-0.027(-0.69)
Price Supported IPO
-0.079(-5.63)***
-0.070(-4.78)***
-0.099(-3.34)***
-0.091(-3.10)***
-0.107(-2.97)***
-0.099(-2.78)***
Adj. R2 0.09 0.09 0.04 0.03 0.05 0.03
35
Profitability
Variable
Three month Nasdaq adjusted return
Lowest Q2 Q3 Highest
Annual Weighted Loan Fee From T+3 to T+24
2.19% 1.35% 1.88% 2.05%
Month Weighted Loan Fee From T+3 to T+24
0.17% 0.11% 0.15% 0.17%
First Day Return 8.08% 6.92% 13.02% 10.31%
Short SalesT+0/Shares Offered
8.38% 7.06% 7.05% 6.53%
Cumulative Short SalesT+0 to
T+21/
Shares Offered
19.98% 18.25% 20.92% 19.34%
Three Month Nasdaq Adjusted Return
-26.34% -6.71% -7.43% 36.30%
Three Month Profit 23.09% 6.30% -7.78% -37.64%
36
Potential market maker activity
Market makers important in aftermarket trading– Krigman, Shaw and Womack (1999), Ellis,
Michaely and O’Hara (2000) and Ellis (2006)
Exempt from locate requirement and some execution rulesUse “exempt” indicator for trading on Nasdaq for Nasdaq IPOs as proxy for market maker activity– 40% of first day short sales in Nasdaq IPOs
are marked exempt
37
Market maker effect on short selling
Table 9
Panel AShort Selling
ExemptOther Short
Sales
Intercept0.054
(7.03)***0.050
(3.75)***
First Day Return fromOffer to Open
0.059(4.99)***
0.106(5.14)***
Price Supported IPO-0.00002(-0.01)
-0.005(-0.72)
Float-0.001(-0.78)
-0.003(-1.13)
Ability to ExecuteT+0
-0.049(-3.87)***
-0.012(-0.52)
Adj. R2 0.12 0.11
38
Market maker effect on fails to deliver
Table 9
Panel BFails to Deliver
ExemptOther Short
Sales
Intercept0.055
(2.31)** 0.069
(2.99)**
Short SalesT+0/Shares Offered
0.214(1.30)
-0.040(-0.43)
Price Supported IPO0.029
(2.86)*** 0.027
(2.64)***
Float-0.048
(-2.23)** -0.046
(-2.15)**
Ability to ExecuteT+0
-0.081(-2.35)***
-0.090(-2.64)***
Log Likelihood 99.87 99.12
39
Market maker effect on return predictability
Table 9
Panel COne Month Return
From First Day Close
Panel DThree Month Return
From First Day Close
Exempt
Other Short
Sales Exempt
Other Short
Sales
Intercept0.134
(3.42)***0.148
(3.90)***0.234
(3.94)***0.228
(4.43)***
Short SalesT+0 /Shares Offered
-0.257(-0.74)
-0.399(1.97)*
-0. 897(-1.32)
-0.766(-2.17)**
Price Supported IPO-0.079
(-4.17)***-0.084
(-4.88)***-0.098
(-2.74)***-0.1.05
(-3.05)***
Float0.0042(2.08)**
0.003(1.74)*
-0.005(1.30)
0.004(1.00)
Ability to ExecuteT+0-0.178
(-3.11)***-0.170
(-3.35)***-0.263
(-3.66)***-0.226
(-2.72)**
Adj. R2 0.09 0.11 0.04 0.05
40
ConclusionShort selling is prevalent in early trading of IPOsConstraints on short selling do not appear binding– No evidence that short sellers fail to borrow shares
Fails to deliver may be due to underwriter price support
– Loan fees positively related to amount of short selling
Short selling is negatively related to short term returns but do not appear to mitigate underpricingResults not due to market making activityUnlikely that short sale constraints are responsible for high underpricing