This is a product assessment for Full Size. Many shoppers don"t straightaway think of eBay when it comes to Full Size, but in fact eBay is amongst the top three retailers in the nations marketplace. Be amazed with the items you will discover here at War Auction.
republic of vietnam campaign full size medal made in usa
us army good conduct full size medal made in usa in original packaging
us national defense full size medal made in usa in original packaging
military us air force badge senior communications full size badge
air force chevrons stripes full size abu cmsgt e 9 1398 free shipping
air force chevrons stripes full size desert cmsgt e 9 1398 free shipping
air force chevrons stripes full size full color cmsgt e 9 1998 free shipping
badge air force master missile operator mirror finish full size 2294 shipped
badge air force chief flight nurse full size mirror finish 1398 free shipping
air force chevrons stripes full size subdued cmsgt e 9 1398 free shipping
defense distinguished service medalfull size
us army achievement medal full size
nip us military basic parachutist wings badge bright metal full size
asiatic pacific campaign combat service full size military medal made in usa
us armydivers badge set stay brite full size set of 5
national guard achievement medal full size
canadian forces airborne para red leaf cotton wing full size
nebraska national guard faithful service medal full size
usmarine corps officers hat emblem full sizescrew back
usn navy marine corps marines fleet marine full size silver qualification badge
usn navy uss ship sub submarine warfare full size silver qualification badge p
usn navy uss ship sub submarine warfare full size silver qualification badge
usn navy uss ship shore air warfare full size silver qualification badge
usn us navy information dominence spooks full size silver qualification badge
usn navy uss ship surface warfare full size silver qualification badge
usn us navy seal team warfare full size gold qualification badge
usn navy uss ship sub submarine warfare full size gold qualification badge
usn navy uss ship surface warfare suppo full size gold qualification badge
usn navy uss ship shore carrier air observer full size gold qualification badge
usn navy uss ship shore carrier air aviator full size gold qualification badge
usn navy uss ship shore carrier air aviator full size gold qualification badge 2
usn navy uss ship shore carrier balloon pilot full size gold qualification badge
usn navy ship shore carrier astronaut pilot full size gold qualification badge
usn navy ship shore carrier space astronaut full size gold qualification badge
usn navy uss ship shore sub deep submergence full size gold qualification badge
usn navy uss ship shore sub deep submergence full size silv qualification badge
usn navy ship shore air ebola medical ready full size gold qualification badge
usn navy ship shore air ebola medical ready full size silver qualification badge
usn us navy uss ship shore air carreer counselor enameled full size pocket badge
us air force senior remotely piloted aircraft sensor operator badge full size
united states southern commandstaff id breast badge full size
us air force master remotely piloted aircraft pilot badge full size
us coast guard taclet tactical law enforcement qual pinfull size
iraq campaign medal current full size with ribbon
us armed forces global war on terrorism expeditionary medal full size w ribbon
usn us navy ship uss arizona bb39 battleship full size ballcap badge
usn us navy ship uss damage control full size ballcap badge
usn us navy ship uss aviation antisub warfare operator full size ballcap badge
usn us navy ship uss electronics technician full size ballcap badge
usmc full size nato article 5 active endeavour medal anodized by officers equip
usmc full size national defense service medal anodized by officers equipment co
army commanders award for public service full size military medal usa made
army commendation full size military medal usa made
army distinguished service medal full size military medal usa made
army exceptional civilian service medal full size military medal usa made
army exceptional civilian service medal 3 piece full size ribbon boxed set
genuine us military full size medal global war on terrorism service new
navy seal badge trident insignia bud gold full size1 inch
meritorious service medal 2 piece full size cased
us armed forces humanitarian service medal full size medal and ribbon 2050
full size riddell football helmet with navy seals logo
full size us medals army commendation armed forces reserve american campaign
lot of 5 coast guard auxiliary full size medal ribbons
national guard medalvirgin islandsmeritorious service medalfull size
operation iraqi freedom unissued us iraq campaign medal new in box full size
usaf operational support medal full size newly authorized 2009
us coast guard coast guardsmen medal full sizefor heroism
us navy us navy flight officer astronaut qualification pin badgefull size
usaf smsgt first sergeant rank abu full size 1 pair
usaf msgt first sergeant rank abu full size 1 pair
usaf msgt master sergeant rank abu full size 1 pair
us army full size male sergeant first class rank e 7 class a pair lot 8
us army full size male sergeant first class rank e 7 class a pair lot 7
us army full size male sergeant first class rank e 7 class a pair lot 6
us army full size male sergeant first class rank e 7 class a pair lot 5
us army full size male sergeant first class rank e 7 class a pair lot 4
us army full size male sergeant first class rank e 7 class a pair lot 3
us army full size male sergeant first class rank e 7 class a pair lot 2
us army full size male sergeant first class rank e 7 class a pair lot 1
vanguard anodized full size national defense medal new
us army senior pilot wings badge bright metal full size
nevada full size military medal of valor made in usa
nevada full size military medal of merit made in usa
nevada commendation full size military medal made in usa
nevada distinguished service full size military medal made in usa
sapper tab full size dress blues
ghana parachutist badge for wear on usmilitary dress uniform full size
full size us service in korea korean merchant marine miltary medal orig pkg
wwii us american theater campaign service medal ribbon full size ml
us cold war victory medal service ribbon full size pin back ml
sheaffer 1002 targa full size black matt ct fp steel medium mint
vintage solid brass full size bird cage with handsome wood finial
sheaffer targa full size steel fp medium w converter mint lable usa
sheaffer targa 1000 full size chrome fp medium w converter brand new lable usa
bic classic full size lighters 40 count assorted
doctor doom statue diamond select full size fs damaged fantastic four
star wars lucasfilm boba fett full size electronic helmet prop
ww2 german panzerfaust 60mm anti tank rock steel full size toy replica wwii
wwii usn us coast guard pilot wings vanguard ny full size gold gilt ww2 d316
old patchwork star hand sewn early mid 1900s 82x68 light green
orgy horn beer drinkers gag gift full size you have to finish your drink
thor hard hero statue avengers rare mps edition marvel comics full size
marvel the mighty thor full size statue 3000 new in box no reserve
mercury hub caps dog dish police 60s 70s full size cars
snk neo geo full size 27x28 red arcade monitor bezel nos
clone trooper phase 2 rots 11 full size helmet kit prop resin
touhou project koishi komeiji top grade cosplay costume full size
vintage antique north brunswick twp nj special police full size obsolete badge
rush the rock alcatraz full size arcade sit down driving game works 100 nice
ka bar 1217 usmc full size straight edge knife w leather sheath
nib avon 2004 mrs albee award full size figurine presidents club
creative license elektra statue 1996 rare full size mib
hand made tobacco pipe model 53 heavy brown shades full size bowl
hand made tobacco pipe model 53 heavy mahogany full size bowl
harley davidson full size knapp sack
vtg cannon floral full size sheet set w 2 pillowshams black purple red yellow
corona full size football helmet snack beer holder
amazing vintage full size all brass childs rocking horse 26 high 44 pounds
Private information and the exercise of executive stock options.
This paper finds strong evidence that executives use private
information when exercising their stock options. The most informed
executives tend to exercise early, do not exercise on the vest date, do
not exercise to capture dividends, exercise a high percentage of their
options, and exercise when the option is the least in-the-money. We also
find that exercises around resignation and retirement are followed by
. Furthermore, the operating
performance affirms following exercises
tr.v. mo·ti·vat·ed, mo·ti·vat·ing, mo·ti·vates
To provide with an incentive; move to action; impel.
by private information
is significantly worse than that of firms in which the exercises are not
motivated by private information.
The use of private information by executives in conducting stock
transactions has been an important and controversial issue to academics,
practitioners, and especially regulators. Stock transactions, however,
are not the only means by which executives can exploit private
information. The extensive use of stock options as compensation and
incentives provides executives with another means of exploiting private
information. In particular, the early exercise of stock options is one
obvious means by which holders can do so. There is substantial evidence
in the literature that a considerable number of stock options are
exercised early. If the executive has negative information, the stock
would then almost surely be sold, and in all likelihood, the stock would
perform poorly for a period of time thereafter. This paper examines the
exercise of executive stock options in which the stock is sold to
determine if these exercises are consistent with the possession of
private information by the executive.
Previous empirical evidence on the use of private information in
exercising stock options is somewhat mixed. The reason could be that
there are a number of justifications for exercising stock options, many
of which have little to do with private information. For example, stock
options are sometimes exercised to simply reduce the executive's
exposure in the firm. Stock options may also be exercised for the
principal reason why traded stock options are exercised early: to
capture dividends. Stock options could also be exercised shortly before
an executive leaves the firm, and, thus, may not be motivated by private
information. In addition, stock option exercises that are associated
with private information are likely to involve the exercise of a large
percentage of exercisable options. By segregating exercises likely to be
motivated by private information from those that are not, we can better
determine whether private information is a factor in the exercise
The primary contribution of this paper is to examine a large sample
of option exercises that are accompanied by immediate sale of the stock,
sorting out those that are likely to be motivated by the use of private
information from those that are likely to be motivated by other reasons.
In doing so, we can obtain a better picture of the various factors that
motivate option exercises accompanied by stock sales, and more
accurately identify if and how private information is driving some or
perhaps much of this exercise activity. Our definitions and
interpretations are not intended to suggest that this type of trading on
private information is illegal. Insiders are always in possession of
private information on which they
1. to state in the form of a formula.
2. to prepare in accordance with a prescribed or specified method.
opinions on the
company's prospects, and every trade an insider does is with
private information. Although our view of the
n. pl. le·gal·i·ties
1. The state or quality of being legal; lawfulness.
2. Adherence to or observance of the law.
3. A requirement enjoined by law. Often used in the plural.
of these trades
a. One who believes that it is impossible to know whether there is a God.
b. One who is skeptical about the existence of God but does not profess true atheism.
, we believe that option exercises using private
religious revolution that took place in Western Europe in the 16th cent. It arose from objections to doctrines and practices in the medieval church (see Roman Catholic Church) and ultimately led to the freedom of dissent (see Protestantism).
are more likely based on general perceptions of the company's
prospects over a period beyond a few weeks. Trades based on information
that becomes public over the very short run are likely to be illegal and
legally quite risky to undertake. Trades made in anticipation of poor
stock price performance over perhaps six to twelve months are likely to
be less exposed to legal risk. While we will focus on performance over
the short run through one year, we anticipate that performance beyond a
few weeks is likely to reflect this private information.
The remainder of the paper is organized as follows. Section I
discusses the relevant literature. In Section II, we discuss our data,
hypothesis development, and methodology. Section III reports the
empirical findings. Section IV presents our conclusions.
I. Previous Research on Stock Option Exercises and Insider Trading
Insider trading is one the most widely examined topics in finance
research. Seyhun (1998) provides an excellent summary of the research on
insider trading, which includes
, Maya Ying Born 1959.
American sculptor and architect whose public works include the Vietnam Veterans Memorial in Washington, D.C. (1982).
and Howe (1990), Jeng, Metrick, and Zeckhauser (2003),
and Lakonishok and Lee (2001). The findings suggest that inside
information has value in earning abnormal returns after transaction
costs, with most of the information contained in insider purchases. (1)
Insider trading activities have also been linked to a wide variety
of corporate activities, including seasoned equity offerings (see
Karpoff and Lee, 1991; Gombola, Lee, and Liu, 1997; Niehaus and Roth,
in biology, any gradual change in a particular characteristic of a population of organisms from one end of the geographical range of the population to the other.
and Fu, 2010). The use of executive stock options to
v to move the teeth into their proper positions to conform to the line of occlusion.
management incentives has been widely researched. For example, academic
literature has investigated how options are used to reduce the
over-investment of free cash flow (
, Buchenroth, and Pilotte,
of agency problems (Zhang, 2009; Billett, Mauer, and
Zhang, 2010), and
to avoid excessive risk-taking (Rogers,
Although there is a large body of literature devoted to insider
trading in general, there have been few studies of insider trading on
the form of executive or employee stock option exercises. We to start
the new sentence, we must first note that before May 1991, insiders were
required to hold shares acquired upon exercise for at least six months.
After May 1991, insiders could
v. dis·posed, dis·pos·ing, dis·pos·es
1. To place or set in a particular order; arrange.
acquired shares immediately
provided that the options had been held for at least six months. Hence,
before May 1991, insiders would not necessarily exercise their options
when they had strong reason to believe that the stock would perform
poorly in the
. If insiders were in possession of such
information after May 1991, they would likely be far more inclined to
exercise, given that they could immediately sell the stock. Thus, it is
in the post- 1991 period that we concentrate our focus.
The frequent occurrence of early exercise of stock options is well
documented in the literature. The early exercise behavior of executives
is studied by
1. An edge or border on a piece of cloth, especially a finished edge, as for a garment or curtain, made by folding an edge under and stitching it down.
, and Shevlin (1996) and Bettis, Bizjak,
adj nonexecutive director →
adj nonexecutive director → ,
employees also exercise options early as
shown by Huddart and
LANG Los Angeles Newspaper Guild
A. Private Information as a Motivating Factor in Early Exercise
Seyhun (1998) examines insider option exercises in the post-1991
period and finds that the shares
the market by 0.8%
following exercises. Carpenter and Remmers (2001) examine periods before
and after May 1991 and find that exercises precede positive abnormal
performance prior to May 1991. Following May 1991, however, they find no
significant abnormal returns for their broad sample but do find some
evidence that top executives at small firms exploit their private
information in exercising their options. In a study of exercises in the
United Kingdom over 1995-1998, Kyriacou, Luintel, and Mase (2010) find
some evidence that these options are exercised based on private
information. Significant negative abnormal returns occur following
exercises in which a relatively high proportion of acquired stock is
sold. Core and Guay (2001) and Huddart and Lang (1996) find no evidence
to support the notion that lower-level employees exercise based on
private information. In a later study, however, Huddart and Lang (2003)
do find such evidence.
(Marcus Tullius Cicero) or 106 B.C.–43 B.C., greatest Roman orator, famous also as a politician and a philosopher.
(2009) examines the interaction of
exercising on private information and
of exercises and finds
that when the shares are sold immediately, there is evidence of the use
of private information. When the shares are not
v. dis·posed, dis·pos·ing, dis·pos·es
1. To place or set in a particular order; arrange.
he finds evidence of both timing and the backdating of the exercise
date, though the incidence of backdating was reduced by the
) Act of 2002.
, and Heitzman
(2009) find evidence that exercises were backdated to days with low
closing stock prices in the pre-SOX era. Cai (2007) finds similar
evidence and concludes that 5% to 12% of exercises involve manipulation
of dates or exercise prices.
Another thread of research has examined the relationship between
option exercises by insiders and the flexibility afforded by accounting
rules. Bergstresser and Philippon (2006), Bartov and Mohanram (2004),
, river, c.450 mi (720 km) long, rising in SE Gansu prov. and flowing E through Gansu and Shaanxi provs. to the Huang He.
(2004), and Safdar (2004) all find evidence that exercises are
commonly associated with earnings manipulation, often in the form of
. Safdar, however, concludes that the degree to
which earnings are manipulated appears to be somewhat small.
There are several explanations other than private information for
the early exercise of these options along with sale of the stock. To
identify exercises that could have been motivated by private
information, it is necessary that we identify those that are motivated
by other reasons. In the following
any of the smaller parts into which a section may be divided
Noun 1. subsection - a section of a section; a part of a part; i.e.
, we discuss the various
justifications given for the early exercise of executive stock options.
B. Factors Motivating Exercise Unrelated to Private Information
Because executives are so heavily compensated with stock and
options, their portfolios are typically poorly
Larcker, and Verrecchia (1991) propose that exercising options, followed
by sale of the stock, is a reasonable strategy to achieve greater
, and this idea is supported by the empirical work of
Hemmer, Matsunaga, and Shevlin (1996).
Cuny and Jorion (1995) note that executive departure typically
forces early exercise of options. If the executive leaves the company,
whether by choice or by force, he/she typically has 90 days to exercise
his/her options or else
their entire value. Regardless of the
reason for leaving the company, in-the-money options would be exercised
Another possible reason for early exercise is to capture dividends
to justify discarding the option's
remaining time value. Although Carpenter and Returners (2001) find that
controlling for exercises that fall between a dividend announcement date
and an ex-dividend date leaves their results unchanged, we attempt to
identify exercises motivated by capture of a dividend in a different
Early exercise could also be attractive because of tax benefits.
Goolsbee (2000) shows empirically that the anticipation of a tax
increase has apparently led to increased exercise of options. Carpenter
and Remmers (2001), and McDonald (2003) demonstrate that there are
superior strategies to exercise-and-hold. McDonald (2003) notes,
however, that exercise in anticipation of a tax increase or moving to a
could be justified. The backdating issues referenced
above from the papers by Cicero (2009), Dhaliwal, Erickson, and Heitzman
(2009), and Cai (2007) consider the possibility that exercises without
immediate disposition of the stock might be part of a tax-minimization
strategy and could
1. To bring about or stimulate the occurrence of something, such as labor.
2. To initiate or increase the production of an enzyme or other protein at the level of genetic transcription.
manipulation of the exercise date, but given
our focus only on exercises followed by sale of the stock, this issue
would not be a concern}
Before proceeding with our analysis, we acknowledge that the use of
private information, while largely viewed as
Causing damage or harm; injurious.
shareholders, is not without some potential benefits. Laux (2010) argues
that early exercise on private information can induce executives to
abandon poorly performing projects. In light of this and other possible
benefits of the use of private information by executives, we limit our
study to an examination of the incidence of exercise using private
information. We do not take a policy position on whether the social
tr.v. out·weighed, out·weigh·ing, out·weighs
1. To weigh more than.
2. To be more significant than; exceed in value or importance:
II. Data, Hypothesis Development, and Methodology
In this section we identify the data set and formally develop
testable hypotheses. We also describe the methodology we use to test
The primary data set used in this study consists of option
exercises by corporate insiders that were obtained from the Table II
file of the
A major provider of information, analytical tools, and consulting services to the financial community. The firm, a division of Thomson Corporation, is best known to investors for its First Call segment, which publishes consensus earnings
Insider Filing Data (
TFI The Franklin Institute
TFI The Fertilizer Institute
TFI Technology Futures, Inc.
). TFI defines
corporate insiders as those that have "access to
material, insider information" who are required to file Securities
and Exchange Commission (SEC) forms 3, 4, or 5 when trading in their
company stock as required by Section 16(a) of the Securities and
Exchange Act of 1934. (3) Prior to the enactment of Sarbanes-Oxley in
2002, insiders were required to report transactions by the tenth day of
the calendar month following the month in which the trade occurred.
Sarbanes-Oxley reduced the reporting time to two business days following
We collect insider
In mathematics, a fundamental concept of differential calculus representing the instantaneous rate of change of a function.
transactions from TFI, which contains
all Table II transactions and holdings information reported on SEC Forms
3, 4, or 5. The information reported consists of derivative transactions
such as options, warrants, and convertible securities. The data fields
include open market derivative transactions as well as information on
the award, such as the type of option received, the number of shares
involved, the strike price, the vest date, and the
begin by restricting our sample to only include option exercises in
which the individual's highest title was one of Chief Executive
Officer, Chief Financial Officer, Chief Investment Officer,
, Chief Technology Officer, Executive Vice President,
Officer, Officer and Director, Officer and
, Officer and
Treasurer, Divisional Officer, President, Secretary, Senior Vice
President, Vice President, or Assistant Vice President. We then match
option exercises reported on Table II of the Form 4 with stock
dispositions reported on Table I of Form 4 and remove all cases in which
the stock is not immediately sold by an executive of the company as well
v. a·mend·ed, a·mend·ing, a·mends
1. To change for the better; improve:
transactions and transactions with insufficient
information on the
Center for Research in Security Prices
CRSP Collaborative Research Support Program
CRSP Center for Research in Security Prices
CRSP Center for Research in Security Prices
This process renders 200,659 exercises of 28,073 executives of 5,211
firms. Certain hypotheses require additional information about the
exercises and the executives. For these purposes, we construct a
of the original data set containing all insider trades in the original
data set that can be matched with the necessary insider compensation
data reported in Standard and Poor's ExecuComp. These restrictions
reduce the total number of exercises to 59,683 and include 6,597
executives from 1,799 firms.
The general approach of our tests is to examine the behavior of
long-term abnormal returns around the exercise date. If the abnormal
returns are significantly negative following exercise, there is support
for, though not confirmation of, the use of private information. Tests
comparing the differences in abnormal returns between two groups---one
of which involves exercises likely to be motivated by private
information, and the other of which involves exercises likely to be
motivated by some other factor--are then used to examine whether there
are significant differences and whether the stock performs significantly
worse for the exercises that would
That can be presumed or taken for granted; reasonable as a supposition:
be motivated by private
information. In the next subsection we describe our methodology for
measuring abnormal returns. At this juncture, let us assume that
abnormal returns can be measured.
A straightforward test of whether executives effectively time
option exercises can be conducted by measuring abnormal returns
following exercise. To conduct a test on the entire sample, however,
reveals only whether the sample is dominated by private
information-motivated exercises or by noninformation motivated
exercises. We examine the overall characteristics of the sample with
respect to abnormal returns, but the formal hypotheses are developed by
stratifying the sample into groups that should be motivated by private
information, or by different degrees of private information, and those
that should not.
There is no reason to believe that private information is
associated with exercises at
. Thus, we should expect
exercises that occur at expiration are conducted merely to capture the
value of expiring in-the-money options, while exercises that occur prior
to expiration could be motivated by private information. Hypothesis 1
addresses this point as follows:
H1: Options exercised early are based on private information and
options exercised at expiration are not based on private information.
To test Hypothesis 1 we first define an exercise at expiration as
any exercise that occurs within 30 days of the expiration date. We
acknowledge that some of these exercises near maturity could be
motivated by private information, but the majority should be motivated
by the expiration. Thus, H 1 is supported if early exercises are
followed by abnormal returns that are negative and significantly lower
than the abnormal returns that follow exercises at expiration.
Executives typically hold investment portfolios that are
suboptimally diversified, with the company's stock constituting an
abnormal proportion of the executive's overall wealth.
executives could rationally choose to exercise their options to
rebalance their portfolios. Since executive stock options cannot be
exercised prior to
, one might expect some
to occur on the vest date. Insiders that hold private negative
information would still choose to exercise at this time and therefore
some negative abnormal returns should be present; however, the private
information effect could be dominated by portfolio-rebalancing
transactions that contain no private information. Indeed, Fu and Ligon
(2009) find that executives with greater diversification needs and
riskier underlying stocks are more likely to exercise on the vest date.
Thus, our hypothesis is stated as follows:
H2: Early exercises on the vest date are less likely to be
motivated by private information than early exercises made after the
To test this hypothesis we define an exercise on the vest date as
one occurring within 30 days after the vest date. H2 is supported if
abnormal returns following exercises not on the vest date are negative
and significantly lower than abnormal returns following exercises on the
The standard and only justification for exercise of a traded call
option on a stock is to capture the underlying stock's dividend
payment. Executive stock options could also be exercised early to
capture a dividend, but these exercises should not contain private
information. (4) This hypothesis is, therefore:
H3: Early exercises occurring prior to but close to the ex-dividend
date are less likely to be motivated by private information than those
exercised just after the ex-dividend date.
To test this hypothesis we examine the pattern of exercises in
relation to the ex-dividend date following a procedure we explain later.
tr.v. clas·si·fied, clas·si·fy·ing, clas·si·fies
1. To arrange or organize according to class or category.
2. To designate (a document, for example) as confidential, secret, or top secret.
exercises into two categories: those not likely to be
motivated by dividends and, therefore, could be motivated by private
information, and those that are likely to be motivated by dividends.
This procedure is a conservative one in that the latter category could
include a number of exercises that are motivated by private information.
Thus, we bias the test against finding evidence of private information.
We then examine whether abnormal returns following exercises not
motivated by dividends are negative and significantly lower than those
following exercises motivated by dividends.
All exercised options will be in-the-money but clearly some more
than others. When an in-the-money option is exercised, the holder
discards the time value. Time value is greatest when the option is
at-the-money and is smaller the more in-the-money the option. Hence, the
most expensive options to exercise are those that are the least
in-the-money and the least expensive are those that are deep
in-the-money. Thus, our next hypothesis is:
H4: Because early exercises that are the least in-the-money result
in the loss of more time value than are those that are the deepest
in-the-money, they are more costly to exercise and would therefore be
more likely to be exercised based on private information.
To test this hypothesis, we determine whether exercises of options
the least in-the-money are followed by abnormal returns that are
negative and significantly lower than those following exercises of
options that are the most in-the-money.
It is also possible that some early exercises are motivated by the
fact that the executive leaves the firm. We
v. hy·poth·e·sized, hy·poth·e·siz·ing, hy·poth·e·siz·es
To assert as a hypothesis.
To form a hypothesis.
exercises are less informative than early exercises in which the
executive remains with the firm. This hypothesis is, therefore:
H5: Early exercises motivated by departure of the executive are
less likely to be motivated by private information than are early
exercises that are not motivated by departure.
The departure date variable in ExecuComp enables us to identify
options that are exercised early because the executive left the company
and would otherwise have had to forfeit the options. We examine whether
abnormal returns following exercises in which the executive does not
leave the firm are negative and significantly lower than those following
exercises in which the executive leaves the firm.
Since executives avoid larger losses by exercising more of their
portfolio, exercises that are large relative to the overall option
portfolio should have a greater likelihood of containing private
information. Small exercises on the other hand are likely to be
1. produced artificially.
2. produced by induction.
adj artificially caused to occur.
for other reason such as the need for cash. Stronger private information
should be revealed in higher negative abnormal postexercise returns,
leading to the following hypothesis:
H6: Evidence of early exercise based on private information is
increasing in the number of options exercises relative to total options.
The TFI data set enables us to obtain a proxy to measure this
effect by using the proportion of options exercised to total options. We
calculate the proportion of options exercised scaled from 0 to 1 in the
following manner: Options Exercised / (Options Remaining + Options
Exercised). Thus, we can examine whether abnormal returns following
exercises in which a large percentage of options are exercised are
negative and significantly lower than those following exercises in which
a small percentage of options are exercised.
We should note that there may be interactions between the various
motivations for exercise that are unrelated to private information. For
example, an executive in need of cash but holding an unvested option
would surely exercise as soon as it is
adj. referring to having an absolute right or title, when previously the holder of the right or title only had an expectation. Examples: after 20 years of employment Larry Loyal's pension rights are now vested. (See: vest, vested remainder)
. If the cash is not needed
at a specific time, the executive might wait until shortly before the
ex-dividend date. These motivations are impossible to disentangle, but
they cause no serious problems. If some exercises based on private
information are contained in samples of exercises that should not be
based on private information, we have merely raised the barrier in
detecting the presence of private information. Nonetheless, we address
these potentially related exercises later in the multivariate portion of
Each hypothesis is tested by examining abnormal performance
following exercise with one-tailed t-tests. A one-tailed test is
appropriate because any private information that leads to exercise
followed by sale of the stock is likely to be only negative in nature.
1. One who usually expects a favorable outcome.
2. A believer in philosophical optimism.
information should not trigger exercise followed by sale of
the stock because any positive appreciation in the stock price would not
be experienced by the executive.
Exercises of stock options, however, pose a special difficulty that
can induce a subtle bias. The parameters required for estimating
expected returns are typically estimated over a period prior to the
event that is assumed to be a term of normal (vis-a-vis
"abnormal") stock price behavior. However stock option
exercises usually occur following a potentially long period of strong
positive performance of the stock. Thus, the alphas would be estimated
over a period of primarily rising stock prices and would be biased
upward, which would bias expected returns upward and would increase the
likelihood of finding negative abnormal returns after exercise. This
momentum-driven bias, as discussed in Jegadeesh and
n. New England & Upstate New York
1. A runt, especially one of a litter of pigs.
2. A small person. See Regional Note at tit1.
lead us to mis-specify the magnitude of any abnormal returns after
exercise. Thus, even if there are truly negative abnormal returns
At a later time; subsequently.
later [Old English æfterweard]
, we could overstate their magnitude, or we could falsely
identify postexercise abnormal returns that are merely a statistical
To avoid this bias, we follow
(1997) and Lyon,
Barber, and Tsai (1999) in adopting an event-time matching-firm
portfolio as a benchmark for calculating buy-and-hold abnormal returns
for our sample firms. Because there is no
method eliminates the bias discussed above. Our sample firm is,
therefore, benchmarked against a portfolio of comparable firms over the
same time period that did not experience exercises.
Specifically, for each exercise event, we identify five firms
1. As stated or indicated by; on the authority of:
2. In keeping with:
industry, size, book-to-market, and prior firm
performance in the following manner. (5) Consider a single exercise
event that is a component of a
A sample drawn from a larger sample.
tr.v. sub·sam·pled, sub·sam·pling, sub·sam·ples
To take a subsample from (a larger sample).
, all exercises of which have a
common characteristic. Let us assume they are all vest date exercises
and the firm is called
Used to indicate to someone that the zipper of his or her pants is open.
[ex(amine) y(our) z(ipper).]
. We identify the industry of the firm and
select all firms in that industry as potential benchmark firms. (6) We
first eliminate XYZ from the benchmark set and then eliminate all other
firms from the benchmark set that also have a vest date exercise within
one year of the event date. Next, we eliminate all firms with a
at the end of the previous year that differs by more than
30% from that of XYZ and any firm that has a previous year stock price
performance that differs by more than 5%. Finally, we select the five
firms in the benchmark set that are closest to XYZ with respect to the
at the end &the previous year. We now have a
benchmark portfolio for each unique exercise and can compare the returns
following exercise of the options to those of a comparable group. (7)
The primary tests are conducted using daily returns, which permits
more precise estimates of the abnormal returns avoided by early
exercise. We observe performance over 252 trading days after exercise.
For a given day, the
daily return of these five matching
firms is calculated to obtain the matching portfolio return. Performance
is then measured by calculating the difference between the buy-and-hold
returns of firms that had options exercised and the corresponding
buy-and-hold matching portfolio returns in the following manner:
[MATHEMATICAL EXPRESSION NOT
v. re·pro·duced, re·pro·duc·ing, re·pro·duc·es
1. To produce a counterpart, image, or copy of.
2. Biology To generate (offspring) by sexual or asexual means.
or a set of codes used to represent letters, numbers, a few symbols, and control characters. Originally designed for teletype operations, it has found wide application in computers.
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (2)
BHAR(T) = [
BHR Birmingham Hip Resurfacing
BHR Bureau for Humanitarian Response
BHR Bronchial Hyper Reactivity
.sub.EX](T) - [
BHM Bachelor in Hospitality Management
BHM British Heavy Metal
EX refers to the firm with the option exercise
and MP refers to the matching portfolio, t = 1 is the first day
following exercise, T is the selected day after exercise or the
day (whichever comes first), [R.sub.t] = return on day t of
the firm with the exercised option, and [r.sub.it] = return of matching
firm i on day t. We use buy-and-hold abnormal returns (BHARs) as our
primary means of comparison; however, we also calculate and report raw
(unadjusted) returns as well as returns adjusted for the returns of the
CRSP equally weighted market portfolio. Significance for the mean BHARs
is assessed using the Lyon, Barber, and Tsai (1999) bootstrapped,
n a value or number that describes a series of quantitative observations or measures; a value calculated from a sample.
a numerical value calculated from a number of observations in order to summarize them.
. (8) Cumulative abnormal returns, on
the other hand, are less
Epidemiology adjective Referring to an asymmetrical distribution of a population or of data
than BHARs. Lyon, Barber, and Tsai
(1999) show that conventional t-tests render well-specified test
statistics in these instances. For this reason, we use traditional test
statistics for all non buy-and-hold related returns.
Fama (1998) argues that BHARs do not account for potential
cross-sectional correlation of stock returns, which is likely to result
tr.v. o·ver·stat·ed, o·ver·stat·ing, o·ver·states
To state in exaggerated terms. See Synonyms at exaggerate.
test statistics. To address this concern, Fama advocates
the calendar-time portfolio method for assessing long-term performance.
Many exercises in our sample
1. A part or portion of a structure that extends or projects over another.
2. The suturing of one layer of tissue above or under another layer to provide additional strength, often used in dental surgery.
in event time, so we are unable to
assume that the events are independent. Hence, to ensure robustness we
also estimate the standard Fama and French (1993) model including a
fourth Carhart (1997) momentum factor:
(4) [R.sub.pt] - [R.sub.ft] = [[alpha].sub.p] +
[[beta].sub.p]([R.sub.mt] - [R.sub.ft]) + [s.sub.p][
HML Hawaii Medical Library
HML High Minus Low
HML Hard Money Lender
HML Human Media Lab
.sub.t] + [m.sub.p][
.sub.t] + [[epsilon].sub.pt], (4)
where [R.sub.pt] is the equal-weighted return for calendar day t
for the portfolio of exercises, [R.sub.ft] is the
[R.sub.mt] is the CRSP equally weighted market portfolio return,
[SMB.sub.pt] the difference between the return on the portfolio of small
minus big stocks, [HML.sub.pt] the difference between the return on the
portfolio of high minus low book-to-market stocks, and [UMD.sub.pt] the
difference between the return on the portfolio of high minus low prior
1. an expression of the change or effect produced by variation in certain factors, or of the ratio between two different quantities.
is the average daily
portfolio of exercises over T-days following exercise. This
calendar-time approach mitigates the problem of cross-sectional
dependence among exercises, but, unlike BHARs, the abnormal return
measures do not accurately reflect the returns experienced by an
investor. Loughran and
n. pl. ritter
[German, from Middle High German riter, from Middle Dutch ridder, from r
(2000) also argue that the calendar-time
approach leads to an under rejection of the
since it smoothes out
high and low event activity.
The first set of tests are primarily
adjective Determined, produced, or caused by only one variable
tests. Hence, there
could be overlaps and interactions between certain groups. We address
this concern later in the paper with a set of multivariate regressions
of the BHARs following exercises on certain variables.
We impose one other
A restriction on the natural degrees of freedom of a system. If n and m are the numbers of the natural and actual degrees of freedom, the difference n - m is the number of constraints.
on the sample. In many cases, there
are multiple exercises by the same or different executives of the same
firm in a single day. To avoid
these exercises we
eliminate multiple exercises on the same day by an executive of the same
firm. (9) Because each exercise event does not necessarily have the same
relationship to the exercise or vest date, however, we make these
eliminations separately on each subsample. For example, assume there are
five exercises for various executives for a firm in a day. Suppose in
constructing our overall sample, we select the first exercise and
the other four. It is possible that the
exercises could be vest
date or departure-motivated. If we deleted them, we would lose these
observations from subsequent tests in which they would be most needed.
Hence, we construct each subsample separately and then choose only the
first exercise per company per day.
This section is divided into seven subsections, according to the
hypotheses presented above. We will refer to the full sample, consisting
of all exercises meeting the matching criteria mentioned above, and a
subset called the merged sample, which includes only exercises for which
certain additional variables are available on ExecuComp, as previously
described. Table I reports summary statistics for the full and merged
samples and the various stratifications described above. These
statistics are frequently referenced throughout this section.
As noted previously, we do not formally develop hypotheses for the
composite full or merged samples. These samples are not
/strat·i·fied/ () formed or arranged in layers.
Arranged in the form of layers or strata.
therefore, do not reveal whether abnormal returns following exercises
that should be motivated by private information are different from those
following exercises that should not be motivated by private information.
We do, however, examine the overall and merged samples. Figure 1
illustrates the mean BHARs from day 0 to day +252 for both samples. Here
and in all remaining figures, we show only postexercise abnormal
returns, because the matching process effectively eliminates any
abnormal returns prior to exercise. Table II reports the mean returns
and the results from the appropriate test statistic. Included are mean
BHARs with significance from the Lyon, Barber, Tsai (1999) bootstrapped
t-test and calendar-time abnormal returns and t-tests. In addition, we
show market-adjusted (MAR) and raw (unadjusted) returns (RAW).
Statistical tests are applied over holding periods of 10, 20, 60, 125,
and 252 trading days. For the BHARs, we also present results based on
monthly returns. (10)
[FIGURE 1 OMITTED]
As Figure 1 shows, BHARs fall steadily over the full 252-day period
following exercise for the full sample. For the full sample, the
risk-adjusted results are significantly negative for virtually every
period following exercise--regardless of whether BHARs, market-adjusted
returns, or calendar-time abnormal returns are used. Interestingly, the
raw returns are positive, which illustrates the importance of risk
adjustment. In addition, the results are highly negatively significant
using monthly returns in the postexercise period. Figure 1 shows that
the merged sample appears to exhibit a smaller decline following
exercise, but almost all of the BHARs, CTARs, and MARs are significant.
One reason for the slightly lower degree of significance could be that
the merged sample, which requires ExecuComp data, will tend to consist
of larger firms. Consistent with Carpenter and Remmers (2001), the use
of private information could be more prevalent in smaller firms.
Of course, these full and merged sample results are not stratified
and merely indicate that exercises motivated by private information
could dominate those not motivated by private information. We next
examine various subsamples comparing those exercises hypothesized to be
motivated by private information to those that are not.
A. Results According to Early Exercise versus Exercise at
In this subsection, we report our tests to determine whether there
is any difference in postexercise abnormal performance according to
whether the options are exercised early or at expiration. Relevant
summary statistics are presented in Panel A of Table I. About 94% of the
exercises in the full sample occur prior to expiration, an average of
3.46 years after vesting and about 4.6 years before expiration, which is
consistent with previous studies.
Figure 2.A shows the post-exercise BHARs for both the early and
maturity exercises in the full sample. Note that following early
exercises, the stock shows considerably worse performance than for
maturity exercises. The first three columns (denoted (1), (2), and
(1)-(2)) of Table III show statistical details of the means and
difference between the means of the BHARs, CTARs, raw returns, and
market-adjusted returns, with all BHARs tested using the Lyon, Barber,
and Tsai (1999) test statistic. Consistent with the visual impression
from Figure 2.A, the BHAR difference results show strong significance,
particularly over the 60-day period, and the CTAR differences are highly
significant over all periods. The MARs are significant over the 60-day
holding periods (at the 1% level), and raw returns are significant over
all periods from 20 to 252 days. The monthly results over 12 months are
also significant. Thus, stock price behavior following early exercises
appears statistically lower following early exercises than following
exercises at maturity. It is notable, however, that the returns for
early exercises show a nearly
1. Having no fixed or regular course; wandering.
2. Lacking consistency, regularity, or uniformity:
3. , initially declining and nearly
v. con·verged, con·verg·ing, con·verg·es
a. To tend toward or approach an intersecting point:
b. to the early exercise series at around 160 days, then
v. di·verged, di·verg·ing, di·verg·es
1. To go or extend in different directions from a common point; branch out.
2. To differ, as in opinion or manner.
3. upward. We investigated this result by considering that the
maturity exercises included all exercises within 30 days of expiration.
Hence, some of these so-called maturity exercises, while close to, are
not strictly at maturity. We re-ran these tests defining a maturity
exercises as one that strictly occurs on the maturity date, which
comprise about 25% of the exercises defined as occurring at maturity.
The returns show much less tendency to decline following exercise,
suggesting that, indeed, many of these near-maturity exercises could
contain some private information.
We conclude that the evidence supports H1, which states that
options exercised early appear to be associated with the use of private
information. Those exercised at expiration indicate some evidence of
private information but are followed by significantly better performance
than options exercised early. Many exercises somewhat near the maturity
date, however, can also contain private information, but in general,
exercises that are well away from maturity show more evidence of private
information than those that are within 30 days of maturity.
B. Results According to Vest Date Exercise
We now partition all early exercises according to whether the
exercise occurred on the vest date. As noted earlier, we define a vest
date exercise as one in which exercise occurs within 30 days after the
vest date. (11) Figure 2.B shows the results for options exercised on
the vest date compared to those not exercised on the vest date. For
about the first 100 days following exercise, both samples show negative
BHARs and are even nearly the same at a point, though the not-vest date
exercise returns are almost always lower. At around day 100, the vest
date returns start turning up, while they continue to fall for not-vest
date exercise. As Table III shows, there is strong negative significance
for the not-vest date returns using all measures. The difference in
BHARs and CTARs show strong significance over multiple periods and there
is evidence of significance for the MARs and even the raw returns. The
monthly return difference is also significant. Like maturity date
exercises, however, the vest date sample almost surely contains some
exercises that are motivated by private information. In fact, this point
is likely to be even stronger with vest date exercises. Whereas
expiration is likely to dominate private information as a reason for
exercising on the last day of an option's life, an executive with
private information prior to the vest date would almost surely exercise
at the vest date or as soon as possible thereafter if that information
has not been incorporated into the stock price. Thus, exercises right on
the vest date may well contain the effect of private information. We
re-ran the tests for vest-date and nonvest date exercises with the
latter strictly defined as any exercise not precisely on the vest date.
The results are almost unchanged from those in Figure 2.B. The strong
divergence continues to be observed well after day 100, suggesting that
if non vest-date exercises are more likely to be based on private
information, the effect does not show until after 100 days. The effect
is strong after 100 days, however, as significance is at the 5% level
for 252 days.
[FIGURE 2 OMITTED]
Thus, evidence from the analysis based on whether the options are
exercised on the vest date is moderately consistent with H2. Options not
exercised on the vest date but before expiration appear to be more
likely motivated by private information than exercises that occur on the
vest date. Conclusive evidence, however, is not found until at least 100
days after exercise.
C. Results According to Dividend-Motivated Exercises
As noted previously, standard option theory demonstrates that one
reason any call option holder could choose to exercise early is to
capture an upcoming dividend payment. Thus, exercises that are motivated
by the capture of dividends should not be based on private information.
To remove exercises that are likely to be motivated by dividends, we
identify the exercises that occur close but prior to the ex-dividend
date. (12) We first capture the ex-dividend dates from CRSP and then
merge them with our full sample of exercises. Using only early
exercises, each exercise event is assigned
tr.v. as·signed, as·sign·ing, as·signs
1. To set apart for a particular purpose; designate:
2. a date measured in weeks
relative to the upcoming dividend in which a week is defined as five
business days. We then examine the number of exercises per week and
observe a convex pattern in which the number of exercises begins to
decrease following the ex-dividend dates, reaches a floor, and then
begins increasing with the number sharply increasing in the last few
weeks before the ex-dividend date.
The observed pattern of exercises prior to an ex-dividend date
suggests that dividend-motivated early exercise could start occurring
much earlier than the day before the ex-dividend date. We define a
dividend-motivated exercise as one occurring within 15 business days
before the ex-dividend date, which we refer to as weeks -1, -2, and -3.
Exercises not motivated by dividends are defined as those occurring in
weeks -7, -8, -9, and -10 as well as exercises of options on stocks that
do not pay dividends. Exercises occurring -4, -5, and -6 weeks relative
to the ex-dividend date could contain both dividend-motivated and
nondividend-motivated exercises. Because these exercises cannot be
clearly classified either way, we omit
tr.v. o·mit·ted, o·mit·ting, o·mits
1. To fail to include or mention; leave out:
a. To pass over; neglect.
b. this group from consideration in
testing this hypothesis. After removing observations with insufficient
data and those where an appropriate match-sample could not be attained
v. at·tained, at·tain·ing, at·tains
1. To gain as an objective; achieve:
there are 3,502 dividend-motivated exercises and 22,417 not-dividend
motivated exercises. Other summary statistics are presented in Table I.
Figure 2.C shows the post-exercise BHARs. Dividend-motivated
exercises show a pattern that distinctly contrasts with the remaining
exercises. While the former are moderately negative, the latter are
sharply negative. Not-dividend motivated exercises show high degrees of
significance for almost all measures except the raw returns.
Dividend-motivated exercises do show some degrees of negative
significance as reported in Table IV. More importantly, the differences
between exercises that are dividend motivated and those that are not are
significant, with Type I errors of less than 1% using BHARs, CTARs, raw,
market-adjusted returns, and BHAR monthly returns. Thus, we see that
some exercises of executive stock options are motivated by dividends,
but those that are not precede poorer stock price, which supports H3.
D. Moneyness Tests
We now examine whether moneyness is a factor in distinguishing
exercises of options based on private information from those exercised
for other reasons. Recall that for in-the-money options, time value is
inversely related to moneyness. Hence, options with low moneyness are
more costly to exercise than options with high moneyness and exercising
them would more likely be motivated by private information. If this
premise is supported, options closest to at-the-money should show the
strongest negative performance following exercise, while those deepest
in-the-money should show the weakest negative or possibly positive
We divide the early exercises into five moneyness quintiles. Money
1 is the group closest to at-the-money, and Money 5 is the group deepest
in-the-money. Summary statistics are in Table I. (14) Figure 2.D shows
the postexercise performance for each group. The BHARs line up exactly
as hypothesized. The most negative performance is in Money 1 (the group
closest to at-the-money), while Money 5 (the most in-the-money group)
shows strong, albeit largely not-significant, positive performance.
Difference tests for Money 1 versus Money 5, as shown in Table IV, are
highly significant for the BHARs, the CTARs, the raw returns, and the
market-adjusted returns. (15) Hence, we see that the options that are
the most costly to exercise are those that show the strongest evidence
of the use of private information. Those that are the least costly to
exercise show virtually no evidence of private information. Thus, H4 is
E. Results According to Executive Departure
In this subsection we compare exercises based on whether the
executive leaves the firm, which we call departure exercises, to those
in which the executive does not leave the firm. We posit that those in
which the executive leaves the firm are not likely to be based on
private information. We define a departure exercise as one that occurs
within plus or minus 270 days of the executive departure date. For this
test, we require the departure date, so we must use the merged sample.
This data set contains 7,700 early exercises not associate with
executive departure and 594 that are associated with executive
[FIGURE 3 OMITTED]
The results are shown in Figure 3.A and in Table V. As we
hypothesized, exercises not associated with executive departure are
followed by significantly negative abnormal returns, but somewhat
strangely we find that exercises around executive departure are also
associated with significantly negative abnormal returns that appear to
be even stronger. While the group differences are not significant in the
one-tailed test in the hypothesized direction, they would be significant
in the opposite direction.
There are several reasons why a departure exercise could precede
poor stock price performance. The executive could have foreseen
tr.v. fore·saw , fore·seen , fore·see·ing, fore·sees
To see or know beforehand: poor
firm-specific performance and chose to "cash out" and leave.
If that is so, there is indeed an element of private information
motivating the exercise as well as the departure. Alternatively, the
board could have anticipated difficult times ahead and felt no
confidence in the executive's ability to lead the firm through this
period. If that is the case, then the executive's replacement did
no better. A third possible explanation is that the firm performed
poorly because the executive left. (17)
To examine this issue further, we observe that the two reasons
identified on ExecuComp for departure are resignation and retirement. Of
course, we do not know whether a resignation is a forced resignation, a
voluntary resignation, or a voluntary resignation that is de facto forced. After applying the buy and hold matching criteria, the sample
contains 440 exercises that indicate as occurring around either
resignation or retirement with 208 exercises around resignation and 232
around retirement. Figure 3.B shows that those exercises due to
retirement show much more negative postexercise performance than those
due to resignation, and the tests, as shown in Table V, are significant
for the BHARs and the CTARs.
It can be argued that an exercise before resignation is more likely
to be motivated by private information than one that occurs after
resignation. In the latter case, the executive is no longer with the
firm, which would reduce the executive's access to private
information. In Panel 3.C and the right three columns in Table V, we
compare the postexercise performance of the resignation sample broken
down into exercises that occur before resignation and those that occur
afterwards. The results are consistent with that explanation. Exercises
that occur before resignation show significantly negative performance
while performance following exercises that occur after resignation is
not significantly different from zero. (18)
Thus, overall, we find that H5 is not supported. Departing
v. de·part·ed, de·part·ing, de·parts
1. To go away; leave.
2. To die.
3. executives exercising options do appear to be motivated by private
information, with the effect stronger for departures motivated by
retirement and departures motivated by resignation in which the options
are exercised before the departure. It is possible that extremely poor
expected performance could motivate both retirement and resignation, but
we leave this subject for future research.
F. Results According to Proportion of Options Exercised
As noted earlier, we are interested in separating exercises that
are motivated by private information from those that are motivated by
other reasons. If an executive is in possession of private information
that suggests poor upcoming firm-specific performance, the executive
would probably exercise a larger proportion of exercisable options, if
not all of them, than if he were exercising for some reason unrelated to
private information. By exercising a larger proportion of his options,
the executive avoids greater losses.
There is no perfect measure of the proportion of options exercised
to those that could have been exercised. The exercise value of
potentially exercisable options (vested and in-the-money) is reported
annually, but that number can give a biased picture given the amount of
time between the exercise date and the end of the year, a period during
which we have found there are typically strong downward stock price
movements. The TFI database, however, gives the number of derivatives held (in terms of shares) at the time of exercise. Thus, we can
calculate the number of options exercised relative to the total number
of options. Obviously, the total number of options will include
not-vested and out-of-the-money options, which clearly are not
candidates for exercise. This problem, however, injects a conservative
bias. For example, if an executive exercises a large portion of his
in-the-money options, we hypothesize that the exercise is likely to be
motivated by private information. If there are a large number of
unexercisable options, our measure will be biased low. If the executive
exercises a small portion of his exercisable options, we hypothesize
that the exercise is unlikely to be motivated by private information. If
there are a small number of unexercisable options, this measure will be
biased high. Thus, both cases would make it harder to distinguish
exercises of a lot of options from exercises of a few and to correlate
those groups with subsequent stock price performance. There is a slight
v. mit·i·gat·ed, mit·i·gat·ing, mit·i·gates
To moderate (a quality or condition) in force or intensity; alleviate. See Synonyms at relieve.
To become milder. factor, which is that companies with exercised stock options
typically experienced strong stock price performance prior to exercise,
thus, tending to push more options in-the-money and making them
exercisable, provided they are vested. There is no perfect measure, but
the one we use is biased against finding the result that leads to
rejection of our null hypothesis
n theoretical assumption that a given therapy will have results not statistically different from another treatment.
n . (19)
The TFI sample contains about 23,000 transactions after screening
for matching firms that can be ranked by the proportion of options
exercised out of the total number of options in the executives'
portfolio. We divide this sample into five groups of approximately 4,600
each. Proportion 1 contains those exercises in which the percentage of
options exercised to those not exercised is greatest, while Proportion 5
contains the exercises in which the percentage of options exercised is
smallest. Thus, Proportion 1 would be the group in which the executive
exercised relatively more options and, thus, is likely to show poorer
stock price performance. As Figure 3. C shows, that is indeed the
result. Long-run performance aligns according to the percentage of
options exercised. Table VI shows the results for each group. We see
that Proportions 1, 2, and 3 all have highly significantly negative
postexercise performance. The difference between Proportion 1 and
Proportion 5 is highly significant in the expected direction using all
measures. Thus, this evidence is consistent with H6 in that exercises in
which the number of options exercised is large relative to those not
exercised are more likely motivated by private information.
G. Multivariate Tests
Since many factors can simultaneously influence the decision of an
executive to exercise stock options, we attempt to control for these
effects in a multivariate framework. We regress the 252-day BHARs on the
variables reflecting the factors we previously examined as well as a set
of dummy and control variables that could influence BHARs. Specifically,
we run the following regressions:
BHAR(0, +252) = [[delta].sub.0] + [[delta].sub.1] Informed +
[[delta].sub.2]LogSize + [[delta].sub.3]BooktoMarket +
[[delta].sub.4]Volatility + [[delta].sub.5]Momentum +
[[delta].sub.6]PriorFirmPerf + [[delta].sub.7]PriorMktPerf +
[[delta].sub.8] YearDummies + [[delta].sub.9]IndustryDummies +
BHAR(0, +252) = [[beta].sub.0] + [[beta].sub.1]Early +
[[beta].sub.2]NotVest + [[beta].sub.3]DivMotiv + [[beta].sub.4]Depart +
[[beta].sub.5]Resign + [[beta].sub.6]Retire + [[beta].sub.7]PropEx +
[[beta].sub.8]Moneyness + [[beta].sub.9]Reload + [[beta].sub.10]ExecRank
+ [[beta].sub.11]PreSOX + [[beta].sub.12]LogSize +
[[beta].sub.13]BooktoMarket + [[beta].sub.14] Volatility +
[[beta].sub.15]Momentum + [[beta].sub.16]PriorFirmPerf +
[[beta].sub.17]PriorMktPerf + [[beta].sub.18]YearDummies +
[[beta].sub.19]IndustryDummies + [epsilon]. (6)
Informed is a composite dummy variable equal to one if the exercise
is early, not vest related, not dividend motivated, proportion exercised
high, and moneyness low. In the second regression in psychology: see defense mechanism.
In statistics, a process for determining a line or curve that best represents the general trend of a data set. , we replace this
variable with the next eight informed variables. Early equals one if the
exercise did not occur within 30 days of expiration. NotVest equals one
if the exercise did not occur between zero and 30 days after the vest
date. DivMotiv equals one if the exercise occurred within 15 days prior
to an ex-dividend date. Depart equals one if the early exercise occurs
within 270 days of the reported departure date. Resign and Retire equal
one if the reported reason for departure was resignation or retirement
respectively. PropEx is the proportion of total options exercised.
Moneyness is the stock price divided by the exercise price.
We also examined some other factors that could motivate exercise.
Reload options are those in which the holder exercises the option by
tendering the stock and receives replacement options. Hence, reload
options might be more likely to be exercised not based on private
information. Thus, we introduce the dummy variable Reload, which equals
one if the option is a reload. ExecRank is a one if the executive is
listed as a C-level executive. PreSOX is a dummy equal to one if the
exercise occurred prior to the Sarbanes-Oxley Act.
The remaining variables capture firm and market characteristics.
LogSize is the log of the firm's market capitalization in the
fiscal year end prior to exercise. BooktoMarket is the ratio of book
value to market value at the fiscal year end prior to exercise.
Volatility is the standard deviation of the stock returns computed over
the 60 months prior to exercise. Momentum is the unadjusted 90-day stock
return prior to exercise, and PriorFirmPerf is the unadjusted
performance of the firm's stock in the one year prior to the
beginning of the momentum window. PriorMktPerf is the performance of the
CRSP equal-weighted index in the year preceding exercise. Year and
Industry dummies are also included. There are 5,121 observations in the
The results of four variations of these models are shown in Table
VII. In Model (1), we include only a single dummy variable for whether
the exercise is believed to be informed, as described above, and year
and industry dummies. The informed variable is, as expected, highly
significant and negative. In Model (2), we include the informed variable
as well as firm-specific variables. The informed variable has
approximately the same effect but volatility and prior firm performance
are also significant. Model (3) includes the specific factors as
described above in place of the informed variable as well as the
firm-specific variables. At the 5% level of significance or better, we
find that BHARs are more negative if the exercise is early, did not
occur at the vest date, the executive is departing, the proportion of
options exercised is highest, and the options are the least
in-the-money. (20) Volatility and prior firm performance are again
significant. In Model (4), we add the reload, executive rank, and preSOX
factors. We achieve virtually the same results for our informed factors,
but the executive rank factor is significant at the 5% level. The
positive significance suggests that if the BHARs are more negative (or
less positive) if the option holder is not a chief officer. We suspect
this result is due to the fact that top executives are far more
scrutinized and may feel less able to get away with exercising based on
H. Operating Performance
In our last set of tests, we examine the operating performance of
firms from one year prior to three years following option exercise.
Based on evidence presented up to this point in the paper, we create a
subsample of exercises most likely to have been based on private
information by selecting early exercises that are not around the vest or
ex-dividend date, moneyness exceeds the sample median, and the
proportion exercised exceeds the median proportion exercised. Operating
performance for these potentially informed exercises is then compared to
that of the other firms. We use only cases where the firm had no other
exercises over the following three-year period and follow Barber and
Lyon (1996) by defining abnormal operating performance as the difference
between the actual operating performance and that of an industry control
group matched on two-digit standard industrial classification (SIC) code
and operating performance in the year prior that is within 90% to 110%
of that of the firm in question. We examined four measures of
performance: return on assets, return on market value, operating return
on assets, and operating return on market value and report mean and
median levels and changes in these measures. The number of observations
varies but ranges from 1,003 to 1,388 for the informed sample and 1,436
to 1,894 for the complimentary sample.
The results are shown in Table VIII. The mean and median level
measures are predominately significant in the year of exercise and year
following exercise, and the changes are predominately significant from
year--1 to 0 and 0 to 1. For return on assets and operating return on
assets, significant differences are observed for several years after
exercise. The signs are also in the expected direction. Thus, the
operating performance of firms that have exercises that are likely to be
motivated by private information is significantly worse than that of
firms whose exercises are not likely motivated by private information.
Hence, the negative private information that we found in these
firms' abnormal stock performance appears consistent with a decline
in operating performance for at least one and possibly more years
in this study, we examine whether early exercises of executive
stock options likely to have been motivated by private information are
followed by significantly lower returns than are those that are likely
to have been motivated by other reasons. Our results confirm that
private information appears to be widely used by executives. It would be
desirable if we could measure the percentage of exercises that are
motivated by private information. We do find that about 94% of all
exercises occur early, about 95% occur after the vest date, and about
88% appear to be not motivated by the capture of dividends. Because many
of these motivations can interact, however, it is not possible to
identify how many exercises are strictly motivated by private
information and how many strictly are not. Indeed some maturity and vest
date exercises are also likely to be motivated by private information.
However we do find distinctly lower BHARs following exercise for samples
that should be motivated by private information in comparison to samples
that should not. In the absence of private information, there appears to
be no other logical explanation.
Of course, these findings do not prove that executives are engaged
in behavior that would meet the legal definition of insider trading, and
we certainly make no such judgments. All executives form opinions about
the future performance of the stock and their ability to manage the firm
successfully. Illegal inside information is but one of many forms of
private and potentially quite accurate information about future company
performance. These exercises and the subsequent stock sales are filed
with the SEC, so executives and the SEC must generally believe the
transactions pass the test of legality. Indeed, the evidence points far
more to the use of private information that reflects an outlook of at
least more than a few weeks. In any case, documenting the use of private
information is an important step toward understanding the costs and
This paper was presented at the Financial Management Association
and the Southern Finance Association meetings and at faculty seminars at
the University of Alabama, West Virginia University mainly at Morgantown; coeducational; land-grant and state supported; est. and opened 1867 as an agricultural college, renamed 1868. , Clemson University at Clemson, S.C.; coeducational; land-grant; state supported; opened in 1893 as a college, gained university status in 1964. The university includes programs in textile and computer research, wildlife biology, and aquaculture and maintains ,
Material such as thread, wire, or gauze that is passed through subcutaneous tissues or through a cyst in order to form a sinus or fistula.
1. a thin woven fabric wick, 6 in × 0. Hall Louisiana State University, the University of Alabama at
1 City (1990 pop. 265,968), seat of Jefferson co., N central Ala., in the Jones Valley near the southern end of the Appalachian system; founded and inc. . the University of North Carolina state in the SE United States. It is bordered by the Atlantic Ocean (E), South Carolina and Georgia (S), Tennessee (W), and Virginia (N).
Facts and Figures
Area, 52,586 sq mi (136,198 sq km). Pop. at Charlotte, and Miami
University main campus at Oxford, Ohio; coeducational; state supported; chartered 1809, opened 1824. The library has extensive collections in literature and American history, including the William Holmes McGuffey Library and Museum and the Edgar W. . The authors appreciate the comments of David Yermack, Anup
Agrawal (Hindi अग्रवाल or अगरवाल) are a community in India. , Bruce Scottish royal family descended from an 11th-century Norman duke, Robert de Brus. He aided William I in his conquest of England (1066) and was given lands in England. Barrett, Brian Carver /car·ver/ () a tool for producing anatomic form in artificial teeth and dental restorations.
n . Doug Cook, John Finnerty, Xudong
Fu, Daniel Bradley, Alexander Kurov, Raman Kumar, Laura Starks, Mike
Lemmon, Qin Lian, John Olagues, Pam Peterson, Harris Schlesinger,
Wei-Ling Song, Jonathan Stanley town (1991 pop. 1,557), capital of the Falkland Islands, S Atlantic Ocean, on East Falkland island. It is the main port and trading center of the islands. The name is sometimes written as Port Stanley. , Jun Wang, Jing () [Chinese] one of the basic substances that according to traditional Chinese medicine pervade the body, usually translated as "essence"; the body reserves or constitutional makeup, replenished by food and rest, that supports Zhao, and an anonymous
Aboody, D., J. Hughes, J. Liu, and W Su, 2008, "Are Executive
Stock Option Exercises Driven by Private Information?" Review of
Accounting Studies 13, 551-570.
Barber, B.M. and J.D. Lyon, 1996, "Detecting Abnormal
Operating Performance: The Empirical Power and Specification of Test
Statistics," Journal of Financial Economics 41, 359-399.
Barber, B.M. and J.D. Lyon, 1997, "Detecting Long-Run Abnormal
Stock Returns: The Empirical Power and Specification of Test
Statistics," Journal of Financial Economics 43, 341-372.
Bartov, E. and P. Mohanram, 2004, "Private Information,
Earnings Manipulations, and Executive Stock Option Exercises,"
Accounting Review 79, 889-920.
Bergstresser, D. and T. Philippon, 2006, "CEO Incentives and
Earnings Management," Journal of Financial Economics 80, 511-529.
Bettis, J.C., J.M. Bizjak, and M. Lemmon, 2005, "Exercise
Behavior, Valuation, and the Incentive Effects of Employee Stock
Options," Journal of Financial Economics 76, 445-470.
Bettis, J.C., J.L. Coles, and M. Lemmon, 2000, "Corporate
Policies Restricting Trading by Insiders," Journal of Financial
Economics 57, 191-220.
Billette, M.T., D.C. Mauer, and Y. Zhang, 2010, "Stockholder
An individual or institution that owns bonds in a corporation or other organization. Wealth Effects of CEO Incentive Grants," Financial
Management 39, 463-487.
Broussard, J.P., S.A. Buchenroth, and E.A. Pilotte, 2004, "CEO
Incentives, Cash Flow, and Investment," Financial Management 33,
Cai, J., 2007, "Executive Stock Option Exercises: Good Timing
or Backdating," Drexel University at Philadelphia, Pa.; coeducational; founded 1891 by Anthony J. Drexel, opened 1892, chartered 1894 as Drexel Institute of Art, Science, and Industry. It was renamed Drexel Institute of Technology in 1936 and gained university status in 1970. Working Paper.
Carhart, M.M., 1997, "On Persistence in Mutual Fund
Performance," Journal of Finance 52, 57-82.
Carpenter, J. and B. Remmers, 2001, "Executive Stock Option
Exercises and Inside Information," The Journal of Business 74,
Cohen, L., C. Malloy, and L. Pomorski, 2012, "Decoding
tr.v. de·cod·ed, de·cod·ing, de·codes
1. To convert from code into plain text.
2. To convert from a scrambled electronic signal into an interpretable one.
Information," Journal of Finance, forthcoming.
Cicero, D.C., 2009, "The Manipulation of Executive Stock
Option Exercise Strategies: Information Timing and Backdating,"
Journal of Finance 64, 2627-2663.
Cline, B.N. and X. Fu, 2010, "Executive Stock Option Exercise
and Seasoned Equity Offerings," Financial Management 39, 1643-1670.
Core, J.E. and W.R. Guay, 2001, "Stock Option Plans for
Non-Executive Employees," Journal of Financial Economics 61,
Cuny, C. and P. Jorion, 1995, "Valuing Executive Stock Options
with Endogenous /en·dog·e·nous/ () produced within or caused by factors within the organism.
1. Originating or produced within an organism, tissue, or cell. Departure," Journal of Accounting and Economics 20,
Dhaliwal, D., M. Erickson, and S. Heitzman, 2009, "Taxes and
the Backdating of Stock Option Exercise Dates," Journal of
Accounting and Economics 47, 27-49.
Fama, E.F. and K.R. French, 1993, "Common Risk Factors in the
Returns on Stocks and Bonds," Journal of Financial Economics 33,
Fama, E.F., 1998, "Market Efficiency, Long-term Returns, and
Behavioral Finance," Journal of Financial Economics 49, 283-306.
Finnerty, J.F., 1974, "Special Information and Insider
Trading," Journal of Finance 31, 1141-1148.
Fu, X. and J.A. Ligon, 2009, "Exercises of Executive Stock
Options on the Vesting Date," Financial Management 39, 1097-1126.
Gombola, M., H.W. Lee, and F. Liu, 1997, "Evidence of Selling
by Managers after Seasoned Equity Offering Announcements,"
Financial Management 26, 37-53.
Goolsbee, A., 2000, "What Happens When You Tax the Rich?
Evidence from Executive Compensation," Journal of Political Economy
Hemmer, T., S. Matsunaga, and T. Shevlin, 1996, "The Influence
of Risk Diversification on Early Exercise of Employee Stock Options by
Executive Officers," Journal of Accounting and Economics 21, 45-68.
Huddart, S. and M. Lang, 1996, "Employee Stock Option
Exercise: An Empirical Analysis," Journal of Accounting and
Economics 21, 5-43.
Huddart, S. and M. Lang, 2003, "Information Distribution
within Firms: Evidence from Stock Option Exercises," Journal of
Accounting and Economics 35, 315-344.
Inci, A.C., B. Lu, and H.N. Seyhun, 2010, "Intraday Behavior
of Stock Prices and Trades Around Insider Trading," Financial
Management 39, 323-363.
Jaffe, J.F., 1974, "Special Information and Insider
Trading," The Journal of Business 47, 410-428.
Jegadeesh, N., 2000, "Long Term Performance of Seasoned Equity
Offerings: Benchmark Errors and Biases in Expectations," Financial
Management 29, 5-30.
Jegadeesh, N. and S. Titman, 1993, "Returns to Buying Winners
and Selling Losers: Implications for Stock Market Efficiency,"
Journal of Finance 48, 65-91.
Jeng, L., A. Metrick, and R. Zeckhauser, 2003, "Estimating the
Returns to Insider Trading: A Performance Evaluation Perspective,"
Review of Economics and Statistics 85, 453-471.
Johnson, N. J., 1978, "Modified t tests and Confidence
Intervals for Asymmetrical or a·sym·met·ric
adj. Abbr. a
Lacking symmetry between two or more like parts; not symmetrical. Populations," Journal of the American
Statistics Association 73, 536-544.
Karpoff, J. and D. Lee, 1991, "Insider Trading Before New
Issue Announcements," Financial Management 20, 18-26.
Kyriacou, K., K.B. Luintel, and B. Mase, 2010, "Private
Information in Executive Stock Option Trades: Evidence of Insider
Trading in the UK," Economica 77, 751-774.
Lakonishok, J. and I. Lee, 2001, "Are Insiders' Trades
Informative?" Review of Financial Studies 14, 79-111.
Lambert, R., D. Larcker, and R. Verrecchia, 1991, "Portfolio
Considerations in Valuing Executive Compensation," Journal of
Accounting Research 29, 129-149.
Laux, V., 2010, "On the Benefits of Allowing CEOs to Time
their Stock Option Exercises," Rand
See Table at currency.
[Afrikaans, after(Witwaters)rand. Journal of Economics 41,
Lin, J-C. and J. Howe, 1990, "Insider Trading in the OTC
Market," Journal of Finance 45, 1273-1284. Loughran, T. and J.
Ritter, 1995, "The New Issue Puzzle," Journal of Finance 50,
Loughran, T. and J. Ritter, 2000, "Uniformly Least Powerful
Tests of Market Efficiency," Journal of Financial Economics 55,
Lyon, J., B. Barber, and C. Tsai, 1999, "Improved Methods for
Tests of Long-Run Abnormal Stock Returns," Journal of Finance 59,
McDonald, R. L., 2003, "Is it Optimal to Accelerate the
Payment of Income Tax on Share-Based Compensation?" Northwestern
University mainly at Evanston, Ill.; coeducational; chartered 1851, opened 1855 by Methodists. In 1873 it absorbed Evanston College for Ladies. Working Paper.
Niehaus, G. and G. Roth, 1999, "Insider Trading, Equity
Issues, and CEO Turnover in Firms Subject to Securities Class
Action," Financial Management 28, 52-72.
Ritter, J.R., 1991, "The Long-run Performance of Initial
Public Offerings," Journal of Finance 46, 3-27.
Rogers, D.A., 2005, "Managerial Risk-Taking Incentives and
Executive Stock Option Repricing: A Study of US Casino or .
1 Card game played with a full deck by two to four players. Its origins are obscure though it probably traces back to the Italian game of Scopa. Executives,"
Financial Management 34, 95-121.
Rozeff, M.S. and M.A. Zaman, 1988, "Overreaction
intr.v. o·ver·re·act·ed, o·ver·re·act·ing, o·ver·re·acts
To react with unnecessary or inappropriate force, emotional display, or violence. and Insider
Trading: New Evidence," Journal of Finance 46, 3-27.
Safdar, I., 2004, "Stock Option Exercise, Earnings Management,
and Abnormal Stock Returns," University of Rochester Working Paper.
Seyhun, H. N., 1998, "Investment Intelligence from Insider
Trading," Cambridge, MIT Press.
Spiess, K.D. and J. Affleck-Graves, 1995, "Underperformance in
Long-run Stock Returns Following Seasoned Equity Offerings,"
Journal of Financial Economics 38, 243-268.
Wei, Y., 2004, "Executive Stock Option Exercises, Insider
Information and Earnings Management," University of Utah Working
Yermack, D., 2006, "Golden Handshakes: Separation Pay for
Retired and Dismissed CEOs," Journal of Accounting and Economics
Zhang, Y., 2009, "Are Debt and Incentive Compensation
Substitutes in Controlling the Free Cash Flow Agency Problem?"
Financial Management 38, 507-541.
(1) Recent work by Cohen, Mallory, and Pomorski (2012) find that
while routine open market transactions on average do not contain private
information, a trading strategy focusing on nonroutine stock
transactions earns significant abnormal returns. Interestingly, they
also find that local insiders have trades that are particularly
predictive of future information events. Inci, Lu, and Seyhun (2010)
also find significant new information about intraday stock price
behavior based on insider trading.
(2) Aboody et al. (2008) present some evidence to support the view
that it may be optimal to exercise and hold based on an optimistic view
of the stock. They argue that dividends (an issue we address), taxes,
and suboptimality could result in exercise and holding of the shares. As
noted, we focus strictly on exercise and sale strategies.
(3) Form 3 is called "Initial Statement of Beneficial
Ownership of Securities." Form 4 is called "Statement of
Changes in Beneficial Ownership," which covers purchases, sales,
option grants, option exercises, gifts, and any other transaction that
changes an insider's ownership position. Form 5 is called
"Annual Statement of Changes in Beneficial Ownership of
Securities" and contains information regarding activity for exempt
transactions, which includes small transactions and small transfers
within the company that are not required in Form 4.
(4) It is possible that the dividends themselves contain
information. A decrease (increase) in the dividend could, for example,
signal the expectation of poor (good) firm-specific performance.
Exercises that appear to be motivated by dividend capture could contain
private information, but consistent with our hypothesis,
dividend-motivated exercises should be less likely to be motivated by
private information than exercises not motivated by dividends.
(5) This procedure requires two important choices: the number of
firms and the matching variables. Barber and Lyon (1997) make no
recommendation on the number of firms. Jegadeesh (2000) uses ten firms.
We attempted to use ten firms but were unable to obtain a sufficiently
high quality match on size, book-to-market, and prior firm performance
for firms without the corresponding type of exercise, probably due to
the fact that executive stock option exercises are such a common event.
Choosing a small number of firms makes the benchmark relatively
undiversified, which will make it more difficult to find consistent
results. To this extent, our results are biased against finding abnormal
returns. With respect to the matching variables, we match on the most
common criteria of industry, size, book-to-market, and prior firm
performance, which should be adequate for obtaining a satisfactory
(6) The use of industry as a matching factor is fairly standard but
comes at the expense of preventing us from examining whether executives
exercise based on information about the industry. We regard this broader
interpretation of private information as an interesting question for
future research that will require careful attention to this empirical
(7) When imposing a sequential matching procedure, there exists the
decision of which matching criteria to hold more/less restricted. For
our purpose, we choose to be most restrictive with prior firm
performance. The associated cost is the loss of sample size and other
matching statistics, which are not as strong. Loughran and Ritter (1995)
point out this problem and they control only for size. For our full
sample, there is no statistically significant difference in prior firm
performance between the mean sample firm and the portfolio of matching
firms, with the actual difference of seven basis points. The sample firm
is statistically smaller than the average of the portfolio of matching
firms; however, the average difference is less than 2%. Since our
procedure takes the best firm with respect to book-to-market after the
other criteria have been met, the sample average book-to-market ratio is
0.46, whereas the ratio of the portfolio of matching firms is 0.37,
which is statistically significant. Jegadeesh (2000) reports similar
differences in matching characteristics with respect to size and
book-to-market when using a similar matching method. He illustrates that
his approach is well-specified. Spiess and Affleck-Graves (1995) report
similar differences regarding size for their size and industry matched
portfolio of firms and for their industry, size, and book-to-market
matching procedure. Ritter (1991) reports significant differences with
respect to industry and size when using an industry and size matching
methodology. The differences in matching characteristics are
qualitatively similar for all stratifications of the full sample.
(8) To calculate the bootstrapped skewness-adjusted test statistic
we closely follow the method laid out in Lyon, Barber, and Tsai (1999).
Their bootstrapped method is performed on the series of abnormal
returns. We apply the same procedure on our set of BHARs, where the
BHARs are estimated instead from a portfolio of control firms. We begin
by calculating the skewness-adjusted test statistic (Johnson, 1978),
which adjusts the conventional test statistic by two terms which are a
function of the skewness of the distribution of abnormal returns. We
next construct a bootstrapped distribution of the skewness-adjusted test
statistic by drawing 1,000 bootstrapped resamples from the original
sample of abnormal returns (each of the of n/4 size). In each resample,
we calculate the skewness-adjusted test statistic (see Lyon, Barber,
Tsai, 1999, Equation (6)). The resulting distribution of the test
statistics is then used to assess significance.
(9) This procedure is also used by Cicero (2009). Carpenter and
Remmers (2001) likewise recognize this problem with their monthly data
and its potential for yielding biased test statistics. They use only the
first exercise of an executive of a firm in a month. In a later set of
tests that focus on the proportion of options exercised we eliminate
multiple exercises only at the executive level since in this context the
measure is specific to each executive and not the firm.
(10) It may be noted that the returns in Table II imply a large
average market return of about 23%. We independently verified
tr.v. ver·i·fied, ver·i·fy·ing, ver·i·fies
1. To prove the truth of by presentation of evidence or testimony; substantiate.
2. that this
was the correct average return on the benchmark CRSP equally weighted
portfolio for the year following all exercises.
(11) Although these exercises occur after the vest date (no more
than 30 days), we refer to them as occurring on the vest date. This
terminology is used only for expositional ease.
(12) In understanding the nature of a dividend-motivated exercise,
note that the ex-dividend decline in the stock price is passed on to the
option, so the option value will fall with certainty. Exercise and
holding of the stock through the ex-dividend date will result in capture
of the dividend, which mitigates the anticipated loss in the option.
Nonetheless, early exercise also results in discarding the remaining
time value. It is the trade-off of these costs and benefits that
determine whether to exercise in anticipation of a dividend, in the case
of executive stock options where the stock is immediately sold, which
are the case for all of our exercises, there is no direct capture of the
dividend. Nonetheless, the stock is sold at the cure-dividend price.
Hence, the executive effectively achieves the same result, except that
exercise may not occur at the last instant before going ex-dividend,
thereby resulting in the loss of greater time value. Standard option
theory suggests that such exercises are suboptimally timed, but
executive stock options possess different characteristics from standard
traded options, and executives are less likely to be as attuned
tr.v. at·tuned, at·tun·ing, at·tunes
1. To bring into a harmonious or responsive relationship:
optimally timing their exercises than are professional option traders.
If the exercises we classify as dividend-motivated are not truly
dividend-motivated, then our tests are likely to fail. As we see here,
they do not.
(13) There are two possible explanations for why some negative
stock price performance was found in the dividend- motivated sample. One
is that blackout periods, which have been studied by Bettis, Coles, and
Lemmon (2000), could interact with dividends. Thus, some apparently
dividend-motivated exercises could occur as option holders come out of
blackout periods. In addition, some apparently dividend-motivated
exercises could occur prior to periods of poor abnormal performance
because the dividend was decreased and provided a negative signal. But
if there is such a bias in our results, then exercises motivated by
private information will simply be included in the dividend-motivated
exercises sample and make it more difficult to find evidence of the use
of private information. Thus, any such bias would be conservative and
would seem relevant only if we had not found significant differences.
(14) Although these samples start off as quintiles with an
equivalent number &components, further restrictions, principally the
creation of the matching portfolio, result in uneven numbers in each
(15) Although not shown here, we also compared Money 2 against
Money 4, and the results are also highly significant in the expected
direction. That is, exercises in which the options are more costly to
exercise are followed by negative and significantly lower abnormal
returns than are those in which the options are less costly to exercise.
(16) Executives are likely to begin liquidating their option
portfolios well in advance of their actual departure. They also
typically have a specified number of days after departure to exercise
before they must forfeit the options. For these reasons, we separate
exercises within plus or minus 270 days of the reported departure date.
For robustness, we also separated exercises within plus or minus 90,
180, 300, 330, and 365 days of departure. Results are consistent with
(17) We should add that it is common for departing executives to
receive large severance packages. Yermack (2006) estimates that, for
CEO's of Fortune 500 companies in virtually the same period studied
here, the average package amounts to $5.4 million. The avoidance of
future losses by exercising options on private information could be
viewed as another element of these "golden handshakes."
(18) We conducted the same test for the full departure sample
(resignations and retirements) and for the retirement subsample but the
differences were not significant.
(19) For these tests, we eliminate multiple exercises at the
executive level rather than firm level as in the other tests since each
proportion is executive-specific and can vary with the cross-section of
executives at a given firm.
(20) The moneyness variable is the stock price divided by the
exercise price. Hence, the positive significance means that the BHARs
are more positive the deeper in-the-money is the option. Hence, they are
more negative (or less positive) the less deep in-the-money they are.
Robert Brooks, Don M. Chance, and Brandon Cline *
* Robert Brooks is a Professor o['Finance at the University of
Alabama in Tuscaloosa, AL. Don M. Chance is a Professor of Finance at
Louisiana State University in Baton Rouge [Fr.,=red stick], city (1990 pop. 219,531), state capital and seat of East Baton Rouge parish, SE La. , LA. Brandon N. Cline is an
Assistant Professor of Finance at Mississippi State University at Mississippi State, near Starkville; land-grant and state supported; coeducational; chartered 1878 as an agricultural and mechanical college, opened 1880. From 1932 to 1958 it was known as Mississippi State College. in
Mississippi , one of the Deep South states of the United States. It is bordered by Alabama (E), the Gulf of Mexico (S), Arkansas and Louisiana, with most of the border formed by State, MS.
Table I. Descriptive Statistics of the Exercise Data Panel A reports statistics for the full sample, which includes option exercises by corporate executives reported to the SEC between 1996 and 2005 on Thomson Financial Insider Filings Data (TFI) where the stock is immediately sold. The full sample is partitioned according to whether the exercise occurred early or at expiration, whether the exercise occurred on the vest date, and whether the exercise was motivated by capture of a dividend. "Exercised early" have more than 30 days to expiration. "Exercised on vest date" are between zero and 30 days after the vest date. "Exercised not on vest date" includes all other early exercises. "Dividend motivated" exercise includes exercises within 15 business days prior to the ex-dividend date. "Not dividend motivated" exercise represents exercises seven to ten weeks before the dividend or exercises of firms that do not pay dividends. "Money 1-Money 5" refers to quintiles based on the moneyness of the options at exercise with Money 1 closest to at- the-money and Money 5 deepest in-the-money. For "Proportion 1- Proportion 5," the number of early options exercised are aggregated by executive per transaction date and ranked according to the amount of options exercised relative to total options and placed into quintiles. Proportion 1 contains exercises with the highest ratio of exercises to total options held, and Proportion 5 contains those with the lowest ratio. Panel B reports statistics for the merged sample, comprised of insider trades reported in the Table II File of TFI for the period 1996 through 2005 in which the transactions can be matched with the insider data in S&P's ExecuComp. The merged sample is partitioned similarly to the full sample and according to whether the exercise was induced by executive departure from the firm. Exercises associated with departure are those early exercises that occur within 270 days of the executive leaving the firm. Resigned (retired) exercises are those in which the reason for departure was resignation (retirement). Before (after) resignation, option exercises are exercises that occur before (after) the announced date of resignation. Panel A. Summary Statistics for Full Sample of Option Exercises Number Number Number of of of exercises executives firms Full Sample 31,505 11,654 3,393 Exercised early 29,730 10,610 3,210 Exercised at expiration 1,100 1,054 525 Exercise not on vest date 29,241 10,360 3,173 Exercised on vest date 1,064 819 499 Not dividend motivated 22,417 8,375 2,923 Dividend motivated 3,502 2,024 807 Money 1 8,929 3,352 1,719 Money 2 8,166 3,428 1,788 Money 3 6,801 3,111 1,512 Money 4 5,628 2,907 1,175 Money 5 3,817 2,264 777 Proportion 1 4,541 2,427 1,472 Proportion 2 4,750 2,333 1,619 Proportion 3 4,707 2,124 1,555 Proportion 4 4,345 2,091 1,257 Proportion 5 4,765 1,575 999 Panel B. Summary Statistics for Merged Sample of Option Exercises Merged Sample 8,907 2,848 1,181 No executive departure 7,700 2,471 1,097 Executive departure 594 207 201 Executive resigned 208 94 76 Executive retired 232 108 95 Before resignation 145 84 72 After resignation 63 13 12 Panel A. Summary Statistics for Full Sample of Option Exercises Mean Mean years Mean years to after moneyness expiration vesting (median) (median) (median) Full Sample 4.39 (4.68) 3.51 (3.02) 44.35(2.65) Exercised early 4.58 (4.87) 3.46 (2.98) 47.69(2.66) Exercised at expiration 0.03 (0.02) 4.86 (4.94) 7.77(2.10) Exercise not on vest date 4.54 (4.82) 3.65 (3.18) 47.64(2.67) Exercised on vest date 5.80 (6.02) 0.01 (0.00) 47.23(2.55) Not dividend motivated 4.63 (4.95) 3.40 (2.92) 62.91(2.88) Dividend motivated 4.16 (4.19) 3.84 (3.56) 3.62(2.20) Money 1 5.24 (5.90) 2.61 (1.85) 1.31(1.33) Money 2 5.10 (5.54) 3.00 (2.41) 1.87(1.85) Money 3 4.66 (4.98) 3.35 (2.76) 2.70(2.67) Money 4 4.39 (4.62) 3.73 (3.34) 4.44(4.27) Money 5 4.34 (4.52) 3.67 (3.22) 227.46(11.80) Proportion 1 4.82 (5.28) 3.05 (2.54) 13.55(2.51) Proportion 2 5.07 (5.62) 3.08 (2.45) 10.69(2.39) Proportion 3 5.16 (5.59) 3.09 (2.57) 19.79(2.77) Proportion 4 4.92 (5.38) 3.37 (2.79) 13.73(3.17) Proportion 5 5.10 (5.49) 3.40 (2.82) 99.78(3.24) Panel B. Summary Statistics for Merged Sample of Option Exercises Merged Sample 3.85 (3.85) 3.96 (3.49) 19.77(2.84) No executive departure 4.02 (4.01) 3.94 (3.46) 22.37(2.91) Executive departure 4.46 (4.81) 3.60 (3.20) 3.97(2.20) Executive resigned 5.07 (5.71) 3.36 (2.93) 4.40(2.15) Executive retired 4.29 (4.76) 3.55 (3.84) 2.46(2.10) Before resignation 4.99 (5.17) 3.17 (2.17) 4.42(2.35) After resignation 5.26 (6.03) 3.72 (2.96) 4.37(1.69) Table II. Abnormal Returns for Full and Merged Samples The sample is drawn from the full sample described in Table I by selecting only the first exercise for a firm in a given day. The merged sample is comprised of the subset of insider trades reported in the Thomson Financial Insider Filings Data for the period 1996 through 2005 in which the transactions can be matched to ExecuComp. There are 31,505 full sample exercises and 8,907 merged sample exercises. The buy-and-hold abnormal returns (BHARs) benchmark is the portfolio of five firms matched on industry, size, book-to-market, and prior performance. Days relative to option exercise are reported in parentheses. Market adjusted returns (MARS) are calculated using the CRSP equally weighted index. CTARs are the intercept of Fama-French four- factor calendar-time portfolio regressions and reported as the average daily abnormal return over the postexercise period. Significance for BHARs is inferred from the Lyon, Barber, and Tsai (1999) bootstrapped t-test and from a traditional t-test for the RAW, MAR, and CTARs. (1) Full Sample (2) Merged Sample Buy and Hold Abnormal Returns BHAR (-12,-1) mo 0.08 0.12 BHAR (0,12) mo -1.27 *** 0.75 BHAR (0,252) -3.31 *** -0.59 * BHAR (0,125) -2.06 *** -0.81 *** BHAR (0,60) -1.77 *** -1.05 *** BHAR (0,20) -0.69 *** -0.66 *** BHAR (0,10) -0.31 *** -0.31 *** Calendar-Time Portfolio Abnormal Returns CTAR (0,252) -0.04 *** -0.04 *** CTAR (0,125) -0.05 *** -0.06 *** CTAR (0,60) -0.06 *** -0.06 *** CTAR (0,20) -0.05 *** -0.06 ** CTAR (0,10) -0.04 ** -0.03 Raw and Market Adjusted Returns RAW (0,252) 14.05 14.79 RAW (0,125) 6.03 6.68 RAW (0,60) 2.18 2.96 RAW (0,20) 0.78 0.92 RAW (0,10) 0.47 0.49 MAR (0,252) -9.38 *** -8.33 *** MAR (0, 125) -5.73 *** -5.01 *** MAR (0,60) -3.83 *** -3.20 *** MAR (0,20) -1.45 *** -1.43 *** MAR (0,10) -0.67 *** -0.70 *** *** Significance at the 0.01 level. ** Significance at the 0.05 level. * Significance at the 0.10 level. Table III. Abnormal Returns for Early vs. Maturity and Vest vs. Not-Vest Exercises The sample is drawn from the full sample described in Table I by selecting only the first exercise for a firm in a given day. An early exercise is defined as an exercise with more than 30 days remaining to expiration. A vest date exercise is an exercise defined as occurring between 0 and 30 days after the vest date. There are 29,730 early exercises, 1,100 maturity exercises, 29,241 not-vest date exercises and 1,064 vest date exercises. The BHAR benchmark is the portfolio of five firms matched on industry, size, book-to-market, and prior performance. Days relative to option exercise are reported in parentheses. MARS are calculated using the CRSP equally weighted index. CTARs are the intercept of Fama-French four-factor calendar-time portfolio regressions and reported as the average daily abnormal return over the postexercise period. Significance for BHARs is inferred from the Lyon, Barber, and Tsai (1999) bootstrapped t-test and from a traditional t-test for the RAW, MAR, and CTARs. (2) Early Maturity exercise exercise Buy and Hold Abnormal Returns BHAR (-12,-1)mo 0.08 0.03 BHAR(0,12)mo -1.38 *** 1.31 BHAR (0,252) -3.47 *** -1.06 BHAR (0,125) -2.28 *** -1.05 BHAR (0,60) -1.89 *** -0.11 BHAR (0,20) -0.72 *** -0.27 BHAR (0,10) -0.34 *** -0.13 Calendar-Time Portfolio Abnormal Returns CTAR (0,252) -0.05 *** -0.01 CTAR (0,125) -0.06 *** -0.01 CTAR (0,60) -0.06 *** -0.02 CTAR (0,20) -0.06 *** -0.03 CTAR (0,10) -0.05 * -0.01 Raw and Market Adjusted Returns RAW (0,252) 13.84 18.95 RAW (0,125) 5.84 8.72 RAW (0,60) 2.03 4.94 RAW (0,20) 0.76 1.73 RAW (0,10) 0.46 0.68 MAR (0,252) -9.34 *** -8.24 *** MAR (0, 125) -5.80 *** -4.81 *** MAR (0,60) -3.94 *** -2.44 *** MAR (0,20) -1.47 *** -0.99 *** MAR (0,10) -0.68 *** -0.71 *** (l)-(2) (3) Difference Not-vest exercise Buy and Hold Abnormal Returns BHAR (-12,-1)mo 0.05 0.08 BHAR(0,12)mo -2.68 ** -1.61 *** BHAR (0,252) -2.41 ** -3.68 *** BHAR (0,125) -1.23 * -2.30 *** BHAR (0,60) -1.79 *** -1.87 *** BHAR (0,20) -0.45 * -0.72 *** BHAR (0,10) -0.21 * -0.36 *** Calendar-Time Portfolio Abnormal Returns CTAR (0,252) -0.04 *** -0.05 *** CTAR (0,125) -0.05 *** -0.06 *** CTAR (0,60) -0.04 *** -0.05 ** CTAR (0,20) -0.03 ** -0.07 ** CTAR (0,10) -0.04 ** -0.05 * Raw and Market Adjusted Returns RAW (0,252) -5.11 *** 13.54 RAW (0,125) -2.88 *** 5.72 RAW (0,60) -2.91 *** 1.97 RAW (0,20) -0.96 *** 0.74 RAW (0,10) -0.23 0.44 MAR (0,252) -1.09 -9.58 *** MAR (0, 125) -1.00 -5.91 *** MAR (0,60) -1.50 *** -4.00 *** MAR (0,20) -0.48 * -1.49 *** MAR (0,10) 0.03 -0.70 *** (4) (3)-(4) Vest Difference exercise Buy and Hold Abnormal Returns BHAR (-12,-1)mo 0.10 -0.02 BHAR(0,12)mo 2.43 -4.04 * BHAR (0,252) -0.26 -3.42 ** BHAR (0,125) -2.18 ** -0.13 BHAR (0,60) -1.47 ** -0.40 *** BHAR (0,20) -0.08 -0.64 * BHAR (0,10) -0.01 -0.35 * Calendar-Time Portfolio Abnormal Returns CTAR (0,252) -0.02 ** -0.03 ** CTAR (0,125) -0.04 ** -0.02 * CTAR (0,60) -0.01 -0.04 *** CTAR (0,20) 0.00 -0.07 *** CTAR (0,10) -0.01 -0.04 * Raw and Market Adjusted Returns RAW (0,252) 16.44 -2.90 ** RAW (0,125) 5.90 -0.18 RAW (0,60) 2.53 -0.56 RAW (0,20) 1.52 -0.77 ** RAW (0,10) 0.94 -0.51 ** MAR (0,252) -6.39 *** -3.20 ** MAR (0, 125) -4.83 *** -1.09 MAR (0,60) -3.07 *** -0.93 * MAR (0,20) -0.67 ** -0.82 ** MAR (0,10) -0.36 * -0.34 * *** Significance at the 0.01 level. ** Significance at the 0.05 level. * Significance at the 0.10 level. Table IV. Abnormal Returns for Dividend vs. Not-Dividend Motivated Exercised and Exercises Classified by Moneyness A dividend-motivated exercise is defined as an early exercise in which the reported exercise date occurs within 15 trading days prior to the ex-dividend date. Not-dividend motivated exercises also include exercises of firms that do not pay dividends. Money I quintile is the 20% of exercises that are closest to at-the- money and Money 5 quintile is the 20% of exercises that are deepest in-the-money. These samples are drawn from the full sample described in Table I by selecting only the first exercise for a firm in a given day. There are 22,417 not-dividend motivated exercises, 3,502 dividend motivated exercises, 8,929 exercises in Money 1, and 3,817 in Money 5. The BHAR benchmark is the portfolio of five firms matched on industry, size, book-to- market, and prior performance. Days relative to option exercise are reported in parentheses. MARS are calculated using the CRSP equally weighted index. CTARs are the intercept of Fama-French four-factor calendar-time portfolio regressions and reported as the average daily abnormal return over the postexercise period. Significance for BHARs is inferred from the Lyon, Barber, and Tsai (1999) bootstrapped t-test and from a traditional t-test for the RAW, MAR, and CTARs. (1) (2) (1)-(2) Not dividend- Dividend- Difference motivated motivated Money exercise exercise Buy and Hold Abnormal Returns BHAR (-12,-1)mo 0.07 0.11 -0.04 BHAR(0,12)mo -2.24 *** 0.48 -2.72 *** BHAR (0,252) -4.61 *** -0.74 * -3.87 *** BHAR (0,125) -3.07 *** -0.03 -3.04 *** BHAR (0,60) -2.25 *** -0.56 *** -1.69 *** BHAR (0,20) -0.82 *** -0.57 *** -0.25 * BHAR (0,10) -0.43 *** 0.17 ** -0.27 ** Calendar-Time Portfolio Abnormal Returns CTAR (0,252) -0.06 *** -0.02 * -0.05 *** CTAR (0,125) -0.07 *** -0.01 -0.06 *** CTAR (0,60) -0.05 ** -0.03 -0.03 *** CTAR (0,20) -0.07 *** -0.03 * -0.04 ** CTAR (0,10) -0.07 ** 0.01 -0.08 *** Raw and Market Adjusted Returns RAW (0,252) 12.47 17.27 -4.79 *** RAW (0,125) 4.89 8.75 -3.86 *** RAW (0,60) 1.41 3.93 -2.52 *** RAW (0,20) 0.49 1.42 -0.93 *** RAW (0,10) 0.28 0.93 -0.65 *** MAR (0,252) -10.70 *** -6.16 *** -4.54 *** MAR (0,125) -6.79 *** -2.82 *** -3.97 *** MAR (0,60) -4.49 *** -2.25 *** -2.24 *** MAR (0,20) -1.64 *** -1.11 *** -0.54 *** MAR (0,10) -0.81 *** -0.36 *** -0 45 *** (3) (4) (3)-(4) Money Money Difference 1 quintile 5 quintile Buy and Hold Abnormal Returns BHAR (-12,-1)mo 0.08 0.10 -0.02 BHAR(0,12)mo -4.57 *** 6.61 -11.18 *** BHAR (0,252) -6.73 *** 3.04 -9.77 *** BHAR (0,125) -3.68 *** 0.28 -3.95 *** BHAR (0,60) -2.77 *** -0.61 * -2.16 *** BHAR (0,20) -1.08 *** 0.02 -1.11 *** BHAR (0,10) -0.45 *** -0.16 -0.30 *** Calendar-Time Portfolio Abnormal Returns CTAR (0,252) -0.06 *** -0.02 ** -0.03 *** CTAR (0,125) -0.07 *** -0.04 ** -0.03 *** CTAR (0,60) -0.07 *** -0.03 -0.04 *** CTAR (0,20) -0.07 ** -0.03 -0.04 * CTAR (0,10) -0.04 -0.02 -0.02 Raw and Market Adjusted Returns RAW (0,252) 10.91 19.40 -8.49 *** RAW (0,125) 5.05 7.73 -2.68 RAW (0,60) 1.63 2.61 -0.98 ** RAW (0,20) 0.58 1.11 -0.53 ** RAW (0,10) 0.40 0.56 -0.17 MAR (0,252) -10.61 *** -4.85 *** -5.76 *** MAR (0,125) -6.05 *** -4.50 *** -1.55 *** MAR (0,60) -4.10 *** -3.47 *** -0.63 * MAR (0,20) -1.59 *** -1.15 *** -0.44 ** MAR (0,10) -0.70 *** -0.63 *** -0.06 *** Significance at the 0.01 level. ** Significance at the 0.05 level. * Significance at the 0.10 level. Table V. Abnormal Returns for Departure vs. Not-Departure and Resigned vs. Retired The sample is drawn from the merged sample described in Table 1 by selecting only the first exercise for a firm in a given day. Exercises associated with departure are those early exercises which occurred within 270 days of the executive leaving the firm. Resigned (retired) exercises are those in which the reason for departure was resignation (retirement). There are 7,700 not departure exercises, 594 departure exercises, 232 retired exercises, 208 resignation exercises, 145 exercises before resignation, and 63 exercises after resignation. The BHAR benchmark is the portfolio of five firms matched on industry, size, book-to-market, and prior performance. Days relative to option exercise are reported in parentheses. MARS are calculated using the CRSP equally weighted index. CTARs are the intercept of Fama-French four-factor calendar-time portfolio regressions and reported as the average daily abnormal return over the postexercise period. Significance for BHARs is inferred from the Lyon, Barber, and Tsai (1999) bootstrapped t-test and from a traditional t-test for the RAW, MAR, and CTARs. (1) (2) (1)-(2) Not-Departure Departure Difference exercise exercise Buy and Hold Abnormal Returns BHAR(-12,-1)mo 0.12 0.06 0.06 BHAR (0,12)wo 0.62 -5.44 *** 6.06 BHAR (0,252) -0.56 * -7.15 *** 6.59 BHAR (0,125) -0.64 ** -5.25 *** 4.66 BHAR (0,60) -1.15 *** -2.86 *** 1.71 BHAR (0,20) -0.70 *** -1.52 ** 0.82 BHAR (0,10) -0.34 *** -0.92 ** 0.58 Calendar-Time Portfolio Abnormal Returns CTAR (0,252) -0.05 *** -0.08 *** 0.03 CTAR (0,125) -0.06 *** -0.10 ** 0.05 CTAR (0,60) -0.07 *** -0.08 ** 0.01 CTAR (0,20) -0.06 ** -0.04 0.02 CTAR (0,10) -0.04 ** -0.07 0.03 Raw and Market Adjusted Returns RAW (0,252) 14.74 7.24 7.50 RAW (0,125) 6.75 1.88 4.87 RAW (0,60) 2.73 1.60 1.13 RAW (0,20) 0.75 1.23 -0.48 RAW (0,10) 0.39 0.38 0.01 MAR (0,252) -8.40 *** -12.59 *** 4.19 MAR(0,125) -4.93 *** -8.60 *** 3.67 MAR (0,60) -3.38 *** -4.87 *** 1.49 MAR (0,20) -1.56 *** -1.58 *** 0.03 MAR(0,10) -0.76 *** -1.05 *** 0.29 (3) (4) (3)-(4) Retired Resigned Difference exercise exercise Buy and Hold Abnormal Returns BHAR(-12,-1)mo 0.06 0.13 -0.07 BHAR (0,12)wo -9.70 *** 0.33 -10.02 *** BHAR (0,252) -10.96 *** -3.13 -7.83 *** BHAR (0,125) -8.42 *** -2.46 * -5.96 *** BHAR (0,60) -4.88 *** -2.22 ** -2.61 *** BHAR (0,20) -2.49 *** -1.51 ** -0.98 ** BHAR (0,10) -0.96 *** -0.97 ** 0.01 Calendar-Time Portfolio Abnormal Returns CTAR (0,252) -0.08 *** -0.04' -0.04 *** CTAR (0,125) -0.10 *** -0.03 -0.07 *** CTAR (0,60) -0.08 ** -0.02 -0.06 ** CTAR (0,20) -0.()8 ** -0.02 -0.06 ** CTAR (0,10) -0.10 ** -0.02 -0.08 *** Raw and Market Adjusted Returns RAW (0,252) 8.41 8.29 0.11 RAW (0,125) 3.98 4.06 -0.08 RAW (0,60) 1.80 2.12 -0.32 RAW (0,20) 0.41 1.90 -1.48 ** RAW (0,10) 0.51 0.71 -0.20 MAR (0,252) -9.60 *** -10.05 *** 0.45 MAR(0,125) -6.67 *** -6.87 *** 0.20 MAR (0,60) -4.12 *** -4.60 *** 0.48 MAR (0,20) -2.37 *** -0.56 -1.81 ** MAR(0,10) -1.13 *** -0.16 -0.97 ** (5) (6) (5)-(6) Exercise Exercise Difference befere after resignation resignation Buy and Hold Abnormal Returns BHAR(-12,-1)mo 0.14 -0.05 0.19 BHAR (0,12)wo -3.49 * 8.34 -11.83 *** BHAR (0,252) -5.75 ** 4.21 -9.95 *** BHAR (0,125) -7.40 *** 10.32 -17.73 *** BHAR (0,60) -3.97 *** 1.46 -5.43 *** BHAR (0,20) -1.90 ** -1.08 -0.82 ** BHAR (0,10) -1.28 ** -0.94 -0.34 Calendar-Time Portfolio Abnormal Returns CTAR (0,252) -0.08 *** -0.01 -0.07 *** CTAR (0,125) -0.11 *** 0.04 -0.15 *** CTAR (0,60) -0.09 ** 0.02 -0.11 *** CTAR (0,20) -0.06 ** 0.02 -0.08 *** CTAR (0,10) -0.09 ** -0.01 -0.08 *** Raw and Market Adjusted Returns RAW (0,252) 9.92 -0.22 10.14 RAW (0,125) 0.52 7.51 -6.99 *** RAW (0,60) 0.73 2.34 -1.61 ** RAW (0,20) 0.84 3.88 -3.04 *** RAW (0,10) 0.07 1.26 -1.19 ** MAR (0,252) -10.97 *** -12.77 * 1.80 MAR(0,125) -11.13 *** -0.59 -10.54 *** MAR (0,60) -6.18 *** -4.17 ** -2.02 *** MAR (0,20) -2.07 ** 2.47 -4.53 *** MAR(0,10) -1.20 * 1.24 -2.44 *** Significance at the 0.01 level. ** Significance at the 0.05 level. * Significance at the 0.10 level. Table VI. Abnormal Returns for Proportion of Options Exercised The sample is drawn from the full sample described in Table 1. The number of options exercised early is divided by the total number of options in the executives' portfolio and placed into quintiles. Proportion 1 contains exercises with the highest ratio of options exercised and Proportion 5 contains those with the lowest. The number of exercises in each group is (1) 4,541, (2) 4,750, (3) 4,707, (4) 4,345, and (5) 4,765. The BHAR benchmark is the portfolio of five firms matched on industry, size, book-to-market, and prior performance. Days relative to option exercise are reported in parentheses. MARS are calculated using the CRSP equally weighted index. CTARs are the intercept of Fama-French four-factor calendar-time portfolio regressions and reported as the average daily abnormal return over the postexercise period. Significance for BHARs is inferred from the Lyon, Barber, and Tsai (1999) bootstrapped t-test and from a traditional t-test for the RAW, MAR, and CTARs. (1) (2) (3) Proportion Proportion Proportion 1 2 3 Buy and Hold Abnormal Returns BHAR (-12,-1)mo 0.09 0.08 0.11 BHAR (0,12)mo -4.00 *** -3.19 *** -1.57 ** BHAR (0,252) -5.35 *** -4.96 *** -3.37 *** BHAR (0,125) -2.82 *** -3.18 *** -1.73 *** BHAR (0,60) -2.20 *** -2.27 *** -1.70 *** BHAR (0,20) -1.10 *** -0.93 *** -0.59 *** BHAR (0,10) -0.36 *** -0.43 *** -0.31 *** Calendar--Time Portfolio Abnormal Returns CTAR (0,252) -0.05 *** -0.07 *** -0.05 *** CTAR (0,125) -0.05 *** -0.09 *** -0.07 *** CTAR (0,60) -0.07 ** -0.09 *** -0.08 *** CTAR (0,20) -0.09 *** -0.08 *** -0.07 *** CTAR (0,10) -0.04 * -0.11 *** -0.11 *** Raw and Market Adjusted Returns RAW (0,252) 9.82 11.86 13.85 RAW (0,125) 4.43 5.50 6.43 RAW (0,60) 1.49 1.96 2.53 RAW (0,20) 0.59 0.75 0.89 RAW (0,10) 0.64 0.37 0.48 MAR (0,252) -11.91 *** -11.37 *** -9.25 *** MAR (0,125) -6.56 *** -6.48 *** -5.78 *** MAR (0,60) -4.26 *** -4.26 *** -3.96 *** MAR (0,20) -1.80 *** -1.70 *** -1.40 *** MAR (0,10) -0.59 *** -0.84 *** -0.70 *** (4) (5) (1)-(5) Proportion Proportion Difference 4 5 Buy and Hold Abnormal Returns BHAR (-12,-1)mo 0.11 0.08 0.01 BHAR (0,12)mo -1.35 * 3.16 -7.16 *** BHAR (0,252) -3.30 *** 0.69 -6.04 *** BHAR (0,125) -0.88 ** -0.30 -2.52 *** BHAR (0,60) -0.95 ** -0.40 -1.79 *** BHAR (0,20) -0.21 -0.02 -1.07 *** BHAR (0,10) -0.29 ** 0.09 -0.45 Calendar--Time Portfolio Abnormal Returns CTAR (0,252) -0.02 * -0.01 -0.04 *** CTAR (0,125) -0.03 ** -0.01 -0.04 *** CTAR (0,60) -0.04 ** -0.02 * -0.05 *** CTAR (0,20) -0.03 -0.03 -0.06 *** CTAR (0,10) -0.04 0.01 -0.05 ** Raw and Market Adjusted Returns RAW (0,252) 14.78 19.23 -9.41 *** RAW (0,125) 7.97 8.15 -3.72 *** RAW (0,60) 3.22 3.37 -1.88 *** RAW (0,20) 1.10 0.99 -0.41 ** RAW (0,10) 0.35 0.55 0.09 MAR (0,252) -8.45 *** -6.21 *** -5.69 *** MAR (0,125) -4.42 *** -4.61 *** -1.94 *** MAR (0,60) -3.18 *** -3.05 *** -1.21 *** MAR (0,20) -1.20 *** -1.17 *** -0.62 MAR (0,10) -0.79 *** -0.47 *** -0.11 *** Significance at the 0.01 level. ** Significance at the 0.05 level. * Significance at the 0.10 level. Table VII. Multivariate Regression for Buy-and-Hold Long-Term Returns This table provides cross-sectional regression models on buy- and-hold abnormal returns (BHARs) from 0 to 252 trading days following exercise. Buy-and-hold abnormal returns (BHAR) are estimated using a benchmark portfolio constructed by matching on industry, size, book-to-market, and prior performance. Informed is a dummy variable equal to one if the exercise was early, not vest related, not dividend motivated, proportion exercised high, and moneyness low. High proportion exercised is defined as any exercise where the proportion exercised exceeds the median. Low moneyness is defined as any exercise where moneyness is less than the sample median moneyness. Early is a dummy equal to one if the exercise was not maturity induced. Not Vest equals one if the exercise did not occur between zero and 30 days after the vest date. DivMotiv equals one if the exercise occurred within 15 days prior to an ex-dividend date. Depart equals one if the exercise occurs with 270 days of the executive departure date. Resign and Retire are both dummies equal to one if the reported reason for departure was resigned or retire. PropEx is the proportion of options exercised relative to total options held. Moneyness is the stock price at exercise divided by the strike price. Reload equals one if the exercise was associated with an option reload. ExecRank is a dummy variable equal to one if the executive is listed as a C-level executive by the TFI. PreSOX is a dummy variable for whether the exercise occurred prior to the Sarbanes- Oxley Act. LogSize is the log of firm market capitalization in the fiscal year ending before exercise. BooktoMarket is ratio of book value to market value at the fiscal year end prior to exercise. Volatility is the standard deviation of stock returns computed 60 months prior to exercise. Momentum is the unadjusted 90-day stock return prior to exercise, and PriorFirmPerf is the firm performance in the year prior to the beginning of the momentum window. PriorMktPerf is the performance of the CRSP equal-weighted index in the one year preceding exercise. All regressions also include year and industry dummy variables. The number of observations in all regressions is 5,121. p-values are reported in parentheses. Model (1) Model (2) Model (3) Model (4) Constant -5.31 -3.10 -2.28 -4.14 (0.26) (0.25) (0.41) (0.26) Option-related variables Informed -4.99 -4.22 -- -- (0.02) (0.04) Early -4.10 -3.48 (0.01) (0.04) NotVest -2.72 -2.80 (0.03) (0.03) DivMotiv 3.26 3.38 (0.13) (0.12) Depart -10.72 -10.87 (0.04) (0.04) Resign 8.14 8.63 (0.22) (0.19) Retire 6.41 7.09 (0.33) (0.28) PropEx -1.26 -1.10 (0.00) (0.01) Moneyness 0.03 0.03 (0.00) (0.00) Other tests/controls. Reload -16.40 (0.09) ExecRank 2.41 (0.05) PreSOX -6.84 (0.10) Firm characteristics LogSize -0.58 -0.47 -0.42 (0.36) (0.46) (0.52) BooktoMarket 2.10 1.95 2.08 (0.35) (0.39) (0.36) Volatility 13.11 12.67 12.60 (0.00) (0.00) (0.00) Momentum -0.02 -0.03 -0.03 (0.49) (0.34) (0.32) PriorFirmPerf 0.10 0.11 0.11 (0.00) (0.00) (0.00) PriorMktPerf -0.01 -0.02 -0.01 (0.66) (0.56) (0.60) Adjusted [R.sup.2] 4.0% 4.84% 6.0% 6.12% Table VIII. Operating Performance This table presents accounting measures of firm performance following option exercises as evidenced by abnormal operating performance from one year prior to three years following exercise using the procedure outlined in Barber and Lyon (1996). The "Informed" sample is a subsample of exercises most likely to have been based on private information and determined by selecting early exercises that are not around the vest or ex-dividend date, moneyness exceeds the sample median, and the proportion exercised exceeds the sample median proportion exercised. The "Other" sample is the compliment of the "Informed" sample. Each observation where the firm had no other exercises over the following three-year period is matched to an industry control group in the year prior to exercise based on 2-digit SIC code and operating performance that is within 90% to 110% of the sample firm. Abnormal operating performance is defined as the difference between the observed operating performance of the sample firm and that of the industry control group. Four measures of operating performance are used. ROA is the return on assets, ROMV the return on market value, OPROA is EBITDA scaled by total assets, and OPROMV is defined as EBITDA scaled by total market capitalization. Panel A reports mean level abnormal operating performance for both the informed sample and the other exercise sample. Panel B reports mean changes in abnormal operating performance for both the informed sample and other exercises. Panel C presents median level abnormal operating performance, and Panel D median changes in abnormal operating performance. p- values from a one-tailed t-test are used for difference in means tests. Nonparametric Wilcoxon signed-rank tests are reported for medians. Panel A. Mean Levels in Operating Performance N ROA Year Informed Other Informed Other p-val -1 1,338 1,894 0.00 0.02 (0.26) 0 1,310 1,881 -0.68 1.54 (0.00) 1 1,222 1,723 -3.23 -0.23 (0.00) 2 1,063 1,558 -2.85 -0.15 (0.00) 3 1,016 1,438 -2.67 -0.05 (0.00) Panel B. Mean Changes in Operating Performance -1 to 0 1,310 1,881 -0.68 1.50 (0.00) 0 to 1 1,222 1,723 -2.95 -1.55 (0.01) 1 to 2 1,162 1,555 0.25 -0.52 (0.86) 2 to 3 1,003 1,436 -0.59 0.00 (0.20) Panel C. Median Levels in Operating Performance -1 1,338 1,894 0.01 0.01 (0.47) 0 1,310 1,881 0.28 0.57 (0.00) 1 1,222 1,723 -0.09 0.29 (0.00) 2 1,063 1,558 -0.17 0.21 (0.00) 3 1016 1438 -0.17 0.28 (0.00) Panel D. Median Changes in Operating Performance -1 to 0 1,310 1,881 0.23 0.54 (0.00) 0 to 1 1,222 1,723 -0.24 -0.08 (0.01) 1 to 2 1,162 1,555 0.00 -0.08 (0.10) 2 to 3 1,003 1,436 0.04 0.09 (0.19) ROMV Year Informed Other p-val -1 0.00 0.00 (0.47) 0 0.41 1.07 (0.04) 1 -2.64 -0.72 (0.00) 2 -3.04 -0.57 (0.00) 3 -4.23 -1.40 (0.00) Panel B. Mean Changes in Operating Performance -1 to 0 0.41 1.05 (0.05) 0 to 1 -3.23 -1.90 (0.03) 1 to 2 -0.92 -0.27 (0.26) 2 to 3 -3.13 -0.89 (0.03) Panel C. Median Levels in Operating Performance -1 0.01 0.01 (0.59) 0 0.33 0.34 (0.26) 1 -0.10 0.22 (0.00) 2 -0.02 0.30 (0.02) 3 0.20 0.51 (0.02) Panel D. Median Changes in Operating Performance -1 to 0 0.32 0.33 (0.29) 0 to 1 -0.57 -0.10 (0.00) 1 to 2 -0.18 -0.07 (0.40) 2 to 3 0.03 0.11 (0.16) OPROA Year Informed Other p-val -1 0.03 0.07 (0.23) 0 0.58 1.77 (0.00) 1 -0.96 0.37 (0.00) 2 -1.60 0.33 (0.00) 3 -1.36 0.48 (0.00) Panel B. Mean Changes in Operating Performance -1 to 0 0.52 1.68 (0.00) 0 to 1 -1.73 -1.38 (0.17) 1 to 2 -0.84 -0.51 (0.09) 2 to 3 0.15 0.01 (0.62) Panel C. Median Levels in Operating Performance -1 0.02 0.01 (0.92) 0 0.28 0.80 (0.00) 1 -0.14 0.21 (0.00) 2 -0.28 0.19 (0.00) 3 -0.17 0.16 (0.02) Panel D. Median Changes in Operating Performance -1 to 0 0.21 0.68 (0.00) 0 to 1 -0.31 -0.17 (0.02) 1 to 2 -0.02 0.00 (0.21) 2 to 3 0.11 0.11 (0.45) OPROMV Year Informed Other p-val -1 0.09 0.07 (0.94) 0 0.30 0.41 (0.35) 1 -1.09 -0.32 (0.04) 2 -0.98 0.42 (0.00) 3 0.12 0.52 (0.27) Panel B. Mean Changes in Operating Performance -1 to 0 0.18 0.33 (0.31) 0 to 1 -1.68 -0.88 (0.02) 1 to 2 -0.34 0.44 (0.05) 2 to 3 0.67 0.26 (0.75) Panel C. Median Levels in Operating Performance -1 0.03 0.02 (0.79) 0 0.16 0.28 (0.23) 1 -0.27 -0.16 (0.09) 2 -0.03 0.10 (0.09) 3 0.47 0.43 (0.80) Panel D. Median Changes in Operating Performance -1 to 0 0.10 0.23 (0.20) 0 to 1 -0.54 -0.32 (0.08) 1 to 2 0.19 0.10 (0.46) 2 to 3 0.32 0.29 (0.33)