Description
The purpose of this paper is to investigate the effects of the 2008 SEC short-sell
moratorium on regional bank risk and return. The paper also examines the decline in “failures to
deliver” securities in the wake of SEC short-sell moratorium.
Journal of Financial Economic Policy
Short-sell moratorium effects on regional bank performance
Michael Devaney William L. Weber
Article information:
To cite this document:
Michael Devaney William L. Weber, (2013),"Short-sell moratorium effects on regional bank performance",
J ournal of Financial Economic Policy, Vol. 5 Iss 2 pp. 92 - 110
Permanent link to this document:http://dx.doi.org/10.1108/17576381311329652
Downloaded on: 24 January 2016, At: 21:46 (PT)
References: this document contains references to 52 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 131 times since 2013*
Users who downloaded this article also downloaded:
J ames R. Barth, Gerard Caprio, Ross Levine, (2013),"Bank regulation and supervision in 180
countries from 1999 to 2011", J ournal of Financial Economic Policy, Vol. 5 Iss 2 pp. 111-219 http://
dx.doi.org/10.1108/17576381311329661
Linyue Li, Nan Zhang, Thomas D. Willett, (2012),"Measuring macroeconomic and financial market
interdependence: a critical survey", J ournal of Financial Economic Policy, Vol. 4 Iss 2 pp. 128-145 http://
dx.doi.org/10.1108/17576381211228989
Sherrill Shaffer, (2012),"External economies in banking", J ournal of Financial Economic Policy, Vol. 4 Iss 4
pp. 354-365http://dx.doi.org/10.1108/17576381211279316
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Short-sell moratorium effects
on regional bank performance
Michael Devaney and William L. Weber
Department of Economics and Finance, Southeast Missouri State University,
Cape Girardeau, Missouri, USA
Abstract
Purpose – The purpose of this paper is to investigate the effects of the 2008 SEC short-sell
moratorium on regional bank risk and return. The paper also examines the decline in “failures to
deliver” securities in the wake of SEC short-sell moratorium.
Design/methodology/approach – In total, six regional bank portfolios are derived and the beta
coef?cients from a CAPM model are estimated using the integrated generalized autoregressive
conditional heteroskedasticity (IGARCH) method accounting for the short-sell moratorium. Data on 110
regional banks in six US regions fromJanuary 2002 to December 30, 2011 are used to estimate the model.
Findings – The ban on naked short selling and the SEC short-sell moratorium signi?cantly increased
individual bank risk for a majority of banks in six geographic regions, but also increased return in
three of three regions. There was also reduced naked short selling as failures to deliver securities
declined sharply after the September 2008 moratorium took effect.
Originality/value – Regional banks have generally not achieved the size needed to be deemed“too big
to fail” bypolicy-makers. Thus, policychanges suchas the SECshort-sell moratoriummight be expected
to have larger effects on regional banks than on larger banks, which might be shielded from the policy
change by having achieved “too big to fail” status. The authors’ results are consistent with research that
has shown that short-sell restrictions increase risk by reducing liquidity and trading volume.
Keywords GARCH, SEC short-sell moratorium, Banks, Risk management, Securities,
United States of America, Economic policy, Securities and Exchange Commission
Paper type Research paper
1. Introduction
For much of the past three decades a central focus of bank regulators has been capital
adequacy. In 1988, central bankers from around the world met in Basel, Switzerland
to recommend minimum capital standards that were enforced by law in the G10
countries and a number of other nations. Known as the Basel Accords or Basel I, the rules
established risk-based capital (RBC) standards that assigned increased capital
requirements for banks with riskier assets. Basel II would follow in 2004, while
Basel III attempted to address de?ciencies in Basel II that were revealed by the
?nancial crisis of 2008. Bank capital as de?ned by regulators in Basel I, II and III derives
largely from accounting de?nitions and departs signi?cantly from the market value of
common equity. It is dif?cult in practice to determine when a bank becomes too big to
fail, but on November 4, Financial Stability Board (2011) issued a release listing 29 banks
that they considered to be systematically important ?nancial institutions whose
“failure, because of their size, complexity and systematic interconnectedness, would
cause signi?cant disruption to the wider ?nancial systemand economic activity.” In the
absence of government guarantees such as “too big to fail” bailouts, market value
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JEL classi?cation – G12, G14, G18, G21
Journal of Financial Economic Policy
Vol. 5 No. 2, 2013
pp. 92-110
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576381311329652
JFEP
5,2
92
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
capital has an advantage over the regulatory de?nition of capital in that private
investors have more incentive to monitor the bank’s actions since in the event of
insolvency, private investors bear its costs.
Although “too big to fail” status truncates insolvency risk, 90 percent of US banks
do not have equity capital that is traded in public markets. The other 10 percent are
exposed to the discipline of markets on a daily basis, yet, regional banks might be too
small to achieve the scale and scope economies of their “too big to fail” counterparts.
The ?nance literature suggests that geographically constrained banks may have fewer
investment opportunities and experience greater risk than those that operate over a
wider geographic region (Deng and Elyasiani, 2008; Hughes et al., 1996; Weber and
Devaney, 1999).
RBC standards assign arbitrary weights to different classes of assets based on their
presumed risk. Because of the low or zero risk weight assigned to real estate and
sovereign debt, Cochrane (2011) and Wallison (2011) argued that RBCcontributed to risk
in the banking system. Similarly, the low risk weights on real estate assets might have
geographically concentrated ?nancial stress in US banking relative to the pre-Basel real
estate bubble of the 1980s. This spatial difference is re?ected in bank failures as reported
by the FDIC. From 2000 to 2007 there were 27 bank failures. Then, in 2008, 25 banks
failed, followed by 140 bank failures in 2009, and 157 failures in 2010 (Federal Deposit
Insurance CorporationFailedBankList (2012a)). From2008 to 2010, Georgia and Florida
had the most bank failures with 51 and 45, followed by Illinois (38), California (34),
Minnesota (15), Washington (14), Missouri (11), Nevada (10), Texas (8), and Arizona (9).
Perhaps a better indicator of the decline in bank services as a result of failures
is the decrease in the number of branches. Figure 1 is a map of US counties which
experienced at least one branch closing as a consequence of bank failures from 2008 to
2010. During the period a total of 1,340 branches of failed banks closed, with 294
occurring in 2008, 678 in 2009, and 368 in 2010. In California, 193 branches closed,
followed by Florida (185) and Georgia (155). Rounding out the top ten states with
branch closings are Texas (90), Illinois (83), Washington (74), Oregon (44), Nebraska (37),
Minnesota (29), and Missouri (28) (Source: Federal Deposit Insurance Corporation
Summary of Deposits (2012b)).
The regulatory de?nition of bank capital may be subject to control by regulators, but
the daily pricing of market equity is widely regarded as being beyond their sphere of
expertise. Nevertheless, in July of 2008, following the demise of Bear Stearns, restrictions
on naked short selling were imposed on a small number of very large ?nancial service
companies that could be characterized as too big to fail institutions. After the collapse of
Lehman Brothers in September of 2008, the SEC short-sell restricted list was expanded
to include approximately 800 ?nancial companies. The regulatory justi?cation for the
restriction was basedon the presumption that it would mitigate “excessive ?uctuation in
the prices” of ?nancial institution securities (US SEC, 2008a, b, 2010).
Attempts to pro?t by shorting beleaguered sectors such as banking have become
more common along with the rise in exchange traded funds. The implementation of
leveraged strategies might have a more pronounced price effect on small market
capitalization sectors that have less trading volume. This effect might be especially felt
toward the end of the trading day when fund managers buy or short stocks to settle their
books. Naked shorting occurs when traders do not actually borrow the shares they sell
short as required by securities law. Naked shorting can result in “failure to deliver”
Short-sell
moratorium
effects
93
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Figure 1.
Countries with branch
of a failed bank, 2008-2010
JFEP
5,2
94
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
which is the number of stocks in which large positions have not been properly delivered
to investors. SEC Regulation SHO was supposed to address the problem of failure to
deliver. Implemented in 2005, the regulation mandated daily compilations by exchanges
of stocks that had experienced at least ?ve consecutive days of delivery failures totaling
at least 10,000 shares and at least one half percent of their outstanding shares. When a
stock was placed on the threshold list, traders were presumably required to close out
failed deliveries by the 13th day after the trade.
In 2006, SEC Commissioner Campos (2006) maintained that Regulation SHOshould be
“targeted at protecting a small universe of thinly-capitalized securities from abusive
trading” a de?nition that would include most regional banks. Regulation SHO was
vigorouslychallengedbya securities industrythat hadgeneratedsigni?cant revenue from
short selling. The daily average number of stocks on the SHOthreshold list grewfrom300
issues in2007 to 400 for the ?rst nine months of 2008. Prior to the short-sell moratoriumthe
threshold list increased to over 500 issues and included many ?nancial stocks. In a 2009
letter posted on the SEC web site, former Commissioner Campos argued, “the majority of
these failures-to-deliver are not the result of honest mistakes or badprocessing rather these
companies are instead targets of illegal and manipulative trading.” (McGinty and
Strasburg, 2009) On September 18, 2008 the SEC implemented a new rule that required
short sellers who had sold shares but not delivered themto the buyer within three days of
the trade, to deliver shares by the start of trading on the fourth day. Following the
new rule and the short-sell moratorium, the daily average of stocks on the threshold list
declined to 100 issues in December of 2008 and 79 issues in March and April of 2009.
Although failure-to-deliver can be attributed to a variety of factors, naked shorting and
the inability to borrowshares was regarded as the most important reason for the late 2008
surge. Bradley et al. (2011) argued that regulator actions in September of 2008 largely
resolved delivery failures in the equity market; however, more recently failures have
increased in debt markets and among ETF shares. They argued that the potential
problems stemming from failures to deliver are much more serious than an increase in
replacement cost risk and maintain that delivery failures create a potentially cumulative
and compounding liquidity risk inthe ?nancial system. If Bradley et al. are correct and the
short-sell moratorium assisted regulators in reducing delivery failures among ?nancial
?rms, it might have prevented problems in the ?nancial system from worsening.
To justify the ban, the SEC asserted that short selling caused “sudden and excessive
?uctuations in the prices of [?nancial] securities.” To investigate these ?uctuations
Ulibarri et al. (2009) incorporated the effects that noise traders have on share prices and
found that the ban dampened volatility persistence in ?ve out of the 21 cases they
studied. In addition, they simulated the effects of an alternative uptick rule and found
that it had “disparate impacts on price stability.” The authors concluded that instead of
a blanket uptick rule on all ?rms, the SEC should instead examine speci?c securities on
a temporary basis, which would limit noise traders and still allow informed trading.
Despite some former regulators who have suggested that in hindsight the short-sell
moratorium was a policy failure, the precise impact of the policy on the risk/return of
smaller regional banks is largely unknown. We test several hypotheses concerning
regional bank risk and return. First, we hypothesize that that the CAPMmodel that does
not allow for a time varying risk component is insuf?cient to model regional bank
returns. Instead, we propose an integrated generalized autoregressive conditional
heteroskedasticity (IGARCH) model. Second, we construct six equally weighted
Short-sell
moratorium
effects
95
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
portfolios comprised of regional banks from the Mid-Atlantic (MA), Midwest (MW),
Northeast (NE), Southeast (SE), Southwest (SW), and Paci?c (PAC) regions and use
the IGARCH model to test for policy induced changes in return and risk. Third,
we re-estimate the IGARCH model for each of 110 individual banks. Then, we use the
estimates of policy induced changes in return/risk to test whether there are systematic
differences of the short-sell ban between different geographic regions. The next section
brie?y examines the literature on geographic risk in banking and constrained short
selling. Section 3 explains the data and models. Section 4 presents the results followed by
a ?nal section describing the policy implications and summarizing our results.
2. Bank geographic diversi?cation and constrained short-selling
In 1994 the Riegle-Neal Act eliminated state restrictions on interstate branch banking,
thus allowing greater risk/return bene?ts from geographic diversi?cation (Rivard and
Thomas, 1997; Deng and Elyasiani, 2008; Hughes et al., 1996). Akhigbe and Whyte
(2003) found a signi?cant decline in total risk and unsystematic risk following the
Riegle-Neal Act, while market risk was unchanged. However, Chong (1991) found that
although interstate branching enhanced pro?tability, it also increased risk. Similarly,
Demsetz and Strahan (1997) reported that bank holding companies offset the risk
reducing bene?ts of diversi?cation by increasing leverage and pursuing riskier and
potentially more pro?table lending opportunities.
The SEC attempt to stabilize the market capital of ?nancial institutions by enforcing a
short selling ban was highly controversial. Short selling has enjoyed a nefarious
reputation in the annals of ?nancial market history. Napoleon described the short seller as
“an enemy of the state” while England outlawed short selling in 1,733 and did not make it
legal again until the middle of the nineteenth century (Surowiecki, 2004). Following the
Great Crash of 1,929 short sellers were denounced on the Senate ?oor as “one of the great
commercial evils of the day.” Jones and Lamont (2002) described the anti-shorting climate
following the Great Crash as “hysterical.” Wearden (2010) contended that short sellers
exacerbatedthe Europeansovereigndebt crisis. However, defenders of the practice argued
that short sellers ?rst alerted investors to problems at Enron and Lehman Brothers and
that short selling results in a more ef?cient market.
The process of short sellingequitysecurities is thought toenhance market ef?ciencyby
incorporating information from pessimistic investors (Miller, 1977; Lamont, 2004b;
Brenner and Subrahmanyam, 2009). If true, banning short sales would exclude the
information those pessimistic investors provide, slow the dissemination of information,
andresult inanupwardbias tostockprices. Banningshort sales might alsocause liquidity
to decline since not only are short sellers banned, but some investors might forego long
positions in stocks if they cannot hedge their positions using short sales, thus reducing
trading volume. Diamond and Verrecchia (1987) argued that consistent with rational
expectations, market participants anticipate short sale constraints when formulating
pricing decisions; rather than mitigate volatility short-sell restrictions cause it to increase
and result in negative abnormal returns. Harrison and Kreps (1978) developed a model
where restrictingshort sales causes the price of the securityto exceedits valuationbyeven
the most optimistic current investors. Lamont and Thaler (2003) provided several case
studies of overvaluations that likely occurred as a consequence of dif?culty in short
selling. In a sample of 300 ?rms, Lamont (2004a) found that attempts to restrict short
selling caused the ?rms’ stocks to under-perform the market. In contrast, the predatory
JFEP
5,2
96
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
tradingmodel of Brunnermeier andPedersen(2005) indicatedthat aggressive short selling
can cause stock price declines greater than those implied by market ef?ciency.
The slow adjustment of prices to negative information results in a left skewness in
the distribution of returns, which should manifest as signi?cant abnormal negative
returns. Bris et al. (2007) compared stock market regulation among different countries
and conclude that prices do appear to incorporate negative information more slowly
in those countries where short selling is either not allowed or not generally practiced.
Reed (2007) found evidence of greater negative price reaction associated with
constrained short selling. If the rising cost of short sales is directly related to greater
price reaction when information ?nally becomes public, then regulations against short
selling will increase, rather than decrease volatility.
A growing body of literature on the 2008 ban on short sales in ?nancial stocks
found evidence of less liquidity since bid-ask spreads increased among banned stocks
and greater market inef?ciency given that stock and option prices decoupled in the
wake of the ban (Autore et al., 2011; Battalio and Schultz, 2011; Boehmer et al., 2011;
Jain et al., 2012; March and Payne, 2012). In this paper we contribute to the ongoing
research on the 2008 short-sell ban by investigating the effects of the ban on regional
bank risk and return using an IGARCH model.
3. Data
To examine the effects of the 2008 short-sell ban we examine 110 regional banks during
the ten year period January 2, 2002-December 31, 2011. This ten year period occurred
during a relatively slow growth phase of the US economy and includes the bubble years
in real estate prices and the subsequent bursting of the real estate price bubble and
ensuing ?nancial crisis. Regional banks have North America Industrial Classi?cation
(NAIC) code of 522110 and are part of the broader Finance and Insurance industry
(NAIC 52). Regional banks are smaller than money center banks and less able to achieve
economies associated with scale and geographic diversi?cation. Figure 1 shows that bank
branchclosings have a spatial dimensionwithclosings concentratedinCalifornia, Florida,
and Georgia. The 110 state banks chosenwere selected if they satis?ed three criteria. First,
the bank had to appear on the SEC short-sell restricted list[1]. Second, the bank was
required to have outstanding short interest in the week preceding the ban. We identify
banks with short interest fromShort-Squeeze (2012), a proprietary data base that provides
short interest for approximately 16,000 publicly traded stocks with outstanding short
positions. Short-Squeeze segments regional banks into six regions: Mid-Atlantic (MA),
Midwest (MW), Northeast (NE), Paci?c (PAC), Southeast (SE), and Southwest (SW). Third,
the bank had to be publicly traded for the ten year period fromJanuary 2002 to December
2011. In total, 110 regional banks met the three criteria and consisted of 21 banks from
the MA region, 19 banks from the MW region, 18 banks from the NE region, 18 banks
from the PAC region, 17 banks from the SE region, and 17 banks from the SW region.
Six regional bank portfolios were constructed. For each regional bank portfolio we
calculated the equally weighted daily excess return for the banks in the particular
region. Daily excess returns were estimated as the difference between the stock
return and the risk free return with the daily return on a 30 day T-bill serving as the
risk-free return. The market index is the daily excess return on the S&P500. The period
January 2, 2002-December 31, 2011 comprises 2,519 trading days which are used to
compute 2,518 days of excess returns.
Short-sell
moratorium
effects
97
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Descriptive statistics for January 2, 2002-December 30, 2011 appear in Panel A of
Table I (n ¼ 2518). Descriptive statistics include mean, standard deviation, skewness,
kurtosis and Jarque-Bera statistics. Banks in all of the regions as well as the market
index exhibited positive excess returns for the period. For illustrative purposes we
break the sample in half. Panel B shows descriptive statistics for the ?rst half period
January 2, 2002-December 29, 2006 while Panel C is fromJanuary 3, 2007 to December 30,
2011. Panel B incorporates the bubble years and shows mean daily returns for each
region approximately double the mean return for the entire sample. Panel C is for the
second half period that incorporates the bursting of the housing price bubble and
subsequent ?nancial crisis. Bank portfolios for the Paci?c, Southeast, and Southwest
regions as well as the market index show negative returns.
Bollerslev (1986) observed that return distributions for high frequency daily data
exhibit non-stationary variance along with leptokurtosis and are inconsistent with the
normal distribution. Sample period returns for each portfolio along with the market
index exhibit leptokurtosis and have Jarque-Bera statistics that reject the normality
assumption. The sample standard deviations for the second half period are more than
twice as high as the ?rst period, and appear to violate the classical linear regression
assumption of constant variance. One method for addressing the problems of
non-normal returns and a non-constant variance in returns is the generalized
autoregressive conditional heteroskedasticity (GARCH) model.
4. Time varying GARCH model
A critical issue in the application of the event method is the increase in stochastic
volatility surrounding the event period. Brown and Warner (1980, 1985) observed
that increases in the variance may cause misspeci?cation of the traditional test
statistics and that the power of the tests can be improved by appropriately modeling
the volatility process.
Brockett et al. (1999) developed an event-study method that assumes a market
model with GARCH effects and time-varying market risk. We use a GARCH model to
determine if the risk/return of regional bank portfolios were adversely impacted by the
short-sell constraint. Unlike traditional constant variance models that implicitly
assume that the event induced variance is the same for all portfolios in the sample,
the GARCH model allows event induced volatility to vary and for each portfolio’s
variance to be stochastic outside the event period. Given the observed increase in
market volatility and the premise that the ban may have had a differential impact on
banks from different regions, this allowance is a particularly important advantage of
GARCH relative to conventional OLS applications.
Time varying models have become an important tool in ?nance and economics.
The ARCH model was ?rst proposed by Engle (1982) and generalized to GARCH by
Bollerslev (1986) and extended to ARCH and GARCH in the mean (GARCH-M) by
Engle et al. (1987). The generalized GARCH model is represented by the following
system of equations:
y
t
¼ ax
t
þ1
t
v
t
¼ d
0
þ
i¼1
X
d
i
1
2
t2i
þ
i¼1
X
b
i
v
t2i
1
t
jV
t21
~
Nð0; v
t
Þ
ð1Þ
JFEP
5,2
98
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
S
&
P
5
0
0
M
i
d
-
A
t
l
a
n
t
i
c
M
i
d
-
W
e
s
t
N
o
r
t
h
e
a
s
t
P
a
c
i
?
c
S
o
u
t
h
e
a
s
t
S
o
u
t
h
w
e
s
t
P
a
n
e
l
A
:
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
s
f
o
r
t
h
e
s
i
x
r
e
g
i
o
n
a
l
b
a
n
k
p
o
r
t
f
o
l
i
o
s
a
M
e
a
n
0
.
0
0
0
0
8
0
.
0
0
0
5
4
0
.
0
0
0
3
0
0
.
0
0
0
3
4
0
.
0
0
0
4
8
0
.
0
0
0
4
4
0
.
0
0
0
4
2
S
D
0
.
0
1
3
8
7
0
.
0
1
3
3
5
0
.
0
1
6
7
5
0
.
0
1
3
4
1
0
.
0
2
0
8
1
0
.
0
1
3
2
8
0
.
0
1
4
0
4
S
k
e
w
n
e
s
s
0
.
0
3
3
3
.
2
7
3
1
.
0
0
7
0
.
1
7
9
1
.
3
1
2
0
.
2
3
8
0
.
4
1
6
K
u
r
t
o
s
i
s
1
1
.
3
3
9
.
4
1
5
.
8
9
.
8
1
8
.
7
9
.
8
1
1
.
4
J
a
r
q
u
e
-
B
e
r
a
7
,
2
6
4
.
4
1
4
3
,
2
2
8
.
0
1
7
,
7
0
6
.
0
4
,
8
6
2
.
3
2
6
,
6
4
2
.
0
4
,
8
5
4
.
9
7
,
5
6
0
.
7
P
a
n
e
l
B
:
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
s
f
o
r
t
h
e
s
i
x
r
e
g
i
o
n
a
l
b
a
n
k
p
o
r
t
f
o
l
i
o
s
b
M
e
a
n
0
.
0
0
0
1
5
0
.
0
0
0
7
0
0
.
0
0
0
4
3
0
.
0
0
0
5
6
0
.
0
0
0
9
2
0
.
0
0
0
8
0
0
.
0
0
0
6
4
S
D
0
.
0
1
0
1
6
0
.
0
0
5
5
0
0
.
0
0
7
9
5
0
.
0
0
6
7
1
0
.
0
0
7
8
7
0
.
0
0
7
0
3
0
.
0
0
7
4
9
S
k
e
w
n
e
s
s
0
.
3
2
1
2
0
.
0
0
5
2
0
.
0
5
2
2
0
.
0
0
7
0
.
2
3
3
0
.
0
1
2
0
.
0
2
1
K
u
r
t
o
s
i
s
6
.
2
4
.
4
3
.
5
3
.
4
4
.
4
3
.
3
3
.
8
J
a
r
q
u
e
-
B
e
r
a
5
7
2
.
0
1
0
2
.
3
1
1
.
4
7
.
3
1
1
3
.
5
5
.
1
3
2
.
9
P
a
n
e
l
C
:
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
s
f
o
r
t
h
e
s
i
x
r
e
g
i
o
n
a
l
b
a
n
k
p
o
r
t
f
o
l
i
o
s
c
M
e
a
n
0
.
0
0
0
0
1
0
.
0
0
0
3
9
0
.
0
0
0
1
6
0
.
0
0
0
1
1
0
.
0
0
0
0
3
0
.
0
0
0
0
8
0
.
0
0
0
2
0
S
D
0
.
0
1
6
7
9
0
.
0
1
8
0
5
0
.
0
2
2
3
1
0
.
0
1
7
7
3
0
.
0
2
8
3
5
0
.
0
1
7
4
1
0
.
0
1
8
3
9
S
k
e
w
n
e
s
s
2
0
.
0
2
5
2
.
6
7
3
0
.
8
7
1
0
.
1
8
7
1
.
0
7
6
0
.
2
6
2
0
.
3
9
9
K
u
r
t
o
s
i
s
9
.
7
2
3
.
6
1
0
.
0
6
.
3
1
0
.
9
6
.
6
7
.
7
J
a
r
q
u
e
-
B
e
r
a
2
,
3
7
7
.
3
2
3
,
7
4
6
.
6
2
,
7
5
2
.
7
5
9
6
.
2
3
,
5
2
9
.
0
6
8
0
.
3
1
,
1
9
4
.
5
N
o
t
e
s
:
a
J
a
n
u
a
r
y
2
,
2
0
0
2
-
D
e
c
e
m
b
e
r
3
0
,
2
0
1
1
,
n
¼
2
,
5
1
8
;
b
J
a
n
u
a
r
y
2
,
2
0
0
2
-
D
e
c
e
m
b
e
r
2
9
,
2
0
0
6
,
n
¼
1
,
2
5
8
;
c
J
a
n
u
a
r
y
3
,
2
0
0
7
-
D
e
c
e
m
b
e
r
3
0
,
2
0
1
1
,
n
¼
1
,
2
6
0
Table I.
Descriptive statistics
Short-sell
moratorium
effects
99
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where y
t
is return, x
t
is an exogenous or predetermined vector of variables, 1
t
is a random
error, and V
t21
is the information set. The parameter vectors are a, b
i
, d
0
, and d
i
and
t is a time index. The conditional variance, v
t
, is linearly dependent on the past behavior
of squared errors and a moving average of past conditional variances. The d
i
values
determine the weights attached to lagged innovations while the squared error term
implies that if innovations have been large in absolute value, they are likely to be large in
the future. Parameters b
i
, d
0
, and d
i
must be non-negative to ensure that the return
generating process is well de?ned.
If all coef?cients in the conditional variance equation (second equation of (1) are zero
(except the intercept) then the model reduces to the traditional constant variance
speci?cation. Both ARCH and GARCH model the conditional variance as a function of
past shocks and allow volatility to evolve and persist. The ARCH model incorporates a
limited number of lags in the derivation of the conditional variance and is a short
memory model. In contrast, the GARCH is a long memory model where past values of
the conditional variance and lagged errors in?uence the current period conditional
variance.
For the GARCHprocess to be stable, the effect of shocks to volatility as measured by
the sumof the coef?cients
P
i
d
i
þb
i
À Á
in the conditional variance equation must be less
than one. If the sum of these coef?cients is greater than or equal to one, Engle and
Bollerslev (1986) suggested using an IGARCH process where the constant term in the
conditional variance equation is suppressed and the restriction that
P
i
d
i
þb
i
¼ 1 is
imposed. To be persistent, the shocks must be signi?cantly different from zero. The
degree of persistence is important in determining the relation between volatility and
returns, since only persistent volatility justi?es changes in returns. The GARCH(1,1) is
more parsimonious relative to higher order models, allows for long memory in the
volatility process, andaccording to Bollerslev et al. (1992), ?ts most economic time series.
Based on Brockett et al. (1999) we estimate the following GARCH model for each of
the six portfolios of bank excess returns:
R
it
¼ a
i
þb
it
£ R
mt
þg
i
£ D
t
þl
i
£ R
it21
þh
it
h
it
¼ a
i
þ b
i
£ h
2
it21
þ c
i
£ h
it21
þ d
i
£ D
t
h
it
jV
t
~
Nð0; h
it
Þ
ð2Þ
where R
it
is the daily excess return on the ith regional bank portfolio, R
mt
is the daily
excess return on the S&P500 market index, and a
i
; b
it
; g
i
; a
i
; b
i
; c
i
; and d
i
are parameters to be estimated. Rather than a GARCH-M model that speci?es h
it
as the
risk measure in the return equation, this approach follows the more traditional market
model and allows for a time varying beta (b
it
). The indicator, D
t
equals 1 if t is an event
day, and 0 otherwise; and V
t
consists of all information available at time t, including all
the current and previous market excess returns, R
mu
, and security excess returns,
R
iu
, for u # t, the current and previous volatility estimates, h
iu
, for u # t, and current
and previous error estimates, h
iu
, for u # t. There are 15 trading days during the
short-sell ban that was in effect from September 19 to October 9, 2008.
During the event period, the estimate of g
i
gives the mean abnormal return. If the null
hypothesis of a zero mean abnormal return cannot be rejected, then the estimate of g
i
must be close to zero. The meaning of “close to zero” depends on the volatility of the
market model residual re?ected in h
it
which incorporates the event-induced variance via
JFEP
5,2
100
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the coef?cient d
i
. The GARCH residuals were subjected to diagnostic tests for the
presence of serial correlation. The Durbin-Watson statistic, Q-statistics and Lagrange
multiplier (LM) tests all indicated signi?cant serial correlation in the residuals of all six
models. The models were re-estimated including the autoregressive term (R
it21
).
The sumof the ARCHand GARCHcoef?cients (b
i
þ c
i
) was greater than or equal to one
indicating an unstable conditional variance. We re-estimated the models using
the IGARCHmethod of Engle and Bollerslev (1986). This method sets the intercept (a
i
) in
the conditional variance equation equal to zero and imposes the restriction that the
ARCH and GARCH coef?cients sum to one (b
i
þ c
i
¼ 1).
The results of the IGARCH models for the six regional portfolios appear in Table II.
Time varying betas ranged from0.34 for the portfolio of Mid-Atlantic banks to 0.77 for the
portfolio of Mid-West banks. The betas (
^
b
i
) for all portfolios were positive and
signi?cantly different from zero. The LM tests whether the residuals of the IGARCH
model dependonthe lagged values of the residuals. The LMtest has a x
2
distributionwith
degrees of freedom equal to the number of lagged squared residuals included as
independent variables. The null hypothesis is that laggedsquaredresiduals have no effect
on the current squared residual. Except for the portfolio of Southeast banks we cannot
reject the null hypothesis for the other ?ve regions. The results suggest that the regional
bank portfolio data for the period examined conform to the IGARCH speci?cation.
For the six regional portfolios the coef?cients for the event indicator variable, g
i
, in the
returnequations are negative, but insigni?cant. Thus, the banonshort sales hadno effect
on excess returns. The coef?cients on event-induced risk in the variance equation (d
i
)
was positive for ?ve of the six portfolios (the Paci?c portfolio had a negative and
insigni?cant coef?cient) but was signi?cantly different from zero only for the
Mid-Atlantic and the Northeast. Together, these estimates indicate that the short-sell ban
had no effect on risk or return in four of the six regional portfolios and increased risk for
only two of the six portfolios.
Earlier in the summer, from July 15 to August 12 of 2008, the SEC imposed a ban on
naked short selling in 19 ?nancial stocks. The small regional banks that we examine had
not been singled out in the July-August ban on naked short selling, but might have
experienced indirect effects via the signal that the SEC was sending ?nancial markets
about the condition and underlying speculation of large ?nancial ?rms. We re-estimated
the IGARCH model assuming that the event days, D
t
¼ 1, occurred July 15-August 12,
2008 and September 19-October 9, 2008. These two periods consist of 21 þ 15 ¼ 36
event days. The results are reportedinTable III. In the return equation, the intercepts (a
i
)
and slopes (b
i
) are little changed, but the effect of the event signi?cantly increased excess
returns for all six regional bank portfolios. In addition, three out of the six regional
portfolios (Mid-Atlantic, Northeast, and Southwest) experienced a signi?cant increase in
risk during the event period and two other portfolios (Mid-West and Paci?c) were
positive and signi?cant at the 10 percent level. Thus, the two short-sell ban periods
worked to prop up the portfolios of regional bank stocks, albeit with an increase in risk.
To further test the effects of the short-sell ban we re-estimated the IGARCH model
for eachof the 110regional banks witheachbankassignedtoits respectiveregion. Table IV
reports the meanestimates of the short-sell banondaily returns (
^ g) and ondaily risk (
^
d) for
banks ina particular region. Onlya small number of banks (betweenone and sixineach of
the six regions) have their returns signi?cantly impacted by the short-sell ban that
occurred from September 19 to October 9, 2008, but a majority of banks in all regions
Short-sell
moratorium
effects
101
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
C
o
e
f
f
.
a
,
b
M
i
d
-
A
t
l
a
n
t
i
c
M
i
d
-
W
e
s
t
N
o
r
t
h
e
a
s
t
P
a
c
i
?
c
S
o
u
t
h
e
a
s
t
S
o
u
t
h
w
e
s
t
a
i
0
.
0
0
0
2
4
*
(
0
.
0
0
0
1
0
)
0
.
0
0
0
0
2
(
0
.
0
0
0
1
1
)
0
.
0
0
0
1
3
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
6
1
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
4
2
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
3
2
(
0
.
0
0
0
1
2
)
b
i
0
.
3
4
1
4
3
*
(
0
.
0
0
8
6
9
)
0
.
7
7
7
0
3
*
(
0
.
0
1
1
1
9
)
0
.
6
2
0
9
3
*
(
0
.
0
0
9
9
6
)
0
.
7
0
8
6
3
*
(
0
.
0
1
2
2
2
)
0
.
6
0
8
5
0
*
(
0
.
0
1
0
0
8
)
0
.
6
9
3
0
2
(
0
.
0
1
1
3
7
)
g
i
2
0
.
0
0
0
3
3
(
0
.
0
0
2
9
8
)
2
0
.
0
0
0
7
2
(
0
.
0
0
9
5
2
)
2
0
.
0
0
2
9
9
(
0
.
0
0
5
7
6
)
2
0
.
0
0
9
2
7
(
0
.
0
0
9
8
3
)
2
0
.
0
0
0
5
8
(
0
.
0
0
3
6
7
)
2
0
.
0
0
2
2
4
(
0
.
0
0
6
2
4
)
l
i
2
0
.
0
5
2
3
8
*
(
0
.
0
1
4
2
7
)
2
0
.
0
9
4
8
*
(
0
.
0
1
0
7
0
)
2
0
.
0
5
5
5
5
*
(
0
.
0
1
2
7
9
)
2
0
.
0
5
7
2
0
*
(
0
.
0
1
3
0
6
)
2
0
.
0
8
9
4
3
*
(
0
.
0
1
2
9
9
)
2
0
.
0
9
3
8
2
(
0
.
0
1
2
0
2
)
b
i
0
.
0
1
3
4
3
*
(
0
.
0
0
0
3
5
)
0
.
0
6
6
5
2
*
(
0
.
0
0
4
1
9
)
0
.
0
4
5
1
1
*
(
0
.
0
0
3
4
2
)
0
.
0
5
5
7
9
*
(
0
.
0
0
3
4
9
)
0
.
0
5
7
3
3
*
(
0
.
0
0
3
3
3
)
0
.
0
4
1
4
6
*
(
0
.
0
0
3
0
4
)
c
i
0
.
9
8
6
5
7
*
(
0
.
0
0
0
3
5
)
0
.
9
3
3
4
8
*
(
0
.
0
0
4
1
9
)
0
.
9
5
4
8
9
*
(
0
.
0
0
3
4
2
)
0
.
9
4
4
2
1
*
(
0
.
0
0
3
4
9
)
0
.
9
4
2
6
7
*
(
0
.
0
0
3
3
3
)
0
.
9
5
8
5
4
*
(
0
.
0
0
3
0
4
)
d
i
0
.
0
0
0
0
1
*
(
3
.
3
1
£
1
0
2
6
)
0
.
0
0
0
0
3
(
0
.
0
0
0
0
2
)
0
.
0
0
0
0
3
*
(
0
.
0
0
0
0
1
)
2
2
.
2
5
£
1
0
2
6
(
0
.
0
0
0
0
3
)
0
.
0
0
0
0
2
(
0
.
0
0
0
0
2
)
0
.
0
0
0
0
2
(
8
.
9
7
£
1
0
2
6
)
A
R
C
H
L
M
-
t
e
s
t
c
0
.
1
1
(
0
.
7
5
)
0
.
1
1
(
0
.
7
4
)
3
.
1
5
(
0
.
0
8
)
0
.
8
9
(
0
.
3
5
)
3
5
.
1
6
(
0
.
0
1
)
1
.
1
0
(
0
.
2
9
)
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
t
a
t
:
*
5
p
e
r
c
e
n
t
;
J
a
n
u
a
r
y
2
,
2
0
0
2
-
D
e
c
e
m
b
e
r
3
0
,
2
0
1
1
c
o
e
f
?
c
i
e
n
t
s
a
n
d
(
s
t
a
n
d
a
r
d
e
r
r
o
r
s
)
;
a
t
h
e
e
x
c
e
s
s
r
e
t
u
r
n
e
q
u
a
t
i
o
n
i
s
:
R
i
t
¼
a
i
þ
b
i
£
R
m
t
þ
g
i
£
D
t
þ
l
i
£
R
i
t
2
1
þ
h
i
t
,
w
h
e
r
e
R
i
t
¼
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
f
o
r
b
a
n
k
i
,
R
m
t
¼
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
o
n
t
h
e
S
&
P
5
0
0
,
a
n
d
D
t
¼
1
f
o
r
t
h
e
?
f
t
e
e
n
t
r
a
d
i
n
g
d
a
y
s
f
r
o
m
S
e
p
t
e
m
b
e
r
1
9
,
2
0
0
8
t
o
O
c
t
o
b
e
r
9
,
2
0
0
8
,
0
o
t
h
e
r
w
i
s
e
;
b
t
h
e
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
e
q
u
a
t
i
o
n
i
s
:
h
i
t
¼
b
i
£
h
2i
t
2
1
þ
c
i
£
h
i
t
2
1
þ
d
i
£
D
t
;
b
i
þ
c
i
¼
1
,
w
h
e
r
e
h
i
t
¼
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
f
o
r
b
a
n
k
i
,
h
2i
t
2
1
i
s
t
h
e
l
a
g
g
e
d
s
q
u
a
r
e
d
r
e
s
i
d
u
a
l
f
r
o
m
t
h
e
r
e
t
u
r
n
e
q
u
a
t
i
o
n
,
h
i
t
2
1
i
s
t
h
e
l
a
g
g
e
d
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
,
a
n
d
D
t
¼
1
f
o
r
t
h
e
1
5
t
r
a
d
i
n
g
d
a
y
s
f
r
o
m
S
e
p
t
e
m
b
e
r
1
9
,
2
0
0
8
t
o
O
c
t
o
b
e
r
9
,
2
0
0
8
,
0
o
t
h
e
r
w
i
s
e
;
c
t
h
e
t
e
s
t
s
t
a
t
i
s
t
i
c
f
o
r
t
h
e
A
R
C
H
-
L
M
t
e
s
t
i
s
d
i
s
t
r
i
b
u
t
e
d
a
s
x
2
w
i
t
h
1
d
e
g
r
e
e
o
f
f
r
e
e
d
o
m
Table II.
IGARCH model estimates
for the six regional
bank portfolios
JFEP
5,2
102
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
C
o
e
f
f
.
a
,
b
M
i
d
-
A
t
l
a
n
t
i
c
M
i
d
-
W
e
s
t
N
o
r
t
h
e
a
s
t
P
a
c
i
?
c
S
o
u
t
h
e
a
s
t
S
o
u
t
h
w
e
s
t
a
i
0
.
0
0
0
2
1
*
(
0
.
0
0
0
1
0
)
0
.
0
0
0
0
2
(
0
.
0
0
0
1
1
)
0
.
0
0
0
1
1
(
0
.
0
0
0
1
1
)
0
.
0
0
0
6
0
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
4
0
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
3
0
*
(
0
.
0
0
0
1
2
)
b
i
0
.
3
4
2
3
5
*
(
0
.
0
0
8
7
4
)
0
.
7
7
7
4
6
*
(
0
.
0
1
1
1
7
)
0
.
6
1
9
9
7
*
(
0
.
0
0
9
8
7
)
0
.
7
1
0
2
1
*
(
0
.
0
1
2
2
7
)
0
.
6
0
8
9
2
*
(
0
.
0
1
0
0
6
)
0
.
6
9
3
9
6
*
(
0
.
0
1
1
4
5
)
g
i
0
.
0
0
4
8
8
*
(
0
.
0
0
1
7
4
)
0
.
0
0
8
7
4
*
(
0
.
0
0
4
1
3
)
0
.
0
0
6
1
7
*
(
0
.
0
0
2
4
2
)
0
.
0
1
0
5
6
*
(
0
.
0
0
4
8
9
)
0
.
0
0
4
6
1
*
(
0
.
0
0
2
3
0
)
0
.
0
0
6
5
7
*
(
0
.
0
0
3
2
5
)
l
i
2
0
.
0
5
2
4
6
*
(
0
.
0
1
4
4
8
)
2
0
.
0
9
5
3
8
*
(
0
.
0
1
0
8
4
)
2
0
.
0
5
7
0
0
*
(
0
.
0
1
2
8
3
)
2
0
.
0
5
6
8
9
*
(
0
.
0
1
3
2
6
)
2
0
.
0
9
0
4
0
*
(
0
.
0
1
2
8
3
)
2
0
.
0
9
3
7
0
*
(
0
.
0
1
2
5
8
)
b
i
0
.
0
1
2
0
6
*
(
0
.
0
0
0
3
8
)
0
.
0
6
5
4
4
*
(
0
.
0
0
4
3
2
)
0
.
0
4
4
3
8
*
(
0
.
0
0
3
4
2
)
0
.
0
5
4
8
7
*
(
0
.
0
0
3
5
6
)
0
.
0
5
7
1
7
*
(
0
.
0
0
3
3
3
)
0
.
0
4
0
4
7
*
(
0
.
0
0
3
3
3
)
c
i
0
.
9
8
7
9
4
*
(
0
.
0
0
0
3
8
)
0
.
9
3
4
5
6
*
(
0
.
0
0
4
3
2
)
0
.
9
5
5
6
2
*
(
0
.
0
0
3
4
2
)
0
.
9
4
5
1
3
*
(
0
.
0
0
3
5
6
)
0
.
9
4
2
8
4
*
(
0
.
0
0
3
3
3
)
0
.
9
5
9
5
3
*
(
0
.
0
0
3
3
3
)
d
i
0
.
0
0
0
0
1
*
(
1
.
3
1
£
1
0
2
6
)
0
.
0
0
0
0
1
(
7
.
8
3
£
1
0
2
6
)
7
.
8
2
£
1
0
2
6
*
(
3
.
1
3
£
1
0
2
6
)
0
.
0
0
0
0
3
(
0
.
0
0
0
0
2
)
4
.
8
2
£
1
0
2
6
(
4
.
2
4
£
1
0
2
6
)
0
.
0
0
0
0
1
*
(
3
.
4
5
£
1
0
2
6
)
A
R
C
H
L
M
-
t
e
s
t
c
0
.
1
4
(
0
.
7
9
)
0
.
0
5
(
0
.
8
2
)
3
.
9
7
(
0
.
0
5
)
0
.
7
3
(
0
.
3
9
)
3
7
.
0
(
0
.
0
1
)
0
.
7
7
(
0
.
3
8
)
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
t
a
t
:
*
5
p
e
r
c
e
n
t
;
J
a
n
u
a
r
y
2
,
2
0
0
2
-
D
e
c
e
m
b
e
r
3
0
,
2
0
1
1
c
o
e
f
?
c
i
e
n
t
s
a
n
d
(
s
t
a
n
d
a
r
d
e
r
r
o
r
s
)
;
a
t
h
e
e
x
c
e
s
s
r
e
t
u
r
n
e
q
u
a
t
i
o
n
i
s
:
R
i
t
¼
a
i
þ
b
i
£
R
m
t
þ
g
i
£
D
t
þ
l
i
£
R
i
t
2
1
þ
h
i
t
,
w
h
e
r
e
R
i
t
¼
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
f
o
r
b
a
n
k
i
,
R
m
t
¼
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
o
n
t
h
e
S
&
P
5
0
0
,
a
n
d
D
t
¼
1
f
o
r
i
f
t
h
e
t
r
a
d
i
n
g
d
a
y
o
c
c
u
r
s
b
e
t
w
e
e
n
J
u
l
y
1
5
,
2
0
0
8
a
n
d
A
u
g
u
s
t
1
2
,
2
0
0
8
o
r
f
r
o
m
S
e
p
t
e
m
b
e
r
1
9
,
2
0
0
8
t
o
O
c
t
o
b
e
r
9
,
2
0
0
8
,
0
o
t
h
e
r
w
i
s
e
;
t
h
e
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
e
q
u
a
t
i
o
n
i
s
:
h
i
t
¼
b
i
£
h
2i
t
2
1
þ
c
i
£
h
i
t
2
1
þ
d
i
£
D
t
;
b
i
þ
c
i
¼
1
,
w
h
e
r
e
h
i
t
¼
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
f
o
r
b
a
n
k
i
,
h
2i
t
2
1
i
s
t
h
e
l
a
g
g
e
d
s
q
u
a
r
e
d
r
e
s
i
d
u
a
l
f
r
o
m
t
h
e
r
e
t
u
r
n
e
q
u
a
t
i
o
n
,
h
i
t
2
1
i
s
t
h
e
l
a
g
g
e
d
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
,
a
n
d
D
t
¼
1
f
o
r
i
f
t
h
e
t
r
a
d
i
n
g
d
a
y
o
c
c
u
r
s
b
e
t
w
e
e
n
J
u
l
y
1
5
,
2
0
0
8
a
n
d
A
u
g
u
s
t
1
2
,
2
0
0
8
o
r
f
r
o
m
S
e
p
t
e
m
b
e
r
1
9
,
2
0
0
8
t
o
O
c
t
o
b
e
r
9
,
2
0
0
8
,
0
o
t
h
e
r
w
i
s
e
,
t
h
e
r
e
s
t
r
i
c
t
i
o
n
,
b
i
þ
c
i
¼
1
,
i
m
p
o
s
e
s
t
h
e
I
G
A
R
C
H
;
c
t
h
e
t
e
s
t
s
t
a
t
i
s
t
i
c
f
o
r
t
h
e
A
R
C
H
-
L
M
t
e
s
t
i
s
d
i
s
t
r
i
b
u
t
e
d
a
s
x
2
w
i
t
h
1
d
e
g
r
e
e
o
f
f
r
e
e
d
o
m
(
p
-
v
a
l
u
e
)
Table III.
IGARCH model estimates
for the six regional
bank portfolios
Short-sell
moratorium
effects
103
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
except the Mid-West experienced a signi?cant increase in risk. When the event period is
expanded to include July 15-August 12, 2008 and September 19-October 9, 2008, more
banks ineach regionexperience a signi?cant increase inreturn andmore banks infour out
of the six regions experience a signi?cant increase in risk due to the event.
Are there signi?cant differences in the effect of the ban on bank risk between the six
regions? Let the kernel density of the individual bank estimates of event induced risk,
^
d
i
,
withina particular regionbe f ð
^
d
Region
Þ andlet the densityof the individual bankestimates
of
^
d
i
for banks outside the region be gð
^
d
Outside region
Þ. We test the null hypothesis that
f ð
^
d
Region
Þ ¼ gð
^
d
Outside region
Þ using Li’s (1996) J
n1,n2
test which has a standard normal
distribution. Details of this test are foundinthe Appendix. Because of the small number of
banks within each region we follow the bootstrap approach described by Pagan and
Ullah (1999) with 500 bootstrap replications. The results of this test are reported in
Table V and indicate whether or not the short-sell ban resulted in systematic regional
effects on individual bank risk. The short-sell ban for the period September 19 to October
9 had a signi?cant regional impact on risk for banks in the Northeast region ( p ¼ 0.05)
and for banks in the Southwest region ( p ¼ 0.01), but no signi?cant impact on banks in
the other four regions. When the event periodis expanded to include the periods July 15 to
August 12 andSeptember 19 to October 9, there is a signi?cant increase inriskfor ?ve out
of the six regions, with the exception being the Mid-Atlantic region. We also performed a
similar battery of tests for regional differences in event induced changes in return,
but found no signi?cant differences across the six regions.
Despite evidence which indicates that the short-sell ban increased risk, it may have
substantially reduced regional bank failures-to-deliver. Figure 2 graphs the cumulative
value of failures-to-deliver for the 110 regional banks from April 2007 through
No. of
banks
^ g
Region
No. of banks with ^ g – 0
at p , 0.10
^
d
Region
No. of banks with
^
d – 0
at p , 0.10
Event period ¼ September 19-October 9, 2008
Mid-Atlantic 21 0.00027 2 0.00029 14
Mid-West 19 0.00043 1 0.00020 9
Northeast 18 20.00273 6 0.00008 12
Paci?c 18 20.00222 3 0.00011 10
Southeast 17 0.00073 3 0.00021 14
Southwest 17 0.01200 1 0.00013 12
Event period ¼ July 15-August 12, 2008 and September 19-October 9, 2008
Mid-Atlantic 21 0.00402 4 0.00011 17
Mid-West 19 0.01041 8 0.00007 13
Northeast 18 0.00587 8 0.00015 13
Paci?c 18 0.01318 7 0.00014 13
Southeast 17 0.00560 9 0.00008 13
Southwest 17 0.00768 7 0.00007 11
Notes: The event induced change in return for bank i is captured by the estimated coef?cient ^ g
i
; the event
induced change in risk is for bank i is captured by the estimated coef?cient
^
d
i
; event induced changes in
return and risk are estimated from the excess returns, R
it
¼ a
i
þb
i
£ R
mt
þg
i
£ D
t
þl
i
£ R
it21
þh
it
,
and conditional variance, h
it
¼ b
i
£ h
2
it21
þ c
i
£ h
it21
þ d
i
£ D
t
; b
i
þ c
i
¼ 1, estimated from an
IGARCHmodel, where R
it
¼daily excess return for bank i, R
mt
¼daily excess return on the S&P500, h
it
¼
conditional variance for bank i, h
2
it21
is the lagged squared residual from the return equation, h
it21
is the
laggedconditional variance, andD
t
¼1 for if the tradingdayoccurs duringthe event periodand0 otherwise
Table IV.
Effects of the short-sell
ban on individual bank
returns ( ^ g) and risk (
^
d)
JFEP
5,2
104
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
December 2011 (Source: US SEC: Fails to Deliver Data-Archive Data (2012)). For a given
bank, the value of failures-to-deliver securities equals the product of the security prices
and the cumulative net failures-to-deliver securities up to that date. We then sum these
values over the 110 banks to get the cumulative value of failures to deliver. Figure 2 shows
that in the 18 months before and including September of 2008 the cumulative value
Event induced change in risk
event ¼ September 19-October 9,
2008
Event induced change in risk
event ¼ July 15-August 12, 2008
and September 19-October 9, 2008
f ðd
Region
Þ ¼ gðd
other regions
Þ
a
f ðd
Region
Þ ¼ gðd
other regions
Þ
Null hypothesis
Bootstrapped J
n1,n2
b
(Prob . J
n1,n2
)
Bootstrapped J
n1,n2
(Prob . J
n1,n2
)
Mid-Atlantic vs other regions 1.31 (0.19) 1.00 (0.32)
Mid-West vs other regions 1.51 (0.13) 4.07 (0.01)
Northeast vs other regions 1.97 (0.05) 20.11 (0.01)
Paci?c vs other regions 0.80 (0.42) 4.06 (0.01)
Southeast vs other regions 1.51 (0.13) 6.78 (0.01)
Southwest vs other regions 3.93 (0.01) 7.88 (0.01)
Notes:
a
The kernel density estimates (empirical distribution functions) of event induced changes in
risk (d ) are given as f(d
Region
) for banks within a particular region; for banks outside the region, the
kernel density estimates of event induced changes in risk are given as g(d
other regions
); the estimates of
event induced changes in risk are from the conditional variance equation of the GARCH model;
b
the
test statistic J
n1,n2
has a standard normal distribution, where there are n
1
bank observations within the
region and n
2
banks observations outside the region; the number of bootstrap replications is 500
Table V.
Tests for differences in
risk between regions,
2 January 2002-
30 December 2011
Figure 2.
Value of failures to deliver
securities for 110 banks
A
p
r
-
0
7
A
p
r
-
0
8
A
p
r
-
0
9
A
p
r
-
1
0
A
p
r
-
1
1
J
u
l
-
0
7
J
u
l
-
0
8
J
u
l
-
0
9
J
u
l
-
1
0
J
u
l
-
1
1
O
c
t
-
0
7
O
c
t
-
0
8
O
c
t
-
0
9
O
c
t
-
1
0
O
c
t
-
1
1
J
a
n
-
0
9
J
a
n
-
1
0
J
a
n
-
1
1
J
a
n
-
0
8
0
500,000,000
1E+09
1.5+09
2E+09
2.5E+09
3.5E+09
3E+09
Short-sell
moratorium
effects
105
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
of failures had increased sharply to approximately $3 billion, but then declined in the
wake of the short-sell ban to remain at less than $130 million each month.
5. Conclusions
This research utilized an IGARCH version of the market model to test for event induced
volatility and excess returns as a consequence of the SEC mandate that restricted short
sales on approximately 800 ?nancial stocks. 110 regional banks were identi?ed as having
outstanding short interest and used to construct equally weighted regional bank
portfolios. The results indicate that the September 19-October 9 short-sell banhadnoeffect
onreturns, but a small positive effect onriskintwo out of the sixregions (Mid-Atlantic and
Northeast). The sample of regional banks was not legally subject to the July 15-August 12
ban that applied to 19 “too big to fail” institutions, however, the regional portfolios were
tested for evidence of “signaling” and found signi?cant positive effects in each of the six
regions. The two short-sell bans were associated with an increase in risk for the portfolios
of banks comprising the Mid-Atlantic and Northeast regions.
The IGARCH model was also estimated for each of the 110 regional banks. For the
expanded event period we found that 43 banks experienced a signi?cant event induced
change in return and 80 banks experienced a signi?cant event induced increase in risk.
The expanded event period found that ?ve of the six regions exhibited signi?cant
event induced changes in risk. The results suggest that rather than mitigate volatility
in the equity prices of regional banks, the SEC short-sell moratorium contributed to
bank risk which could be attributed to declines in trading volume and liquidity
identi?ed by Autore et al. (2011), Battalio and Schultz (2011), Jain et al. (2012) and
March and Payne (2012).
The results are consistent with Diamond and Verrecchia’s (1987) model, which
predicts that market participants anticipate the rising costs of constrained short
selling causing a greater price reaction. Despite ?nding signi?cant event induced
effects on individual bank risk and return, the magnitudes were small. Whether or not
sophisticated investors were able to engage in regulatory arbitrage from the policy,
a circumstance that has been described by Lowenstein (2000) as “picking up nickels in
front of a steamroller,” is not clear.
The lifting of the SEC short-sell restriction after a scant three weeks is interpreted as
an admission that the policy failed to achieve the stated result. In his assessment of the
ban, former SEC Chairman Christopher Cox said on December 31, 2008: “Knowing what
we know now, [we] would not do it again. The costs appear to outweigh the bene?ts.”
(Younglai, 2008) Our ?nding of event induced volatility would appear to substantiate
this interpretation.
However, in light of “too big to fail” and the relatively modest impact of the ban on
regional bank risk/return we offer an alternative assessment that judges the policy less
harshly. If delivery failures caused primarily by naked shorting pose a potentially
compounding and cumulative liquidity risk to the ?nancial system as argued by
Bradley et al. (2011), then the short-selling ban along with the four day delivery rule
may have provided a “cooling off period” that helped reduce that risk. In turn, Figure 2
shows that the ban might have helped to substantially reduce delivery failures among
regional bank stocks and the corresponding liquidity risk. If the ?nancial crisis has
taught us anything it is that ?nancial markets and institutions are far more complex
than many of us believed.
JFEP
5,2
106
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Note
1. On Friday, September 19, 2008, the SEC Released an Emergency Order prohibiting short sales
in 799 securities (www.sec.gov/rules/other/2008/34-58592.pdf). On the following Monday the
order was amended with important material differences: www.sec.gov/news/press/2008/
2008-218.htm that included the SEC’s delegation of authority to each national securities
exchange to identify and add listed companies that qualify for inclusion on the covered list. As
the additions and removals on the list are subject to change on a daily basis, the NYSE has
maintained and provided daily end-of-day list updates to the covered list since the amended
order went into effect.
As of Friday, September 26, 2008 the NYSE and NYSE Arca short-sell prohibition list
had been enhanced as part of a collaborative effort with NASDAQ and AMEX, to include
securities that will be covered for the next trading day from all four exchanges in one
convenient document. Daily updates are provided at: www.nyse.com/attachment/
CONSOLIDATED-SSPROHIBTION.xls or once can subscribe to NYSE System Status
Trader e-mail updates: www.nysenet.com/subscription/smLogin to receive notice as soon as it
becomes available along with other NYSE Cash Equities Market System and Trading
Information news.
References
Akhigbe, A. and Whyte, A. (2003), “Changes in market assessments of bank risk following the
Riegle-Neal Act of 1994”, Journal of Banking & Finance, Vol. 27, pp. 87-102.
Autore, D.M., Billingsley, R.S. and Kovacs, T. (2011), “The 2008 short sale Ban: liquidity,
dispersion of opinion, and the cross-section of returns of US ?nancial stocks”, Journal of
Banking & Finance, Vol. 35 No. 9, pp. 2252-2266.
Battalio, R. and Schultz, P. (2011), “Regulatory uncertainty and market liquidity: the 2008 Short
Sale Ban’s impact on equity option markets”, The Journal of Finance, Vol. 66 No. 6,
pp. 2013-2053.
Boehmer, E., Jones, C.M. and Zhang, X. (2011), “Shackling short sellers: the 2008 shorting Ban”,
available at:http://ssrn.com/abstract¼1412944
Bollerslev, T. (1986), “Generalized autoregressive conditional heteroscedasticity”, Journal of
Econometrics, Vol. 31, pp. 307-327.
Bollerslev, T., Chou, Y. and Kroner, K. (1992), “ARCH modeling in ?nance: a review of the theory
and empirical evidence”, Journal of Econometrics, Vol. 52, pp. 5-59.
Bradley, R., Fawls, H., Litan, R. and Sommers, F. (2011), Canaries in the Coal Mine:
How Settlement Fails Create Systemic Risk for Financial Firms and Investors,
Ewing Marion Kauffman Foundation, Kansas City, MO.
Brenner, M. and Subrahmanyam, M. (2009), “Short selling”, in Acharya, V. and
Richardson, M. (Eds), Restoring Financial Stability: How to Repair a Failed System,
Wiley, New York, NY.
Bris, A., Goetzmann, W. and Zhu, N. (2007), “Ef?ciency and the bear: short sales and markets
around the world”, Journal of Finance, Vol. 62 No. 3, pp. 1029-1079.
Brockett, R., Chen, H. and Garven, J. (1999), “A new stochastically ?exible event methodology
with application to proposition 103”, Insurance, Mathematics and Economics, Vol. 25,
pp. 197-217.
Brown, S. and Warner, J. (1980), “Measuring security price performance”, Journal of Financial
Economics, Vol. 8, pp. 205-258.
Brown, S. and Warner, J. (1985), “Using daily stock returns: the case of event studies”, Journal of
Financial Economics, Vol. 14, pp. 3-31.
Short-sell
moratorium
effects
107
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Brunnermeier, M.K. and Pedersen, L.H. (2005), “Predatory trading”, Journal of Finance, Vol. 60,
pp. 1825-1863.
Campos, R. (2006), Speech by SEC Commissioner: Remarks on Regulation SHO, US Securities and
Exchange Commission, Washington, DC, July 12.
Chong, B.S. (1991), “Effects of interstate banking on commercial banks risk and pro?tability”,
Review of Economics and Statistics, Vol. 73, pp. 78-84.
Cochrane, J. (2011), “The fed’s mission impossible”, The Wall Street Journal, December 29.
Demsetz, R.S. and Strahan, P.E. (1997), “Diversi?cation, size and risk at bank holding
companies”, Journal of Money, Credit and Banking, Vol. 29, pp. 300-313.
Deng, S. and Elyasiani, E. (2008), “Geographic diversi?cation, bank holding company value,
and risk”, Journal of Money, Credit and Banking, Vol. 40, pp. 1217-1238.
Diamond, D. and Verrecchia, R. (1987), “Constraints on short selling and asset price adjustments
to private information”, Journal of Financial Economics, Vol. 18, pp. 277-311.
Engle, R.F. (1982), “Autoregressive conditional heteroscedasticity with estimates of the variance
of UK in?ation”, Econometrica, Vol. 50, pp. 987-1007.
Engle, R.F. and Bollerslev, T. (1986), “Modeling the persistence of conditional variances”,
Econometric Reviews, Vol. 5, pp. 1-50.
Engle, R.F., Lilien, D. and Robins, R. (1987), “Estimating time varying term premium in the term
structure: the ARCH-M model”, Econometrica, Vol. 55, pp. 391-407.
Federal Deposit Insurance Corporation (2012a), Failed Bank List, available at: www.fdic.gov/
bank/individual/failed/banklist.html (accessed March 1, 2012).
Federal Deposit Insurance Corporation (2012b), Summary of Deposits, available at:http://www2.
fdic.gov/sod/ (accessed March 1, 2012).
Financial Stability Board (2011), Policy Measures to Address Systematically Important Financial
Institutions, Financial Stability Board, Basel, November 4.
Harrison, J.M. and Kreps, D.M. (1978), “Speculative behavior in a stock market with
heterogeneous expectations”, Quarterly Journal of Economics, Vol. 92 No. 2, pp. 323-336.
Hughes, J.P., Lang, W., Mester, L. and Moon, C. (1996), “Ef?cient banking under interstate
branching”, Journal of Money, Credit and Banking, Vol. 28, pp. 1045-1071.
Jain, C., Jain, P. and McInish, T.H. (2012), “Short-selling: the impact of SEC rule 201 of 2010”,
The Financial Review, Vol. 47 No. 1, pp. 37-64.
Jones, C. and Lamont, O. (2002), “Short sale constraints and stock returns”, Journal of Financial
Economics, Vol. 66, pp. 207-239.
Lamont, O. (2004a), “Go down ?ghting: short sellers vs ?rms”, National Bureau of Economic
Research Working Paper 10659, August.
Lamont, O. (2004b), Short Sale Constraints and Overpricing: The Theory and Practice of Short
Selling, Frank J. Fabozzi Associates, New Hope, PA.
Lamont, O. and Thaler, R. (2003), “Can the market add and subtract: mispricing in tech stock
carve-outs”, Journal of Political Economy, Vol. 111, pp. 227-268.
Li, Q. (1996), “Nonparametric testing of closeness between two unknown distribution functions”,
Econometric Reviews, Vol. 15, pp. 261-274.
Lowenstein, R. (2000), When Genius Failed: The Rise and Fall of Long-term Capital Management,
Random House, New York, NY.
McGinty, T. and Strasburg, J. (2009), “Short sellers squeezed all around”, The Wall Street Journal,
April 7, p. C1.
JFEP
5,2
108
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
March, I.W. and Payne, R. (2012), “Banning short sales and market quality: the UK’s experience”,
Journal of Banking & Finance, Vol. 36 No. 7, pp. 1975-1986.
Miller, E. (1977), “Risk, uncertainty and divergence of opinion”, Journal of Finance, Vol. 32,
pp. 1151-1168.
Pagan, A. and Ullah, A. (1999), Nonparametric Econometrics, Cambridge University Press,
Cambridge.
Reed, A. (2007), “Costly short selling and stock price adjustments to earnings announcements”,
Unpublished manuscript, University of North Carolina, Chapel Hill, NC.
Rivard, R. and Thomas, C. (1997), “The effect of interstate banking on large bank
holding company pro?tability and risk”, Journal of Economics and Business, Vol. 49,
pp. 61-76.
Short Squeeze (2012), available at: www.shortsqueeze.com/
Surowiecki, J. (2004), The Wisdom of Crowds, Doubleday, New York, NY.
Ulibarri, C., Florescu, I. and Eidsath, J.M. (2009), “Noise traders and short-selling troubled ?rms”,
Journal of Financial Economic Policy, Vol. 1 No. 3, pp. 227-245.
US SEC (2008a), “Emergency order pursuant to section 12(k)(2) of the securities and exchange act
of 1934 taking temporary action to respond to market developments”, Release 34-58592
September 18, 2008, US Securities and Exchange Commission.
US SEC (2008b), “SEC halts short selling of ?nancial stocks to protect investors and markets”,
For Immediate Release 2008-211, US Securities and Exchange Commission, available at:
www.sec.gov/news/press/2008/2008-211.htm (accessed February 15, 2012).
US SEC (2010), “Alternative up-tick rule”, Amendment to Regulation SHO, US Securities and
Exchange Commission, February 24, available at: www.sec.gov/news/press/2010/2010-26.
htm
US SEC (2012), “Fails to deliver data”, Archive data, US Securities and Exchange Commission,
available at: www.sec.gov/foia/docs/failsdata-archive.htm
Wallison, P. (2011), “How regulators herded banks into trouble”, The Wall Street Journal,
December 3.
Wearden, G. (2010), “European debt crisis: markets fall as Germany bans naked shorting”,
The Guardian, May 20, p. A1.
Weber, W.L. and Devaney, M. (1999), “Bank ef?ciency, risk-based capital, and real estate
exposure: the credit crunch revisited”, Real Estate Economics, Vol. 27, pp. 1-25.
Younglai, R. (2008), SEC Chief Has Regrets over Short Selling Ban, Reuters US edition,
December 31, available at: www.reuters.com/article/idUSTRE4BU3GG20081231
Further reading
Congressional Oversight Panel (2010), Commercial Real Estate Losses and the Risk to Financial
Stability, available at:http://cop.senate.gov/reports/library/report-021110-cop.cfm
Appendix
Let random variables x and y have estimated density functions given by
^
fðxÞ and ^ gð yÞ. The
standard normal kernel function is given as kðCÞ ¼ ð2pÞ
20:5
expð20:5CÞ where C ¼ ðx
j
2xÞ=h
with h representing the bandwidth or smoothing parameter. The kernel function has the
property that
R
1
21
kðCÞdC ¼ 1. Qi Li’s test for the difference between two distributions
^
fðxÞ and
^ gð yÞ is based on the integrated squared error I ¼
R
x
ð f ðxÞ 2gðxÞÞ
2
dx. With sample size of n
1
for x
and n
2
for y, Li’s J
n1,n2
test is given as:
Short-sell
moratorium
effects
109
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
J
n1;n2
¼
n
1
h
0:5
I
^ s
, Nð0; 1Þ;
where:
I ¼
1
h
X
n1
i¼1
X
n2
j ¼1;
j – i
1
n
1
ðn
1
21Þ
K
x
ij
þ
1
n
2
ðn
2
21Þ
K
y
ij
2
1
n
1
ðn
2
21Þ
K
x;y
ij
2
1
n
2
ðn
1
21Þ
K
y;x
ij
and:
^ s
2
¼
2
h
X
n1
i¼1
X
n2
j¼1
1
n
2
1
K
x
ij
þ
l
2
n
2
2
K
y
ij
þ 2
l
n
1
n
2
K
x;y
ij
for l ¼
n
1
n
2
:
In the above formulation we use K
x
ij
¼ kððx
i
2x
j
Þ=h), K
y
ij
¼ kðð y
i
2y
j
Þ=h), K
x;y
ij
¼ kððx
i
2y
j
Þ=h),
and K
y;x
ij
¼ kðð y
i
2x
j
Þ=h). The optimal bandwidth is found as h ¼ 1:06 £ min s
1
=n
0:2
1
; s
2
=n
0:2
2
À Á
where s
i
is the standard deviation of random variable i ¼ x, y. Pagan and Ullah (1999) provide
an in depth discussion of kernel density estimation and the bootstrap method we use for Li’s
J
n1;n2
test.
About the authors
Michael Devaney is Professor of Finance at Southeast Missouri State University where he
teaches courses in Financial Institutions and Real Estate Finance. His research examines the
effects of public policy changes for various ?nancial institutions including US and Japanese
banks and real estate investment trusts.
William L. Weber is Professor of Economics at Southeast Missouri State University where he
teaches courses inFinancial Institutions, Econometrics, andEnvironmental Economics. His research
examines productivity growth when ?rms produce desirable and undesirable outputs, such as
pollution or risk and the risk/return tradeoff for US and Japanese banks, Japanese securities ?rms,
credit cooperatives, andinsurance companies. WilliamL. Weber is the correspondingauthor andcan
be contacted at: [email protected]
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
JFEP
5,2
110
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_561483486.pdf
The purpose of this paper is to investigate the effects of the 2008 SEC short-sell
moratorium on regional bank risk and return. The paper also examines the decline in “failures to
deliver” securities in the wake of SEC short-sell moratorium.
Journal of Financial Economic Policy
Short-sell moratorium effects on regional bank performance
Michael Devaney William L. Weber
Article information:
To cite this document:
Michael Devaney William L. Weber, (2013),"Short-sell moratorium effects on regional bank performance",
J ournal of Financial Economic Policy, Vol. 5 Iss 2 pp. 92 - 110
Permanent link to this document:http://dx.doi.org/10.1108/17576381311329652
Downloaded on: 24 January 2016, At: 21:46 (PT)
References: this document contains references to 52 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 131 times since 2013*
Users who downloaded this article also downloaded:
J ames R. Barth, Gerard Caprio, Ross Levine, (2013),"Bank regulation and supervision in 180
countries from 1999 to 2011", J ournal of Financial Economic Policy, Vol. 5 Iss 2 pp. 111-219 http://
dx.doi.org/10.1108/17576381311329661
Linyue Li, Nan Zhang, Thomas D. Willett, (2012),"Measuring macroeconomic and financial market
interdependence: a critical survey", J ournal of Financial Economic Policy, Vol. 4 Iss 2 pp. 128-145 http://
dx.doi.org/10.1108/17576381211228989
Sherrill Shaffer, (2012),"External economies in banking", J ournal of Financial Economic Policy, Vol. 4 Iss 4
pp. 354-365http://dx.doi.org/10.1108/17576381211279316
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Short-sell moratorium effects
on regional bank performance
Michael Devaney and William L. Weber
Department of Economics and Finance, Southeast Missouri State University,
Cape Girardeau, Missouri, USA
Abstract
Purpose – The purpose of this paper is to investigate the effects of the 2008 SEC short-sell
moratorium on regional bank risk and return. The paper also examines the decline in “failures to
deliver” securities in the wake of SEC short-sell moratorium.
Design/methodology/approach – In total, six regional bank portfolios are derived and the beta
coef?cients from a CAPM model are estimated using the integrated generalized autoregressive
conditional heteroskedasticity (IGARCH) method accounting for the short-sell moratorium. Data on 110
regional banks in six US regions fromJanuary 2002 to December 30, 2011 are used to estimate the model.
Findings – The ban on naked short selling and the SEC short-sell moratorium signi?cantly increased
individual bank risk for a majority of banks in six geographic regions, but also increased return in
three of three regions. There was also reduced naked short selling as failures to deliver securities
declined sharply after the September 2008 moratorium took effect.
Originality/value – Regional banks have generally not achieved the size needed to be deemed“too big
to fail” bypolicy-makers. Thus, policychanges suchas the SECshort-sell moratoriummight be expected
to have larger effects on regional banks than on larger banks, which might be shielded from the policy
change by having achieved “too big to fail” status. The authors’ results are consistent with research that
has shown that short-sell restrictions increase risk by reducing liquidity and trading volume.
Keywords GARCH, SEC short-sell moratorium, Banks, Risk management, Securities,
United States of America, Economic policy, Securities and Exchange Commission
Paper type Research paper
1. Introduction
For much of the past three decades a central focus of bank regulators has been capital
adequacy. In 1988, central bankers from around the world met in Basel, Switzerland
to recommend minimum capital standards that were enforced by law in the G10
countries and a number of other nations. Known as the Basel Accords or Basel I, the rules
established risk-based capital (RBC) standards that assigned increased capital
requirements for banks with riskier assets. Basel II would follow in 2004, while
Basel III attempted to address de?ciencies in Basel II that were revealed by the
?nancial crisis of 2008. Bank capital as de?ned by regulators in Basel I, II and III derives
largely from accounting de?nitions and departs signi?cantly from the market value of
common equity. It is dif?cult in practice to determine when a bank becomes too big to
fail, but on November 4, Financial Stability Board (2011) issued a release listing 29 banks
that they considered to be systematically important ?nancial institutions whose
“failure, because of their size, complexity and systematic interconnectedness, would
cause signi?cant disruption to the wider ?nancial systemand economic activity.” In the
absence of government guarantees such as “too big to fail” bailouts, market value
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JEL classi?cation – G12, G14, G18, G21
Journal of Financial Economic Policy
Vol. 5 No. 2, 2013
pp. 92-110
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576381311329652
JFEP
5,2
92
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
capital has an advantage over the regulatory de?nition of capital in that private
investors have more incentive to monitor the bank’s actions since in the event of
insolvency, private investors bear its costs.
Although “too big to fail” status truncates insolvency risk, 90 percent of US banks
do not have equity capital that is traded in public markets. The other 10 percent are
exposed to the discipline of markets on a daily basis, yet, regional banks might be too
small to achieve the scale and scope economies of their “too big to fail” counterparts.
The ?nance literature suggests that geographically constrained banks may have fewer
investment opportunities and experience greater risk than those that operate over a
wider geographic region (Deng and Elyasiani, 2008; Hughes et al., 1996; Weber and
Devaney, 1999).
RBC standards assign arbitrary weights to different classes of assets based on their
presumed risk. Because of the low or zero risk weight assigned to real estate and
sovereign debt, Cochrane (2011) and Wallison (2011) argued that RBCcontributed to risk
in the banking system. Similarly, the low risk weights on real estate assets might have
geographically concentrated ?nancial stress in US banking relative to the pre-Basel real
estate bubble of the 1980s. This spatial difference is re?ected in bank failures as reported
by the FDIC. From 2000 to 2007 there were 27 bank failures. Then, in 2008, 25 banks
failed, followed by 140 bank failures in 2009, and 157 failures in 2010 (Federal Deposit
Insurance CorporationFailedBankList (2012a)). From2008 to 2010, Georgia and Florida
had the most bank failures with 51 and 45, followed by Illinois (38), California (34),
Minnesota (15), Washington (14), Missouri (11), Nevada (10), Texas (8), and Arizona (9).
Perhaps a better indicator of the decline in bank services as a result of failures
is the decrease in the number of branches. Figure 1 is a map of US counties which
experienced at least one branch closing as a consequence of bank failures from 2008 to
2010. During the period a total of 1,340 branches of failed banks closed, with 294
occurring in 2008, 678 in 2009, and 368 in 2010. In California, 193 branches closed,
followed by Florida (185) and Georgia (155). Rounding out the top ten states with
branch closings are Texas (90), Illinois (83), Washington (74), Oregon (44), Nebraska (37),
Minnesota (29), and Missouri (28) (Source: Federal Deposit Insurance Corporation
Summary of Deposits (2012b)).
The regulatory de?nition of bank capital may be subject to control by regulators, but
the daily pricing of market equity is widely regarded as being beyond their sphere of
expertise. Nevertheless, in July of 2008, following the demise of Bear Stearns, restrictions
on naked short selling were imposed on a small number of very large ?nancial service
companies that could be characterized as too big to fail institutions. After the collapse of
Lehman Brothers in September of 2008, the SEC short-sell restricted list was expanded
to include approximately 800 ?nancial companies. The regulatory justi?cation for the
restriction was basedon the presumption that it would mitigate “excessive ?uctuation in
the prices” of ?nancial institution securities (US SEC, 2008a, b, 2010).
Attempts to pro?t by shorting beleaguered sectors such as banking have become
more common along with the rise in exchange traded funds. The implementation of
leveraged strategies might have a more pronounced price effect on small market
capitalization sectors that have less trading volume. This effect might be especially felt
toward the end of the trading day when fund managers buy or short stocks to settle their
books. Naked shorting occurs when traders do not actually borrow the shares they sell
short as required by securities law. Naked shorting can result in “failure to deliver”
Short-sell
moratorium
effects
93
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Figure 1.
Countries with branch
of a failed bank, 2008-2010
JFEP
5,2
94
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
which is the number of stocks in which large positions have not been properly delivered
to investors. SEC Regulation SHO was supposed to address the problem of failure to
deliver. Implemented in 2005, the regulation mandated daily compilations by exchanges
of stocks that had experienced at least ?ve consecutive days of delivery failures totaling
at least 10,000 shares and at least one half percent of their outstanding shares. When a
stock was placed on the threshold list, traders were presumably required to close out
failed deliveries by the 13th day after the trade.
In 2006, SEC Commissioner Campos (2006) maintained that Regulation SHOshould be
“targeted at protecting a small universe of thinly-capitalized securities from abusive
trading” a de?nition that would include most regional banks. Regulation SHO was
vigorouslychallengedbya securities industrythat hadgeneratedsigni?cant revenue from
short selling. The daily average number of stocks on the SHOthreshold list grewfrom300
issues in2007 to 400 for the ?rst nine months of 2008. Prior to the short-sell moratoriumthe
threshold list increased to over 500 issues and included many ?nancial stocks. In a 2009
letter posted on the SEC web site, former Commissioner Campos argued, “the majority of
these failures-to-deliver are not the result of honest mistakes or badprocessing rather these
companies are instead targets of illegal and manipulative trading.” (McGinty and
Strasburg, 2009) On September 18, 2008 the SEC implemented a new rule that required
short sellers who had sold shares but not delivered themto the buyer within three days of
the trade, to deliver shares by the start of trading on the fourth day. Following the
new rule and the short-sell moratorium, the daily average of stocks on the threshold list
declined to 100 issues in December of 2008 and 79 issues in March and April of 2009.
Although failure-to-deliver can be attributed to a variety of factors, naked shorting and
the inability to borrowshares was regarded as the most important reason for the late 2008
surge. Bradley et al. (2011) argued that regulator actions in September of 2008 largely
resolved delivery failures in the equity market; however, more recently failures have
increased in debt markets and among ETF shares. They argued that the potential
problems stemming from failures to deliver are much more serious than an increase in
replacement cost risk and maintain that delivery failures create a potentially cumulative
and compounding liquidity risk inthe ?nancial system. If Bradley et al. are correct and the
short-sell moratorium assisted regulators in reducing delivery failures among ?nancial
?rms, it might have prevented problems in the ?nancial system from worsening.
To justify the ban, the SEC asserted that short selling caused “sudden and excessive
?uctuations in the prices of [?nancial] securities.” To investigate these ?uctuations
Ulibarri et al. (2009) incorporated the effects that noise traders have on share prices and
found that the ban dampened volatility persistence in ?ve out of the 21 cases they
studied. In addition, they simulated the effects of an alternative uptick rule and found
that it had “disparate impacts on price stability.” The authors concluded that instead of
a blanket uptick rule on all ?rms, the SEC should instead examine speci?c securities on
a temporary basis, which would limit noise traders and still allow informed trading.
Despite some former regulators who have suggested that in hindsight the short-sell
moratorium was a policy failure, the precise impact of the policy on the risk/return of
smaller regional banks is largely unknown. We test several hypotheses concerning
regional bank risk and return. First, we hypothesize that that the CAPMmodel that does
not allow for a time varying risk component is insuf?cient to model regional bank
returns. Instead, we propose an integrated generalized autoregressive conditional
heteroskedasticity (IGARCH) model. Second, we construct six equally weighted
Short-sell
moratorium
effects
95
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
portfolios comprised of regional banks from the Mid-Atlantic (MA), Midwest (MW),
Northeast (NE), Southeast (SE), Southwest (SW), and Paci?c (PAC) regions and use
the IGARCH model to test for policy induced changes in return and risk. Third,
we re-estimate the IGARCH model for each of 110 individual banks. Then, we use the
estimates of policy induced changes in return/risk to test whether there are systematic
differences of the short-sell ban between different geographic regions. The next section
brie?y examines the literature on geographic risk in banking and constrained short
selling. Section 3 explains the data and models. Section 4 presents the results followed by
a ?nal section describing the policy implications and summarizing our results.
2. Bank geographic diversi?cation and constrained short-selling
In 1994 the Riegle-Neal Act eliminated state restrictions on interstate branch banking,
thus allowing greater risk/return bene?ts from geographic diversi?cation (Rivard and
Thomas, 1997; Deng and Elyasiani, 2008; Hughes et al., 1996). Akhigbe and Whyte
(2003) found a signi?cant decline in total risk and unsystematic risk following the
Riegle-Neal Act, while market risk was unchanged. However, Chong (1991) found that
although interstate branching enhanced pro?tability, it also increased risk. Similarly,
Demsetz and Strahan (1997) reported that bank holding companies offset the risk
reducing bene?ts of diversi?cation by increasing leverage and pursuing riskier and
potentially more pro?table lending opportunities.
The SEC attempt to stabilize the market capital of ?nancial institutions by enforcing a
short selling ban was highly controversial. Short selling has enjoyed a nefarious
reputation in the annals of ?nancial market history. Napoleon described the short seller as
“an enemy of the state” while England outlawed short selling in 1,733 and did not make it
legal again until the middle of the nineteenth century (Surowiecki, 2004). Following the
Great Crash of 1,929 short sellers were denounced on the Senate ?oor as “one of the great
commercial evils of the day.” Jones and Lamont (2002) described the anti-shorting climate
following the Great Crash as “hysterical.” Wearden (2010) contended that short sellers
exacerbatedthe Europeansovereigndebt crisis. However, defenders of the practice argued
that short sellers ?rst alerted investors to problems at Enron and Lehman Brothers and
that short selling results in a more ef?cient market.
The process of short sellingequitysecurities is thought toenhance market ef?ciencyby
incorporating information from pessimistic investors (Miller, 1977; Lamont, 2004b;
Brenner and Subrahmanyam, 2009). If true, banning short sales would exclude the
information those pessimistic investors provide, slow the dissemination of information,
andresult inanupwardbias tostockprices. Banningshort sales might alsocause liquidity
to decline since not only are short sellers banned, but some investors might forego long
positions in stocks if they cannot hedge their positions using short sales, thus reducing
trading volume. Diamond and Verrecchia (1987) argued that consistent with rational
expectations, market participants anticipate short sale constraints when formulating
pricing decisions; rather than mitigate volatility short-sell restrictions cause it to increase
and result in negative abnormal returns. Harrison and Kreps (1978) developed a model
where restrictingshort sales causes the price of the securityto exceedits valuationbyeven
the most optimistic current investors. Lamont and Thaler (2003) provided several case
studies of overvaluations that likely occurred as a consequence of dif?culty in short
selling. In a sample of 300 ?rms, Lamont (2004a) found that attempts to restrict short
selling caused the ?rms’ stocks to under-perform the market. In contrast, the predatory
JFEP
5,2
96
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
tradingmodel of Brunnermeier andPedersen(2005) indicatedthat aggressive short selling
can cause stock price declines greater than those implied by market ef?ciency.
The slow adjustment of prices to negative information results in a left skewness in
the distribution of returns, which should manifest as signi?cant abnormal negative
returns. Bris et al. (2007) compared stock market regulation among different countries
and conclude that prices do appear to incorporate negative information more slowly
in those countries where short selling is either not allowed or not generally practiced.
Reed (2007) found evidence of greater negative price reaction associated with
constrained short selling. If the rising cost of short sales is directly related to greater
price reaction when information ?nally becomes public, then regulations against short
selling will increase, rather than decrease volatility.
A growing body of literature on the 2008 ban on short sales in ?nancial stocks
found evidence of less liquidity since bid-ask spreads increased among banned stocks
and greater market inef?ciency given that stock and option prices decoupled in the
wake of the ban (Autore et al., 2011; Battalio and Schultz, 2011; Boehmer et al., 2011;
Jain et al., 2012; March and Payne, 2012). In this paper we contribute to the ongoing
research on the 2008 short-sell ban by investigating the effects of the ban on regional
bank risk and return using an IGARCH model.
3. Data
To examine the effects of the 2008 short-sell ban we examine 110 regional banks during
the ten year period January 2, 2002-December 31, 2011. This ten year period occurred
during a relatively slow growth phase of the US economy and includes the bubble years
in real estate prices and the subsequent bursting of the real estate price bubble and
ensuing ?nancial crisis. Regional banks have North America Industrial Classi?cation
(NAIC) code of 522110 and are part of the broader Finance and Insurance industry
(NAIC 52). Regional banks are smaller than money center banks and less able to achieve
economies associated with scale and geographic diversi?cation. Figure 1 shows that bank
branchclosings have a spatial dimensionwithclosings concentratedinCalifornia, Florida,
and Georgia. The 110 state banks chosenwere selected if they satis?ed three criteria. First,
the bank had to appear on the SEC short-sell restricted list[1]. Second, the bank was
required to have outstanding short interest in the week preceding the ban. We identify
banks with short interest fromShort-Squeeze (2012), a proprietary data base that provides
short interest for approximately 16,000 publicly traded stocks with outstanding short
positions. Short-Squeeze segments regional banks into six regions: Mid-Atlantic (MA),
Midwest (MW), Northeast (NE), Paci?c (PAC), Southeast (SE), and Southwest (SW). Third,
the bank had to be publicly traded for the ten year period fromJanuary 2002 to December
2011. In total, 110 regional banks met the three criteria and consisted of 21 banks from
the MA region, 19 banks from the MW region, 18 banks from the NE region, 18 banks
from the PAC region, 17 banks from the SE region, and 17 banks from the SW region.
Six regional bank portfolios were constructed. For each regional bank portfolio we
calculated the equally weighted daily excess return for the banks in the particular
region. Daily excess returns were estimated as the difference between the stock
return and the risk free return with the daily return on a 30 day T-bill serving as the
risk-free return. The market index is the daily excess return on the S&P500. The period
January 2, 2002-December 31, 2011 comprises 2,519 trading days which are used to
compute 2,518 days of excess returns.
Short-sell
moratorium
effects
97
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Descriptive statistics for January 2, 2002-December 30, 2011 appear in Panel A of
Table I (n ¼ 2518). Descriptive statistics include mean, standard deviation, skewness,
kurtosis and Jarque-Bera statistics. Banks in all of the regions as well as the market
index exhibited positive excess returns for the period. For illustrative purposes we
break the sample in half. Panel B shows descriptive statistics for the ?rst half period
January 2, 2002-December 29, 2006 while Panel C is fromJanuary 3, 2007 to December 30,
2011. Panel B incorporates the bubble years and shows mean daily returns for each
region approximately double the mean return for the entire sample. Panel C is for the
second half period that incorporates the bursting of the housing price bubble and
subsequent ?nancial crisis. Bank portfolios for the Paci?c, Southeast, and Southwest
regions as well as the market index show negative returns.
Bollerslev (1986) observed that return distributions for high frequency daily data
exhibit non-stationary variance along with leptokurtosis and are inconsistent with the
normal distribution. Sample period returns for each portfolio along with the market
index exhibit leptokurtosis and have Jarque-Bera statistics that reject the normality
assumption. The sample standard deviations for the second half period are more than
twice as high as the ?rst period, and appear to violate the classical linear regression
assumption of constant variance. One method for addressing the problems of
non-normal returns and a non-constant variance in returns is the generalized
autoregressive conditional heteroskedasticity (GARCH) model.
4. Time varying GARCH model
A critical issue in the application of the event method is the increase in stochastic
volatility surrounding the event period. Brown and Warner (1980, 1985) observed
that increases in the variance may cause misspeci?cation of the traditional test
statistics and that the power of the tests can be improved by appropriately modeling
the volatility process.
Brockett et al. (1999) developed an event-study method that assumes a market
model with GARCH effects and time-varying market risk. We use a GARCH model to
determine if the risk/return of regional bank portfolios were adversely impacted by the
short-sell constraint. Unlike traditional constant variance models that implicitly
assume that the event induced variance is the same for all portfolios in the sample,
the GARCH model allows event induced volatility to vary and for each portfolio’s
variance to be stochastic outside the event period. Given the observed increase in
market volatility and the premise that the ban may have had a differential impact on
banks from different regions, this allowance is a particularly important advantage of
GARCH relative to conventional OLS applications.
Time varying models have become an important tool in ?nance and economics.
The ARCH model was ?rst proposed by Engle (1982) and generalized to GARCH by
Bollerslev (1986) and extended to ARCH and GARCH in the mean (GARCH-M) by
Engle et al. (1987). The generalized GARCH model is represented by the following
system of equations:
y
t
¼ ax
t
þ1
t
v
t
¼ d
0
þ
i¼1
X
d
i
1
2
t2i
þ
i¼1
X
b
i
v
t2i
1
t
jV
t21
~
Nð0; v
t
Þ
ð1Þ
JFEP
5,2
98
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
S
&
P
5
0
0
M
i
d
-
A
t
l
a
n
t
i
c
M
i
d
-
W
e
s
t
N
o
r
t
h
e
a
s
t
P
a
c
i
?
c
S
o
u
t
h
e
a
s
t
S
o
u
t
h
w
e
s
t
P
a
n
e
l
A
:
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
s
f
o
r
t
h
e
s
i
x
r
e
g
i
o
n
a
l
b
a
n
k
p
o
r
t
f
o
l
i
o
s
a
M
e
a
n
0
.
0
0
0
0
8
0
.
0
0
0
5
4
0
.
0
0
0
3
0
0
.
0
0
0
3
4
0
.
0
0
0
4
8
0
.
0
0
0
4
4
0
.
0
0
0
4
2
S
D
0
.
0
1
3
8
7
0
.
0
1
3
3
5
0
.
0
1
6
7
5
0
.
0
1
3
4
1
0
.
0
2
0
8
1
0
.
0
1
3
2
8
0
.
0
1
4
0
4
S
k
e
w
n
e
s
s
0
.
0
3
3
3
.
2
7
3
1
.
0
0
7
0
.
1
7
9
1
.
3
1
2
0
.
2
3
8
0
.
4
1
6
K
u
r
t
o
s
i
s
1
1
.
3
3
9
.
4
1
5
.
8
9
.
8
1
8
.
7
9
.
8
1
1
.
4
J
a
r
q
u
e
-
B
e
r
a
7
,
2
6
4
.
4
1
4
3
,
2
2
8
.
0
1
7
,
7
0
6
.
0
4
,
8
6
2
.
3
2
6
,
6
4
2
.
0
4
,
8
5
4
.
9
7
,
5
6
0
.
7
P
a
n
e
l
B
:
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
s
f
o
r
t
h
e
s
i
x
r
e
g
i
o
n
a
l
b
a
n
k
p
o
r
t
f
o
l
i
o
s
b
M
e
a
n
0
.
0
0
0
1
5
0
.
0
0
0
7
0
0
.
0
0
0
4
3
0
.
0
0
0
5
6
0
.
0
0
0
9
2
0
.
0
0
0
8
0
0
.
0
0
0
6
4
S
D
0
.
0
1
0
1
6
0
.
0
0
5
5
0
0
.
0
0
7
9
5
0
.
0
0
6
7
1
0
.
0
0
7
8
7
0
.
0
0
7
0
3
0
.
0
0
7
4
9
S
k
e
w
n
e
s
s
0
.
3
2
1
2
0
.
0
0
5
2
0
.
0
5
2
2
0
.
0
0
7
0
.
2
3
3
0
.
0
1
2
0
.
0
2
1
K
u
r
t
o
s
i
s
6
.
2
4
.
4
3
.
5
3
.
4
4
.
4
3
.
3
3
.
8
J
a
r
q
u
e
-
B
e
r
a
5
7
2
.
0
1
0
2
.
3
1
1
.
4
7
.
3
1
1
3
.
5
5
.
1
3
2
.
9
P
a
n
e
l
C
:
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
s
f
o
r
t
h
e
s
i
x
r
e
g
i
o
n
a
l
b
a
n
k
p
o
r
t
f
o
l
i
o
s
c
M
e
a
n
0
.
0
0
0
0
1
0
.
0
0
0
3
9
0
.
0
0
0
1
6
0
.
0
0
0
1
1
0
.
0
0
0
0
3
0
.
0
0
0
0
8
0
.
0
0
0
2
0
S
D
0
.
0
1
6
7
9
0
.
0
1
8
0
5
0
.
0
2
2
3
1
0
.
0
1
7
7
3
0
.
0
2
8
3
5
0
.
0
1
7
4
1
0
.
0
1
8
3
9
S
k
e
w
n
e
s
s
2
0
.
0
2
5
2
.
6
7
3
0
.
8
7
1
0
.
1
8
7
1
.
0
7
6
0
.
2
6
2
0
.
3
9
9
K
u
r
t
o
s
i
s
9
.
7
2
3
.
6
1
0
.
0
6
.
3
1
0
.
9
6
.
6
7
.
7
J
a
r
q
u
e
-
B
e
r
a
2
,
3
7
7
.
3
2
3
,
7
4
6
.
6
2
,
7
5
2
.
7
5
9
6
.
2
3
,
5
2
9
.
0
6
8
0
.
3
1
,
1
9
4
.
5
N
o
t
e
s
:
a
J
a
n
u
a
r
y
2
,
2
0
0
2
-
D
e
c
e
m
b
e
r
3
0
,
2
0
1
1
,
n
¼
2
,
5
1
8
;
b
J
a
n
u
a
r
y
2
,
2
0
0
2
-
D
e
c
e
m
b
e
r
2
9
,
2
0
0
6
,
n
¼
1
,
2
5
8
;
c
J
a
n
u
a
r
y
3
,
2
0
0
7
-
D
e
c
e
m
b
e
r
3
0
,
2
0
1
1
,
n
¼
1
,
2
6
0
Table I.
Descriptive statistics
Short-sell
moratorium
effects
99
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where y
t
is return, x
t
is an exogenous or predetermined vector of variables, 1
t
is a random
error, and V
t21
is the information set. The parameter vectors are a, b
i
, d
0
, and d
i
and
t is a time index. The conditional variance, v
t
, is linearly dependent on the past behavior
of squared errors and a moving average of past conditional variances. The d
i
values
determine the weights attached to lagged innovations while the squared error term
implies that if innovations have been large in absolute value, they are likely to be large in
the future. Parameters b
i
, d
0
, and d
i
must be non-negative to ensure that the return
generating process is well de?ned.
If all coef?cients in the conditional variance equation (second equation of (1) are zero
(except the intercept) then the model reduces to the traditional constant variance
speci?cation. Both ARCH and GARCH model the conditional variance as a function of
past shocks and allow volatility to evolve and persist. The ARCH model incorporates a
limited number of lags in the derivation of the conditional variance and is a short
memory model. In contrast, the GARCH is a long memory model where past values of
the conditional variance and lagged errors in?uence the current period conditional
variance.
For the GARCHprocess to be stable, the effect of shocks to volatility as measured by
the sumof the coef?cients
P
i
d
i
þb
i
À Á
in the conditional variance equation must be less
than one. If the sum of these coef?cients is greater than or equal to one, Engle and
Bollerslev (1986) suggested using an IGARCH process where the constant term in the
conditional variance equation is suppressed and the restriction that
P
i
d
i
þb
i
¼ 1 is
imposed. To be persistent, the shocks must be signi?cantly different from zero. The
degree of persistence is important in determining the relation between volatility and
returns, since only persistent volatility justi?es changes in returns. The GARCH(1,1) is
more parsimonious relative to higher order models, allows for long memory in the
volatility process, andaccording to Bollerslev et al. (1992), ?ts most economic time series.
Based on Brockett et al. (1999) we estimate the following GARCH model for each of
the six portfolios of bank excess returns:
R
it
¼ a
i
þb
it
£ R
mt
þg
i
£ D
t
þl
i
£ R
it21
þh
it
h
it
¼ a
i
þ b
i
£ h
2
it21
þ c
i
£ h
it21
þ d
i
£ D
t
h
it
jV
t
~
Nð0; h
it
Þ
ð2Þ
where R
it
is the daily excess return on the ith regional bank portfolio, R
mt
is the daily
excess return on the S&P500 market index, and a
i
; b
it
; g
i
; a
i
; b
i
; c
i
; and d
i
are parameters to be estimated. Rather than a GARCH-M model that speci?es h
it
as the
risk measure in the return equation, this approach follows the more traditional market
model and allows for a time varying beta (b
it
). The indicator, D
t
equals 1 if t is an event
day, and 0 otherwise; and V
t
consists of all information available at time t, including all
the current and previous market excess returns, R
mu
, and security excess returns,
R
iu
, for u # t, the current and previous volatility estimates, h
iu
, for u # t, and current
and previous error estimates, h
iu
, for u # t. There are 15 trading days during the
short-sell ban that was in effect from September 19 to October 9, 2008.
During the event period, the estimate of g
i
gives the mean abnormal return. If the null
hypothesis of a zero mean abnormal return cannot be rejected, then the estimate of g
i
must be close to zero. The meaning of “close to zero” depends on the volatility of the
market model residual re?ected in h
it
which incorporates the event-induced variance via
JFEP
5,2
100
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the coef?cient d
i
. The GARCH residuals were subjected to diagnostic tests for the
presence of serial correlation. The Durbin-Watson statistic, Q-statistics and Lagrange
multiplier (LM) tests all indicated signi?cant serial correlation in the residuals of all six
models. The models were re-estimated including the autoregressive term (R
it21
).
The sumof the ARCHand GARCHcoef?cients (b
i
þ c
i
) was greater than or equal to one
indicating an unstable conditional variance. We re-estimated the models using
the IGARCHmethod of Engle and Bollerslev (1986). This method sets the intercept (a
i
) in
the conditional variance equation equal to zero and imposes the restriction that the
ARCH and GARCH coef?cients sum to one (b
i
þ c
i
¼ 1).
The results of the IGARCH models for the six regional portfolios appear in Table II.
Time varying betas ranged from0.34 for the portfolio of Mid-Atlantic banks to 0.77 for the
portfolio of Mid-West banks. The betas (
^
b
i
) for all portfolios were positive and
signi?cantly different from zero. The LM tests whether the residuals of the IGARCH
model dependonthe lagged values of the residuals. The LMtest has a x
2
distributionwith
degrees of freedom equal to the number of lagged squared residuals included as
independent variables. The null hypothesis is that laggedsquaredresiduals have no effect
on the current squared residual. Except for the portfolio of Southeast banks we cannot
reject the null hypothesis for the other ?ve regions. The results suggest that the regional
bank portfolio data for the period examined conform to the IGARCH speci?cation.
For the six regional portfolios the coef?cients for the event indicator variable, g
i
, in the
returnequations are negative, but insigni?cant. Thus, the banonshort sales hadno effect
on excess returns. The coef?cients on event-induced risk in the variance equation (d
i
)
was positive for ?ve of the six portfolios (the Paci?c portfolio had a negative and
insigni?cant coef?cient) but was signi?cantly different from zero only for the
Mid-Atlantic and the Northeast. Together, these estimates indicate that the short-sell ban
had no effect on risk or return in four of the six regional portfolios and increased risk for
only two of the six portfolios.
Earlier in the summer, from July 15 to August 12 of 2008, the SEC imposed a ban on
naked short selling in 19 ?nancial stocks. The small regional banks that we examine had
not been singled out in the July-August ban on naked short selling, but might have
experienced indirect effects via the signal that the SEC was sending ?nancial markets
about the condition and underlying speculation of large ?nancial ?rms. We re-estimated
the IGARCH model assuming that the event days, D
t
¼ 1, occurred July 15-August 12,
2008 and September 19-October 9, 2008. These two periods consist of 21 þ 15 ¼ 36
event days. The results are reportedinTable III. In the return equation, the intercepts (a
i
)
and slopes (b
i
) are little changed, but the effect of the event signi?cantly increased excess
returns for all six regional bank portfolios. In addition, three out of the six regional
portfolios (Mid-Atlantic, Northeast, and Southwest) experienced a signi?cant increase in
risk during the event period and two other portfolios (Mid-West and Paci?c) were
positive and signi?cant at the 10 percent level. Thus, the two short-sell ban periods
worked to prop up the portfolios of regional bank stocks, albeit with an increase in risk.
To further test the effects of the short-sell ban we re-estimated the IGARCH model
for eachof the 110regional banks witheachbankassignedtoits respectiveregion. Table IV
reports the meanestimates of the short-sell banondaily returns (
^ g) and ondaily risk (
^
d) for
banks ina particular region. Onlya small number of banks (betweenone and sixineach of
the six regions) have their returns signi?cantly impacted by the short-sell ban that
occurred from September 19 to October 9, 2008, but a majority of banks in all regions
Short-sell
moratorium
effects
101
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
C
o
e
f
f
.
a
,
b
M
i
d
-
A
t
l
a
n
t
i
c
M
i
d
-
W
e
s
t
N
o
r
t
h
e
a
s
t
P
a
c
i
?
c
S
o
u
t
h
e
a
s
t
S
o
u
t
h
w
e
s
t
a
i
0
.
0
0
0
2
4
*
(
0
.
0
0
0
1
0
)
0
.
0
0
0
0
2
(
0
.
0
0
0
1
1
)
0
.
0
0
0
1
3
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
6
1
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
4
2
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
3
2
(
0
.
0
0
0
1
2
)
b
i
0
.
3
4
1
4
3
*
(
0
.
0
0
8
6
9
)
0
.
7
7
7
0
3
*
(
0
.
0
1
1
1
9
)
0
.
6
2
0
9
3
*
(
0
.
0
0
9
9
6
)
0
.
7
0
8
6
3
*
(
0
.
0
1
2
2
2
)
0
.
6
0
8
5
0
*
(
0
.
0
1
0
0
8
)
0
.
6
9
3
0
2
(
0
.
0
1
1
3
7
)
g
i
2
0
.
0
0
0
3
3
(
0
.
0
0
2
9
8
)
2
0
.
0
0
0
7
2
(
0
.
0
0
9
5
2
)
2
0
.
0
0
2
9
9
(
0
.
0
0
5
7
6
)
2
0
.
0
0
9
2
7
(
0
.
0
0
9
8
3
)
2
0
.
0
0
0
5
8
(
0
.
0
0
3
6
7
)
2
0
.
0
0
2
2
4
(
0
.
0
0
6
2
4
)
l
i
2
0
.
0
5
2
3
8
*
(
0
.
0
1
4
2
7
)
2
0
.
0
9
4
8
*
(
0
.
0
1
0
7
0
)
2
0
.
0
5
5
5
5
*
(
0
.
0
1
2
7
9
)
2
0
.
0
5
7
2
0
*
(
0
.
0
1
3
0
6
)
2
0
.
0
8
9
4
3
*
(
0
.
0
1
2
9
9
)
2
0
.
0
9
3
8
2
(
0
.
0
1
2
0
2
)
b
i
0
.
0
1
3
4
3
*
(
0
.
0
0
0
3
5
)
0
.
0
6
6
5
2
*
(
0
.
0
0
4
1
9
)
0
.
0
4
5
1
1
*
(
0
.
0
0
3
4
2
)
0
.
0
5
5
7
9
*
(
0
.
0
0
3
4
9
)
0
.
0
5
7
3
3
*
(
0
.
0
0
3
3
3
)
0
.
0
4
1
4
6
*
(
0
.
0
0
3
0
4
)
c
i
0
.
9
8
6
5
7
*
(
0
.
0
0
0
3
5
)
0
.
9
3
3
4
8
*
(
0
.
0
0
4
1
9
)
0
.
9
5
4
8
9
*
(
0
.
0
0
3
4
2
)
0
.
9
4
4
2
1
*
(
0
.
0
0
3
4
9
)
0
.
9
4
2
6
7
*
(
0
.
0
0
3
3
3
)
0
.
9
5
8
5
4
*
(
0
.
0
0
3
0
4
)
d
i
0
.
0
0
0
0
1
*
(
3
.
3
1
£
1
0
2
6
)
0
.
0
0
0
0
3
(
0
.
0
0
0
0
2
)
0
.
0
0
0
0
3
*
(
0
.
0
0
0
0
1
)
2
2
.
2
5
£
1
0
2
6
(
0
.
0
0
0
0
3
)
0
.
0
0
0
0
2
(
0
.
0
0
0
0
2
)
0
.
0
0
0
0
2
(
8
.
9
7
£
1
0
2
6
)
A
R
C
H
L
M
-
t
e
s
t
c
0
.
1
1
(
0
.
7
5
)
0
.
1
1
(
0
.
7
4
)
3
.
1
5
(
0
.
0
8
)
0
.
8
9
(
0
.
3
5
)
3
5
.
1
6
(
0
.
0
1
)
1
.
1
0
(
0
.
2
9
)
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
t
a
t
:
*
5
p
e
r
c
e
n
t
;
J
a
n
u
a
r
y
2
,
2
0
0
2
-
D
e
c
e
m
b
e
r
3
0
,
2
0
1
1
c
o
e
f
?
c
i
e
n
t
s
a
n
d
(
s
t
a
n
d
a
r
d
e
r
r
o
r
s
)
;
a
t
h
e
e
x
c
e
s
s
r
e
t
u
r
n
e
q
u
a
t
i
o
n
i
s
:
R
i
t
¼
a
i
þ
b
i
£
R
m
t
þ
g
i
£
D
t
þ
l
i
£
R
i
t
2
1
þ
h
i
t
,
w
h
e
r
e
R
i
t
¼
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
f
o
r
b
a
n
k
i
,
R
m
t
¼
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
o
n
t
h
e
S
&
P
5
0
0
,
a
n
d
D
t
¼
1
f
o
r
t
h
e
?
f
t
e
e
n
t
r
a
d
i
n
g
d
a
y
s
f
r
o
m
S
e
p
t
e
m
b
e
r
1
9
,
2
0
0
8
t
o
O
c
t
o
b
e
r
9
,
2
0
0
8
,
0
o
t
h
e
r
w
i
s
e
;
b
t
h
e
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
e
q
u
a
t
i
o
n
i
s
:
h
i
t
¼
b
i
£
h
2i
t
2
1
þ
c
i
£
h
i
t
2
1
þ
d
i
£
D
t
;
b
i
þ
c
i
¼
1
,
w
h
e
r
e
h
i
t
¼
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
f
o
r
b
a
n
k
i
,
h
2i
t
2
1
i
s
t
h
e
l
a
g
g
e
d
s
q
u
a
r
e
d
r
e
s
i
d
u
a
l
f
r
o
m
t
h
e
r
e
t
u
r
n
e
q
u
a
t
i
o
n
,
h
i
t
2
1
i
s
t
h
e
l
a
g
g
e
d
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
,
a
n
d
D
t
¼
1
f
o
r
t
h
e
1
5
t
r
a
d
i
n
g
d
a
y
s
f
r
o
m
S
e
p
t
e
m
b
e
r
1
9
,
2
0
0
8
t
o
O
c
t
o
b
e
r
9
,
2
0
0
8
,
0
o
t
h
e
r
w
i
s
e
;
c
t
h
e
t
e
s
t
s
t
a
t
i
s
t
i
c
f
o
r
t
h
e
A
R
C
H
-
L
M
t
e
s
t
i
s
d
i
s
t
r
i
b
u
t
e
d
a
s
x
2
w
i
t
h
1
d
e
g
r
e
e
o
f
f
r
e
e
d
o
m
Table II.
IGARCH model estimates
for the six regional
bank portfolios
JFEP
5,2
102
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
C
o
e
f
f
.
a
,
b
M
i
d
-
A
t
l
a
n
t
i
c
M
i
d
-
W
e
s
t
N
o
r
t
h
e
a
s
t
P
a
c
i
?
c
S
o
u
t
h
e
a
s
t
S
o
u
t
h
w
e
s
t
a
i
0
.
0
0
0
2
1
*
(
0
.
0
0
0
1
0
)
0
.
0
0
0
0
2
(
0
.
0
0
0
1
1
)
0
.
0
0
0
1
1
(
0
.
0
0
0
1
1
)
0
.
0
0
0
6
0
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
4
0
*
(
0
.
0
0
0
1
1
)
0
.
0
0
0
3
0
*
(
0
.
0
0
0
1
2
)
b
i
0
.
3
4
2
3
5
*
(
0
.
0
0
8
7
4
)
0
.
7
7
7
4
6
*
(
0
.
0
1
1
1
7
)
0
.
6
1
9
9
7
*
(
0
.
0
0
9
8
7
)
0
.
7
1
0
2
1
*
(
0
.
0
1
2
2
7
)
0
.
6
0
8
9
2
*
(
0
.
0
1
0
0
6
)
0
.
6
9
3
9
6
*
(
0
.
0
1
1
4
5
)
g
i
0
.
0
0
4
8
8
*
(
0
.
0
0
1
7
4
)
0
.
0
0
8
7
4
*
(
0
.
0
0
4
1
3
)
0
.
0
0
6
1
7
*
(
0
.
0
0
2
4
2
)
0
.
0
1
0
5
6
*
(
0
.
0
0
4
8
9
)
0
.
0
0
4
6
1
*
(
0
.
0
0
2
3
0
)
0
.
0
0
6
5
7
*
(
0
.
0
0
3
2
5
)
l
i
2
0
.
0
5
2
4
6
*
(
0
.
0
1
4
4
8
)
2
0
.
0
9
5
3
8
*
(
0
.
0
1
0
8
4
)
2
0
.
0
5
7
0
0
*
(
0
.
0
1
2
8
3
)
2
0
.
0
5
6
8
9
*
(
0
.
0
1
3
2
6
)
2
0
.
0
9
0
4
0
*
(
0
.
0
1
2
8
3
)
2
0
.
0
9
3
7
0
*
(
0
.
0
1
2
5
8
)
b
i
0
.
0
1
2
0
6
*
(
0
.
0
0
0
3
8
)
0
.
0
6
5
4
4
*
(
0
.
0
0
4
3
2
)
0
.
0
4
4
3
8
*
(
0
.
0
0
3
4
2
)
0
.
0
5
4
8
7
*
(
0
.
0
0
3
5
6
)
0
.
0
5
7
1
7
*
(
0
.
0
0
3
3
3
)
0
.
0
4
0
4
7
*
(
0
.
0
0
3
3
3
)
c
i
0
.
9
8
7
9
4
*
(
0
.
0
0
0
3
8
)
0
.
9
3
4
5
6
*
(
0
.
0
0
4
3
2
)
0
.
9
5
5
6
2
*
(
0
.
0
0
3
4
2
)
0
.
9
4
5
1
3
*
(
0
.
0
0
3
5
6
)
0
.
9
4
2
8
4
*
(
0
.
0
0
3
3
3
)
0
.
9
5
9
5
3
*
(
0
.
0
0
3
3
3
)
d
i
0
.
0
0
0
0
1
*
(
1
.
3
1
£
1
0
2
6
)
0
.
0
0
0
0
1
(
7
.
8
3
£
1
0
2
6
)
7
.
8
2
£
1
0
2
6
*
(
3
.
1
3
£
1
0
2
6
)
0
.
0
0
0
0
3
(
0
.
0
0
0
0
2
)
4
.
8
2
£
1
0
2
6
(
4
.
2
4
£
1
0
2
6
)
0
.
0
0
0
0
1
*
(
3
.
4
5
£
1
0
2
6
)
A
R
C
H
L
M
-
t
e
s
t
c
0
.
1
4
(
0
.
7
9
)
0
.
0
5
(
0
.
8
2
)
3
.
9
7
(
0
.
0
5
)
0
.
7
3
(
0
.
3
9
)
3
7
.
0
(
0
.
0
1
)
0
.
7
7
(
0
.
3
8
)
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
t
a
t
:
*
5
p
e
r
c
e
n
t
;
J
a
n
u
a
r
y
2
,
2
0
0
2
-
D
e
c
e
m
b
e
r
3
0
,
2
0
1
1
c
o
e
f
?
c
i
e
n
t
s
a
n
d
(
s
t
a
n
d
a
r
d
e
r
r
o
r
s
)
;
a
t
h
e
e
x
c
e
s
s
r
e
t
u
r
n
e
q
u
a
t
i
o
n
i
s
:
R
i
t
¼
a
i
þ
b
i
£
R
m
t
þ
g
i
£
D
t
þ
l
i
£
R
i
t
2
1
þ
h
i
t
,
w
h
e
r
e
R
i
t
¼
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
f
o
r
b
a
n
k
i
,
R
m
t
¼
d
a
i
l
y
e
x
c
e
s
s
r
e
t
u
r
n
o
n
t
h
e
S
&
P
5
0
0
,
a
n
d
D
t
¼
1
f
o
r
i
f
t
h
e
t
r
a
d
i
n
g
d
a
y
o
c
c
u
r
s
b
e
t
w
e
e
n
J
u
l
y
1
5
,
2
0
0
8
a
n
d
A
u
g
u
s
t
1
2
,
2
0
0
8
o
r
f
r
o
m
S
e
p
t
e
m
b
e
r
1
9
,
2
0
0
8
t
o
O
c
t
o
b
e
r
9
,
2
0
0
8
,
0
o
t
h
e
r
w
i
s
e
;
t
h
e
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
e
q
u
a
t
i
o
n
i
s
:
h
i
t
¼
b
i
£
h
2i
t
2
1
þ
c
i
£
h
i
t
2
1
þ
d
i
£
D
t
;
b
i
þ
c
i
¼
1
,
w
h
e
r
e
h
i
t
¼
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
f
o
r
b
a
n
k
i
,
h
2i
t
2
1
i
s
t
h
e
l
a
g
g
e
d
s
q
u
a
r
e
d
r
e
s
i
d
u
a
l
f
r
o
m
t
h
e
r
e
t
u
r
n
e
q
u
a
t
i
o
n
,
h
i
t
2
1
i
s
t
h
e
l
a
g
g
e
d
c
o
n
d
i
t
i
o
n
a
l
v
a
r
i
a
n
c
e
,
a
n
d
D
t
¼
1
f
o
r
i
f
t
h
e
t
r
a
d
i
n
g
d
a
y
o
c
c
u
r
s
b
e
t
w
e
e
n
J
u
l
y
1
5
,
2
0
0
8
a
n
d
A
u
g
u
s
t
1
2
,
2
0
0
8
o
r
f
r
o
m
S
e
p
t
e
m
b
e
r
1
9
,
2
0
0
8
t
o
O
c
t
o
b
e
r
9
,
2
0
0
8
,
0
o
t
h
e
r
w
i
s
e
,
t
h
e
r
e
s
t
r
i
c
t
i
o
n
,
b
i
þ
c
i
¼
1
,
i
m
p
o
s
e
s
t
h
e
I
G
A
R
C
H
;
c
t
h
e
t
e
s
t
s
t
a
t
i
s
t
i
c
f
o
r
t
h
e
A
R
C
H
-
L
M
t
e
s
t
i
s
d
i
s
t
r
i
b
u
t
e
d
a
s
x
2
w
i
t
h
1
d
e
g
r
e
e
o
f
f
r
e
e
d
o
m
(
p
-
v
a
l
u
e
)
Table III.
IGARCH model estimates
for the six regional
bank portfolios
Short-sell
moratorium
effects
103
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
except the Mid-West experienced a signi?cant increase in risk. When the event period is
expanded to include July 15-August 12, 2008 and September 19-October 9, 2008, more
banks ineach regionexperience a signi?cant increase inreturn andmore banks infour out
of the six regions experience a signi?cant increase in risk due to the event.
Are there signi?cant differences in the effect of the ban on bank risk between the six
regions? Let the kernel density of the individual bank estimates of event induced risk,
^
d
i
,
withina particular regionbe f ð
^
d
Region
Þ andlet the densityof the individual bankestimates
of
^
d
i
for banks outside the region be gð
^
d
Outside region
Þ. We test the null hypothesis that
f ð
^
d
Region
Þ ¼ gð
^
d
Outside region
Þ using Li’s (1996) J
n1,n2
test which has a standard normal
distribution. Details of this test are foundinthe Appendix. Because of the small number of
banks within each region we follow the bootstrap approach described by Pagan and
Ullah (1999) with 500 bootstrap replications. The results of this test are reported in
Table V and indicate whether or not the short-sell ban resulted in systematic regional
effects on individual bank risk. The short-sell ban for the period September 19 to October
9 had a signi?cant regional impact on risk for banks in the Northeast region ( p ¼ 0.05)
and for banks in the Southwest region ( p ¼ 0.01), but no signi?cant impact on banks in
the other four regions. When the event periodis expanded to include the periods July 15 to
August 12 andSeptember 19 to October 9, there is a signi?cant increase inriskfor ?ve out
of the six regions, with the exception being the Mid-Atlantic region. We also performed a
similar battery of tests for regional differences in event induced changes in return,
but found no signi?cant differences across the six regions.
Despite evidence which indicates that the short-sell ban increased risk, it may have
substantially reduced regional bank failures-to-deliver. Figure 2 graphs the cumulative
value of failures-to-deliver for the 110 regional banks from April 2007 through
No. of
banks
^ g
Region
No. of banks with ^ g – 0
at p , 0.10
^
d
Region
No. of banks with
^
d – 0
at p , 0.10
Event period ¼ September 19-October 9, 2008
Mid-Atlantic 21 0.00027 2 0.00029 14
Mid-West 19 0.00043 1 0.00020 9
Northeast 18 20.00273 6 0.00008 12
Paci?c 18 20.00222 3 0.00011 10
Southeast 17 0.00073 3 0.00021 14
Southwest 17 0.01200 1 0.00013 12
Event period ¼ July 15-August 12, 2008 and September 19-October 9, 2008
Mid-Atlantic 21 0.00402 4 0.00011 17
Mid-West 19 0.01041 8 0.00007 13
Northeast 18 0.00587 8 0.00015 13
Paci?c 18 0.01318 7 0.00014 13
Southeast 17 0.00560 9 0.00008 13
Southwest 17 0.00768 7 0.00007 11
Notes: The event induced change in return for bank i is captured by the estimated coef?cient ^ g
i
; the event
induced change in risk is for bank i is captured by the estimated coef?cient
^
d
i
; event induced changes in
return and risk are estimated from the excess returns, R
it
¼ a
i
þb
i
£ R
mt
þg
i
£ D
t
þl
i
£ R
it21
þh
it
,
and conditional variance, h
it
¼ b
i
£ h
2
it21
þ c
i
£ h
it21
þ d
i
£ D
t
; b
i
þ c
i
¼ 1, estimated from an
IGARCHmodel, where R
it
¼daily excess return for bank i, R
mt
¼daily excess return on the S&P500, h
it
¼
conditional variance for bank i, h
2
it21
is the lagged squared residual from the return equation, h
it21
is the
laggedconditional variance, andD
t
¼1 for if the tradingdayoccurs duringthe event periodand0 otherwise
Table IV.
Effects of the short-sell
ban on individual bank
returns ( ^ g) and risk (
^
d)
JFEP
5,2
104
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
December 2011 (Source: US SEC: Fails to Deliver Data-Archive Data (2012)). For a given
bank, the value of failures-to-deliver securities equals the product of the security prices
and the cumulative net failures-to-deliver securities up to that date. We then sum these
values over the 110 banks to get the cumulative value of failures to deliver. Figure 2 shows
that in the 18 months before and including September of 2008 the cumulative value
Event induced change in risk
event ¼ September 19-October 9,
2008
Event induced change in risk
event ¼ July 15-August 12, 2008
and September 19-October 9, 2008
f ðd
Region
Þ ¼ gðd
other regions
Þ
a
f ðd
Region
Þ ¼ gðd
other regions
Þ
Null hypothesis
Bootstrapped J
n1,n2
b
(Prob . J
n1,n2
)
Bootstrapped J
n1,n2
(Prob . J
n1,n2
)
Mid-Atlantic vs other regions 1.31 (0.19) 1.00 (0.32)
Mid-West vs other regions 1.51 (0.13) 4.07 (0.01)
Northeast vs other regions 1.97 (0.05) 20.11 (0.01)
Paci?c vs other regions 0.80 (0.42) 4.06 (0.01)
Southeast vs other regions 1.51 (0.13) 6.78 (0.01)
Southwest vs other regions 3.93 (0.01) 7.88 (0.01)
Notes:
a
The kernel density estimates (empirical distribution functions) of event induced changes in
risk (d ) are given as f(d
Region
) for banks within a particular region; for banks outside the region, the
kernel density estimates of event induced changes in risk are given as g(d
other regions
); the estimates of
event induced changes in risk are from the conditional variance equation of the GARCH model;
b
the
test statistic J
n1,n2
has a standard normal distribution, where there are n
1
bank observations within the
region and n
2
banks observations outside the region; the number of bootstrap replications is 500
Table V.
Tests for differences in
risk between regions,
2 January 2002-
30 December 2011
Figure 2.
Value of failures to deliver
securities for 110 banks
A
p
r
-
0
7
A
p
r
-
0
8
A
p
r
-
0
9
A
p
r
-
1
0
A
p
r
-
1
1
J
u
l
-
0
7
J
u
l
-
0
8
J
u
l
-
0
9
J
u
l
-
1
0
J
u
l
-
1
1
O
c
t
-
0
7
O
c
t
-
0
8
O
c
t
-
0
9
O
c
t
-
1
0
O
c
t
-
1
1
J
a
n
-
0
9
J
a
n
-
1
0
J
a
n
-
1
1
J
a
n
-
0
8
0
500,000,000
1E+09
1.5+09
2E+09
2.5E+09
3.5E+09
3E+09
Short-sell
moratorium
effects
105
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
of failures had increased sharply to approximately $3 billion, but then declined in the
wake of the short-sell ban to remain at less than $130 million each month.
5. Conclusions
This research utilized an IGARCH version of the market model to test for event induced
volatility and excess returns as a consequence of the SEC mandate that restricted short
sales on approximately 800 ?nancial stocks. 110 regional banks were identi?ed as having
outstanding short interest and used to construct equally weighted regional bank
portfolios. The results indicate that the September 19-October 9 short-sell banhadnoeffect
onreturns, but a small positive effect onriskintwo out of the sixregions (Mid-Atlantic and
Northeast). The sample of regional banks was not legally subject to the July 15-August 12
ban that applied to 19 “too big to fail” institutions, however, the regional portfolios were
tested for evidence of “signaling” and found signi?cant positive effects in each of the six
regions. The two short-sell bans were associated with an increase in risk for the portfolios
of banks comprising the Mid-Atlantic and Northeast regions.
The IGARCH model was also estimated for each of the 110 regional banks. For the
expanded event period we found that 43 banks experienced a signi?cant event induced
change in return and 80 banks experienced a signi?cant event induced increase in risk.
The expanded event period found that ?ve of the six regions exhibited signi?cant
event induced changes in risk. The results suggest that rather than mitigate volatility
in the equity prices of regional banks, the SEC short-sell moratorium contributed to
bank risk which could be attributed to declines in trading volume and liquidity
identi?ed by Autore et al. (2011), Battalio and Schultz (2011), Jain et al. (2012) and
March and Payne (2012).
The results are consistent with Diamond and Verrecchia’s (1987) model, which
predicts that market participants anticipate the rising costs of constrained short
selling causing a greater price reaction. Despite ?nding signi?cant event induced
effects on individual bank risk and return, the magnitudes were small. Whether or not
sophisticated investors were able to engage in regulatory arbitrage from the policy,
a circumstance that has been described by Lowenstein (2000) as “picking up nickels in
front of a steamroller,” is not clear.
The lifting of the SEC short-sell restriction after a scant three weeks is interpreted as
an admission that the policy failed to achieve the stated result. In his assessment of the
ban, former SEC Chairman Christopher Cox said on December 31, 2008: “Knowing what
we know now, [we] would not do it again. The costs appear to outweigh the bene?ts.”
(Younglai, 2008) Our ?nding of event induced volatility would appear to substantiate
this interpretation.
However, in light of “too big to fail” and the relatively modest impact of the ban on
regional bank risk/return we offer an alternative assessment that judges the policy less
harshly. If delivery failures caused primarily by naked shorting pose a potentially
compounding and cumulative liquidity risk to the ?nancial system as argued by
Bradley et al. (2011), then the short-selling ban along with the four day delivery rule
may have provided a “cooling off period” that helped reduce that risk. In turn, Figure 2
shows that the ban might have helped to substantially reduce delivery failures among
regional bank stocks and the corresponding liquidity risk. If the ?nancial crisis has
taught us anything it is that ?nancial markets and institutions are far more complex
than many of us believed.
JFEP
5,2
106
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Note
1. On Friday, September 19, 2008, the SEC Released an Emergency Order prohibiting short sales
in 799 securities (www.sec.gov/rules/other/2008/34-58592.pdf). On the following Monday the
order was amended with important material differences: www.sec.gov/news/press/2008/
2008-218.htm that included the SEC’s delegation of authority to each national securities
exchange to identify and add listed companies that qualify for inclusion on the covered list. As
the additions and removals on the list are subject to change on a daily basis, the NYSE has
maintained and provided daily end-of-day list updates to the covered list since the amended
order went into effect.
As of Friday, September 26, 2008 the NYSE and NYSE Arca short-sell prohibition list
had been enhanced as part of a collaborative effort with NASDAQ and AMEX, to include
securities that will be covered for the next trading day from all four exchanges in one
convenient document. Daily updates are provided at: www.nyse.com/attachment/
CONSOLIDATED-SSPROHIBTION.xls or once can subscribe to NYSE System Status
Trader e-mail updates: www.nysenet.com/subscription/smLogin to receive notice as soon as it
becomes available along with other NYSE Cash Equities Market System and Trading
Information news.
References
Akhigbe, A. and Whyte, A. (2003), “Changes in market assessments of bank risk following the
Riegle-Neal Act of 1994”, Journal of Banking & Finance, Vol. 27, pp. 87-102.
Autore, D.M., Billingsley, R.S. and Kovacs, T. (2011), “The 2008 short sale Ban: liquidity,
dispersion of opinion, and the cross-section of returns of US ?nancial stocks”, Journal of
Banking & Finance, Vol. 35 No. 9, pp. 2252-2266.
Battalio, R. and Schultz, P. (2011), “Regulatory uncertainty and market liquidity: the 2008 Short
Sale Ban’s impact on equity option markets”, The Journal of Finance, Vol. 66 No. 6,
pp. 2013-2053.
Boehmer, E., Jones, C.M. and Zhang, X. (2011), “Shackling short sellers: the 2008 shorting Ban”,
available at:http://ssrn.com/abstract¼1412944
Bollerslev, T. (1986), “Generalized autoregressive conditional heteroscedasticity”, Journal of
Econometrics, Vol. 31, pp. 307-327.
Bollerslev, T., Chou, Y. and Kroner, K. (1992), “ARCH modeling in ?nance: a review of the theory
and empirical evidence”, Journal of Econometrics, Vol. 52, pp. 5-59.
Bradley, R., Fawls, H., Litan, R. and Sommers, F. (2011), Canaries in the Coal Mine:
How Settlement Fails Create Systemic Risk for Financial Firms and Investors,
Ewing Marion Kauffman Foundation, Kansas City, MO.
Brenner, M. and Subrahmanyam, M. (2009), “Short selling”, in Acharya, V. and
Richardson, M. (Eds), Restoring Financial Stability: How to Repair a Failed System,
Wiley, New York, NY.
Bris, A., Goetzmann, W. and Zhu, N. (2007), “Ef?ciency and the bear: short sales and markets
around the world”, Journal of Finance, Vol. 62 No. 3, pp. 1029-1079.
Brockett, R., Chen, H. and Garven, J. (1999), “A new stochastically ?exible event methodology
with application to proposition 103”, Insurance, Mathematics and Economics, Vol. 25,
pp. 197-217.
Brown, S. and Warner, J. (1980), “Measuring security price performance”, Journal of Financial
Economics, Vol. 8, pp. 205-258.
Brown, S. and Warner, J. (1985), “Using daily stock returns: the case of event studies”, Journal of
Financial Economics, Vol. 14, pp. 3-31.
Short-sell
moratorium
effects
107
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Brunnermeier, M.K. and Pedersen, L.H. (2005), “Predatory trading”, Journal of Finance, Vol. 60,
pp. 1825-1863.
Campos, R. (2006), Speech by SEC Commissioner: Remarks on Regulation SHO, US Securities and
Exchange Commission, Washington, DC, July 12.
Chong, B.S. (1991), “Effects of interstate banking on commercial banks risk and pro?tability”,
Review of Economics and Statistics, Vol. 73, pp. 78-84.
Cochrane, J. (2011), “The fed’s mission impossible”, The Wall Street Journal, December 29.
Demsetz, R.S. and Strahan, P.E. (1997), “Diversi?cation, size and risk at bank holding
companies”, Journal of Money, Credit and Banking, Vol. 29, pp. 300-313.
Deng, S. and Elyasiani, E. (2008), “Geographic diversi?cation, bank holding company value,
and risk”, Journal of Money, Credit and Banking, Vol. 40, pp. 1217-1238.
Diamond, D. and Verrecchia, R. (1987), “Constraints on short selling and asset price adjustments
to private information”, Journal of Financial Economics, Vol. 18, pp. 277-311.
Engle, R.F. (1982), “Autoregressive conditional heteroscedasticity with estimates of the variance
of UK in?ation”, Econometrica, Vol. 50, pp. 987-1007.
Engle, R.F. and Bollerslev, T. (1986), “Modeling the persistence of conditional variances”,
Econometric Reviews, Vol. 5, pp. 1-50.
Engle, R.F., Lilien, D. and Robins, R. (1987), “Estimating time varying term premium in the term
structure: the ARCH-M model”, Econometrica, Vol. 55, pp. 391-407.
Federal Deposit Insurance Corporation (2012a), Failed Bank List, available at: www.fdic.gov/
bank/individual/failed/banklist.html (accessed March 1, 2012).
Federal Deposit Insurance Corporation (2012b), Summary of Deposits, available at:http://www2.
fdic.gov/sod/ (accessed March 1, 2012).
Financial Stability Board (2011), Policy Measures to Address Systematically Important Financial
Institutions, Financial Stability Board, Basel, November 4.
Harrison, J.M. and Kreps, D.M. (1978), “Speculative behavior in a stock market with
heterogeneous expectations”, Quarterly Journal of Economics, Vol. 92 No. 2, pp. 323-336.
Hughes, J.P., Lang, W., Mester, L. and Moon, C. (1996), “Ef?cient banking under interstate
branching”, Journal of Money, Credit and Banking, Vol. 28, pp. 1045-1071.
Jain, C., Jain, P. and McInish, T.H. (2012), “Short-selling: the impact of SEC rule 201 of 2010”,
The Financial Review, Vol. 47 No. 1, pp. 37-64.
Jones, C. and Lamont, O. (2002), “Short sale constraints and stock returns”, Journal of Financial
Economics, Vol. 66, pp. 207-239.
Lamont, O. (2004a), “Go down ?ghting: short sellers vs ?rms”, National Bureau of Economic
Research Working Paper 10659, August.
Lamont, O. (2004b), Short Sale Constraints and Overpricing: The Theory and Practice of Short
Selling, Frank J. Fabozzi Associates, New Hope, PA.
Lamont, O. and Thaler, R. (2003), “Can the market add and subtract: mispricing in tech stock
carve-outs”, Journal of Political Economy, Vol. 111, pp. 227-268.
Li, Q. (1996), “Nonparametric testing of closeness between two unknown distribution functions”,
Econometric Reviews, Vol. 15, pp. 261-274.
Lowenstein, R. (2000), When Genius Failed: The Rise and Fall of Long-term Capital Management,
Random House, New York, NY.
McGinty, T. and Strasburg, J. (2009), “Short sellers squeezed all around”, The Wall Street Journal,
April 7, p. C1.
JFEP
5,2
108
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
March, I.W. and Payne, R. (2012), “Banning short sales and market quality: the UK’s experience”,
Journal of Banking & Finance, Vol. 36 No. 7, pp. 1975-1986.
Miller, E. (1977), “Risk, uncertainty and divergence of opinion”, Journal of Finance, Vol. 32,
pp. 1151-1168.
Pagan, A. and Ullah, A. (1999), Nonparametric Econometrics, Cambridge University Press,
Cambridge.
Reed, A. (2007), “Costly short selling and stock price adjustments to earnings announcements”,
Unpublished manuscript, University of North Carolina, Chapel Hill, NC.
Rivard, R. and Thomas, C. (1997), “The effect of interstate banking on large bank
holding company pro?tability and risk”, Journal of Economics and Business, Vol. 49,
pp. 61-76.
Short Squeeze (2012), available at: www.shortsqueeze.com/
Surowiecki, J. (2004), The Wisdom of Crowds, Doubleday, New York, NY.
Ulibarri, C., Florescu, I. and Eidsath, J.M. (2009), “Noise traders and short-selling troubled ?rms”,
Journal of Financial Economic Policy, Vol. 1 No. 3, pp. 227-245.
US SEC (2008a), “Emergency order pursuant to section 12(k)(2) of the securities and exchange act
of 1934 taking temporary action to respond to market developments”, Release 34-58592
September 18, 2008, US Securities and Exchange Commission.
US SEC (2008b), “SEC halts short selling of ?nancial stocks to protect investors and markets”,
For Immediate Release 2008-211, US Securities and Exchange Commission, available at:
www.sec.gov/news/press/2008/2008-211.htm (accessed February 15, 2012).
US SEC (2010), “Alternative up-tick rule”, Amendment to Regulation SHO, US Securities and
Exchange Commission, February 24, available at: www.sec.gov/news/press/2010/2010-26.
htm
US SEC (2012), “Fails to deliver data”, Archive data, US Securities and Exchange Commission,
available at: www.sec.gov/foia/docs/failsdata-archive.htm
Wallison, P. (2011), “How regulators herded banks into trouble”, The Wall Street Journal,
December 3.
Wearden, G. (2010), “European debt crisis: markets fall as Germany bans naked shorting”,
The Guardian, May 20, p. A1.
Weber, W.L. and Devaney, M. (1999), “Bank ef?ciency, risk-based capital, and real estate
exposure: the credit crunch revisited”, Real Estate Economics, Vol. 27, pp. 1-25.
Younglai, R. (2008), SEC Chief Has Regrets over Short Selling Ban, Reuters US edition,
December 31, available at: www.reuters.com/article/idUSTRE4BU3GG20081231
Further reading
Congressional Oversight Panel (2010), Commercial Real Estate Losses and the Risk to Financial
Stability, available at:http://cop.senate.gov/reports/library/report-021110-cop.cfm
Appendix
Let random variables x and y have estimated density functions given by
^
fðxÞ and ^ gð yÞ. The
standard normal kernel function is given as kðCÞ ¼ ð2pÞ
20:5
expð20:5CÞ where C ¼ ðx
j
2xÞ=h
with h representing the bandwidth or smoothing parameter. The kernel function has the
property that
R
1
21
kðCÞdC ¼ 1. Qi Li’s test for the difference between two distributions
^
fðxÞ and
^ gð yÞ is based on the integrated squared error I ¼
R
x
ð f ðxÞ 2gðxÞÞ
2
dx. With sample size of n
1
for x
and n
2
for y, Li’s J
n1,n2
test is given as:
Short-sell
moratorium
effects
109
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
J
n1;n2
¼
n
1
h
0:5
I
^ s
, Nð0; 1Þ;
where:
I ¼
1
h
X
n1
i¼1
X
n2
j ¼1;
j – i
1
n
1
ðn
1
21Þ
K
x
ij
þ
1
n
2
ðn
2
21Þ
K
y
ij
2
1
n
1
ðn
2
21Þ
K
x;y
ij
2
1
n
2
ðn
1
21Þ
K
y;x
ij
and:
^ s
2
¼
2
h
X
n1
i¼1
X
n2
j¼1
1
n
2
1
K
x
ij
þ
l
2
n
2
2
K
y
ij
þ 2
l
n
1
n
2
K
x;y
ij
for l ¼
n
1
n
2
:
In the above formulation we use K
x
ij
¼ kððx
i
2x
j
Þ=h), K
y
ij
¼ kðð y
i
2y
j
Þ=h), K
x;y
ij
¼ kððx
i
2y
j
Þ=h),
and K
y;x
ij
¼ kðð y
i
2x
j
Þ=h). The optimal bandwidth is found as h ¼ 1:06 £ min s
1
=n
0:2
1
; s
2
=n
0:2
2
À Á
where s
i
is the standard deviation of random variable i ¼ x, y. Pagan and Ullah (1999) provide
an in depth discussion of kernel density estimation and the bootstrap method we use for Li’s
J
n1;n2
test.
About the authors
Michael Devaney is Professor of Finance at Southeast Missouri State University where he
teaches courses in Financial Institutions and Real Estate Finance. His research examines the
effects of public policy changes for various ?nancial institutions including US and Japanese
banks and real estate investment trusts.
William L. Weber is Professor of Economics at Southeast Missouri State University where he
teaches courses inFinancial Institutions, Econometrics, andEnvironmental Economics. His research
examines productivity growth when ?rms produce desirable and undesirable outputs, such as
pollution or risk and the risk/return tradeoff for US and Japanese banks, Japanese securities ?rms,
credit cooperatives, andinsurance companies. WilliamL. Weber is the correspondingauthor andcan
be contacted at: [email protected]
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
JFEP
5,2
110
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_561483486.pdf