University of Southern California Dissertations and Theses

Disclosure distance and earnings announcement returns

(USC Thesis Other)

Disclosure distance and earnings announcement returns

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content DISCLOSURE DIST ANCE AND EARNINGS ANNOUNCEMENT RETURNS
b y
R y an D. Erhard
A Dissertation Presen ted to the
F A CUL TY OF THE USC GRADUA TE SCHOOL
UNIVERSITY OF SOUTHERN CALIF ORNIA
In P artial F ulfillmen t of the
Requiremen ts for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRA TION)
A ugust 2022
Cop yrigh t 2022 R y an D. Erhard
Dedication
T o m y wife, Bro ok elyn; m y paren ts, Doug and Lorie; m y sib lings, T yler and Bro ok e;
and m y paren ts-in-la w, Douglas and Brenda.
ii
A c kno wledgemen ts
I thank the mem b ers of m y dissertation c ommittee: Ric hard Sloan (c hair), P atricia
Dec ho w, Christopher Jones, Maria Ognev a, and Mark Soliman for man y helpful
discussions and suggestions for this dissertation. I also thank Inna Abramo v a, Aleksander
Aleszczyk, Marion Boisseau-Sierra, Katherine Bruere, John Campb ell, Gino Cenedese, AJ
Chen, T ed Christensen, Jenn y Ch u, John Core, Jonathan Crask e, Christine Cun y , Yiw ei
Dou, Bro ok elyn Erhard, Jesse Gardner, Rac hel Geoffro y , Ilan Guttman, F rank Heflin,
Moritz Hiemann, Eric Holzman, Alan Jagolinzer, Sudarshan Ja y araman, Jan Jindra,
Marcin Kacp erczyk, Jung K o o Kang, April Klein, Cliv e Lenno x, R ussell Lundholm, Tim
Martens, Brian Mittendorf, Marcel Olb ert, Viv ek P andey , P eter P op e, Rob ert Resutek,
Scott Ric hardson, Rafael Rogo, Darren Roulstone, Matthew Shaffer, Am y Sheneman,
Lakshmanan Shiv akumar, Nikki Skinner, Ew a Sletten, Eric So, V en tsisla v Stameno v,
Erin T o w ery , Da vid T sui, Irem T una, Kristen V alen tine, Andrew V an Buskirk, Florin
V asv ari, Lynn W ang, Charles W asley , Ben Whipple, Regina Witten b e rg-Mo erman, T.J.
W ong, Joanna W u, Jason Xiao, Jun Y o on, and w orkshop participan ts at F ulcrum Asset
Managemen t, London Business Sc ho ol, The Ohio State Univ ersit y , the U.S. Securities
and Exc hange Commission’s Division of Economic and Risk Analysis, Univ ersità Bo cconi,
Univ ersit y of Cam bridge, Univ ersit y of Georgia, Univ ersit y of Ro c hester, and Univ ersit y
of Southern California for constructiv e feedbac k. I thank USC Marshall Sc ho ol of
Business and USC Lev en thal Sc ho ol of A ccoun ting for researc h funding.
iii
T able of Con ten ts
Dedication ii
A c kno wledgemen ts iii
List of T ables vi
List of Figures vii
Abstract viii
Chapter 1: In tro duction 1
Chapter 2: Prior Literature and Hyp othesis Dev elopmen t 8
2.1 Prior Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1.1 Disclosure Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1.2 Disclosure and Asset Prices . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.3 Pricing of Idiosyncratic Risk . . . . . . . . . . . . . . . . . . . . . . 10
2.1.4 Earnings Announcemen t Premium . . . . . . . . . . . . . . . . . . . 12
2.2 Hyp otheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Chapter 3: Data and Measuremen t 15
3.1 Data Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2 Measuring Disclosure Distance . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.3 Defining Jumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.4 Defining Con trol V ariables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.5 Earnings Announcemen t Dates . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.6 Earnings Announcemen t Returns . . . . . . . . . . . . . . . . . . . . . . . . 20
3.7 Sample Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Chapter 4: Main Empirical Results 22
4.1 Descriptiv e Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.2 Determinan ts of Disclosure Distance . . . . . . . . . . . . . . . . . . . . . . 23
4.3 Earnings Announcemen t Jump Risk . . . . . . . . . . . . . . . . . . . . . . 25
4.4 Earnings Announcemen t Sto c k Returns . . . . . . . . . . . . . . . . . . . . 27
4.5 In v estor Recognition and the Pricing of I diosyncratic Risk . . . . . . . . . . 29
iv
Chapter 5: A dditional Empirical Results 31
5.1 In v estor Recognition and Jump Risk . . . . . . . . . . . . . . . . . . . . . . 31
5.2 In tertemp oral Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.3 Size Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.4 Con trolling for Beta-shifts at Earnings Announc emen ts . . . . . . . . . . . . 33
5.5 Alternativ e Measures of Jumps . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.6 Con tin uous Idiosyncratic V olatilit y . . . . . . . . . . . . . . . . . . . . . . . 34
5.7 Alternativ e Measures of Disclosure Distanc e . . . . . . . . . . . . . . . . . . 35
Chapter 6: Conclusion 36
Bibliograph y 38
App endix A 45
V ariable Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
v
List of T ables
1 Descriptiv e Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2 Determinan ts of Disclosure Distance . . . . . . . . . . . . . . . . . . . 59
3 Earnings Announcemen t Jump Risk . . . . . . . . . . . . . . . . . . . 60
4 Earnings Announcemen t Sto c k Returns . . . . . . . . . . . . . . . . 64
5 In v estor Recognition and Earnings Announcemen t Sto c k Returns 66
6 In v estor Recognition and Earnings Announcemen t Jump Risk . . 69
7 Earnings Announcemen t Sto c k Returns: Subp erio ds . . . . . . . . . 72
8 Earnings Announcemen t Sto c k Returns: Size Subsamples . . . . . 75
9 Con trolling for Beta-Shifts at Earnings Announcemen ts . . . . . . 78
10 Alternativ e Measures of Jumps . . . . . . . . . . . . . . . . . . . . . . 79
11 Earnings Announcemen t Idiosyncratic V olatilit y . . . . . . . . . . . 83
12 Earnings Announcemen t Idiosyncratic V olatilit y Excluding J umps 86
13 Alternativ e Measure of Disclosure Distance: Da ys Since Disclosure 89
14 Alternativ e Measure of Disclosure Distance: Rank ed Disclosure
Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
vi
List of Figures
1 Disclosure Arriv al o v er the Quarter . . . . . . . . . . . . . . . . . . . . 48
2 Num b er of In terim 8-Ks Filed Eac h Firm-Quarter . . . . . . . . . . 49
3 Disclosure Distance Distribution . . . . . . . . . . . . . . . . . . . . . 50
4 Disclosure Distance o v er Time . . . . . . . . . . . . . . . . . . . . . . 51
5 Cum ulativ e Jump Rates around Earnings Announcemen ts . . . . 52
6 Cum ulativ e Abnormal Returns around Earnings Announcemen ts 53
7 Distan t Min us Recen t Hedge P ortfolio Returns and Difference in
Jump Rates around Earnings Announcemen ts . . . . . . . . . . . . 54
8 P ortfolio Returns Conditional on News . . . . . . . . . . . . . . . . . 55
9 Hedge P ortfolio Returns Conditional on News . . . . . . . . . . . . 56
vii
Abstract
I h yp othesize that in v estors view earnings annou ncemen ts preceded b y distan t
disclosure as b eing riskier than announcemen ts preceded b y recen t disclosure. Consisten t
with this h yp othesis, distan t disclosing sto c ks ha v e greater idiosyncratic jump risk at
earnings announcemen ts. Distan t disclosing sto c ks also earn p ositiv e abnormal
announcemen t returns, suggesting that in v estors an ticipate the greater idiosyncratic
jump risk and demand a risk premium to hold these sto c ks at earnings announcemen ts.
A dditional empirical tests dra wing on the theoretical mo del in Merton (1987) supp ort the
in v estor pricing of idiosyncratic jump risk explanation.
viii
Chapter 1
In tro duction
In v estors rely on discrete corp orate disclos ures to learn priv ate information ab out
firm v alue, while managers observ e this information in real time. A t a minim um, SEC
regulation requires disclosure at quarterly in terv als, prev en ting managers from engaging
in nondisclosure for longer than a quarter. This pap er asks: what are the capital mark et
consequences of elapsed corp orate silence (disclosure distance) b et w een mandatory
rep orting p erio ds? More sp ecifically , do es disclosure distance impact jump risk at
earnings announcemen ts? Do in v estors an ticipate and price this risk?
Corp orate silence b et w een mandatory re p orting p erio ds creates a more discrete
(jump y) disclosure en vironmen t where firm-pro duced information arriv es at few er p oin ts
spread out o v er time. An earnings announcemen t preceded b y in terim
1
silence represen ts
a jumpier information flo w compared to an earnings announcemen t preceded b y other
disclosures, suc h as 8-K filings or issuance of earnings guidance. Empirically , I find that a
1
“In terim” refers to nonearnings announcemen t p erio ds.
1
more discrete information en vironmen t with silence b et w een rep orting p erio ds leads to
increased idiosyncratic jump risk at earnings announcemen ts.
2
Managers face trade-offs when making disclos ure decisions, pro viding sev eral reasons
wh y managers engage in nondisclosure. While disclosure has capital mark et b enefits (e.g.,
Diamond and V errecc hia , 1991 ; Heflin et al. , 2005 ; Balakrishnan et al. , 2014 ; Billings
et al. , 2015 ), disclosure can also b e costly . These costs range from proprietary costs of
disclosure (e.g., V errecc hia , 1983 ) where disclosure causes com p etitiv e harm, to more
concrete costs of disclosure suc h as compliance costs asso ciated with making a disclosure
(e.g., Ka jüter et al. , 2018 ). Increased disclosure can also cro wd out trading from informed
in v estors, inhibiting managerial learning from sto c k price ( Ja y araman and W u , 2018 ;
Chen et al. , 2021 ). Another practical consideration is whether routine v olun tary
disclosure of short-term forecasts, suc h as quarterly earnings guidance, in tro duces
incen tiv es to engage in managerial m y opia (e.g., Chen et al. , 2011 ; Kim et al. , 2017 ).
F urther, short-term metrics can b e noisy . F or example, in a break from industry practice,
automotiv e man ufacturer General Motors stopp ed pro viding mon thly sales n um b ers to
in v estors, citing the noisiness of the underlying mon thly sales metric. GM’s sales
op e rations c hief, Kurt McNeil, said in an in terview with The W al l Str e et Journal :
“Thirt y da ys is not enough time to separate real sales trends from short-term fluctuations
in a v ery dynamic, highly comp etitiv e mark et,” (“GM Scraps a Standard in Sales
Rep orting,” WSJ , April 3, 2018).
2
Idiosyncratic jump risk is the risk of a sto c k-sp ecific sharp up or do wn mo v emen t in sto c k price. I use
“idiosyncratic jump risk” and “jum p risk” in terc hangeably throughout.
2
Managers ma y also c ho ose to defer issuance of v olun tary disclosure while engaged in
uncertain business transactions suc h as M&A transactions un til the resolution of
uncertain t y . The National In v estor Relations Institute (NIRI) illustrates the balancing
act managers face in this regard when determining the timing of disclosure:
“Premature disclosure can b e extremely damaging if future ev en ts c hange the
p erception o f the ev en t. These circumstances call for judgmen t — disclosure
to o so on ma y mean a compan y has failed its curren t sto c kholders b y failing to
include all re lev an t information; disclosure to o late ma y mean the compan y
has missed a dis closure deadline or extended the class p erio d for a p oten tial
class action la wsuit,” (Standards of Practice for In v estor Relations, NIRI ,
2016 ).
Classic disclosure mo dels imply unra v eling, where full disclosure is optimal. More
recen t theoretical w ork iden tifies equilibria where b oth go o d and bad news are not
immediately disclosed. Extending the static mo del in Dy e ( 1985 ), Guttman et al. ( 2014 )
find in a dynamic mo del that managers p oten tially endo w ed with m ultiple signals in a
m ulti-p erio d game can maximize firm v alue b y disclosing priv ate information later in the
game rather than so oner, sho wing an option v alue of nondisclosure. Marino vic and V aras
( 2016 ) mo del that litigation risk can cro wd out p ositiv e disclosures, leading to a w orld
where “no news is go o d news. ” Bond and Ze ng ( 2021 ) mo del optimal disclosure for a
manager who do es not kno w shareholder preferences and find an equilibrium where
nondisclosure is optimal, pro viding another explanation for wh y unra v eling mo dels do not
fit the data in some settings.
T o empirically study the capital mark et conseque nces of in terim silence, I measure
elapsed silence via 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 . A t eac h earnings announcemen t date, I first
calculate the n um b er of da ys since the previous material disclosure, with material
disclosures iden tified using 8-K filings. T o calculate 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 , I scale the
n um b e r of trading da ys since disclosure b y the in terim length in trading da ys b et w een
adjacen t earnings announcemen ts to neutralize the effects of early or late rep orting.
3
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 ranges from (0, 1] with a lo w v alue indicating a recen t disclosing
firm and a v alue of 1 indicating a distan t disclosing firm (i.e., a silen t firm with no
material in terim disclosures).
Once in v estors learn information at earnings announc emen ts, the sto c k price can
jump in resp onse to the announcemen t. F ollo wing Kapadia and Zekhnini ( 2019 ), I define
a jump in daily sto c k price as an absolute idiosyncratic return greater than 3 historic
standard deviations. Consisten t with silence b et w een mandatory rep orting p erio ds
increasing jump risk, I find that 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 forecasts idiosyncratic jumps in
sto c k price around earnings announcemen t dates: 40% of distan t disclosing sto c ks
exp erience a jump in the earnings announcemen t windo w compared to 33% of recen t
disclosing sto c ks.
3
Next, I build on prior researc h finding that idios yncratic jump risk is priced ( Kapadia
and Zekhnini , 2019 ) and only the jump comp onen t of idios yncratic risk is priced ( Bégin
et al. , 2019 ).
4
Because in v estors in distan t disclosing sto c ks are exp osed to greater
idiosyncratic jump risk at earnings announcemen ts, I h yp othesize that jump-risk-a v erse
in v estors demand a risk premium for holding distan t disclosing sto c ks at earnings
announcemen ts. Consisten t with this h yp othesis, 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is p ositiv ely
asso ciated with earnings announcemen t returns. In p ortfolios sorted b y
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 , only the distan t disclosure p ortfolio earns a significan t, p ositiv e
alpha on earnings announcemen t dates, suggesting that in v estors price in terim silence as
increasing idiosyncratic jump risk.
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is p ositiv ely asso ciated with earnings surprises, consisten t with
the mo del in Marino vic and V aras ( 2016 ) where “no news is go o d news” on a v erage due
to litigation risk, and consisten t with managers’ tendency to not preannounce extreme
p o sitiv e earnings surprises ( K o ester et al. , 2016 ). Firms with go o d news can w ait to
3
Hereafter, I refer to the earnings announcemen t ev en t windo w [0,1] as the earnings announcemen t date
for brevit y .
4
Both Kapadia and Zekhnini ( 2019 ) and Bégin et al. ( 2019 ) do not explore the disclosure en vironmen t
when examining in v estor pricing of jump risk.
4
disclose p ositiv e information at earnings announcemen ts, while firms with bad news ha v e
incen tiv es to disclose it ahead of time (e.g., Skinner , 1994 ). In m ultiv ariate regressions,
con trolling for earnings surprises eliminates the explanatory p o w er of
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 for earnings announcemen t returns. The presence of abnormal
returns to the disclosure distance strategy (in regressions not con trolling for concurren tly
announced earnings surprises) in b oth halv es of the sample suggests that in v estors are
a v erse to the idiosyncratic risk asso ciated with pursuing the disclosure distance strategy
ahead of earnings announcemen ts, resulting in idiosyncratic risk premium realized at
earnings announcemen t dates. Kapadia and Zekhnini ( 2019 ) conclude that the jump risk
premium, in general, is lik ely due to costly-to-arbitrage mispricing where the limit to
arbitrage is in v estor a v ersion to idiosyncratic risk.
In v estor pricing of idiosyncratic risk is at o dd s with traditional asset pricing theory
(e.g., Sharp e , 1964 ) that p osits exp osure to only systematic risk should b e priced in
equilibrium. Merton ( 1987 ), ho w ev er, sho ws that exp osure to idiosyncratic risk can b e
priced when in v estors are under-div ersified. Prior researc h finds supp orting evidence for
in v estor under-div ersification for b oth activ e m utual fund managers ( Kacp erczyk, Sialm,
and Zheng , 2005 ) and individual in v estors ( Barb er and Odean , 2008 ), whic h pro vides an
explanation for idiosyncratic risk b eing priced unconditionally . More sp ecifically , Merton
( 1987 ) coins the term “in v estor recognition” as the n um b er of in v estors who kno w ab out
a sto c k in order to illustrate a friction that could lead to under-div ersification in practice.
Idiosyncratic risk can b e priced for sto c ks that ha v e lo w in v estor recognition b ecause
some in v estors m ust tak e large p ositions in sto c ks with lo w in v estor recognition, resulting
in under-div ersification and exp osure to idiosyncratic risk. Leha vy and Sloan ( 2008 ) and
5
Bo dnaruk and Ostb erg ( 2009 ) find empiri cal results consisten t with the predictions in
Merton’s mo del.
5
T o pro vide additional evidence for the idiosyncratic risk explanation for the p ositiv e
relation b et w een 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 and earnings announcemen t returns, I emplo y a
cross-sectional test based on the theory in Merton ( 1987 ). Because I h yp othesize that the
p o sitiv e relation b et w een earnings announcemen t returns and 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 reflects pricing of idiosyncratic risk, I exp ect that the p ositiv e relation b et w een earnings
announcemen t returns and 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 atten uates as in v estor recognition
increases. I find empirical results consisten t with this prediction: 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 has stronger explanatory p o w er for earnings announcemen t returns in the cross section of
sto c ks w ith lo w c hanges in in v estor recognition and almost no explanatory p o w er in the
cross section of sto c ks with high c hanges in in v estor recognition.
6
P ortfolio sorts con trolling for b eta-shifts at earnings annou ncemen ts further
corrob orate the idiosyncratic risk explanation. Ball and K othari ( 1991 ) examine the
unconditionally p ositiv e sto c k returns on earnings announcemen t dates and attribute the
p o sitiv e returns, in part, to increased systematic risk at earnings announcemen ts. Using
high frequency return data, P atton and V erardo ( 2012 ) also find that CAPM b etas shift
at earnings announcemen ts. As an additional test to con trol for elev ated systematic risk
at earnings announcemen ts, I re-estimate CAPM b etas using earnings announcemen t
returns. While distan t disclosing sto c ks do exp erience an up w ard shift in b eta,
con trolling for this shift do es not explain the p ositiv e abnormal returns to distan t
disclosing sto c ks. Estimated alphas are nearly iden tical to those used in the main tests.
5
I emplo y Merton ( 1987 )’s theory b ecause in v estor recognition has an observ able empirical pro xy es-
tablished in the literature. There are other theories of in v estor pricing of idiosyncratic risk, ho w ev er. F or
example, Go y al and San ta-Clara ( 2003 ) argue in v estors are exp osed to “bac kground” risk in non traded assets,
whic h is unobserv able to the econometrician. The bac kground risk theory p osits that in v estors holding risky
non t raded assets are less willing to hold risky traded assets, resulting in idiosyncratic risk b eing priced for
traded assets.
6
The explanatory p o w er of 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 for jump risk is similar in b oth subsamples, sug-
gesting differing exp osure to jump risk b et w een subsamples is not an alternativ e explanation for the differing
pricing results b et w een subsamples.
6
The empirical results are robust to a host of alternativ e researc h design c hoices.
Defining jumps as absolute idiosyncratic returns greater than 5%, 10%, or 15% yield
similar inferences to those in the main tests where jumps are defined relativ e to trailing
firm-sp ecific v olatilit y . 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is also a predictor of con tin uous
idiosyncratic v olatilit y , and these results are primarily driv en b y idiosyncratic jump
v olatilit y , consisten t with the jump risk explanation. The sto c k return results are
significan t among b oth small and big sto c ks, and among b oth halv es of the sample p erio d.
Redefining 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 as the n um b er of da ys since disclosure (unscaled) yields
similar results, as do es utilizing a measure of rank ed 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 .
Jump risk at earnings announcemen ts is esp ec ially relev an t in the curren t financial
rep orting en vironmen t. Recen t w ork b y Bea v er et al. ( 2020 ) sho ws that the mark et
resp onse to earnings announcemen ts is increasing o v er the p erio d 2001-2016. The authors
attribute this trend to managers making earnings announcemen ts more informativ e to
in v estors b y issuing forecasts of earnings and pro viding financial statemen t line items on
earnings announcemen t dates. In a similar v ein, Shao et al. ( 2021 ) highligh t that returns
around fundamen tal information releases suc h as earnings announcemen ts, 8-K filings,
and managemen t forecasts explain a greater prop ortion of ann ual sto c k returns in recen t
y ears.
This dissertation con tributes to the disclosure literat ure b y sho wing that there is a
capital mark et consequence for the gro wing trend of sta ying silen t in-b et w een earnings
announcemen ts: elev ated jump v olatilit y at earnings announcemen ts. This pap er also
con tributes to the gro wing literature on the pricing of idiosyncratic risk b y highligh ting
the disclosure en vironmen t as a source of idiosyncratic risk that in v estors price at
earnings announcemen ts. Lastly , this pap er con tributes to the literature on earnings
announcemen t returns b y sho wing that exp osure to idiosyncratic risk is imp ortan t for
explaining announcemen t returns.
7
Chapter 2
Prior Literature and Hyp othesis Dev elopmen t
2.1 Prior Literature
In this section, I summarize the m ain areas of researc h I build on: the literatures on
disclosure timing, disclosure and asset prices, the pricing of idiosyncratic risk, and
earnings announcemen t premium.
2.1.1 Disclosure Timing
Most of the literature on disclosure timing studies managers using timing strategically .
Questions in the literature include whether managers prefer to disclose bad news early to
a v oid litigati on risk (e.g., Skinner , 1994 ) or whether they prefer to hold on to it (e.g.,
K o thari et al. , 2009 ; Zh u , 2016 ; Baginski et al. , 2017 ). Other pap ers argue that managers
target lo w p erio ds of in v estor atten tion to disclose bad news (e.g., deHaan et al. , 2015 ) or
manipulate disclosure timing strategically to extract priv ate b enefits via insider trading
( Niessner , 2015 ; Billings and Cedergren , 2015 ). K o ester et al. ( 2016 ) pro vide evidence
that managers withhold extreme p ositive news in order to attract in v estor atten tion at
earnings announcemen ts leading to capital mark et b enefits in the forms of increased
analyst follo wing, greater institutional o wnership, and higher trading v olume.
8
Recen t literature explores disclosure timing not conditional on the information con ten t
of the underlying disclosure. Chapman et al. ( 2019 ) pro vide evidence that some managers
attempt to prev en t information o v erload b y spreading out disclosures of ev en ts that o ccur
on the same date, and these firms ha v e lo w er v olatilit y . Noh et al. ( 2021 ) iden tify firms
that follo w a pattern for sc heduling earnings announcemen ts (e.g., a pattern firm alw a ys
announces Q4 earnings on the first T uesda y in F ebruary). “Calendar rotations” pro vide
quasi-exogenous v ariation in the relativ e timing of earnings announcemen ts for pattern
firms. P attern firms whose earnings announcemen ts are mo v ed up exogenously
exp erience greater media atten tion and greater earnings announcemen t premium.
1
2.1.2 Disclosure and Asset Prices
Increased disclosure is link ed to cost of capital b oth theoretically and empirically .
Increased information ab out a sto c k reduces estimation risk, resulting in a lo w er cost of
capital for firms with increased disclosure (e.g., Barry and Bro wn , 1985 ; Lam b ert et al. ,
2007 ). Zhao ( 2017 ) finds a negativ e relation b et w een information in tensit y (based on the
n u m b e r of 8-Ks filed) and future sto c k returns, and pro vides an estimation risk
explanation for this result. V an Buskirk ( 2012 ) finds that retail firms that issue m on thly
sales figures ha v e lo w er sto c k price impacts of earnings announcemen ts, suggesting more
frequen t disclosure mak es prices more informativ e. Similarly , McMullin et al. ( 2019 ) and
Noh et al. ( 2019 ) find in terim 8-K filings are asso ciated with increased in trap erio d
timeliness. Zhou and Zhou ( 2020 ) find lo w er retur ns around earnings announcemen ts for
firms that do not issue guidance, arguing that firms that do not issue guidance ha v e p o or
fundamen tals that are mispriced due to mark et frictions. Lenno x and P ark ( 2006 ) find
1
My measure of disclosure distance is unaffected b y calendar rotations b ecause I scale the n um b er of
da y s since disclosure b y the n um b er of da ys since the last earnings announcemen t. Noh et al. ( 2019 ) use
calendar rotations to iden tify causal effects of the relativ e ordering of earnings announcemen ts due to p oten tial
endogeneit y asso ciated with the timing of earnings news, suc h as dela ying bad news and accelerating p ositiv e
news, whic h w ould affect the relativ e ordering (e.g., P enman , 1987 ; Begley and Fisc her , 1998 ; Bagnoli et al. ,
2002 ; Johnson and So , 2018b ). This p oten tial source of endogeneit y is wh y I scale disclosure distance b y
the in terim length. That is, I don’t distinguish b et w een a silen t firm that go es 80 da ys without announcing
earnings and a silen t firm that go es 100 da ys without announcing earnings.
9
that firms with ex an te higher earnings resp onse co efficien ts are more lik ely to b e guiders,
suggesting firms pro vide guidance when EPS is a k ey determinan t of firm v alue.
2.1.3 Pricing of Idiosyncratic Risk
Classical asset pricing theory p osits that in v estors can div ersify a w a y idiosyncratic
risk, therefore only systematic risk should b e priced in equilibrium. Merton ( 1987 ),
ho w ev er, sho ws analytically that when a sto c k has lo w in v estor recognition (the n um b er
of in v estors who kno w ab out a sto c k), idiosyncratic risk can b e priced. The in tuition is
that when few in v estors kno w ab out a sto c k, some in v estors m ust tak e large p ositions in
a sto c k, resulting in exp osure to idiosyncratic risk that requires a risk premium to hold
sto c ks with lo w in v estor recognition when idiosyncratic risk is high. Leha vy and Sloan
( 2008 ) pro vide empirical evidence consisten t with the negativ e relation b et w een c hanges
in in v estor recognition and exp ected returns, and find the relation is strongest for sto c ks
with high idiosyncratic risk, based on a mark et measure of idiosyncratic risk.
Exp ected idiosyncratic risk that maps in to risk pr emium is difficult to measure
empirically without using implied v olatilit y from option prices. F or example, sto c ks with
high past idiosyncratic risk ha v e lo w er exp ected returns, whic h is inconsisten t with
idiosyncratic risk b eing priced ( Ang et al. , 2006 ). F u ( 2009 ) reconciles the negativ e
relation b et w een historic idiosyncratic risk and future returns established in Ang et al.
( 2006 ) b y sho wing that past idiosyncratic risk is a p o or pro xy for exp ected idiosyncratic
risk; when using a measure of exp ected idiosyncratic risk, there is a p ositiv e relation with
future sto c k returns. Go y al and San ta-Clara ( 2003 ) find that the mark et p ortfolio earns
higher returns when idiosyncratic risk of the underlying sto c ks is high, suggesting
idiosyncratic risk is priced at the firm lev el. Kapadia and Zekhnini ( 2019 ) find b oth an
ex p ost and ex an te jump risk premium in the cross section of equit y returns using
implied v olatilit y as a forecasting v ariable for jumps. Bégin et al. ( 2019 ) also find that
idiosyncratic jump risk is priced. Crucially , they sho w that the link b et w een idiosyncratic
10
risk and future returns is driv en en tirely b y idiosyncratic jump v olatilit y . That is,
extreme idiosyncratic v olatilit y driv es the idiosyncratic risk premium. These studies are
largely silen t on the source of idiosyncratic risk b ecause past idiosyncratic v olatilit y and
implied v olatilit y are somewhat of a blac k b o x. They do not explore the disclosure
en vironmen t as a source of idiosyncratic jump risk.
The rationale in Bégin et al. ( 2019 ) for idiosyncratic jump v olatilit y commanding a
risk premium is that it is difficult for in v estors to hedge discon tin uous c hanges in sto c k
price. Also arguing that in v estors price risks that are hard to hedge, Bollerslev et al.
( 2016 ) estimate “rough” b etas using discon tin uous mark et returns and find that
systematic exp osure to discon tin uous mark et returns is priced, while exp osure to
con tin uous mark et returns is not priced. Amiram et al. ( 2019 ) ec ho this hard-to-hedge
rationale, and find that mark et mak ers reduce liquidit y for sto c ks with high jump
v olatilit y . They also find that firms with b etter information en vironmen ts ha v e lo w er
jump v olatilit y , but they do not study earnings announcemen ts or examine in v estor
pricing of jump v olatilit y . Stoum b os ( 2019 ) finds that illiquidit y rises as the next
earnings announcemen t dra ws near and that in terim disclosure helps to coun teract the
rise in illiquidit y .
Extreme do wnside idiosyncratic v olatilit y (negativ e jumps) i s the fo cus of the crash
risk literature. This literature explores ho w rep orting decisions of managers can lead to
o v erv a lued equit y that corrects via a large decline in sto c k price (a crash). Some
determinan ts of sto c k price crashes include gro wth exp ectations ( Skinner and Sloan ,
2002 ), opaque financial rep orting ( Hutton et al. , 2009 ), high accoun ting accruals ( Zh u,
2016 ), and non-GAAP rep orting (Hsu et al. , 2021 ). The literature on jump risk differs
from the literature on crash risk in that it considers extreme price mo v emen ts in b oth
directions (b oth p ositiv e and negativ e jumps). P ositiv e jumps are also of in terest to
in v estors with mean-v ariance preferences due to their negativ e effect on the Sharp e ratio,
ceteris paribus.
11
2.1.4 Earnings Announcemen t Premium
The main finding in the earnings announceme n t premium (EAP) literature is that
sto c ks exp ected to announce earnings ha v e higher exp ected returns relativ e to sto c ks not
exp ected to announce. Belo w, I summarize the literature on the EAP .
The EAP is do cumen ted in man y studies b eginning w ith Bea v er ( 1968 ) and has
attracted man y explanations. Ball and K othari ( 1991 ) sho w that CAPM b etas increase
around earnings announcemen ts, but ev en after con trolling for the up w ard shift in b eta,
abnormal returns still p ersist at earnings announcemen ts. F razzini and Lamon t ( 2007 )
argue that announcemen ts are atten tion-grabbing ev en ts, and up w ard price pressure from
high trading v olume explains the rise in price on earnings announcemen t dates. Cohen
et al. ( 2007 ) argue that the EAP is due to mispricing arising from costly arbitrage.
Barb er et al. ( 2013 ) find the EAP is presen t globally and is strongest in coun tries with
high idiosyncratic risk. Johnson and So ( 2018a) pro vide evidence that asymmetric
liquidit y pro vision b y mark et mak ers driv es the p ositiv e abnormal returns leading up to
and at the earnings announcemen t due to increased costs of trading on negativ e news.
Sa v or and Wilson ( 2016 ) offer a systematic risk explanation: announcing firm returns
predict future aggregate earnings, making announcing firms systematically risky . Chan
and Marsh ( 2021 ) find a p ositiv e relation b et w een b eta and sto c k returns on lead
earnings announcemen t da ys (firms announcing early in the earnings announcemen t
season) and a flat relation outside of those da ys. Johnson et al. ( 2020 ) find that earnings
announcemen t mon th returns are increasing in managers’ incen tiv es to manage in v estor
exp ectations, suggesting some managers “man ufacture” p ositiv e earnings surprises. Heitz
et al. ( 2020 ) find the EAP has atten uated in recen t y ears follo wing c hanging accoun ting
regulation for filing form 8-K, where firms m ust disclose certain material ev en ts within
four business da ys, up dating in v estors b et w een fiscal quarters on a timelier basis. In a
review of the literature on asset prices and recurring firm ev en ts, Hartzmark and Solomon
( 2018 ) conclude that “Risk-based explanations, particularly those relating to idiosyncratic
12
risk, ha v e the most p oten tial for explaining announcemen t returns. ” I answ er this call b y
emphasizing that despite increased 8-K filings econom y wide as a result of regulatory
c hange, there still exist firms that do not issue an y in terim 8-Ks, and these firms ha v e
high idiosyncratic risk and earn p ositiv e abnormal returns on announcemen t dates.
A related literature examines option prices and v olatilit y around earnings
announcemen t dates. P atell and W olfson ( 1979 , 1981 ) do cumen t that preannouncemen t
option prices reflect an ticipated v olatilit y at earnings announcemen ts. Barth and So
( 2014 ) argue that earnings announcemen ts ha v e non-div ersifiable v olatilit y risk and
preannouncemen t options prices reflect this risk. Dubinsky et al. ( 2018 ) further
corrob orate that preannouncemen t option prices reflect earnings announcemen t risk, and
the authors incorp orate announcemen t risk in to option pricing mo dels. Ov erall, the
literature examining option prices around earnings announcemen ts suggests that option
traders price v olatilit y risk at earnings announcemen ts, but these pap ers do not lo ok at
returns on the underlying equit y around earnings announcemen ts.
2.2 Hyp otheses
I h yp othesize that sto c ks with distan t disclosure are more sensitiv e to new information
relativ e to sto c ks with recen t disclosure, resulting in an increased lik eliho o d of a sharp
c hange in sto c k price once in v estors learn information at earnings announcemen ts.
H1: Jump risk at e arnings announc ements is incr e asing in Disclosur e Distanc e.
While H1 ma y app ear in tuitiv e, it is not without tension giv en incen tiv es for in v estors
to engage in costly information acquisition in the absence of firm disclosure (e.g.,
Ja y araman and W u , 2018 ). F or example, in v estors can use altern ativ e data sources suc h
as satellite data to learn ab out asset prices (e.g., Mukherjee et al. , 2021 ), or can use
F O IA requests to learn previously undisclosed information ab out firm v alue ( Gargano
et al. , 2016 ).
13
Next, building on the literature that sho ws exp osure t o idiosyncratic jump risk is
priced, I h yp othesize that in v estors demand a jump risk premium for holding sto c ks with
distan t disclosure at earnings announcemen ts, where the risk premium is realized once
firms announce earnings.
H2: Earnings announc ement r eturns ar e incr e asing in Disclosur e Distanc e.
Kapadia and Zekhnini ( 2019 ) find that exp ected idiosyncratic jump risk is priced
(using options data to forecast jumps) and they find that earnings announcemen ts are a
frequen t source of jumps, ho w ev er, they do not examine the cross section of earnings
announcemen t returns or examine the disclosure en vironmen t. Ex-an te, it is unclear
whether in v estors pa y atten tion to disclosure distance and whether it influences the ir
short-term discoun t rates. Evidence in supp ort of H2 w ould suggest that in v estors learn
ab o ut jump risk at earnings announcemen ts via the in terim disclosure en vironmen t and
set short-term discoun t rates accordingly .
T o distinguish b et w een the idiosyncratic risk and s ystematic risk explanations for the
abnormal returns to distan t disclosing sto c ks, the theory in Merton ( 1987 ) w ould suggest
a cross-sectional test based on a pro xy for in v estor recognition. Merton sho ws
idiosyncratic risk is priced when in v estors are underdiv ersified. As in v estor recognition
increases, the idiosyncratic jump risk premium should decrease due to increased risk
sharing among in v estors.
H3: The p ositive r elation b etwe en e arnings announc ement r eturns and Disclosur e
Distanc e attenuates as investor r e c o gnition i ncr e ases.
14
Chapter 3
Data and Measuremen t
3.1 Data Sources
I obtain daily sto c k returns and daily trading v olume from the Cen ter for Researc h in
Securit y Prices (CRSP), financial statemen t data and earnings announcemen t dates from
Compustat, managemen t forecasts, analyst co v erage, and earnings announcemen t dates
from Institutional Brok ers Estimate System (I/B/E/S), 13-F filing data from Thompson
Reuters (using WRDS Thomson Reuters Institutional (13-F) Holdings - Sto c k Ownership
Summary File), form 8-K filing dates from SEC EDGAR, and factor returns and risk-free
rate data from Kenneth F renc h’s w ebsite.
3.2 Measuring Disclosure Distance
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is the elapsed time in trading da ys since the last material
disclosure ( 𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝑞 ) scaled b y the in terim length ( 𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐸 𝐴 𝑞 −1
).
Scaling b y the in terim length b et w een adjacen t earnings announcemen ts neutralizes
factors suc h as late/early rep orting whic h can correlate with the direction of earnings
news (e.g., P enman , 1987 ; Begley and Fisc her , 1998 ; Bagnoli et al. , 2002 ; Johnson and So ,
15
2018b ). 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 ranges from (0, 1] where a lo w v alue represen ts a firm with
recen t disclosure and a v alue of one represen ts a firm with distan t disclosure.
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 =
𝐸 𝐴 𝑞 − 𝑀 𝐴𝑋 (𝐸 𝐴 𝑞 −1
, 8 - 𝐾 𝑞 −1,𝑞
)
𝐸 𝐴 𝑞 − 𝐸 𝐴 𝑞 −1
(3.1)
=
𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝑞 𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐸 𝐴 𝑞 −1
Exhibit 1 illustrates the measuremen t of 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 via a timeline plotting
the time in terv al b et w een adjacen t earnings announcemen ts. EA
𝑞 is the fo cal earnings
announcemen t studied and EA
𝑞 −1
is the earnings announcemen t one quarter prior. T o
calculate 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 at eac h EA
𝑞 , first, I lo ok bac k in time to iden tify the
most re cen t material disclosure ( 𝑀 𝐴𝑋 (𝐸 𝐴 𝑞 −1
, 8𝐾
𝑞 −1,𝑞
) ). Second, I calculate the n um b er
of da ys b et w een EA
𝑞 and the most recen t disclosure ( 𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝑞 ). Third, I
calculate the n um b er of da ys b et w een EA
𝑞 and EA
𝑞 −1
( 𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐸 𝐴 𝑞 −1
). Finally , I
calculate 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 as the ratio of 𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝑞 to
𝐷 𝑎 𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐸 𝐴 𝑞 −1
. In p ortfolio tests, I sort sto c ks in to three p ortfolios: recen t, mid, and
distan t disclos ure p ortfolios based on the cross-sectional distribution of
𝐷 𝑖𝑠𝑐 𝑙 𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 eac h quarter.
EA
𝑞 −1
1 0 0.5
EA
𝑞 Distan t Disclosure Mid Disclosure Recen t Disclosure
Exhibit 1: Measuring Disclosure Distance
I use 8-K filing dates to iden tify material in terim disclosure dates. SEC regulation
requires companies to file 8-Ks follo wing material v olun tary disclosure ev en ts suc h as
managemen t guidance, and mandatory disclosure ev en ts suc h as en try in to a material
agreemen t. This regulatory requiremen t mak es 8-K filing dates an ideal source for
iden tifying material in terim disclosures. As noted b y Campb ell et al. ( 2021 ), other
corp orate disclosures unaccompanied b y 8-K filings presen t a c hallenge for researc hers in
16
determining managers’ b eliefs ab out disclosure materialit y . Managers ha v e some
discretion in determining disclosure materialit y thresholds for filing 8-Ks, and not
capturing v alue-relev an t disclosures unaccompanied b y 8-Ks is a limitation of the
measure.
F or data qualit y , I exclude 8-Ks where filing date s precede rep ort (ev en t) dates, and
8-Ks filed more than 10 da ys after the rep ort date. I also exclude amended 8-Ks and
earnings announcemen t 8-Ks when calculating 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 . Firms m ust ha v e
nonmissing earnings announcemen t dates in either Compustat or I/B/E/S for earnings
announcemen ts for fiscal quarters 𝑞 and 𝑞 − 1 to ha v e a v alid measure of
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 . I drop observ ations where the n um b er of calendar da ys b et w een
adjacen t earnings announcemen ts 𝑞 and 𝑞 − 1 is outside the range of 60-120 da ys to
eliminate uncommon observ ations.
3.3 Defining Jumps
I follo w Kapadia and Zekhnini ( 2019 ) in defining idiosyncratic jum p da ys,
idiosyncratic sto c k returns, and idiosyncratic v olatilit y .
Jump
𝑖𝑡 =
⎧
{
⎨
{
⎩
1 if |𝑟
𝑖𝑡 | > 3𝜎
𝑖𝑡 0 else
(3.2)
Idiosyncratic sto c k returns, 𝑟 𝑖𝑡 , are estimated eac h da y out of sample relativ e to a
Carhart ( 1997 ) four-factor mo del (4F) with factor loadings estimated in a rolling 120 da y
windo w from trading da ys t -150 to t -31. 𝑟 𝑒 𝑖𝑡 is the daily sto c k return in excess of the ris k
free rate,
̂ 𝛽 ′4𝐹
𝑖𝑡−150,𝑖𝑡−31
is a v ector of factor loadings and 𝐹 4𝐹
𝑡 is a v ector of f actor returns
(MKT, SMB, HML, UMD).
1
F actor loadings are winsorized at the 5% tails of the daily
cross-sectional distribution.
1
MKT is the mark et factor, SMB is the size factor, HML is the v alue factor, and UMD is the momen tum
factor.
17
𝑟 𝑖𝑡 = 𝑟 𝑒 𝑖𝑡 −
̂ 𝛽 ′
4𝐹
𝑖𝑡−150,𝑖𝑡−31
× 𝐹 4𝐹
𝑡 (3.3)
Conditional idiosyncratic v olatilit y , 𝜎 𝑖𝑡 , is estimated via an exp onen tially w eigh ted
mo ving a v erage mo del using 𝑟 𝑖𝑡 in Equation 3.3 with 𝜆 =0.94.
𝜎 𝑖𝑡 =
√
√
√
⎷
(1 − 𝜆)
150
∑
𝑠=1
𝜆 𝑠 (𝑟
𝑖𝑡−𝑠
)
2
(3.4)
I create 3 v ariables to measure idiosyncratic jumps:
1. 𝐽 𝑢𝑚𝑝 0|1
= 1 if there is a jump in sto c k price on earnings announcemen t da y zero or
da y one, 0 e lse.
2. 𝑃 𝑜 𝑠𝑖𝑡𝑖𝑣 𝑒 𝐽 𝑢𝑚𝑝 0|1
= 1 if there is a p ositiv e jump in sto c k price on earnings
announcemen t da y z ero or da y one, 0 else.
3. 𝑁 𝑒𝑔 𝑎𝑡𝑖𝑣 𝑒 𝐽 𝑢𝑚𝑝 0|1
= 1 if there is a negativ e jump in sto c k price on earnings
announcemen t d a y zero or da y one, 0 else.
3.4 Defining Con trol V ariables
𝑀 𝑎𝑟 𝑘 𝑒𝑡 𝐸 𝑞 𝑢𝑖𝑡𝑦 𝑞 is the natural logarithm of the mark et v alue of equit y at fiscal
quarter 𝑞 end. 𝐵 𝑜 𝑜 𝑘 - 𝑡𝑜 - 𝑚𝑎𝑟 𝑘 𝑒𝑡 𝑞 −1
is the natural logarithm of the b o ok-to-mark et ratio
calculated using 𝑞 − 1 b o ok v alues and mark et v alues (with negativ e b o ok v alues
excluded from the sample).
2
R O A
𝑞 −4,𝑞 −1
is op erating income after depreciati on summed
o v er the trailing four quarters deflated b y 𝑞 − 1 total assets. 𝐷 𝑆 𝑈 𝐸 𝑞 is the decile rank of
earnings surprises where earnings surprises are calculated relativ e to analyst forecasts
when a v ailable (
𝐸 𝑃 𝑆 𝑞 −𝐹 𝐸 𝑃 𝑆 𝑞 𝑃 −20
where 𝐸 𝑃 𝑆 is rep orted EPS in I/B/E/S, 𝐹 𝐸 𝑃 𝑆 is the
2
Bo ok Equit y is calculated using the metho d in Da vis et al. ( 2000 ). Bo ok Equit y = Equit y - Preferred
Sto c k. Equit y is the first nonmissing v alue in the set: {SEQQ, CEQQ+PSTK Q, A TQ-L TQ}. Preferred
Sto c k is the first nonmissing v alue in the set: {PSTKR Q, PSTK Q, 0}.
18
median forecast of EPS in a sample of most-recen t-p er-analyst forecasts no older than 90
da y s, and 𝑃 −20
is the sto c k price lagged 20 trading da ys, all split adjusted) and re lativ e
to a seasonal random w alk otherwise (
𝐸 𝑃 𝑆 𝑃 𝑋 𝑞 −𝐸 𝑃 𝑆 𝑃 𝑋 𝑞−4
𝑃 𝑞 where 𝐸 𝑃 𝑆 𝑃 𝑋 is basic
earnings p er share excluding extraordinary items and 𝑃 𝑞 is the share price at the end of
fiscal quarter 𝑞 , all split-adjusted). 𝑅 &𝐷 𝐸 𝑥𝑝 𝑒𝑛𝑠𝑒 𝑞 −4,𝑞 −1
is researc h and dev elopmen t
exp ense summed o v er the trailing four quarters deflated b y 𝑞 − 1 total assets (set to zero
if missing). 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑎𝑙 𝐿𝑒𝑣 𝑒𝑟 𝑎𝑔 𝑒 𝑞 −1
is 𝑞 − 1 total debt deflated b y total assets (set to zero
if missing). 𝑁 𝑒𝑡 𝐸 𝑥𝑡𝑒𝑟 𝑛𝑎𝑙 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑛𝑔 𝑞 −4, 𝑞 −1
is calculated follo wing Bradsha w et al.
( 2006 ) using the summed trailing four quarters of data as: sale of common and preferred
sto c k – purc hase of common and preferred sto c k – cash dividends + long-term debt
issuance – long-term debt reduction + c hanges in curren t debt (with v alues set to zero if
missing, all scaled b y 𝑞 − 1 total assets). 𝐴𝑛𝑎𝑙𝑦 𝑠𝑡 𝐶 𝑜 𝑣 𝑒𝑟 𝑎𝑔 𝑒 𝑞 is the n um b er of analysts
issuing at least one EPS forecast for fiscal quarter 𝑞 no older than 90 da ys at the earnings
announcemen t (with missing v alues set equal to zero). 𝐸 𝑃 𝑆 𝐺𝑢𝑖𝑑 𝑎𝑛𝑐 𝑒 𝑞 𝑞 −1]
is an indicator
v ariable for firms that issue at least one EPS forecast for fiscal quarter 𝑞 issued no later
than the 𝑞 − 1 earnings announcemen t date.
𝑟 −1,−1
is the abnormal sto c k return one da y prior to the e arnings announcemen t.
𝑟 −20,−2
is the da y -20 to -2 comp ounded abnormal sto c k return. 𝑟 0,1 𝑞 −4
is the 𝑞 − 4
earnings announcemen t abnormal sto c k return. 𝐼 𝑑 𝑖𝑜 𝑠𝑦 𝑛𝑐 𝑟 𝑎𝑡𝑖𝑐 𝑉 𝑜 𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦 −150,−1
is the
natural logarithm of p ercen t conditional idiosyncratic v olatilit y , ln( 𝜎 𝑖𝑡 × 100), estimated
using Equation 3.4 . Abnormal sto c k returns are calculated relativ e to the Carhart ( 1997 )
mo del as in Equation 3.3 .
𝐵 𝑒𝑡𝑎 (𝑞 −1,𝑞 )
is the slop e from a regression of daily sto c k returns on daily m ark et
returns estimated o v er the trading da ys b et w een 𝑞 − 1 and 𝑞 earnings announcemen ts.
𝑆 𝑦 𝑛𝑐 ℎ𝑟 𝑜 𝑛𝑖𝑐 𝑖𝑡𝑦 (𝑞 −1,𝑞 )
is the natural logarithm of
𝑅 2
1−𝑅
2
using the 𝑅 2
from the mark et
mo del used to calculate 𝐵 𝑒𝑡𝑎 (𝑞 −1,𝑞 )
. 𝑙𝑛(𝐼 𝑙𝑙𝑖𝑞 𝑢𝑖𝑑 𝑖𝑡𝑦 (𝑞 −1,𝑞 )
) is the natural logarithm of the
a v erage daily price impact of trades,
|𝑜𝑝𝑒𝑛−𝑡𝑜−𝑐 𝑙𝑜𝑠𝑒−𝑟 𝑒𝑡𝑢𝑟 𝑛|
$ 𝑇 𝑟 𝑎𝑑 𝑖𝑛𝑔 𝑉 𝑜𝑙𝑢𝑚𝑒 , calculated o v er the in terim
19
in t erv a l (𝑞 − 1, 𝑞 ) ( Amih ud , 2002 ; Barardehi et al. , 2021 ). 𝑙𝑛(𝐴𝑣 𝑔 . 𝑇 𝑟 𝑎𝑑 𝑖𝑛𝑔 𝑉 𝑜 𝑙𝑢𝑚𝑒 (𝑞 −1,𝑞 )
)
is the natural logarithm of the a v erage dollar trading v olume calculated o v er the in terim
in t erv a l ( 𝑞 − 1 , 𝑞 ).
3.5 Earnings Announcemen t Dates
When Compustat and I/B/E/S announcemen t dates diff er, I use the earlier earnings
announcemen t date b ecause Johnson and So ( 2018a) find the earlier date to b e more
accurate on a v erage. I incremen t after-hours earnings announcemen ts b y one trading da y
so that da y 0 represen ts the mark et reaction to the announcemen t for after-hours
announcemen ts. If the I/B/E/S time is missing, I use the earnings 8-K filing time to
iden tify after hours filings if the rep ort date, SEC acceptance date, and MIN(Compustat
date, I/B/E/S date) agree.
3.6 Earnings Announcemen t Returns
T o b oth capture mark et reactions to an y unsuccessfully iden tified after hours earnings
announcemen ts (e.g., Berkman and T ruong , 2009 ) and to a v oid p ote n tial microstructure
frictions that confound earnings announcemen t returns (e.g., Johnson and So , 2018a), I
comp ound da y zero and da y one returns in m y tests. I calculate earnings announcemen t
returns, 𝑟 0,1
, as the c omp ounded da y zero and da y one idiosyncratic return with daily
idiosyncratic returns computed as in Equation 3.3 .
3.7 Sample Selection
My sample spans earnings announcemen ts from Jan uary 2006 to Decem b er 2019. The
b e ginning of m y sample starts in 2006 b ecause it is the second complete y ear of the
20
curren t 8-K regulatory regime, and I require lagged 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 −4
in m y tests.
3
In m y sample, I include common shares (CRSP shrcd of 10 or 11) of sto c ks listed on
NYSE, Amex, or Nasdaq (CRSP exc hcd of 1, 2, or 3). Lastly , to main tain a constan t
sample across tables, I require nonmissing v alues of 𝑟 0,1
, 𝑟 0,1 𝑞 −4
, 𝑟 −1,−1
, 𝑟 −20,−2
,
𝑀 𝑎 𝑟 𝑘 𝑒𝑡 𝐸 𝑞 𝑢𝑖𝑡𝑦 𝑞 , 𝑙𝑛(𝐵 𝑜 𝑜 𝑘 - 𝑡𝑜 - 𝑚𝑎𝑟 𝑘 𝑒𝑡 𝑞 −1
), 𝑅 𝑂 𝐴 𝑞 −4, 𝑞 −1
, 𝐷 𝑆 𝑈 𝐸 𝑞 ,
𝐼 𝑑 𝑖𝑜 𝑠𝑦 𝑛𝑐 𝑟 𝑎𝑡𝑖𝑐 𝑉 𝑜 𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦 −150,−1
, 𝐵 𝑒𝑡𝑎 (𝑞 −1,𝑞 )
, 𝑆 𝑦 𝑛𝑐 ℎ𝑟 𝑜 𝑛𝑖𝑐 𝑖𝑡𝑦 (𝑞 −1,𝑞 )
, 𝑙𝑛(𝐼 𝑙𝑙𝑖𝑞 𝑢𝑖𝑑 𝑖𝑡𝑦 (𝑞 −1,𝑞 )
,
and 𝑙𝑛(𝐴𝑣 𝑔 . 𝑇 𝑟 𝑎𝑑 𝑖𝑛𝑔 𝑉 𝑜 𝑙𝑢𝑚𝑒 (𝑞 −1,𝑞 )
), and Δ Num b er of 13-F Filers
𝑞 −1,𝑞
. I set the follo wing
v ariables equal to zero if missing: 𝑅 &𝐷 𝐸 𝑥𝑝 𝑒𝑛𝑠𝑒 𝑞 −4, 𝑞 −1
, 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑎𝑙 𝐿𝑒𝑣 𝑒𝑟 𝑎𝑔 𝑒 𝑞 −1
,
𝑁 𝑒𝑡 𝐸 𝑥𝑡𝑒𝑟 𝑛𝑎𝑙 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑛𝑔 𝑞 −4, 𝑞 −1
, and 𝐴𝑛𝑎𝑙𝑦 𝑠𝑡 𝐶 𝑜 𝑣 𝑒𝑟 𝑎𝑔 𝑒 𝑞 . I winsorize non-return
con tin uous v ariables at the 1% tails of the cross-sectional distribution.
3
I b egin m y sample in the curren t 8-K rep orting regime for sev eral reasons. First, limiting the sample
increases the relev ance of this study for the curren t financial rep orting en vironmen t. Second, 8-K disclosures
are timelier in the new regime with firms ha ving four da ys to file most in terim 8-Ks compared to deadlines
ranging from fiv e to fifteen da ys in the earlier regime with filing deadlines dep ending on the 8-K item t yp e
( Lerman and Livnat , 2010 ). Lastly , the curren t 8-K regime p ostdates b oth Regulation F air Disclosure and
the Global Analyst Researc h Settlemen t, reducing concerns that firms selectiv ely disclose information to
in v estors or analysts. F urther, firms m ust file an 8-K if a priv ate meeting acciden tally yields a material
disclosure. See Lerman and Livnat ( 2010 ) and He and Plumlee ( 2020 ) for a thorough bac kground on the
curren t 8-K regime.
21
Chapter 4
Main Empirical Results
4.1 Descriptiv e Statistics
Figure 1 sho ws the distribut ion of all 8-K filings o v er the in terim in terv al. P anel A
con tains a sample of all 8-Ks, and P anel B excludes un timely filed earnings
announcemen t 8-Ks (when the SEC acceptance date is greater than the rep ort date).
Firms are required to file their earnings announcemen t 8-Ks concurren tly with or ahead
of their earnings announcemen t, y et ≈ 10% of earnings announcemen t 8-Ks are filed late
(computation un tabulated). In P anel B, the elev ated n um b er of observ ations near EA
𝑞 −1
in P anel A disapp ears when late-filed earnings 8-Ks are excluded. P anel B illustrates
that firms file non-earnings 8-Ks almost uniformly across the in terim.
Figure 2 plots the distribut ion of the n um b er of non-earnings in terim 8-Ks filed eac h
firm-quarter. Appro ximately 18% of firm-quarters do not ha v e an y 8-Ks filed b et w een
earnings announcemen ts (distan t disclosing firms), while the most common filing
frequency for firms is one in terim 8-K filed. Figure 3 plots the distribution of
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 . Outside of the in terim nondisclosure bin at 1, the distribution of
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is decreasing in densit y across 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 bins.
Figure 4 sho ws the distribut ional statistics for 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 o v er time. P anel
A plots median 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 eac h y ear, while P anel B plots the median, 25th
22
p e rcen tile, and the 75th p ercen tile. All distributions p oin t to w ard 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 increasing o v er time, reflecting the gro wing trend in disclosure bundling at earnings
announcemen ts where firms issue man y disclosures on earnings announcemen t dates
rather than spreading them out o v er time (e.g., Rogers and V an Buskirk , 2013 ; Arif et al. ,
2019 ; Bea v er e t al. , 2020 ).
T able 1 presen ts descriptiv e statistics. P anel A sho ws the time series a v erage of
cross-sectional distributions. 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 has a m ean of .514 and a standard
deviation of .346. The a v erage realized earnings announcemen t jump rate in this sample
is .377, meaning 37.7% of sto c ks jump on earnings announcemen t dates. Similar to
Kapadia and Zekhnini ( 2019 ), I also find that p ositiv e jumps are more lik ely to o ccur
than negativ e jumps, reflecting the p ositiv ely sk ew ed distribution of sto c k returns.
P anel B sho ws the time series a v erage of c ross-sectional correlations. P earson
correlations are ab o v e the diagonal of 1’s and Sp earman correlations are b elo w the
diagonal. 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is p ositiv ely correlated with 𝑅 𝑂 𝐴 , 𝑅 &𝐷 𝐸 𝑥𝑝 𝑒𝑛𝑠𝑒 ,
𝐷 𝑆 𝑈 𝐸 , 𝐼 𝑙𝑙𝑖𝑞 𝑢𝑖𝑑 𝑖𝑡𝑦 , 𝐼 𝑑 𝑖𝑜 𝑠𝑦 𝑛𝑐 𝑟 𝑎𝑡𝑖𝑐 𝑉 𝑜 𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦 , earnings announcemen t returns, and jumps.
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is negativ ely correlated with 𝑀 𝑎𝑟 𝑘 𝑒𝑡 𝐸 𝑞 𝑢𝑖𝑡𝑦 , 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑎𝑙 𝐿𝑒𝑣 𝑒𝑟 𝑎𝑔 𝑒 ,
𝑁 𝑒𝑡 𝐸 𝑥𝑡𝑒𝑟 𝑛𝑎𝑙 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑛𝑔 , 𝐵 𝑒𝑡𝑎 , 𝐴𝑣 𝑒𝑟 𝑎𝑔 𝑒 $𝑇 𝑟 𝑎𝑑 𝑖𝑛𝑔 𝑉 𝑜 𝑙𝑢𝑚𝑒 , 𝐸 𝑃 𝑆 𝐺𝑢𝑖𝑑 𝑎𝑛𝑐 𝑒 , and
𝐴𝑛𝑎𝑙𝑦 𝑠𝑡 𝐶 𝑜 𝑣 𝑒𝑟 𝑎𝑔 𝑒 . The p ositiv e correlations b et w een 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 and b oth
𝐼 𝑙 𝑙𝑖𝑞 𝑢𝑖𝑑 𝑖𝑡𝑦 and 𝐼 𝑑 𝑖𝑜 𝑠𝑦 𝑛𝑐 𝑟 𝑎𝑡𝑖𝑐 𝑉 𝑜 𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦 , and the negativ e correlation with
𝐴𝑣 𝑒𝑟 𝑎𝑔 𝑒 $𝑇 𝑟 𝑎𝑑 𝑖𝑛𝑔 𝑉 𝑜 𝑙𝑢𝑚𝑒 agree with the findings in Stoum b os ( 2019 ) who finds
illiquidit y increases gradually o v er the in terim, but firms who file 8-Ks during the in terim
ha v e a less steep increase in illiquidit y .
4.2 Determinan ts of Disclosure Distance
What t yp es of firms are distan t disclosers? V errecc hia ( 1983 ) w ould suggest that
proprietary cost of disclosure w ould mak e distan t disclosure an attractiv e option for firms
23
with high proprietary cost of disclosure. Heinle et al. ( 2020 ) surv eys the literature on
proprietary cost of disclosure and finds that man y empirical studies pro xy for proprietary
cost of disclosure using profitabilit y (e.g., Dedman and Lenno x , 2009 ), R &D in tensit y
(e.g., Ellis et al. , 2012 ), and t he mark et-to-b o ok ratio (e.g., K w ak et al. , 2012 ). The
predicted relation b et w een these v ariables and 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 under the
proprietary cost h yp othesis w ould b e p ositiv e for 𝑅 𝑂 𝐴 and 𝑅 &𝐷 𝐸 𝑥𝑝 𝑒𝑛𝑠𝑒 , and negativ e
for 𝐵 𝑜 𝑜 𝑘 - 𝑡𝑜 - 𝑚𝑎𝑟 𝑘 𝑒𝑡 (i.e., gro wth firms ha v e high proprietary costs of disclosure).
Firms highly relian t on external financing ha v e inc en tiv es to issue more disclosures in
order to reduce information asymmetry with in v estors (e.g., F rank el et al. , 1995 ),
resulting in a negativ e predicted relation b et w een 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 and b oth
𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑎𝑙 𝐿𝑒𝑣 𝑒𝑟 𝑎𝑔 𝑒 and 𝑁 𝑒𝑡 𝐸 𝑥𝑡𝑒𝑟 𝑛𝑎𝑙 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑛𝑔 .
Ka jüter et al. ( 2018 ) find that compliance cost of disclosure is a deterren t for small
firms, whic h suggests a negativ e relation b et w een 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 and
𝑀 𝑎 𝑟 𝑘 𝑒𝑡 𝐸 𝑞 𝑢𝑖𝑡𝑦 . Small firms also tend to ha v e no or lo w analyst co v erage, whic h leads to
the same negativ e prediction for 𝐴𝑛𝑎𝑙𝑦 𝑠𝑡 𝐶 𝑜 𝑣 𝑒𝑟 𝑎𝑔 𝑒 .
Skinner ( 1994 ) do cumen ts that some firms pre-announce bad news lik ely due to
litigation risk concerns and Marino vic and V aras ( 2016 ) mo del that litigation risk cro wds
out p ositiv e disclosures. Litigation risk suggests a p ositiv e relation b et w een earnings
surprise ( 𝐷 𝑆 𝑈 𝐸 𝑞 ) and 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 b ecause firms with p ositiv e earnings news
could remain quiet while firms with bad news ha v e litigation risk-induced incen tiv es to
announce bad news in a timely manner.
Dra wing on these determinan ts of disclosure from p rior literature, in T able 2 , I
estimate T obit regressions of 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 on the aforemen tioned explanatory
v ariables. I estimate T obit regressions b ecause 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is b ounded b et w een
zero and one and T obit mo dels are appropriate for b ounded dep enden t v ariables. In
column (1), I regress 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 on v ariables publicly a v ailable ahead of
earnings announcemen ts: 𝑀 𝑎𝑟 𝑘 𝑒𝑡 𝐸 𝑞 𝑢𝑖𝑡𝑦 , 𝐴𝑛𝑎𝑙𝑦 𝑠𝑡 𝐶 𝑜 𝑣 𝑒𝑟 𝑎𝑔 𝑒 , 𝐸 𝑃 𝑆 𝐺𝑢𝑖𝑑 𝑎𝑛𝑐 𝑒 ,
24
𝐵 𝑜 𝑜 𝑘 - 𝑡𝑜 - 𝑚𝑎𝑟 𝑘 𝑒𝑡 , 𝑅 𝑂 𝐴 , 𝑅 &𝐷 𝐸 𝑥𝑝 𝑒𝑛𝑠𝑒 , 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖 𝑎𝑙 𝐿𝑒𝑣 𝑒𝑟 𝑎𝑔 𝑒 , and
𝑁 𝑒𝑡 𝐸 𝑥𝑡𝑒𝑟 𝑛𝑎𝑙 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑛𝑔 . The estimates are consisten t with firms with high proprietary
cost of disclosure b eing distan t disclosing sto c ks, while sto c ks with reliance on external
financing and larger sto c ks with more analyst co v erage b eing recen t disclosers. In column
(2), I augmen t the mo del with earnings news ( 𝐷 𝑆 𝑈 𝐸 𝑞 ) to examine the theory of
litigation risk cro wding out p ositiv e disclosures. I find evidence consisten t with firms
with bad earnings news issuing in terim disclosures, reducing their 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 .
𝐸 𝑃 𝑆 𝐺𝑢𝑖𝑑 𝑎𝑛𝑐 𝑒 has an insignifican t relation with 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 , suggesting
𝐷 𝑖𝑠 𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 do es not differ b et w een guiders and nonguiders after con trolling for
other fir m c haracteristics.
In column (3), I examine the p ersistence of 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 b y regressing
𝐷 𝑖𝑠 𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 on lagged 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 −4
. 𝐷 𝑖𝑠𝑐 𝑙 𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 −4
has
p ositiv e p ersistence, but its p ersistence is less than one with a co efficien t of 0.333 and a
Pseudo 𝑅 2
of 6%, suggesting it is not a firm fixed disclosure c haracteristic for most firms.
Column (4) [5] augmen ts the mo del from column (1) [2] with lagged
𝐷 𝑖𝑠 𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 −4
and the relations remain the same b et w een
𝐷 𝑖𝑠 𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 and the explanatory v ariables from columns (2) and (3) with the
exception of 𝑁 𝑒𝑡 𝐸 𝑥𝑡𝑒𝑟 𝑛𝑎𝑙 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑛𝑔 losing significance, s uggesting that trailing
𝑁 𝑒𝑡 𝐸 𝑥𝑡𝑒𝑟 𝑛𝑎𝑙 𝐹 𝑖𝑛𝑎𝑛𝑐 𝑖𝑛𝑔 is not a go o d explanatory v ariable for inno v ations in
𝐷 𝑖𝑠 𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 . I include these v ariables in m y main tests to con trol for the
determinan ts of 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 .
4.3 Earnings Announcemen t Jump Risk
Next, I test H1 that examines the link b e t w een 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 and
idiosyncratic jump risk in T able 3 . P anel A sho ws a v erage idiosyncratic jump rates for
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 sorted p ortfolios. I sort sto c ks in to three p ortfolios eac h quarter
25
based on the distribution of 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 for that quarter. Quarters are based
on earnings announcemen t calendar quarters (e.g., earnings announcemen ts in Jan uary
20YY – Marc h 20YY constitute a quarter). T able 3 , P anel A pro vides univ ariate
evidence in supp ort of H1. Distan t disclosing firms ha v e higher jump risk at earnings
announcemen ts than recen t disclosing firms and this relation is statistically significan t.
The spread is wider for p ositiv e jumps compared to negativ e jumps, l ik ely due to
litigation risk concerns discouraging systematic hoarding of bad news that w ould in vite a
negativ e jump at earnings announcemen ts.
T able 3 , P anels B-D demonstrate robustness to con trolling for other firm
c haracteristics in logit regressions. P anel B presen ts results for all jumps, P anel C
presen ts results for p ositiv e jumps, and P anel D presen ts results for negativ e jumps. I
con trol for the determinan ts of 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 from column (4) in T able 2 .
A dditionally , I include con trols for mark et v ariables calculated o v er the in terim: 𝐵 𝑒𝑡𝑎 ,
𝑆 𝑦 𝑛𝑐 ℎ𝑟 𝑜 𝑛𝑖𝑐 𝑖𝑡𝑦 , 𝐼 𝑙𝑙𝑖𝑞 𝑢𝑖𝑑 𝑖𝑡𝑦 , 𝐴𝑣 𝑔 . 𝑇 𝑟 𝑎𝑑 𝑖𝑛𝑔 𝑉 𝑜 𝑙𝑢𝑚𝑒 , 𝑟 −20,−2
, and 𝑟 −1,−1
. These regressions
also include time indicators to con trol for y ear-quarter fixed effects in most sp ecifications,
and F ama-F renc h 12 indicators to con trol for industry fixed effects in column (6).
The p ositiv e co efficien t for 𝐸 𝑃 𝑆 𝐺𝑢𝑖𝑑 𝑎𝑛𝑐 𝑒 𝑞 𝑞 −1]
is consisten t with prior literature that
guiding firms tend to ha v e higher v olatilit y . Billings et al. ( 2015 ) pro vide evidence that
managers initiate guidance in resp onse to v olatilit y run-ups, arguing that it is not
guidance p er se that increases v olatilit y . Ho w ev er, past guidance is a stic ky c haracteristic,
and guiding firms predominan tly bundle guidance with earnings announcemen ts (e.g.,
Rogers and V an Buskirk , 2013 ), whic h c ould increase the lik eliho o d of a jump.
The results in T able 3 pro vide evidence consisten t with H1: distan t disclosing sto c ks
ha v e elev ated idiosyncratic jump risk at earnings announcemen ts. Figure 5 illustrates the
T able 3 results in ev en t time, plotting the cum ulativ e prop ortion of sto c ks with at least
one jump for eac h 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 sorted p ortfolio. Prior to earnings
announcemen ts, the three p ortfolios mo v e in unison with the recen t disclosing p ortfolio
26
ha v ing the highest jump rate, but once the announcemen t o ccurs, the distan t disclosing
p o rtfolio has the highest jump rate, consisten t with distan t disclosing sto c ks b eing more
sensitiv e to earnings news.
4.4 Earnings Announcemen t Sto c k Returns
Next, I test whether in v estors price this r isk. If in v estors price jump risk for distan t
disclosing sto c ks, these sto c ks should earn p ositiv e abnormal returns at earnings
announcemen ts. Eac h quarter, I sort sto c ks in to three p ortfolios based on their
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 calculated at eac h earnings announcemen t. T able 4 , P anel A
presen ts Carhart ( 1997 ) alphas for 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 sorted p ortfolios. Distan t
disclosing sto c ks earn an alpha of 11.1 basis p oin ts at earnings announcemen ts.
Figure 6 plots the cum ulativ e abnormal returns to 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 sorted
p o rtfolios in ev en t time for the elev en trading da ys cen tered on the earnings
announcemen t date. Ahead of the announcemen t, prices mo v e up w ard, consisten t with
mark et mak ers pro viding liquidit y asymmetrically ahead of earnings announcemen ts
( Johnson and So , 2018a). Once firms announce earnings, distan t disclosing sto c ks earn
p o sitiv e abnormal returns that do not rev erse, in con trast to the sharp rev ersal of the
pre-announcemen t return run-up for recen t disclosing sto c ks. Johnson and So ( 2018a)
also do cumen t an unconditional return rev ersal after earnings announcemen ts.
Figure 7 com bines the i nsigh ts from Figure 5 and Figure 6 , plotting o nly the distan t
min us recen t difference in sto c k returns and jump rates. Ahead of earnings
announcemen ts, the difference is insignifican t, but once firms announce earnings, the
cum ulativ e abnormal returns and cum ulativ e jump rates mo v e in tandem, pro viding
further evidence that the abnormal returns are comp ensation for jump risk at earnings
announcemen ts.
27
T able 4 , P anel B presen ts panel regressions of abnormal earnings announcemen t
returns, 𝑟 0,1
, on explan atory v ariables. 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 remains a significan t
predictor of earnings announcemen t returns after con trolling for other firm
c h aracteristics. Column (1) presen ts a univ ariate regression of returns on
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 . Column (2) adds y ear-quarter fixed effects. Column (3) is a
regression with 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 −4
. 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 −4
is insignifican t,
suggesting that the p ositiv e earnings announcemen t returns are not driv en b y firms that
routinely remain silen t. Column (4) adds bac k curren t p erio d 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 and
sho ws it is robust to con trolling for lagged 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 −4
, pro viding further
evidence it is abnormal 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 that driv es the risk premium, and not a
stic k y dis closure c haracteristic. Next, in column (5), I add the ex an te determinan ts of
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 from T able 2 as con trols. A dditionally , I con trol for mark et
v ariables including 𝑟 −20,−2
, 𝑟 −1,−1
, 𝑟 0,1 𝑞 −4
, 𝐼 𝑙𝑙𝑖𝑞 𝑢𝑖𝑑 𝑖𝑡𝑦 , 𝐴𝑣 𝑒𝑟 𝑎𝑔 𝑒 𝑇 𝑟 𝑎𝑑 𝑖𝑛𝑔 𝑉 𝑜 𝑙𝑢𝑚𝑒 ,
𝑆 𝑦 𝑛𝑐 ℎ𝑟 𝑜 𝑛𝑖𝑐 𝑖𝑡𝑦 , 𝐼 𝑑 𝑖𝑜 𝑠𝑦 𝑛𝑐 𝑟 𝑎𝑡𝑖𝑐 𝑉 𝑜 𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦 , and 𝐵 𝑒𝑡𝑎 . The results are robust to including
these v ariables. Next, in column (6), I add industry fixed effects based on the F ama and
F renc h 12 industry classification. The estimates remain similar. Lastly , in column (7), I
add inform ation not a v ailable prior to the announcemen t: 𝐷 𝑆 𝑈 𝐸 𝑞 . The estimate on
𝐷 𝑖𝑠𝑐 𝑙 𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 b ecomes statistically insignifican t. This suggests that distan t
disclosing fir ms rep ort b etter earnings news than recen t disclosing sto c ks on a v erage, and
this explain s the abnormal returns. Higher idiosyncratic risk for distan t disclosing sto c ks
discourages preannouncemen t arbitrage activit y , resulting in the jump-risk premium for
distan t disclos ing sto c ks, in line with the rationale in Kapadia and Zekhnini ( 2019 ). The
results in T able 4 pro vide evidence consisten t with H2 that in v estors price
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 as increasing risk at earnings announcemen ts.
Figure 8 examines the c um ulativ e abnormal returns for 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 sorted
p o rtfolios for go o d news in P anel A and bad news in P anel B. Go o d news is ab o v e
median earnings surprises and bad news is b elo w median earnings surprises based on
28
𝐷 𝑆 𝑈 𝐸 𝑞 . This figure demonstrates that distan t disclosing sto c ks are more sensitiv e to
b o th go o d and bad news relativ e to recen t disclosing sto c ks.
Figure 9 plots the hedge p ortfolios conditional on news as in Figure 8 . The absolute
magnitude is greater for go o d news relativ e to bad news. The p oten tial for negativ e news
for distan t disclosing firms lik ely creates the asymmetry in magnitudes do cumen ted here.
If in v estors price idiosyncratic risk ahead of earnings announcemen ts, preannouncemen t
prices will b e lo w er than exp ected p ost announcemen t prices.
4.5 In v estor Recognition and the Pricing of
Idiosyncratic Risk
T able 5 dra ws on Merton ( 1987 )’s in v estor recognition theory to he lp distinguish
b e t w een the idiosyncratic risk and the systematic risk explanation for the p ositiv e
abnormal returns to distan t disclosing sto c ks. Under an idiosyncratic risk explanation,
the returns to distan t disclosing sto c ks should b e concen trated in sto c ks with lo w in v estor
recognition. T o test this, I partition the sample in to lo w and high c hanges in in v estor
recognition subsamples based on a cross-sectional median spl it on the quarterly c hange in
the n um b er of 13-F filing institutions that hold the sto c k, follo wing Leha vy and Sloan
( 2008 ). Leha vy and Sloan ( 2008 ) note that examining c hanges in in v estor recognition,
instead of lev els, creates a more p o w erful test of the theory in Merton ( 1987 ) due to it
reducing omitted v ariable problems that are difficult to con trol for in the cross section.
1
The n um b er of observ ations in eac h subsample is not equal b ecause there is no c hange
in the n um b er of institutional o wners for ≈ 7% of firm-quarters. I lag institutional
1
Another de terminan t of the sensitivit y of exp ected returns to idiosyncratic risk in Merton’s mo del is
the size of the firm. In Merton’s mo del, the p ositiv e relation b et w een idiosyncratic risk and exp ected returns
is increasing in firm size, ceteris paribus, while the p ositiv e relation b et w een idiosyncratic risk and exp ected
returns is decreasing in in v estor recognition, ceteris paribus. A cross-sectional test based on the lev el of
in v estor recognition also partitions on the size of the firm due to the p ositiv e correlation b et w een in v estor
recognition and firm size. Therefore, a cross-sectional cut on the lev el of in v estor recognition do es not yield
a p o w erful test for examining pricing of idiosyncratic risk due to the offsetting in teractiv e effects in equation
(31.a) in Merton ( 1987 ) arising from the p ositiv e correlation b et w een firm size and i n v estor recognition.
29
o wnership data b y t w o mon ths from the rep ort date to ensure public a v ailabilit y of data
due to the 45 da y rep orting deadline for filing form 13-F. That is, o wnership data with a
rep ort date of 12/31/Y0 w ould b e a v ailable for earnings announcemen ts b eginning in
03/01/Y1.
T able 5 , P anel A presen ts the earnings announcemen t return results for the lo w
c hanges in in v estor recognition subsample. H3 predicts 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 has the
strongest explanatory p o w er for earnings announcemen t returns when in v estor
recognition is lo w. Consisten t with H3, 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 has stronger explanatory
p o w er for sto c k returns in the lo w c hanges in in v estor recognition subsample. In the high
c h anges in in v estor recognition subsample in P anel B, 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 has almost
no e xplanatory p o w er for earnings announcemen t sto c k returns. This is consisten t with
Merton ( 1987 )’s mo del where the sensitivit y of the p ositiv e relation b et w een exp ec ted
returns and idiosyncratic risk decreases as in v estor recognition increases.
30
Chapter 5
A dditional Empirical Results
5.1 In v estor Recognition and Jump Risk
An alternativ e explanation for 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 ’s stronger explanatory p o w er for
sto c k returns in the lo w c hanges in in v estor recognition cross section could b e that jump
risk is higher for sto c ks with lo w c hanges in in v estor recognition. Under this explanation,
the increased exp osure to idiosyncratic risk explains the stronger return results for the
lo w c hanges in in v estor recognition sample, not the in v estor recognition h yp othesis. T o
examine this explanation, I re-estimate the logit regressions from T able 3 using the same
in v estor recognition cross-sectional cut emplo y ed in T able 5 . T able 6 , P anel A estimates
the relation b et w een jump risk and 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 for sto c ks with lo w c hanges in
in v estor recognition and P anel B estimates the same relation for sto c ks with high c hanges
in in v estor recognition. Comparing the estimates in P anels A and B rev eals that the
relation b et w een jump risk and 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is similar for b oth c hanges in
in v estor recognition subsamples. In the most stringen t sp ecification, column (6), the
co e fficien t on 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is 0.326 in P anel A and 0.352 i n P anel B. Distan t
disclosing sto c ks with high c hanges in in v estor recognition ha v e similar, y et sligh tly
higher jump risk. This similarit y in jump risk b et w een subsamples pro vides evidence that
31
differing exp osure to jump risk is not a comp elling alternativ e explanation for the
differing sto c k return results b et w een subsamples in T able 5 .
5.2 In tertemp oral Analysis
If the abnormal returns to the 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 trading strategy represen t
correction of mispricing, in v estors should learn o v er time and eliminate this trading
opp ortunit y . T o examine this argumen t, T able 7 splits the sample in to t w o halv es. P anel
A presen ts the first half (Jan uary 2006–Decem b er 2012) and P anel B presen ts the second
half (Jan uary 2013–Decem b er 2019). The results are stable across subp erio ds, pro viding
further evidence for the idiosyncratic risk explanation.
5.3 Size Analysis
T able 8 tests for cross-sectional differences in the pricing effects for small and big
sto c ks, based on a median split on mark et v alue of equit y measured at fiscal quarter end.
P a nel A presen ts regressions for sto c ks b elo w the median and P anel B presen ts results for
sto c ks ab o v e the median. The magnitudes of 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 are larger for small
sto c ks in P anel A relativ e to big sto c ks in P anel B, y et the results are still statistically
significan t for big sto c ks. Big sto c ks ha v e b etter alternativ e sources of information and
lo w er exp osure to jump risk at earnings announcemen ts, whic h could explain the differing
magnitudes of idiosyncratic risk pricing.
32
5.4 Con trolling for Beta-shifts at Earnings
Announcemen ts
Sev eral pap ers do cumen t elev ated systematic ris k at earnings announcemen t dates
(e.g., Ball and K othari , 1991 ; P atton and V erardo , 2012 ; Sa v or and Wilson , 2016 ; Chan
and Marsh , 2021 ). In the main tests, the risk factor adjustmen t I use is relativ e to factor
loadings estimated out of sample. T o address the concern that the abnormal returns are
due to incomplete risk adjustmen t as a result of b eta-shifts at earnings announcemen ts, I
estimate p o oled CAPM factor loadings using earnings announcemen t returns for eac h
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 sorted p ortfolio in T able 9 .
The first ro w of T able 9 sho ws the sto c k return in excess of the risk-free rate for a
baseline magnitude (Excess Return
0,1
). Distan t disclosing sto c ks earn excess r eturns of
16.1 basis p oin ts o v er announcemen t da ys [0,1]. Next, I regress excess returns
0,1
on
con temp oraneous mark et returns
0,1
in a p o oled regression across all earnings
announcemen ts in the sample for eac h disclosure p ortfolio. Using the P o oled CAPM Beta
as a measure of announcemen t exp osure to systematic risk, the P o oled CAPM A djusted
Return
0,1
decreases to 11.2 basi s p oin ts, and remains as statistically significan t as the
excess return. This magnitude is nearly iden tical to the 11.1 basis p oin t Carhart ( 1997 )
alpha estimated in T able 4 . Comparing the a v erage Out-of-Sample CAPM b eta
estimated o v er the in terim in terv al to the P o oled CAPM b eta estimated using earnings
announcemen t returns, the b eta increases from 1.004 to 1.042 for distan t disclosing
sto c ks. While b eta increases, it do es not increase b y enough to explain the p ositiv e
abnormal returns.
33
5.5 Alternativ e Measures of Jumps
T able 10 sho w cases robustness to the results in T able 3 , P anel B to alternativ e
definitions of jumps. In T able 3 , jumps are defined relativ e to eac h sto c k’s historic
idiosyncratic v olatilit y , as in Kapadia and Zekhnini ( 2019 ). In T able 10 , I define jumps as
absolute abnormal returns of v arying magnitudes to ensure the jump results are not
driv en b y the c hoice to define jumps relativ e to historic v olatilit y (i.e., if firms with recen t
disclosure ha v e higher conditional v olatilit y due to making a recen t disclosure, they
w ould ha v e a higher v olatilit y h urdle to clear for jump classification). P anel A estimate s
logit mo dels with indicator v ariables for an absolute abnormal return of 5% or higher,
P a nel B estimates logit mo dels with indicator v ariables for an absolute abnormal return
of 10% or higher, and P anel C estimates logit mo dels with indicator v ariables for an
absolute abnormal return of 15% or higher. The results for 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 are
consisten t with those rep orted in T able 3 , pro viding evidence that the results are not
driv en b y defining jumps relativ e to historic v olatilit y .
5.6 Con tin uous Idiosyncratic V olatilit y
Man y studies in accoun ting and finance study idiosyncratic v olatilit y , y et the
literature on idiosyncratic jump v olatilit y is relativ ely nascen t. T able 11 estimates
regressions using con tin uous measures of idiosyncratic v olatilit y commonly used in ev en t
studies, absolute returns in P anel A and squared returns in P anel B. The results for
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 are consisten t with those in T able 3 .
T o illustrate the imp ortance of jump (extrem e) v olatilit y for 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 ’s
p o sitiv e relation with idiosyncratic v olatilit y , in T able 12 , I estimate the same m o dels in
T able 11, but exclude observ ations that a re categorized as jumps (based on the main
Kapadia and Zekhnini ( 2019 ) classification of jumps). The 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 estimates flip signs in sp ecifications without con trol v ariables, and the p ositiv e relation is
34
w eak er when including con trols. F or absolute (squared) returns, the column ( 6) results
shrink from 0.802 (16.527) in the full sample to 0.204 (2.165) in the subsample excluding
jumps. These results demonstrate that jump v olatilit y is the main driv er of the v olatilit y
results, consisten t with a more refined jump risk explanation rather than a more general
idiosyncratic risk explanation.
5.7 Alternativ e Measures of Disclosure Distance
The results th us far ha v e defined 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 as the n um b er of trading da ys
since the last disclosure scaled b y the n um b er of trading da ys since the last earnings
announcemen t. T able 13 replicates the m ain results using an unscaled measure of
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 : 𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝑞 . 𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝑞 calculates the
n um b er of trading da ys since the last 8-K filing, or the n um b er of da ys since the last
earnings announcemen t for firms that do not file an y in terim 8-Ks. P anel A demonstrates
robustness for the main jump logit mo dels in T able 3 , P anel B when using
𝐷 𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝑞 in place of the main scale d measure. P anel B demonstrates
robustness for the main earnings announcemen t sto c k return results in T able 4 , P anel B.
T able 14 uses tercile p ortfolio ranks to address concerns ab out
𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 ’s at ypical distribution driving the results. Here,
𝑅 𝑎𝑛𝑘 𝑒𝑑 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 ranges from [0,1] via the transformation:
𝑅 𝑎𝑛𝑘 ∈0,1,2
2
. P anel
A demonstrates robustness for the main jump logit m o dels in T able 3 , P anel B when
using 𝑅 𝑎𝑛𝑘 𝑒𝑑 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 . P anel B demonstrates robustness for the main
earnings announcemen t sto c k return results in T able 4 , P anel B when using
𝑅 𝑎𝑛𝑘 𝑒𝑑 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 . The co e fficien t on 𝑅 𝑎𝑛𝑘 𝑒𝑑 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 resem bles a tercile hedge p ortfolio; for example, the estimate in column (1) is 0.291 whic h
corresp onds to the 0.291% hedge return in T able 4 , P anel A. T able 14 illustrates that the
at y pical distribution of 𝐷 𝑖𝑠𝑐 𝑙𝑜 𝑠𝑢𝑟 𝑒 𝐷 𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 is not the driv er of the regression results.
35
Chapter 6
Conclusion
This dissertation pro vides empirical evidence c onsisten t with elapsed p erio ds of
nondisclosure b et w een mandatory rep orting p erio ds increasing idiosyncratic jump risk at
earnings announcemen ts. Distan t disclosing sto c ks earn a p ositiv e alpha at earnings
announcemen ts, suggesting in v estors an ticipate elev ated idiosyncratic jump risk for
distan t disclosing sto c ks and demand a risk premium to hold these sto c ks at earnings
announcemen ts.
These findings ha v e implications for in v estors, manage rs, and regulators. F or
in v estors, the higher earnings announcemen t returns for distan t disclosing sto c ks
illustrate ho w short-term discoun t rates can v ary as a function of the elapsed time since
disclosure. This suggests that in v estors consider disclosure distance when c ho osing
short-term discoun t rates for sto c k v aluation. Risk-a v erse individual in v estors ma y w an t
to a v oid en tering in to a p osition in distan t disclosing sto c ks ahead of earnings
announcemen ts due to their high jump risk.
F or managers, this dissertation highligh ts increased jum p risk as a capital mark et
consequence for the recen t trend in disclosure “bundling” (e.g., Rogers and V an Buskirk ,
2013 ; Arif et al. , 2019 ; Bea v er et al. , 2020 ) where firms release man y disclosure items at
earnings announcemen ts, rather than spreading them out o v er time. This disclosure
p o licy can lead to a more “jump y” information en vironmen t, increasing jump risk for
36
in v estors. Managers trade off increased disclosure against other factors, so more frequen t
disclosure ma y not b e optimal for all firms despite p oten tial capital mark et b enefits.
The shift to w ard more frequen t and timely dis closure as a result of the 2004 SEC
regulation for 8-K disclosure created a more lev el information pla ying field for in v estors.
Y et, appro ximately 18% of firms do not file in terim 8-Ks and the se firms ha v e increased
jump risk at earnings announcemen ts. This finding suggests that the list of triggering
corp orate ev en ts could b e expanded to further up date in v estors in the in terim.
37
Bibliograph y
Amih ud, Y., 2002. Illiquidit y and sto c k returns: cross-section and time-series effects.
Journal of Financial Mark ets 5, 31–56.
Amiram, D., Cserna, B., Kala y , A., Levy , A., 2019. The information en vironmen t,
v olatilit y structure, and liquidit y . Colum bia Business Sc ho ol Researc h P ap er No. 15-62,
A v ailable at SSRN: h ttps://ssrn.com/abstract=2618424.
Ang, A., Ho dric k, R., Xing, Y., Zhang, X., 2006. The cross-section of v olatilit y and
exp ected returns. Journal of Finance 61, 259–99.
Arif, S., Marshall, N. T., Sc hro eder, J. H., Y ohn, T. L., 2019. A gro wing disparit y in
earnings disclosure mec hanisms: The rise of concurren tly released earnings
announcemen ts and 10-ks. Journal of A ccoun ting and Economics 68, 1–28.
Baginski, S. P ., Campb ell, J. L., Hinson, L. A., K o o, D. S., 2017. Do Career Concerns
Affect the Dela y of B ad News Disclosure? The A ccoun ting Review 93, 61–95.
Bagnoli, M., Kross, W., W atts, S. G., 2002. The information in managemen t’s exp ected
earnings rep ort date: A da y late, a p enn y short. Journal of A ccoun ting Researc h 40,
1275–1296.
Balakrishnan, K., Billings, M. B., Kelly , B., Ljungqvist, A., 2014. Shaping liquidit y: On
the causal effects of v olun tary disclosure. The Journal of Finance 69, 2237–2278.
Ball, R., K othari, S. P ., 1991. Securit y returns around earnings announcemen ts. The
A ccoun ting Review 66, 7 18–738.
Barardehi, Y. H., Bernhardt, D., R uc h ti, T. G., W eidenmier, M., 2021. The Nigh t and
Da y of Amih ud’s (2002) Liquidit y Measure. The Review of Asset Pricing Studies 11,
269–308.
Barb er, B., George, E. D., Leha vy , R., T rueman., B., 2013. The earnings announcemen t
premium around the glob e. Journal of Financial Economics 108, 118–138.
Barb er, B., Odean, T., 2008. All that glitters: the effect of atten tion and news on the
buying b eha vior of individual and institutional in v estors. Review of Financial Studies
21, 785–818.
Barry , C. B., Bro wn, S. J., 1985. Differen tial information and securit y mark et
equilibrium. Journal of Financial and Quan titativ e Analysis 20, 407–422.
38
Barth, M. E., So, E. C., 2014. Non-div ersifiable v olatilit y risk and risk premiums at
earnings announcemen ts. The A ccoun ting Review 89, 1579–1607.
Bea v er, W., 1968. The information con ten t of ann ual earnings announceme n ts. Journal of
A ccoun ting Researc h 6, 67–92.
Bea v er, W. H., McNic hols, M. F., W ang, Z. Z., 2020. Increased mark et resp onse to
earnings announcemen ts in the 21st cen tury: An empirical in v estigation. Journal of
A ccoun ting and Economics 69, 1 01244.
Begley , J., Fisc her, P . E., 1998. Is there information in an earnings announcemen t dela y?
Review of A ccoun ting Studies 3, 347–363.
Berkman, H., T ruong, C., 2009. Ev en t da y 0? after-hours earnings announcemen ts.
Journal of A ccoun ting Researc h 47, 71–103.
Billings, M. B., Cedergren, M. C., 2015. Strategic silence, insider selling and litigation
risk. Journal of A ccoun ting and Economics 59, 119–142.
Billings, M. B., Jennings, R., Lev, B., 2015. On guidance and v olatilit y . Journal of
A ccoun ting and Economics 60, 1 61–180.
Bo dnaruk, A., Ostb erg, P ., 2009. Do es in v estor recognition predict returns? Journal of
Financial Economics 91, 208–226.
Bollerslev, T., Li, S. Z., T o doro v, V., 2016. Roughing up b eta: Con tin uous v ersus
discon tin uous b etas and the cross section of exp ected sto c k returns. Journal of
Financial Economics 120, 464–490.
Bond, P ., Zeng, Y., 2021. Silence is safest: Information disclosure when the audiences
preferences are uncertain. Journal of Financial Economics 145, 178–193.
Bradsha w, M., Ric hardson, S., Sloan., R., 2006. The relation b et w een corp orate financing
activities, analysts’ forecasts and sto c k returns. Journal of A ccoun ting and Economics
42, 53–85.
Bégin, J.-F., Dorion, C., Gauthier, G., 2019. Idiosyncratic Jump Risk Matters: Evidence
from Equit y Returns and Options. The Review of Financial Studies 33, 155–211.
Campb ell, J. L., T w edt, B. J., Whipple, B. C., 2021. T rading prior to the disclosure of
material information: Evidence from regulation fair disclosure form 8-ks.
Con temp orary A ccoun ting Researc h 38, 412–442.
Carhart, M. M., 1997. On p ersistence in m utual fund p erformance. The Journal of
Finance 52, 57–82.
Chan, K. F., Marsh, T., 2021. Asset pricing on earnings announcemen t da ys. Journal of
Financial Economics 144, 1022–1042.
39
Chapman, K. L., Reiter, N., White, H. D., Williams, C. D., 2019. Information o v erload
and disclosure smo othing. Review of A ccoun ting Studies 24, 1486–1522.
Chen, S., Matsumoto, D., Ra jgopal, S., 2011. Is silence golden? an empirical analysis of
firms that stop giving quarterly earnings guidance. Journal of A ccoun ting and
Economics 51, 134–150.
Chen, Y., Ng, J., Y ang, X., 2021. T alk less, learn more: Strategic disclosure in resp onse
to managerial learning from the options mark et. Journal of A ccoun ting Researc h 59,
1609–1649.
Cohen, D. A., Dey , A., Lys, T. Z., Sunder, S. V., 2007. Earnings announcemen t premia
and the limits to arbitrage. Journal of A ccoun ting and Economics 43, 153 – 180.
Da v is, J. L., F ama, E. F., F renc h, K. R., 2000. Characteristics, co v ariances, and a v erage
returns: 1929 to 1997. Journal of Finance 55, 389–406.
Dedman, E., Lenno x, C., 2009. P erceiv ed comp etition, profitabilit y and the withholding
of information ab out sales and the cost of sales. Journal of A ccoun ting and Economics
48, 210–230.
deHaan, E., Shevlin, T., Thorno c k, J., 2015. Mark et (in)atten tion and the strategic
sc heduling and timing of earnings announcemen ts. Journal of A ccoun ting and
Economics 60, 36–55.
Diamond, D. W., V errecc hia, R. E., 1991. Disclosure, liquidit y , and the cost of capital.
The Journal of Finance 46, 1325–1359.
Dubinsky , A., Johannes, M., Kaec k, A., Seeger, N. J., 2018. Option Pricing of Earnings
Announcemen t Risks. T he Review of Financial Studies 32, 646–687.
Dy e, R. A., 1985. Disclosure of nonproprietary information. Journal of A ccoun ting
Researc h 23, 123–145.
Ellis, J. A., F ee, C. E., Thomas, S. E., 2012. Proprietary costs and the disclosure of
information ab out customers. Journal of A ccoun ting Researc h 50, 685–727.
F ama, E., MacBeth, J., 1973. Risk, return, and equilibrium: Empirical tests. Journal of
P olitical Econom y 81, 607–636.
F rank el, R., McNic hols, M., Wilson, G. P ., 1995. Discretionary disclosure and external
financing. The A ccoun ting Review 70, 135–150.
F razzini, A., Lamon t, O., 2007. The earnings announcemen t premium and trading
v olume. NBER W orking P ap er No. w13090 A v ailable at SSRN:
h ttps://ssrn.com/abstract=986940 .
F u, F., 2009. Idiosyncratic risk and the cross-section of exp ected sto c k returns. Journal of
Financial Economics 91, 24–37.
40
Gargano, A., Rossi, A. G., W ermers, R., 2016. The F reedom of Information A ct and the
Race T o w ard Information A cquisition. The Review of Financial Studies 30, 2179–2228.
Go y al, A., S an ta-Clara, P ., 2003. Idiosyncratic risk matters! The Journal of Finance 58,
975–1007.
Guttman, I., Kremer, I., Skrzypacz, A., 2014. Not only what but also when: A theory of
dynamic v olun tary disclosure. American Economic Review 104, 2400–2420.
Hartzmark, S. M., Solomon, D. H., 2018. Recurring firm ev en ts and predictable returns:
The within-firm time series. Ann ual Review of Financial Economics 10, 499–517.
He, J., Plumlee, M. A., 2020. Measuring disclosure using 8-k filings. Review of
A ccoun ting Studies 25, 903–962.
Heflin, F. L., Sha w, K. W., Wild, J. J., 2005. Disclosure p olicy and mark et liquidit y:
Impact of depth quotes and order sizes. Con temp orary A ccoun ting Researc h 22,
829–865.
Heinle, M. S., Sam uels, D., T a ylor, D. J., 2020. Proprietary costs and disclosure
substitution: Theory and empirical evidence. W orking P ap er A v ailable at SSRN:
h ttps://ssrn.com/abstract=3173664 .
Heitz, A. R., Nara y anamo orth y , G., Zekhnini, M., 2020. The disapp earing earnings
announcemen t premium. W orking P ap er A v ailable at SSRN:
h ttps://ssrn.com/abstract=3296537 .
Hsu, C., W ang, R., Whipple, B. C., 2021. Non-gaap earnings and sto c k price crash risk.
Journal of A ccoun ting and Economics 73, 101473.
Hutton, A. P ., Marcus, A. J., T ehranian, H., 2009. Opaque financial rep orts, r2, and
crash risk. Journal of Financial Economics 94, 67–86.
Ja y araman, S., W u, J. S., 2018. Is Silence Golden? Real Effects of Mandatory Disclosure.
The Review of Financial Studies 32, 2 225–2259.
Johnson, T., Kim, J., So, E., 2020. Exp ectations managemen t and sto c k returns. The
Review of Financial Studies 33, 4580–4626.
Johnson, T. L., So, E. C., 2018a. Asymmetric trading costs prior to earnings
announcemen ts: Implications for price disco v ery and returns. Journal of A ccoun ting
Researc h 56, 217–263.
Johnson, T. L., So, E. C., 2018b. Time will tell: Information in the timing of sc heduled
earnings news. Journal of Financial and Quan titativ e Analysis 53, 2431–2464.
Kacp erczyk, M., Sialm, C., Zheng, L., 2005. On the industry concen tration of activ ely
managed equit y m utual funds. Journal of Finance 60, 1983–2011.
41
Ka j üter, P ., Klassmann, F., Nienhaus, M., 2018. The Effect of Mandatory Quarterly
Rep orting on Firm V alue. The A ccoun ting Review 94, 251–277.
Kapadia, N., Zekhnini, M., 2019. Do idiosyncratic jumps matter? Journal of Financial
Economics 131, 666–692.
Kim, Y., Su, L. N., Zh u, X. K., 2017. Do es the cessation of quarterly earnings guidance
reduce in v estors’short-termism? Review of A ccoun ting Studies 22, 715–752.
K o e ster, A., Lundholm, R., Soliman, M., 2016. A ttracting atten tion in a limited
atten tion w orld: Exploring the causes and consequences of extreme p ositiv e earnings
surprises. Managemen t Science 62, 2871–2896.
K o thari, S. P ., Sh u, S., W yso c ki, P . D., 2009. Do managers withhold bad news? Journal
of A ccoun ting Researc h 47, 241–276.
K w ak, B., Ro, B., Suk, I., 2012. The comp osition of top managemen t with general counsel
and v olun tary information disclosure. Journal of A ccoun ting and Economics 54, 19–41.
Lam b ert, R., Leuz, C., V errecc hia, R. E., 2007. A ccoun ting information, disclosure, and
the cost of capital. Journal of A ccoun ting Researc h 45, 385–420.
Leha vy , R ., Sloan, R., 2008. In v estor recognition and sto c k returns. Review of
A ccoun ting Studies 13, 327–361.
Lenno x, C. S., P ark, C. W., 2006. The informativ eness of earnings and managemen t’s
issuance of earnings forecasts. Journal of A ccoun ting and Economics 42, 439–458.
Lerman, A., Livnat, J., 2010. The new form 8-k disclosures. Review of A ccoun ting
Studies 15, 752–778.
Marino vic, I., V aras, F., 2016. No news is go o d news: v olun tary disclosure in the face of
litigation. The RAND Journal of Economics 47, 822–856.
McMullin, J. L., Miller, B. P ., T w edt, B. J., 2019. Increased mandated disclosure
frequency and price formation: evidence from the 8-k expansion regulation. Review of
A ccoun ting Studies 24, 1–33.
Merton, R., 1987. A simple mo del of capital mark et equilibrium with incomplete
information. Journal of Finance 42, 483–510.
Mukherjee, A., P ana y oto v, G., Shon, J., 2021. Ey e in the sky: Priv ate satellites and
go v ernmen t macro data. Journal of Financial Economics 141, 234–254.
Niessner, M., 2015. Strategic disclosure timing and insider trading. W orking P ap er
A v ailable at SSRN: h ttps://ssrn.com/abstract=2439040 .
NIRI, 2016. Standards of practice for in v estor relations: Disclosure. A v ailable at
h ttps://www.niri.org/NIRI/media/NIRI/Professinal%20Dev elopmen t/Seminars/niri_-
standardspractice_up dated_lr.p df
.
42
Noh, S., So, E. C., V erdi, R. S., 2021. Calendar rotations: A new approac h for studying
the impact of timing using earnings announcemen ts. Journal of Financial Economics
140, 865–893.
Noh, S., So, E. C., W eb er, J. P ., 2019. V olun tary and mandatory disclosures: Do
managers view them as substitutes? Journal of A ccoun ting and Economics 68, 101243.
P a tell, J., W olfson, M., 1979. An ticipated information releases reflected in call option
prices. Journal of A ccoun ting and Economics 2, 117–140.
P a tell, J., W olfson, M., 1981. The ex an te and ex p ost price effects of quarterly earnings
announcemen ts reflected in option and sto c k prices. Journal of A ccoun ting Researc h 2,
434–458.
P a tton, A. J., V erardo, M., 2012. Do es Beta Mo v e with News? Firm-Sp ecific Information
Flo ws and Learning ab out Profitabilit y . The Review of Financial Studies 25,
2789–2839.
P enman, S., 1987. The distribution of earnings news o v er time and seasonalities in
aggregate sto c k returns. Journal of Financial Economics 18, 199–228.
Rogers, J. L., V an Buskirk, A., 2013. Bundled forecasts in empirical accoun ting researc h.
Journal of A ccoun ting and Economics 55, 43–65.
Sa v or, P ., Wilson, M., 2016. Earnings announcemen ts and systematic risk. Journal of
Finance 71, 83–138.
Shao, S., Stoum b os, R., Zhang, X. F., 2021. The p o w er of firm fundamen tal information
in explaining sto c k returns. Review of A ccoun ting Studies 26, 1249–1289.
Sharp e, W., 1964. Capital asset prices: A theory of mark et equilibrium under conditions
of risk. Journal of Finance 19, 425–442.
Skinner, D., Sloan, R., 2002. Earnings surprises, gro wth exp ectations, and sto c k returns
or don’t let an earnings torp edo sink y our p ortfolio. Review of A ccoun ting Studies 7,
289–312.
Skinner, D. J., 1994. Wh y firms v olun tarily disclose bad news. Journal of A ccoun ting
Researc h 32, 38–60.
Stoum b os, R. C., 2019. Ho w do es more frequen t rep orting reduce information
asymmetry? Colum bia Business Sc ho ol Researc h P ap er No. 17-75 A v ailable at SSRN:
h ttps://ssrn.com/abstract=3008708 .
V an B uskirk, A., 2012. Disclosure frequency and information asymmetry . Review of
Quan titativ e Finance and A ccoun ting 38, 411–440.
V errecc hia, R. E., 1983. Discretionary disclosure. Journal of A ccoun ting and Economics 5,
179 – 194.
43
Zhao, X., 2017. Do es information in tensit y matter for sto c k returns? evidence from form
8-k filings. Managemen t Science 63, 1382–1404.
Zhou, F. S., Zhou, Y., 2020. The dog that did not bark: Limited price efficiency and
strategic nondisclosure. Journal of A ccoun ting Researc h 58, 155–197.
Zh u, W., 2016. A ccruals and price crashes. Review of A ccoun ting Studies 21, 349–399.
44
App endix A
V ariable Definitions
V ariables Definition Source
Dep enden t V ariables
Jump
0|1
An indicator v ariable for sto c ks with a
jump in sto c k price on earnings announce-
men ts da y zero or da y one. F ollo wing
Kapadia and Zekhnini ( 2019 ), jump da ys
are da ys when the absolute idiosyncratic
sto c k return is greater than 3 conditional
standard deviations ( Equation 3.2 ).
CRSP
P ositiv e Jump
0|1
An indicator v ariable for an earnings an-
nouncemen t with a p ositiv e jump in sto c k
price on da y ze ro or da y one.
CRSP
Negativ e Jump
0|1
An indicator v ariable for an earnings
announcemen t with a negativ e jump in
sto c k price on d a y zero or da y one.
CRSP
𝑟 0,1
Earnings announcemen t da y zero and da y
one comp ounded abnormal sto c k return
p er the Carhart ( 1997 ) 4-factor mo del
( Equation 3.3 ).
CRSP
Indep enden t V ariables
Disclosure Distance
𝑞 𝐸 𝐴 𝑞 −𝑀 𝐴𝑋 (𝐸 𝐴 𝑞−1
,8 - 𝐾 𝑞−1,𝑞
)
𝐸 𝐴 𝑞 −𝐸 𝐴 𝑞−1
=𝐷𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝑞 =𝐷𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐸 𝐴 𝑞−1
SEC
EDGAR,
Com-
pustat,
I/B/E/S
𝑟 −1,−1
Earnings announcemen t da y -1 abnormal
sto c k return p er the Carhart ( 1997 ) 4-
factor mo del.
CRSP
45
𝑟 −20,−2
Earnings announcemen t da y [-20,-2] com-
p ounded abnormal sto c k return p er the
Carhart ( 1997 ) 4-factor mo del.
CRSP
ln(Mark et Equit y
𝑞 ) The natural logarithm of the mark et
v alue of equit y at the end of fiscal quarter
𝑞 .
CRSP
ln(Bo ok-to-mark et
𝑞 −1
) The natural logarithm of the b o ok-to-
mark et ratio, with the b o ok and mark et
v alue of equit y measured at the end of
fiscal quarter 𝑞 − 1 .
CRSP and
Compus-
tat
R O A
𝑞 −4,𝑞 −1
Op erating income after depreciation
(OIADPQ) summed o v er the trailing four
quarters, deflated b y 𝑞 − 1 total assets.
Compustat
R&D Exp ense
𝑞 −4,𝑞 −1
R&D exp ense (XRDQ) summed o v er the
trailing four quarters (with missing v al-
ues set to zero), deflated b y 𝑞 − 1 total
assets (A TQ).
Financial Lev erage
𝑞 −1
T otal debt (DL TTQ+DLCQ) at 𝑞 − 1 de-
flated b y 𝑞 − 1 total assets (A TQ) with
missing v alues set e qual to zero.
DSUE
𝑞 Decile rank ∈ [ 1, 10] of standardized un-
exp ected earnings based on median an-
alyst forecasts if a v ailable (
𝐸 𝑃 𝑆 𝑞 −𝐹 𝐸 𝑃 𝑆 𝑞 𝑃 −20
)
or based on a seasonal random w alk earn-
ings exp ectation mo del for firms not co v-
ered b y analysts (
𝐸 𝑃 𝑆 𝑃 𝑋 𝑄 𝑞 −𝐸 𝑃 𝑆 𝑃 𝑋 𝑄 𝑞−4
𝑃 𝑞 ).
Decile ranks are computed relativ e to
the distribution of eac h surprise measure
eac h quarter.
Compustat
Net External Financing
𝑞 −4, 𝑞 −1
Net external financing summed o v er
the trailing four quarters from Brad-
sha w et al. ( 2006 ) scaled b y quarter
𝑞 − 1 total assets. ∑
𝑞 −1
𝑞 −4
(SSTKY-
PRSTK CY-D VY+DL TISY-
DL TR Y+DLCCHY)/A TQ.
Compustat
Analyst Co v erage
𝑞 The n um b er of analysts pro viding at least
one forecast for quarter q EPS no older
than 90 da ys at the earnings announce-
men t date.
I/B/E/S
EPS Guidance
𝑞 𝑞 −1]
An indicator v ariable for a quarter q with
at least one managemen t forecast of EPS
issued no later than the 𝑞 − 1 earnings
announcemen t.
I/B/E/S
46
Idiosyncratic V olatilit y
−150,−1
The natural logarithm of p ercen t condi-
tional v olatilit y estimated via the form ula
in Equation 3.4 (ln( 𝜎 −150,−1
× 100 )).
CRSP
Beta
(𝑞 −1,𝑞 )
The slop e co efficien t from a regression of
in terim (trading da ys b et w een t w o adja-
cen t earnings announcemen ts) firm sp e-
cific sto c k returns on the CRSP v alue-
w eigh ted mark et factor.
CRSP
Sync hronicit y
(𝑞 −1,𝑞 )
ln(
𝑅 2
1−𝑅
2
) where the 𝑅 2
is from a regres-
sion of in terim firm sp ecific sto c k returns
on the CRSP v alue-w eigh ted mark et fac-
tor.
CRSP
ln(Illiquidit y
(𝑞 −1,𝑞 )
) Amih ud ( 2002 ) illiquidit y mo dified b y
Barardehi et al. ( 2021 ) to use op en-
to-close prices to compute returns,
measured o v er the in terim ( 𝑞 − 1, 𝑞 ).
( 𝑙𝑛(
1
𝑇 ∑
𝑇 𝑡 |𝑟 𝑒𝑡 𝑜𝑐 𝑡 |
$𝑇 𝑟 𝑎𝑑 𝑖𝑛𝑔 𝑉 𝑜𝑙𝑢𝑚𝑒 𝑡 ))
CRSP
ln(A vg. T rading V olume
(𝑞 −1,𝑞 )
) ln(
1
𝑇 ∑
𝑇 𝑡 $𝑇 𝑟 𝑎𝑑 𝑖𝑛𝑔 𝑉 𝑜 𝑙𝑢𝑚𝑒 𝑡 ), measured
o v er the in terim ( 𝑞 − 1, 𝑞 ).
CRSP
47
P a nel A: All 8-Ks
EA
q
EA
q-1
0 .1 .2 .3
Fraction of 8-Ks

-1 -.8 -.6 -.4 -.2 0

Interim Interval
P a nel B: Excluding Late-filed Earnings 8-Ks
EA
q
EA
q-1
0 .1 .2 .3
Fraction of 8-Ks

-1 -.8 -.6 -.4 -.2 0

Interim Interval
Figure 1: Disclosure Arriv al o v er the Quarter
This figure presen ts the distribution of 8-Ks filed o v er the in terv al (EA
𝑞−1
, EA
𝑞 ]. In terim In terv al =
−
𝐸𝐴 𝑞 −𝐷𝑎𝑡𝑒8𝐾
𝐸𝐴 𝑞 −𝐸𝐴
𝑞−1
, and bin widths are 0.05. P anel A includes all 8-Ks (excluding amended 8-Ks), and P anel
B excludes late-filed earnings 8-Ks for EA
𝑞−1
. T he sample spans Jan uary 2006–Decem b er 2019.
48
≥ 10
0 .1 .2 .3
Fraction of Firm-quarters

0 1 2 3 4 5 6 7 8 9
Number of Interim 8-Ks Filed
Figure 2: Num b er of In terim 8-Ks Filed Eac h Firm-Quarter
This figure presen ts the distribution of the n um b er of non-earnings 8-Ks filed eac h firm-quarter. The sample
spans Jan uary 2006–Decem b er 2019.
49
0 .05 .1 .15 .2
Fraction of Firm-quarters

0 .2 .4 .6 .8 1
Disclosure Distance
Figure 3: Disclosure Distance Distribution
This figure presen ts the distribution of disclosure distance calculated at eac h earnings announcemen t date.
𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 =
𝐸𝐴 𝑞 −𝑀 𝐴𝑋(𝐷𝑎𝑡𝑒8𝐾,𝐸𝐴
𝑞−1
)
𝐸𝐴 𝑞 −𝐸𝐴
𝑞−1
. A v alue of 1 represen ts a silen t firm in-b et w een
adjacen t earnings announcemen ts (a distan t disclosing firm), while a lo w v alue represen ts a firm that issues a
disclosure close to the curren t earnings announcemen t (a recen t disclosing firm). T he sample spans Jan uary
2006–Decem b e r 2019.
50
P a nel A: Median Disclosure Distance
.4 .42 .44 .46 .48 .5
Median Disclosure Distance

2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
Year
P a nel B: 25th P ercen tile, Median, and 75th P ercen tile Disclosure Distance
.2 .4 .6 .8 1
Disclosure Distance

2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
Year
P25 P50 P75
Figure 4: Disclosure Distance o v er Time
This figure presen ts distributional statistics for 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 o v er time. P anel A plots median
𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 . P anel B plots P25, P50 (median), and P75 distributional cutoffs. The sample
spans Jan uary 2006–Decem b er 2019.
51
0 .1 .2 .3 .4 .5
Cumulative Jump Rates (Fraction)

-5 -4 -3 -2 -1 0 1 2 3 4 5
Trading Days Relative to Earnings Announcement
EA Window Recent Mid Distant
Figure 5: Cum ulativ e Jump Rates around Earnings Announcemen ts
This figure presen ts cum ulativ e jump rates around earnings announcemen ts for p ortfolios sorted b y
𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 eac h quarter. Jump da ys are extreme idiosyncratic v olatilit y da ys where the abso-
lute idiosyncratic return is greater than 3 × the historic idiosyncratic v olatilit y . The sample spans Jan uary
2006–Decem b e r 2019.
52
-.2 -.1 0 .1 .2 .3
Avg. Cumulative Abnormal Returns (%)

-5 -4 -3 -2 -1 0 1 2 3 4 5
Trading Days Relative to Earnings Announcement
EA Window Recent Mid Distant
Figure 6: Cum ulativ e Abnormal Returns around Earnings Announcemen ts
This figure presen ts a v erage cum ulativ e abnormal returns (p er the Carhart ( 1997 ) mo del) around earnings
announcemen ts for p ortfolios sorted b y 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 eac h quarter. The sample spans Jan uary
2006–Decem b e r 2019.
53
0 .02 .04 .06

Difference in Cumulative Jump Rates
-.1 0 .1 .2 .3 .4
Difference in Cumulative Abnormal Returns (%)

-5 -4 -3 -2 -1 0 1 2 3 4 5
Trading Days Relative to Earnings Announcement
EA Window Returns Jumps
Figure 7: Distan t Min us Recen t Hedge P ortfolio Returns and Difference in Jump Rates around
Earnings Announcemen ts
This figure plots the difference in returns and jump rates b et w een distan t disclosing sto c ks and recen t
disclosing sto c ks around earnings announcemen ts. The left y-axis scales the dashed line plotting the difference
in cum ulativ e abnormal returns (p er the Carhart ( 1997 ) mo del), and the righ t axis scales the dotted line
plotting the difference in cum ulativ e jump rates. The sample spans Jan uary 2006–Decem b er 2019.
54
P a nel A: Go o d News
0 1 2 3 4
Avg. Cumulative Abnormal Returns (%)

-5 -4 -3 -2 -1 0 1 2 3 4 5
Trading Days Relative to Earnings Announcement
EA Window Recent Mid Distant
P a nel B: Bad News
-3 -2 -1 0
Avg. Cumulative Abnormal Returns (%)

-5 -4 -3 -2 -1 0 1 2 3 4 5
Trading Days Relative to Earnings Announcement
EA Window Recent Mid Distant
Figure 8: P ortfolio Returns Conditional on News
This figure plots cum ulativ e abnormal returns for 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 sorted p ortfolios conditional on
earnings news. P anel A is the go o d news subsample (DSUE >= 6 ) and P anel B is the bad news subsample
(DSUE <= 5 ). The y-axis scales cum ulativ e abnormal returns (p er the Carhart ( 1997 ) mo del). The sample
spans Jan uary 2006–Decem b er 2019.
55
-1 -.5 0 .5 1
Avg. Cumulative Abnormal Returns (%)

-5 -4 -3 -2 -1 0 1 2 3 4 5
Trading Days Relative to Earnings Announcement
EA Window Good News Hedge Bad News Hedge
Figure 9: Hedge P ortfolio Returns Conditional on News
This figure plots cum ulativ e abnormal returns for 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 hedge p ortfolios conditional on
earnings news. The dashed line is the go o d news subsample (DSUE >= 6 ) and the dotted line is the bad news
subsample (DSUE <= 5 ). The y-axis scales cum ulativ e abnormal returns (p er the Carhart ( 1997 ) mo del).
The sample spans Jan uary 2006–Decem b er 2019.
56
T able 1: Descriptiv e Statistics
This table presen ts the time series a v erage of descriptiv e statistics ( F ama and MacBeth , 1973 ). P anel A
is the time series a v erage of quarterly distributional statistics. P anel B presen ts the time series a v erage of
quarterly correlations. Con tin uous indep enden t v ariables are winsorized at the 1% tails of the distribution
eac h quarter. See App endix A for v ariable definitions. The sample spans Jan uary 2006–Decem b er 2019.
P anel A: Distributions
V ariable Mean SD P1 P25 P50 P75 P 99
Disclosure Distance
𝑞 0.514 0.346 0.015 0.195 0.467 0.885 1.000
ln(Bo ok-to-Mark et
𝑞 −1
) -0.770 0.811 -3.324 -1.241 -0.677 -0.210 0.984
ln(Mark et Equit y
𝑞 ) 6.719 1.995 2.343 5.307 6.727 8.062 11.595
R O A
𝑞 −4,𝑞 −1
0.031 0.166 -0.806 0.012 0.054 0.106 0.340
R&D Exp ense
𝑞 −4,𝑞 −1
0.044 0.096 0.000 0.000 0.000 0.041 0.549
Financial Lev erage
𝑞 −1
0.193 0.181 0.000 0.028 0.154 0.308 0.722
Net External Financing
𝑞 −4,𝑞 −1
0.039 0.358 -0.743 -0.109 -0.014 0.079 1.791
DSUE
𝑞 5.503 2.967 1.000 3.000 5.464 8.000 10.000
Beta
(𝑞 −1,𝑞 )
1.043 0.634 -0.544 0.643 1.039 1.431 2.766
Sync hronicit y
(𝑞 −1 ,𝑞 )
-1.812 1.919 -9.088 -2.546 -1.276 -0.519 0.746
ln(Illiquidit y
(𝑞 −1 ,𝑞 )
) -19.269 3.039 -25.009 -21.556 -19.675 -17.166 -12.134
ln(A vg. T rading V olume
(𝑞 −1,𝑞 )
) 15.331 2.610 9.465 13.503 15.650 17.291 20.334
ln(Idiosyncratic V olatilit y
−120,−1
) 0.563 0.542 -0.512 0.169 0.522 0.911 2.046
EPS Guidance
𝑞 𝑞 −1]
0.206 0.395 0.000 0.000 0.000 0.250 1.000
Analyst Co v erage
𝑞 2.814 4.107 0.000 0.000 1.214 3.750 21.125
𝑟 −20,−2
-0.020 10.719 -26.264 -4.929 -0.269 4.302 31.502
𝑟 −1,−1
0.083 3.103 -7.662 -1.136 0.024 1.195 8.686
𝑟 0,1
-0.010 8.934 -24.665 -3.920 -0.039 3.802 25.017
Jump
0|1
0.377 0.480 0.000 0.000 0.000 0.982 1.000
P ositiv e Jump
0|1
0.206 0.402 0.000 0.000 0.000 0.125 1.000
Negativ e Jump
0|1
0.154 0.358 0.000 0.000 0.000 0.000 1.000
57
T able 1 (con tin ued)
P anel B: Correlations (P earson c orrelations are ab o v e the diagonal of ones; Sp earman correlations are b elo w the diagonal of
ones)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
1 Disclosure Distance
𝑞 1.00 0.01 -0.09 0.06 -0.02 -0.09 -0.06 0.02 -0.04 -0.01 0.10 -0.12 0.01 -0.12 -0.05 0.00 0.00 0.01 0.06 0.04 0.02
2 ln(Bo ok-to-Mark et
𝑞 −1
) 0.00 1.00 -0.36 -0.06 -0.29 -0.10 -0.08 -0.01 -0.14 -0.20 0.37 -0.38 -0.16 -0.13 0.10 0.01 0.01 0.00 -0.11 -0.07 -0.07
3 ln(Mark et Equit y
𝑞 ) -0.08 -0.38 1.00 0.38 -0.17 0.21 -0.18 0.00 0.24 0.59 -0.96 0.94 0.24 0.57 -0.59 -0.02 -0.00 0.01 0.16 0.11 0.12
4 R O A
𝑞 −4,𝑞 −1
0.05 -0.37 0.47 1.00 -0.60 0.09 -0.56 -0.05 -0.01 0.26 -0.37 0.32 0.16 0.17 -0.42 0.01 0.03 0.02 0.14 0.09 0.09
5 R&D Exp ense
𝑞 −4,𝑞 −1
0.03 -0.34 -0.06 -0.13 1.00 -0.21 0.39 0.06 0.07 -0.13 0.14 -0.09 0.04 -0.11 0.32 0.00 -0.02 -0.01 -0.03 -0.02 -0.01
6 Financial Lev erage
𝑞 −1
-0.09 -0.03 0.26 0.12 -0.24 1.00 0.07 -0.01 0.08 0.11 -0.22 0.22 -0.01 0.13 -0.06 -0.00 -0.00 0.00 0.02 0.02 0.01
7 Net External Financing
𝑞 −4,𝑞 −1
-0.05 0.05 -0.19 -0.40 0.12 0.09 1.00 0.01 0.08 -0.12 0.15 -0.11 -0.09 -0.09 0.29 -0.01 -0.03 -0.02 -0.07 -0.05 -0.04
8 DSUE
𝑞 0.02 -0.01 0.00 -0.02 0.05 -0.02 -0.00 1.00 0.01 0.01 -0.00 0.01 0.04 -0.01 0.01 0.05 0.06 0.26 0.02 0.19 -0.18
9 Beta
(𝑞 −1,𝑞 )
-0.03 -0.14 0.23 0.03 0.10 0.06 0.07 0.02 1.00 0.64 -0.30 0.35 0.08 0.14 0.08 -0.01 -0.04 -0.01 0.11 0.07 0.07
10 Sync hronicit y
(𝑞 − 1,𝑞 )
0.00 -0.19 0.61 0.31 -0.07 0.15 -0.15 0.01 0.62 1.00 -0.61 0.58 0.16 0.26 -0.49 -0.01 -0.03 0.01 0.19 0.13 0.12
11 ln(Illiquidit y
(𝑞 − 1,𝑞 )
) 0.09 0.39 -0.97 -0.47 0.02 -0.26 0.18 -0.00 -0.23 -0.57 1.00 -0.99 -0.26 -0.56 0.55 0.02 -0.00 -0.00 -0.20 -0.13 -0.14
12 ln(A vg. T rading V olume
(𝑞 −1,𝑞 )
) -0.11 -0.40 0.95 0.43 0.01 0.26 -0.15 0.01 0.29 0.54 -0.99 1.00 0.26 0.57 -0.44 -0.01 0.01 -0.00 0.19 0.12 0.14
13 ln(Idiosyncratic V olatilit y
−120,−1
) 0.01 -0.18 0.24 0.20 0.17 -0.01 -0.08 0.04 0.08 0.16 -0.26 0.26 1.00 0.09 -0.13 0.01 0.00 0.01 0.13 0.09 0.08
14 EPS Guidance
𝑞 𝑞 −1]
-0.11 -0.21 0.65 0.27 -0.08 0.20 -0.09 -0.01 0.20 0.37 -0.68 0.68 0.15 1.00 -0.21 -0.01 -0.01 -0.00 0.04 0.02 0.05
15 Analyst Co v erage
𝑞 -0.04 0.10 -0.59 -0.36 0.21 -0.12 0.26 0.01 0.10 -0.56 0.54 -0.43 -0.12 -0.29 1.00 0.03 0.01 -0.02 -0.19 -0.13 -0.10
16 𝑟 −20,−2
0.01 -0.00 0.01 0.03 0.00 -0.00 -0.02 0.05 0.00 0.01 -0.01 0.01 0.01 0.00 -0.00 1.00 -0.03 -0.04 -0.01 -0.02 0.01
17 𝑟 −1,−1
0.01 -0.01 0.04 0.06 -0.02 0.01 -0.04 0.07 -0.04 0.01 -0.04 0.04 0.01 0.01 -0.05 -0.03 1.00 -0.01 -0.02 -0.02 -0.01
18 𝑟 0,1
0.01 -0.00 0.02 0.02 -0.01 0.01 -0.01 0.27 -0.00 0.02 -0.02 0.01 0.02 0.01 -0.03 -0.06 -0.00 1.00 0.03 0.51 -0.57
19 Jump
0|1
0.06 -0.12 0.17 0.17 0.07 0.02 -0.06 0.02 0.11 0.19 -0.19 0.18 0.13 0.10 -0.18 0.00 -0.01 0.04 1.00 0.65 0.55
20 P ositiv e Jump
0|1
0.04 -0.07 0.11 0.11 0.04 0.02 -0.04 0.19 0.07 0.12 -0.12 0.11 0.09 0.06 -0.12 -0.02 -0.01 0.54 0.65 1.00 -0.16
21 Negativ e Jump
−
0|1
0.03 -0.08 0.12 0.11 0.04 0.01 -0.04 -0.18 0.07 0.12 -0.14 0.13 0.08 0.09 -0.10 0.03 0.00 -0.59 0.55 -0.16 1.00
58
T able 2: Determinan ts of Disclosure Distance
This table presen ts T obit regressions of Disclosure Distance
𝑞 on determinan ts. t -v alues are in brac k ets based
on standard errors clustered b y sto c k (p < .10*, p < .05**, p < .01***). See App endix A for v ariable definitions.
The sample spans Jan uary 2006–Decem b er 2019.
Disclosure Distance
𝑞 (1) (2) (3) (4) (5)
Disclosure Distance
𝑞 −4
0.333
∗∗∗
0.306
∗∗∗
0.307
∗∗∗
[56.79] [53.69] [53.78]
ln(Mark et Equit y
𝑞 ) -0.017
∗∗∗
-0.018
∗∗∗
-0.012
∗∗∗
-0.012
∗∗∗
[-10.11] [-10.19] [-9.06] [-9.18]
Analyst Co v erage
𝑞 -0.008
∗∗∗
-0.008
∗∗∗
-0.007
∗∗∗
-0.007
∗∗∗
[-12.21] [-12.17] [-13.22] [-13.16]
EPS Guidance
𝑞 𝑞 −1]
0.002 0.001 0.003 0.001
[0.41] [0.24] [0.57] [0.31]
ln(Bo ok-to-mark et
𝑞 −1
) -0.014
∗∗∗
-0.014
∗∗∗
-0.008
∗∗∗
-0.008
∗∗∗
[-4.22] [-4.26] [-3.26] [-3.32]
R O A
𝑞 −4, 𝑞 −1
0.266
∗∗∗
0.270
∗∗∗
0.210
∗∗∗
0.214
∗∗∗
[14.34] [14.53] [14.05] [14.32]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.081
∗∗∗
0.077
∗∗
0.056
∗∗
0.051
∗∗
[2.64] [2.52] [2.33] [2.15]
Financial Lev erage
𝑞 −1
-0.152
∗∗∗
-0.152
∗∗∗
-0.109
∗∗∗
-0.109
∗∗∗
[-10.68] [-10.68] [-9.98] [-9.97]
Net External Financing
𝑞 −4, 𝑞 −1
-0.022
∗∗∗
-0.021
∗∗∗
-0.006 -0.005
[-4.23] [-4.05] [-1.45] [-1.21]
DSUE
𝑞 0.003
∗∗∗
0.004
∗∗∗
[8.16] [9.87]
Pseudo 𝑅 2
0.023 0.024 0.060 0.073 0.074
N 144,847 144,847 144,847 144,847 144,847
59
T able 3: Earnings Announcemen t Jump Risk
This table presen ts results for earnings announcemen t jump risk. P anel A presen ts the prop ortion
of eac h p ortfolio that jumps on earnings announcemen t da y zero or da y one for p ortfolios sorted b y
𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 eac h quarter. t -v alues are in brac k ets and standard errors are clustered b y sto c k
and trading date in P anel A. P anel B (C) [D] presen ts logit regressions of indicators for a da y zero or da y
one idiosyncratic (p ositiv e) [negativ e] jump on explanatory v ariables. In P anels B-D, z -v alues are in brac k ets
based on standard errors clustered b y sto c k and trading date (p < .10*, p < .05**, p < .01***). See App endix A
for v ariable definitions. The sample spans Jan uary 2006–Decem b er 2019.
P anel A: Disclosure Distance Sorted P ortfolios
Disclosure Distance
Recen t Mid Distan t Distan t-Recen t
Jump
0|1
0.334 0.385 0.401 0.068***
[82.559] [90.476] [82.939] [15.440]
P ositiv e Jump
0|1
0.178 0.211 0.222 0.044***
[69.244] [74.232] [70.772] [14.345]
Negativ e Jump
0|1
0.140 0.156 0.160 0.021***
[59.751] [66.351] [61.729] [7.554]
𝑁 48,536 48,166 48,145
60
T able 3 (con tin ued)
P anel B: Earnings Announcemen t Jum p Indicator Logit Regressions
Jump
0|1
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 0.343
∗∗∗
0.343
∗∗∗
0.352
∗∗∗
0.383
∗∗∗
0.338
∗∗∗
[15.09] [15.09] [16.59] [18.45] [16.48]
Disclosure Distance
𝑞 −4
0.083
∗∗∗
-0.012 0.021 -0.018
[3.82] [-0.59] [1.12] [-0.98]
ln(Bo ok-to-mark et
𝑞 −1
) -0.086
∗∗∗
-0.022
∗
[-6.41] [-1.72]
R O A
𝑞 −4, 𝑞 −1
1.324
∗∗∗
0.789
∗∗∗
[15.21] [9.45]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.598
∗∗∗
-0.405
∗∗∗
[4.19] [-2.81]
Financial Lev erage
𝑞 −1
-0.068 0.044
[-1.25] [0.83]
Net External Financing
𝑞 −4, 𝑞 −1
-0.051
∗∗
-0.012
[-2.10] [-0.49]
ln(Mark et Equit y
𝑞 ) -0.447
∗∗∗
-0.386
∗∗∗
[-25.02] [-22.48]
Analyst Co v erage
𝑞 -0.027
∗∗∗
-0.019
∗∗∗
[-11.26] [-7.65]
EPS Guidance
𝑞 𝑞 −1]
0.301
∗∗∗
0.153
∗∗∗
[13.47] [6.88]
Beta
(𝑞 −1, 𝑞 )
-0.031
∗
-0.043
∗∗
[-1.78] [-2.49]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.146
∗∗∗
0.143
∗∗∗
[18.67] [19.12]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.394
∗∗∗
-0.489
∗∗∗
[-17.18] [-22.03]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.024 -0.166
∗∗∗
[-1.03] [-7.26]
𝑟 −1,−1
-0.010
∗∗∗
-0.011
∗∗∗
[-4.95] [-5.47]
𝑟 −20,−2
-0.006
∗∗∗
-0.006
∗∗∗
[-9.41] [-9.22]
Pseudo 𝑅 2
0.003 0.003 0.013 0.016 0.078 0.092
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No Y es
N 144,847 144,847 144,847 144,847 144,847 144,847
61
T able 3 (con tin ued)
P anel C: Earnings Announcemen t P ositiv e Jump Indicator Logit Regressions
P ositiv e Jump
0|1
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 0.307
∗∗∗
0.307
∗∗∗
0.317
∗∗∗
0.313
∗∗∗
0.277
∗∗∗
[13.64] [13.64] [14.23] [14.14] [12.63]
Disclosure Distance
𝑞 −4
0.071
∗∗∗
-0.015 0.000 -0.032
[3.17] [-0.68] [0.00] [-1.50]
ln(Bo ok-to-mark et
𝑞 −1
) -0.055
∗∗∗
-0.010
[-4.37] [-0.79]
R O A
𝑞 −4, 𝑞 −1
1.008
∗∗∗
0.597
∗∗∗
[11.57] [6.86]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.505
∗∗∗
-0.262
∗
[3.73] [-1.79]
Financial Lev erage
𝑞 −1
0.023 0.123
∗∗
[0.43] [2.35]
Net External Financing
𝑞 −4, 𝑞 −1
-0.070
∗∗∗
-0.044
∗
[-2.65] [-1.65]
ln(Mark et Equit y
𝑞 ) -0.329
∗∗∗
-0.287
∗∗∗
[-19.69] [-17.47]
Analyst Co v erage
𝑞 -0.023
∗∗∗
-0.016
∗∗∗
[-9.41] [-6.40]
EPS Guidance
𝑞 𝑞 −1]
0.253
∗∗∗
0.150
∗∗∗
[11.91] [7.05]
Beta
(𝑞 −1, 𝑞 )
-0.019 -0.029
[-0.96] [-1.45]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.125
∗∗∗
0.120
∗∗∗
[14.55] [14.17]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.359
∗∗∗
-0.432
∗∗∗
[-15.43] [-18.78]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.099
∗∗∗
-0.203
∗∗∗
[-4.11] [-8.34]
𝑟 −1,−1
-0.026
∗∗∗
-0.027
∗∗∗
[-10.35] [-10.61]
𝑟 −20,−2
-0.006
∗∗∗
-0.006
∗∗∗
[-7.85] [-7.73]
Pseudo 𝑅 2
0.002 0.002 0.009 0.010 0.046 0.055
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No No
N 144,847 144,847 144,847 144,847 144,847 144,847
62
T able 3 (con tin ued)
P anel D: Earnings Announcemen t Negat iv e Jump Indicator Logit Regressions
Negativ e Jump
0|1
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 0.205
∗∗∗
0.205
∗∗∗
0.202
∗∗∗
0.237
∗∗∗
0.201
∗∗∗
[8.23] [8.23] [8.23] [9.69] [8.29]
Disclosure Distance
𝑞 −4
0.053
∗∗
-0.002 0.046
∗∗
0.017
[2.20] [-0.07] [2.01] [0.76]
ln(Bo ok-to-mark et
𝑞 −1
) -0.066
∗∗∗
-0.016
[-4.48] [-1.06]
R O A
𝑞 −4, 𝑞 −1
1.108
∗∗∗
0.722
∗∗∗
[10.82] [6.94]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.496
∗∗∗
-0.255
[3.02] [-1.44]
Financial Lev erage
𝑞 −1
-0.164
∗∗∗
-0.116
∗
[-2.78] [-1.95]
Net External Financing
𝑞 −4, 𝑞 −1
0.033 0.062
∗∗
[1.09] [2.07]
ln(Mark et Equit y
𝑞 ) -0.435
∗∗∗
-0.385
∗∗∗
[-22.58] [-20.39]
Analyst Co v erage
𝑞 -0.013
∗∗∗
-0.006
∗∗
[-4.56] [-2.12]
EPS Guidance
𝑞 𝑞 −1]
0.193
∗∗∗
0.082
∗∗∗
[8.32] [3.48]
Beta
(𝑞 −1, 𝑞 )
-0.024 -0.025
[-1.14] [-1.16]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.120
∗∗∗
0.118
∗∗∗
[12.19] [12.09]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.347
∗∗∗
-0.412
∗∗∗
[-13.70] [-16.32]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
0.044
∗
-0.063
∗∗
[1.71] [-2.39]
𝑟 −1,−1
0.013
∗∗∗
0.013
∗∗∗
[5.60] [5.51]
𝑟 −20,−2
-0.003
∗∗∗
-0.003
∗∗∗
[-3.79] [-3.53]
Pseudo 𝑅 2
0.001 0.001 0.009 0.010 0.053 0.060
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No No
N 144,847 144,847 144,847 144,847 144,847 144,847
63
T able 4: Earnings Announcemen t Sto c k Returns
This table presen ts results for earnings announcemen t abnormal sto c k returns comp ounded o v er ev en t da ys
zero to one ( 𝑟 0,1
). P anel A presen ts a v erage Carhart ( 1997 ) four-factor alphas for p ortfolios sorted b y
𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 𝑞 eac h quarter. P anel B presen ts panel regressions of Carhart ( 1997 ) four-factor
alphas on explanatory v ariables. See App endix A for v ariable definitions. t -v alues are in brac k ets based
on standard errors clustered b y sto c k and trading date (p < .10*, p < .05**, p < .01***). The sample spans
Jan u ary 2006–Decem b er 2019.
P anel A: Disclosure Distance Sorted P ortfolios
Disclosure Distance
Recen t Mid Distan t Distan t-Recen t
𝑟 0,1
-0.180 -0.006 0.111 0.291***
[-4.021] [-0.129] [2.425] [4.943]
𝑁 48,536 48,166 48,145
64
T able 4 (con tin ued)
P anel B: Earnings Announcemen t Sto c k Return Regressions
𝑟 0,1
(1) (2) (3) (4) (5) (6) (7)
Disclosure Distanc e
𝑞 0.329
∗∗∗
0.338
∗∗∗
0.355
∗∗∗
0.254
∗∗∗
0.251
∗∗∗
0.053
[4.74] [4.86] [4.95] [3.55] [3.49] [0.76]
Disclosure Distance
𝑞 −4
0.032 -0.063 -0.131
∗
-0.133
∗
0.020
[0.46] [-0.88] [-1.82] [-1.85] [0.28]
ln(Bo ok -to-mark et
𝑞 −1
) 0.088
∗∗
0.101
∗∗
0.063
[2.00] [2.28] [1.46]
R O A
𝑞 − 4, 𝑞 −1
1.226
∗∗∗
1.186
∗∗∗
2.292
∗∗∗
[4.01] [3.76] [7.39]
R&D Exp ense
𝑞 −4, 𝑞 −1
1.141
∗∗
1.122
∗∗
0.362
[2.26] [2.07] [0.67]
Financial Lev erage
𝑞 −1
0.394
∗∗
0.479
∗∗∗
0.455
∗∗∗
[2.24] [2.62] [2.59]
Net External Financing
𝑞 −4, 𝑞 −1
-0.312
∗∗∗
-0.334
∗∗∗
-0.114
[-3.07] [-3.27] [-1.13]
ln(Mark et Equit y
𝑞 ) 0.226
∗∗∗
0.234
∗∗∗
0.210
∗∗∗
[4.26] [4.33] [3.98]
Analyst Co v erage
𝑞 0.002 0.001 0.006
[0.31] [0.18] [0.85]
EPS Guidance
𝑞 𝑞 −1]
0.204
∗∗∗
0.219
∗∗∗
0.000
[2.95] [3.09] [0.00]
Beta
(𝑞 −1, 𝑞 )
0.061 0.051 0.067
[0.77] [0.64] [0.86]
Sync h ronicit y
(𝑞 −1, 𝑞 )
-0.005 -0.008 -0.023
[-0.19] [-0.29] [-0.93]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.170
∗
-0.179
∗∗
-0.229
∗∗∗
[-1.96] [-2.06] [-2.65]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.388
∗∗∗
-0.399
∗∗∗
-0.456
∗∗∗
[-4.05] [-4.12] [-4.76]
𝑟 −1,−1
-0.119
∗∗∗
-0.119
∗∗∗
-0.151
∗∗∗
[-6.24] [-6.24] [-7.56]
𝑟 −20,−2
-0.009
∗∗
-0.009
∗
-0.022
∗∗∗
[-1.97] [-1.96] [-4.87]
𝑟 0,1 𝑞 −4
-0.006
∗
-0.007
∗
-0.002
[-1.84] [-1.85] [-0.65]
ln(Idiosyncratic V olatilit y
−150, −1
) -0.082 -0.087 -0.023
[-0.91] [-0.94] [-0.27]
DSUE
𝑞 0.817
∗∗∗
[65.04]
A djusted 𝑅 2
0.000 0.001 0.001 0.001 0.005 0.005 0.074
Quarter FE No Y es Y es Y es Y es Y es Y es
Industry FE No No No No No Y es Y es
N 144,847 144,847 144,847 144,847 144,847 144,847 144,847
65
T able 5: In v estor Recognition and Earnings Announcemen t Sto c k Returns
This table presen ts results for earnings announcemen t abnormal sto c k returns comp ounded o v er ev en t da ys
zero to one ( 𝑟 0,1
). P anel A presen ts panel regressions of Carhart ( 1997 ) four-factor alphas on explanatory
v ariables for sto c ks with lo w Δ in v estor recognition and P anel B presen ts results for sto c ks with high Δ in v estor
recognition. Δ In v estor recognition is based on a cross-sectional median partition on the quarterly c hange
in the n um b er of 13-F filing institutions holding the sto c k with o wnership data lagged 2 mon ths. See
App e ndix A for v ariable definitions. t -v alues are in brac k ets based on standard errors clustered b y sto c k
and trading date (p < .10*, p < .05**, p < .01***). The sample spans earnings announcemen ts o v er Jan uary
2006–Decem b er 2019.
66
T able 5 (con tin ued)
P anel A: Lo w Δ In v estor Recognition Sample
𝑟 0,1
(1) (2) (3) (4) (5) (6) (7)
Disclosure Distance
𝑞 0.504
∗∗∗
0.510
∗∗∗
0.498
∗∗∗
0.384
∗∗∗
0.379
∗∗∗
0.142
[4.98] [5.03] [4.71] [3.63] [3.57] [1.39]
Disclosure Distance
𝑞 −4
0.179
∗
0.045 -0.031 -0.036 0.155
[1.82] [0.44] [-0.31] [-0.35] [1.56]
ln(Bo ok-to-mark et
𝑞 −1
) 0.153
∗∗
0.167
∗∗∗
0.121
∗∗
[2.52] [2.75] [2.03]
R O A
𝑞 −4, 𝑞 −1
1.572
∗∗∗
1.540
∗∗∗
2.974
∗∗∗
[4.05] [3.84] [7.61]
R&D Exp ense
𝑞 −4, 𝑞 −1
1.689
∗∗∗
1.651
∗∗
0.802
[2.70] [2.42] [1.17]
Financial Lev erage
𝑞 −1
0.496
∗∗
0.572
∗∗
0.538
∗∗
[2.08] [2.31] [2.25]
Net External Financing
𝑞 −4, 𝑞 −1
-0.341
∗∗∗
-0.364
∗∗∗
-0.169
[-2.63] [-2.79] [-1.31]
ln(Mark et Equit y
𝑞 ) 0.311
∗∗∗
0.310
∗∗∗
0.232
∗∗∗
[4.29] [4.21] [3.25]
Analyst Co v erage
𝑞 0.002 0.004 0.004
[0.20] [0.40] [0.38]
EPS Guidance
𝑞 𝑞 −1]
0.102 0.126 -0.156
[1.07] [1.28] [-1.62]
Beta
(𝑞 −1, 𝑞 )
0.002 -0.004 0.100
[0.02] [-0.04] [1.00]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.001 0.000 -0.034
[0.02] [0.00] [-1.05]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.230
∗∗
-0.235
∗∗
-0.330
∗∗∗
[-2.31] [-2.36] [-3.40]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.480
∗∗∗
-0.482
∗∗∗
-0.545
∗∗∗
[-4.35] [-4.31] [-4.99]
𝑟 −1,−1
-0.120
∗∗∗
-0.120
∗∗∗
-0.151
∗∗∗
[-6.78] [-6.76] [-8.59]
𝑟 −20,−2
-0.011
∗∗
-0.011
∗∗
-0.024
∗∗∗
[-2.06] [-2.07] [-4.74]
𝑟 0,1 𝑞 −4
-0.009
∗∗
-0.009
∗∗
-0.003
[-2.00] [-2.00] [-0.74]
ln(Idiosyncratic V olatilit y
−150, −1
) 0.015 0.012 0.137
[0.12] [0.10] [1.15]
DSUE
𝑞 0.812
∗∗∗
[52.95]
A djusted 𝑅 2
0.000 0.001 0.001 0.001 0.006 0.006 0.075
Quarter FE No Y es Y es Y es Y es Y es Y es
Industry FE No No No No No Y es Y es
N 76,626 76,626 76,626 76,626 76,626 76,626 76,626
67
T able 5 (con tin ued)
P anel B: High Δ In v estor Recognition Sam ple
𝑟 0,1
(1) (2) (3) (4) (5) (6) (7)
Disclosure Distance
𝑞 0.134 0.145 0.195
∗∗
0.107 0.103 -0.049
[1.43] [1.55] [2.01] [1.09] [1.05] [-0.52]
Disclosure Distance
𝑞 −4
-0.135 -0.186
∗
-0.247
∗∗
-0.248
∗∗
-0.142
[-1.42] [-1.90] [-2.48] [-2.49] [-1.48]
ln(Bo ok-to-mark et
𝑞 −1
) 0.028 0.043 -0.031
[0.45] [0.68] [-0.50]
R O A
𝑞 −4, 𝑞 −1
0.584 0.494 1.230
∗∗∗
[1.37] [1.13] [2.83]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.497 0.427 -0.283
[0.70] [0.55] [-0.37]
Financial Lev erage
𝑞 −1
0.336 0.430
∗
0.329
[1.47] [1.80] [1.41]
Net External Financing
𝑞 −4, 𝑞 −1
-0.342
∗∗
-0.359
∗∗
-0.081
[-2.28] [-2.36] [-0.54]
ln(Mark et Equit y
𝑞 ) 0.174
∗∗
0.192
∗∗∗
0.224
∗∗∗
[2.45] [2.64] [3.14]
Analyst Co v erage
𝑞 0.005 0.002 0.010
[0.55] [0.22] [1.05]
EPS Guidance
𝑞 𝑞 −1]
0.323
∗∗∗
0.320
∗∗∗
0.156
∗
[3.52] [3.36] [1.70]
Beta
(𝑞 −1, 𝑞 )
0.104 0.093 -0.013
[0.91] [0.80] [-0.11]
Sync hronicit y
(𝑞 −1, 𝑞 )
-0.015 -0.021 -0.006
[-0.37] [-0.52] [-0.16]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.007 -0.024 -0.013
[-0.04] [-0.16] [-0.09]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.208 -0.235 -0.266
[-1.27] [-1.42] [-1.61]
𝑟 −1,−1
-0.118
∗∗∗
-0.118
∗∗∗
-0.151
∗∗∗
[-2.92] [-2.93] [-3.43]
𝑟 −20,−2
-0.008 -0.008 -0.020
∗∗∗
[-0.99] [-0.98] [-2.68]
𝑟 0,1 𝑞 −4
-0.004 -0.004 -0.001
[-0.78] [-0.78] [-0.20]
ln(Idiosyncratic V olatilit y
−150, −1
) -0.239
∗
-0.249
∗
-0.256
∗∗
[-1.80] [-1.87] [-1.98]
DSUE
𝑞 0.827
∗∗∗
[52.23]
A djusted 𝑅 2
0.000 0.001 0.001 0.001 0.003 0.004 0.073
Quarter FE No Y es Y es Y es Y es Y es Y es
Industry FE No No No No No Y es Y es
N 68,221 68,221 68,221 68,221 68,221 68,221 68,221
68
T able 6: In v estor Recognition and Earnings Announcemen t Jump Risk
This table presen ts results for earnings announcemen t jump risk. P anel A presen ts logit regressions of jump
indicators on explanatory v ariables for sto c ks with lo w Δ in v estor recognition, and P anel B presen ts results
for sto c ks with high Δ in v estor recognition. Δ In v estor recognition is based on a cross-sectional median
partition on the quarterly c hange in the n um b er of 13-F filing institutions holding the sto c k with o wnership
data lagged 2 mon ths. z -v alues are in brac k ets based on standard errors clustered b y sto c k and trading
date (p < .10*, p < .05**, p < .01***). See App endix A for v ariable definitions. The sample spans Jan uary
2006–Decem b e r 2019.
69
T able 6 (con tin ued)
P anel A: Lo w Δ In v estor Recognition Sample
Jump
0|1
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 0.312
∗∗∗
0.312
∗∗∗
0.334
∗∗∗
0.371
∗∗∗
0.326
∗∗∗
[11.23] [11.23] [12.50] [13.60] [11.97]
Disclosure Distance
𝑞 −4
0.050
∗
-0.040 0.002 -0.029
[1.88] [-1.57] [0.09] [-1.17]
ln(Bo ok-to-mark et
𝑞 −1
) -0.084
∗∗∗
-0.019
[-5.45] [-1.24]
R O A
𝑞 −4, 𝑞 −1
1.371
∗∗∗
0.861
∗∗∗
[13.27] [8.65]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.656
∗∗∗
-0.240
[4.01] [-1.44]
Financial Lev erage
𝑞 −1
-0.025 0.086
[-0.40] [1.38]
Net External Financing
𝑞 −4, 𝑞 −1
-0.050 -0.015
[-1.62] [-0.50]
ln(Mark et Equit y
𝑞 ) -0.443
∗∗∗
-0.371
∗∗∗
[-21.22] [-18.07]
Analyst Co v erage
𝑞 -0.029
∗∗∗
-0.020
∗∗∗
[-10.10] [-6.83]
EPS Guidance
𝑞 𝑞 −1]
0.307
∗∗∗
0.164
∗∗∗
[11.40] [6.10]
Beta
(𝑞 −1 , 𝑞 )
-0.030 -0.041
∗
[-1.39] [-1.90]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.134
∗∗∗
0.134
∗∗∗
[15.38] [15.69]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.384
∗∗∗
-0.461
∗∗∗
[-15.03] [-18.07]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.010 -0.141
∗∗∗
[-0.37] [-5.31]
𝑟 −1,−1
-0.009
∗∗∗
-0.009
∗∗∗
[-3.39] [-3.71]
𝑟 −20,−2
-0.006
∗∗∗
-0.006
∗∗∗
[-7.80] [-7.62]
Pseudo 𝑅 2
0.002 0.002 0.014 0.016 0.082 0.096
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No Y es
N 76,626 76,626 76,626 76,626 76,626 76,626
70
T able 6 (con tin ued)
P anel B: High Δ In v estor Recognition Sam ple
Jump
0|1
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 0.383
∗∗∗
0.383
∗∗∗
0.395
∗∗∗
0.396
∗∗∗
0.352
∗∗∗
[13.48] [13.48] [14.18] [14.17] [12.66]
Disclosure Distance
𝑞 −4
0.146
∗∗∗
0.041 0.039 -0.010
[5.32] [1.55] [1.47] [-0.36]
ln(Bo ok-to-mark et
𝑞 −1
) -0.095
∗∗∗
-0.032
∗
[-5.68] [-1.91]
R O A
𝑞 −4, 𝑞 −1
1.267
∗∗∗
0.697
∗∗∗
[11.19] [6.34]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.508
∗∗∗
-0.611
∗∗∗
[2.73] [-3.18]
Financial Lev erage
𝑞 −1
-0.127
∗
-0.011
[-1.88] [-0.17]
Net External Financing
𝑞 −4, 𝑞 −1
-0.044 -0.002
[-1.30] [-0.07]
ln(Mark et Equit y
𝑞 ) -0.455
∗∗∗
-0.404
∗∗∗
[-20.68] [-18.83]
Analyst Co v erage
𝑞 -0.026
∗∗∗
-0.018
∗∗∗
[-8.74] [-5.91]
EPS Guidance
𝑞 𝑞 −1]
0.296
∗∗∗
0.144
∗∗∗
[10.83] [5.18]
Beta
(𝑞 −1, 𝑞 )
-0.040 -0.052
∗∗
[-1.54] [-1.99]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.164
∗∗∗
0.156
∗∗∗
[13.98] [13.68]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.415
∗∗∗
-0.532
∗∗∗
[-12.62] [-16.27]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.047 -0.206
∗∗∗
[-1.40] [-6.12]
𝑟 −1,−1
-0.012
∗∗∗
-0.013
∗∗∗
[-3.68] [-4.12]
𝑟 −20,−2
-0.006
∗∗∗
-0.005
∗∗∗
[-5.84] [-5.58]
Pseudo 𝑅 2
0.003 0.003 0.018 0.021 0.071 0.086
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No Y es
N 68,221 68,221 68,221 68,221 68,221 68,221
71
T able 7: Earnings Announcemen t Sto c k Returns: Subp erio ds
This table presen ts results for earnings announcemen t abnormal sto c k returns comp ounded o v er ev en t da ys
zero to one ( 𝑟 0,1
). P anel A presen ts panel regressions of Carhart ( 1997 ) four-factor alphas on explanatory
v ariables for the first half of the sample p erio d (Jan uary 2006 – Decem b er 2012) and P anel B presen ts
results for the second half of the sample p erio d (Jan uary 2013–Decem b er 2019). See App endix A for v ariable
definitions. t -v alues are in brac k ets based on standard errors clustered b y sto c k and trading date (p < .10*,
p < .05**, p < .01***). The sample spans earnings announcemen ts o v er Jan uary 2006–Decem b er 2019.
72
T able 7 (con tin ued)
P anel A: Jan uary 2006–Decem b er 2012 Subp erio d
𝑟 0,1
(1) (2) (3) (4) (5) (6) (7)
Disclosure Distance
𝑞 0.350
∗∗∗
0.354
∗∗∗
0.361
∗∗∗
0.230
∗∗
0.216
∗∗
-0.028
[3.51] [3.53] [3.55] [2.26] [2.13] [-0.29]
Disclosure Distance
𝑞 −4
0.061 -0.030 -0.116 -0.120 0.031
[0.60] [-0.29] [-1.10] [-1.14] [0.30]
ln(Bo ok-to-mark et
𝑞 −1
) 0.027 0.061 0.083
[0.42] [0.97] [1.37]
R O A
𝑞 −4, 𝑞 −1
1.502
∗∗∗
1.313
∗∗∗
2.502
∗∗∗
[3.34] [2.82] [5.63]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.291 -0.223 -1.051
[0.43] [-0.29] [-1.36]
Financial Lev erage
𝑞 −1
0.539
∗∗
0.631
∗∗
0.515
∗∗
[2.08] [2.35] [2.01]
Net External Financing
𝑞 −4, 𝑞 −1
-0.382
∗∗
-0.412
∗∗∗
-0.135
[-2.42] [-2.63] [-0.87]
ln(Mark et Equit y
𝑞 ) 0.262
∗∗∗
0.294
∗∗∗
0.156
∗∗
[3.41] [3.72] [2.04]
Analyst Co v erage
𝑞 0.013 0.014 0.028
∗∗∗
[1.38] [1.38] [2.80]
EPS Guidance
𝑞 𝑞 −1]
0.352
∗∗∗
0.325
∗∗∗
0.067
[3.60] [3.28] [0.71]
Beta
(𝑞 −1, 𝑞 )
-0.055 -0.053 -0.052
[-0.53] [-0.49] [-0.50]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.048 0.047 0.036
[1.35] [1.31] [1.06]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.146 -0.157 -0.243
∗∗
[-1.23] [-1.30] [-2.06]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.426
∗∗∗
-0.462
∗∗∗
-0.500
∗∗∗
[-3.29] [-3.49] [-3.85]
𝑟 −1,−1
-0.115
∗∗∗
-0.115
∗∗∗
-0.144
∗∗∗
[-4.33] [-4.33] [-5.18]
𝑟 −20,−2
-0.010 -0.010 -0.025
∗∗∗
[-1.61] [-1.60] [-3.97]
𝑟 0,1 𝑞 −4
0.002 0.002 0.006
[0.34] [0.35] [1.21]
ln(Idiosyncratic V olatilit y
−150, −1
) -0.075 -0.111 -0.105
[-0.60] [-0.87] [-0.87]
DSUE
𝑞 0.865
∗∗∗
[51.03]
A djusted 𝑅 2
0.000 0.001 0.001 0.001 0.006 0.006 0.078
Quarter FE No Y es Y es Y es Y es Y es Y es
Industry FE No No No No No Y es Y es
N 74,408 74,408 74,408 74,408 74,408 74,408 74,408
73
T able 7 (con tin ued)
P anel B: Jan uary 2013–Decem b er 2019 Subp er io d
𝑟 0,1
(1) (2) (3) (4) (5) (6) (7)
Disclosure Distance
𝑞 0.303
∗∗∗
0.322
∗∗∗
0.350
∗∗∗
0.272
∗∗∗
0.284
∗∗∗
0.117
[3.15] [3.36] [3.45] [2.68] [2.80] [1.21]
Disclosure Distance
𝑞 −4
0.001 -0.099 -0.155 -0.148 -0.014
[0.01] [-0.96] [-1.53] [-1.47] [-0.15]
ln(Bo ok-to-mark et
𝑞 −1
) 0.139
∗∗
0.131
∗∗
0.062
[2.32] [2.12] [1.04]
R O A
𝑞 −4, 𝑞 −1
0.965
∗∗
1.046
∗∗
2.044
∗∗∗
[2.40] [2.56] [4.94]
R&D Exp ense
𝑞 −4, 𝑞 −1
1.853
∗∗∗
1.979
∗∗∗
1.320
∗
[2.70] [2.78] [1.87]
Financial Lev erage
𝑞 −1
0.310 0.452
∗
0.418
∗
[1.30] [1.81] [1.74]
Net External Financing
𝑞 −4, 𝑞 −1
-0.364
∗∗∗
-0.396
∗∗∗
-0.215
[-2.68] [-2.87] [-1.59]
ln(Mark et Equit y
𝑞 ) 0.217
∗∗∗
0.209
∗∗∗
0.258
∗∗∗
[2.99] [2.88] [3.58]
Analyst Co v erage
𝑞 -0.010 -0.012 -0.014
[-1.02] [-1.25] [-1.44]
EPS Guidance
𝑞 𝑞 −1]
0.033 0.089 -0.102
[0.35] [0.93] [-1.10]
Beta
(𝑞 −1, 𝑞 )
0.186
∗
0.163 0.181
∗
[1.66] [1.46] [1.66]
Sync hronicit y
(𝑞 −1, 𝑞 )
-0.061
∗
-0.066
∗
-0.081
∗∗
[-1.69] [-1.80] [-2.25]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.194 -0.214
∗
-0.212
∗
[-1.54] [-1.69] [-1.69]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.371
∗∗∗
-0.372
∗∗∗
-0.411
∗∗∗
[-2.61] [-2.60] [-2.90]
𝑟 −1,−1
-0.127
∗∗∗
-0.127
∗∗∗
-0.164
∗∗∗
[-5.76] [-5.75] [-7.54]
𝑟 −20,−2
-0.005 -0.005 -0.016
∗∗∗
[-0.85] [-0.85] [-2.72]
𝑟 0,1 𝑞 −4
-0.015
∗∗∗
-0.015
∗∗∗
-0.010
∗∗
[-2.91] [-2.95] [-2.10]
ln(Idiosyncratic V olatilit y
−150, −1
) -0.090 -0.030 0.023
[-0.70] [-0.23] [0.18]
DSUE
𝑞 0.767
∗∗∗
[47.24]
A djusted 𝑅 2
0.000 0.001 0.001 0.001 0.004 0.004 0.070
Quarter FE No Y es Y es Y es Y es Y es Y es
Industry FE No No No No No Y es Y es
N 70,439 70,439 70,439 70,439 70,439 70,439 70,439
74
T able 8: Earnings Announcemen t Sto c k Returns: Size Subsamples
This table presen ts results for earnings announcemen t abnormal sto c k returns comp ounded o v er ev en t da ys
zero to one ( 𝑟 0,1
). P anel A presen ts panel regressions of Carhart ( 1997 ) four-factor alphas on explanatory
v ariables for small sto c ks and P anel B presen ts results for big sto c ks. Size is determined b y a cross-sectional
median partition on the mark et v alue of equit y measured at fiscal quarter end. See App endix A for v ariable
definitions. t -v alues are in brac k ets based on standard errors clustered b y sto c k and trading date (p < .10*,
p < .05**, p < .01***). The sample spans earnings announcemen ts o v er Jan uary 2006–Decem b er 2019.
75
T able 8 (con tin ued)
P anel A: Small Sto c ks
𝑟 0,1
(1) (2) (3) (4) (5) (6) (7)
Disclosure Distance
𝑞 0.471
∗∗∗
0.468
∗∗∗
0.478
∗∗∗
0.295
∗∗
0.288
∗∗
0.035
[4.16] [4.13] [4.08] [2.51] [2.46] [0.31]
Disclosure Distance
𝑞 −4
0.090 -0.036 -0.163 -0.161 0.016
[0.80] [-0.31] [-1.40] [-1.39] [0.14]
ln(Bo ok-to-mark et
𝑞 −1
) 0.244
∗∗∗
0.259
∗∗∗
0.312
∗∗∗
[3.48] [3.70] [4.56]
R O A
𝑞 −4, 𝑞 −1
1.030
∗∗∗
0.997
∗∗
2.347
∗∗∗
[2.69] [2.54] [6.23]
R&D Exp ense
𝑞 −4, 𝑞 −1
1.501
∗∗
1.385
∗∗
0.909
[2.36] [2.06] [1.37]
Financial Lev erage
𝑞 −1
0.505
∗
0.601
∗∗
0.653
∗∗
[1.81] [2.08] [2.41]
Net External Financing
𝑞 −4, 𝑞 −1
-0.416
∗∗∗
-0.450
∗∗∗
-0.283
∗∗
[-3.11] [-3.34] [-2.16]
ln(Mark et Equit y
𝑞 ) 0.679
∗∗∗
0.704
∗∗∗
0.623
∗∗∗
[6.24] [6.36] [5.73]
Analyst Co v erage
𝑞 0.040 0.039 0.071
∗∗∗
[1.54] [1.48] [2.69]
EPS Guidance
𝑞 𝑞 −1]
0.259
∗
0.289
∗∗
-0.071
[1.77] [1.96] [-0.49]
Beta
(𝑞 −1, 𝑞 )
-0.138 -0.137 -0.079
[-1.31] [-1.30] [-0.76]
Sync hronicit y
(𝑞 − 1, 𝑞 )
-0.002 -0.001 -0.011
[-0.05] [-0.03] [-0.37]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.012 -0.005 -0.076
[-0.11] [-0.04] [-0.73]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.381
∗∗∗
-0.393
∗∗∗
-0.446
∗∗∗
[-3.40] [-3.46] [-3.99]
𝑟 −1,−1
-0.115
∗∗∗
-0.115
∗∗∗
-0.142
∗∗∗
[-5.08] [-5.09] [-5.99]
𝑟 −20,−2
-0.010
∗
-0.010
∗
-0.022
∗∗∗
[-1.87] [-1.85] [-3.98]
𝑟 0,1 𝑞 −4
-0.008
∗
-0.008
∗
0.001
[-1.68] [-1.72] [0.15]
ln(Idiosyncratic V olatilit y
−150, −1
) -0.177 -0.195 -0.078
[-1.41] [-1.52] [-0.64]
DSUE
𝑞 0.812
∗∗∗
[49.30]
A djusted 𝑅 2
0.000 0.001 0.001 0.001 0.007 0.008 0.071
Quarter FE No Y es Y es Y es Y es Y es Y es
Industry FE No No No No No Y es Y es
N 72,424 72,424 72,424 72,424 72,424 72,424 72,424
76
T able 8 (con tin ued)
P anel B: Big Sto c ks
𝑟 0,1
(1) (2) (3) (4) (5) (6) (7)
Disclosure Distance
𝑞 0.222
∗∗∗
0.254
∗∗∗
0.264
∗∗∗
0.225
∗∗∗
0.215
∗∗∗
0.066
[2.82] [3.24] [3.26] [2.74] [2.62] [0.82]
Disclosure Distance
𝑞 −4
0.030 -0.039 -0.089 -0.100 0.007
[0.38] [-0.47] [-1.08] [-1.22] [0.08]
ln(Bo ok-to-mark et
𝑞 −1
) -0.098
∗
-0.085 -0.263
∗∗∗
[-1.89] [-1.60] [-4.96]
R O A
𝑞 −4, 𝑞 −1
-0.268 -0.477 -0.371
[-0.60] [-1.02] [-0.81]
R&D Exp ense
𝑞 −4, 𝑞 −1
1.145 0.776 -0.355
[1.53] [0.91] [-0.42]
Financial Lev erage
𝑞 −1
0.310 0.329 0.038
[1.57] [1.60] [0.19]
Net External Financing
𝑞 −4, 𝑞 −1
-0.266
∗
-0.265
∗
0.061
[-1.86] [-1.84] [0.44]
ln(Mark et Equit y
𝑞 ) 0.063 0.066 0.100
[0.98] [1.01] [1.58]
Analyst Co v erage
𝑞 0.011 0.013
∗
0.018
∗∗
[1.55] [1.80] [2.47]
EPS Guidance
𝑞 𝑞 −1]
0.172
∗∗
0.159
∗∗
0.011
[2.41] [2.13] [0.15]
Beta
(𝑞 −1, 𝑞 )
0.023 0.017 -0.035
[0.20] [0.14] [-0.30]
Sync hronicit y
(𝑞 −1, 𝑞 )
-0.036 -0.047 -0.088
∗∗
[-0.82] [-1.06] [-2.02]
Illiquidit y
(𝑞 −1, 𝑞 )
0.010 -0.006 -0.026
[0.07] [-0.04] [-0.17]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.177 -0.199 -0.278
∗
[-1.06] [-1.18] [-1.66]
𝑟 −1,−1
-0.126
∗∗∗
-0.126
∗∗∗
-0.181
∗∗∗
[-6.23] [-6.23] [-9.30]
𝑟 −20,−2
-0.005 -0.005 -0.024
∗∗∗
[-0.83] [-0.82] [-4.13]
𝑟 0,1 𝑞 −4
-0.007 -0.007 -0.012
∗∗
[-1.35] [-1.32] [-2.45]
ln(Idiosyncratic V olatilit y
−150, −1
) -0.049 -0.064 -0.113
[-0.38] [-0.49] [-0.89]
DSUE
𝑞 0.835
∗∗∗
[51.71]
A djusted 𝑅 2
0.000 0.001 0.001 0.001 0.003 0.003 0.086
Quarter FE No Y es Y es Y es Y es Y es Y es
Industry FE No No No No No Y es Y es
N 72,423 72,423 72,423 72,423 72,423 72,423 72,423
77
T able 9: Con trolling for Beta-Shifts at Earnings Announcemen ts
This table presen ts earnings announcemen t returns adjusted for b eta-shifts at earnings announcemen ts. P ort-
folios are sorted in to three p ortfolios eac h quarter based on disclosure distance. Excess Return
0,1
is the
earnings announcemen t return less the risk-free rate comp ounded o v er ev en t da y 0 and 1. P o oled CAPM
A djusted Return
0,1
is the in tercept from a p o oled regression of Excess Return
0,1
on the con temp oraneous
mark et factor return. P o oled CAPM Beta
0,1
is the slop e co efficien t on the mark et factor from the p o oled
mark et mo del regression. Out-of-Sample CAPM Beta
𝑞−1,𝑞
is the slop e co efficien t on the mark et factor from
a time series regression of firm sp ecific excess sto c k returns on the mark et factor o v er the in terim da ys b e-
t w een adjacen t earnings announcemen ts (𝑞 − 1, 𝑞 ) . t -v alues are in brac k ets based on standard errors clustered
b y sto c k and trading date (p < .10*, p < .05**, p < .01***). The sample spans earnings announcemen ts o v er
Jan uary 2006 - Decem b er 2019.
Disclosure Distance
Recen t Mid Distan t Distan t-Recen t
Excess Return
0,1
-0.164 0.045 0.161 0.324***
[-2.154] [0.622] [2.226] [5.268]
P o oled CAPM A djusted Return
0,1
-0.189 -0.005 0.112 0.300***
[-3.867] [-0.104] [2.264] [5.046]
P o oled CAPM Beta
0,1
1.178 1.049 1.042 -0.136***
[23.969] [24.074] [24.246] [-2.745]
Out-of-Sample CAPM Be ta
(𝑞 −1,𝑞 )
1.062 1.055 1.004 -0.058***
[124.400] [124.414] [108.354] [-7.334]
𝑁 48,536 48,166 48,145
78
T able 10: Alternativ e Measures of Jumps
This table presen ts results for earnings announcemen t jumps for alternativ e jump definitions. P anel A (B)
[C] presen ts logit regressions for an indicator v ariable for absolute abnormal sto c k returns greater than 5%
(10%) [15%] where abnormal returns are re lativ e to the Carhart ( 1997 ) four-factor mo del. See App endix A
for v ariable definitions. z -v alues are in brac k ets based on standard errors clustered b y sto c k and trading date
(p < .10*, p < .05**, p < .01***). The sample spans Jan uary 2006–Decem b er 2019.
79
T able 10 (con tin ued)
P anel A: Jump = Absolute Ret urn > 5%
𝐼 (|𝑟
0,1
| > 5%)
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 0.175
∗∗∗
0.175
∗∗∗
0.186
∗∗∗
0.290
∗∗∗
0.244
∗∗∗
[7.88] [7.88] [9.24] [15.03] [12.95]
Disclosure Distance
𝑞 −4
0.017 -0.033 -0.019 -0.052
∗∗∗
[0.78] [-1.62] [-1.03] [-2.82]
ln(Bo ok-to-mark et
𝑞 −1
) -0.126
∗∗∗
-0.045
∗∗∗
[-10.01] [-3.69]
R O A
𝑞 −4, 𝑞 −1
0.887
∗∗∗
0.456
∗∗∗
[11.94] [6.53]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.632
∗∗∗
-0.251
∗∗
[4.92] [-2.06]
Financial Lev erage
𝑞 −1
0.128
∗∗
0.178
∗∗∗
[2.46] [3.48]
Net External Financing
𝑞 −4, 𝑞 −1
0.015 0.045
∗∗
[0.67] [2.08]
ln(Mark et Equit y
𝑞 ) -0.644
∗∗∗
-0.555
∗∗∗
[-36.61] [-32.86]
Analyst Co v erage
𝑞 -0.014
∗∗∗
-0.009
∗∗∗
[-5.79] [-3.49]
EPS Guidance
𝑞 𝑞 −1]
0.270
∗∗∗
0.114
∗∗∗
[12.58] [5.35]
Beta
(𝑞 −1, 𝑞 )
0.330
∗∗∗
0.329
∗∗∗
[18.66] [18.97]
Sync hronicit y
(𝑞 −1, 𝑞 )
-0.043
∗∗∗
-0.044
∗∗∗
[-6.75] [-7.14]
Illiquidit y
(𝑞 −1, 𝑞 )
0.344
∗∗∗
0.267
∗∗∗
[16.31] [13.48]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
0.812
∗∗∗
0.668
∗∗∗
[35.08] [30.86]
𝑟 −1,−1
0.002 0.001
[0.90] [0.61]
𝑟 −20,−2
-0.006
∗∗∗
-0.005
∗∗∗
[-10.13] [-9.54]
Pseudo 𝑅 2
0.001 0.001 0.013 0.014 0.081 0.095
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No No
N 144,847 144,847 144,847 144,847 144,847 144,847
80
T able 10 (con tin ued)
P anel B: Jump = Absolute Retur n > 10%
𝐼 (|𝑟
0,1
| > 10%)
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 0.212
∗∗∗
0.212
∗∗∗
0.224
∗∗∗
0.344
∗∗∗
0.303
∗∗∗
[7.81] [7.81] [8.80] [13.77] [12.29]
Disclosure Distance
𝑞 −4
0.035 -0.025 -0.022 -0.045
∗
[1.34] [-1.00] [-0.94] [-1.93]
ln(Bo ok-to-mark et
𝑞 −1
) -0.131
∗∗∗
-0.066
∗∗∗
[-8.48] [-4.38]
R O A
𝑞 −4, 𝑞 −1
0.937
∗∗∗
0.600
∗∗∗
[10.38] [6.86]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.425
∗∗∗
-0.302
∗∗
[2.84] [-2.00]
Financial Lev erage
𝑞 −1
0.106
∗
0.156
∗∗
[1.71] [2.46]
Net External Financing
𝑞 −4, 𝑞 −1
-0.001 0.031
[-0.03] [1.22]
ln(Mark et Equit y
𝑞 ) -0.797
∗∗∗
-0.711
∗∗∗
[-37.36] [-34.11]
Analyst Co v erage
𝑞 -0.022
∗∗∗
-0.016
∗∗∗
[-6.79] [-4.82]
EPS Guidance
𝑞 𝑞 −1]
0.314
∗∗∗
0.156
∗∗∗
[10.99] [5.45]
Beta
(𝑞 −1, 𝑞 )
0.443
∗∗∗
0.455
∗∗∗
[21.89] [22.20]
Sync hronicit y
(𝑞 −1 , 𝑞 )
-0.093
∗∗∗
-0.098
∗∗∗
[-12.10] [-12.76]
Illiquidit y
(𝑞 −1, 𝑞 )
0.258
∗∗∗
0.219
∗∗∗
[11.09] [9.43]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
0.778
∗∗∗
0.680
∗∗∗
[29.54] [26.01]
𝑟 −1,−1
-0.000 -0.001
[-0.09] [-0.33]
𝑟 −20,−2
-0.006
∗∗∗
-0.005
∗∗∗
[-9.10] [-8.65]
Pseudo 𝑅 2
0.001 0.001 0.018 0.019 0.103 0.115
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No No
N 144,847 144,847 144,847 144,847 144,847 144,847
81
T able 10 (con tin ued)
P anel C: Jump = Absolute Retur n > 15%
𝐼 (|𝑟
0,1
| > 15%)
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 0.229
∗∗∗
0.229
∗∗∗
0.246
∗∗∗
0.382
∗∗∗
0.347
∗∗∗
[6.72] [6.72] [7.48] [11.76] [10.75]
Disclosure Distance
𝑞 −4
0.021 -0.045 -0.047 -0.061
∗
[0.60] [-1.38] [-1.45] [-1.90]
ln(Bo ok-to-mark et
𝑞 −1
) -0.113
∗∗∗
-0.055
∗∗∗
[-6.15] [-2.97]
R O A
𝑞 −4, 𝑞 −1
1.007
∗∗∗
0.731
∗∗∗
[9.24] [6.77]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.311
∗
-0.347
∗
[1.78] [-1.90]
Financial Lev erage
𝑞 −1
0.164
∗∗
0.223
∗∗∗
[2.10] [2.82]
Net External Financing
𝑞 −4, 𝑞 −1
-0.025 0.005
[-0.77] [0.15]
ln(Mark et Equit y
𝑞 ) -0.857
∗∗∗
-0.779
∗∗∗
[-33.23] [-30.34]
Analyst Co v erage
𝑞 -0.028
∗∗∗
-0.021
∗∗∗
[-6.08] [-4.67]
EPS Guidance
𝑞 𝑞 −1]
0.348
∗∗∗
0.186
∗∗∗
[9.55] [5.06]
Beta
(𝑞 −1, 𝑞 )
0.509
∗∗∗
0.532
∗∗∗
[20.89] [21.37]
Sync hronicit y
(𝑞 − 1, 𝑞 )
-0.125
∗∗∗
-0.132
∗∗∗
[-13.06] [-13.55]
Illiquidit y
(𝑞 −1, 𝑞 )
0.238
∗∗∗
0.217
∗∗∗
[9.01] [8.12]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
0.771
∗∗∗
0.700
∗∗∗
[26.38] [23.52]
𝑟 −1,−1
0.001 0.001
[0.51] [0.29]
𝑟 −20,−2
-0.006
∗∗∗
-0.006
∗∗∗
[-7.29] [-7.09]
Pseudo 𝑅 2
0.001 0.001 0.022 0.023 0.115 0.125
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No No
N 144,847 144,847 144,847 144,847 144,847 144,847
82
T able 11: Earnings Announcemen t Idiosyncratic V olatilit y
This table presen ts results for earnings announcemen t idiosyncratic v olatilit y . P anel A (B) presen ts panel
regressions for absolute (squared) abnormal sto c k returns where abnormal returns are relativ e to the Carhart
( 1997 ) four-factor mo del. See App endix A for v ariable definitions. t -v alues are in brac k ets based on standard
errors clustered b y sto c k and trading date (p < .10*, p < .05**, p < .01***). The sample spans Jan uary 2006–
Decem b er 2019.
83
T able 11 (con tin ued)
P anel A: Absolute Returns
|𝑟
0,1
|
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 0.563
∗∗∗
0.564
∗∗∗
0.596
∗∗∗
0.893
∗∗∗
0.802
∗∗∗
[8.06] [8.23] [9.66] [17.38] [15.89]
Disclosure Distance
𝑞 −4
0.041 -0.119
∗∗
-0.131
∗∗∗
-0.186
∗∗∗
[0.60] [-1.97] [-2.68] [-3.86]
ln(Bo ok-to-mark et
𝑞 −1
) -0.136
∗∗∗
-0.005
[-3.84] [-0.14]
R O A
𝑞 −4, 𝑞 −1
1.929
∗∗∗
1.042
∗∗∗
[8.75] [4.72]
R&D Exp ense
𝑞 − 4, 𝑞 −1
0.872
∗∗
-0.965
∗∗
[2.24] [-2.42]
Financial Lev erage
𝑞 −1
0.424
∗∗∗
0.548
∗∗∗
[3.06] [3.83]
Net External Financing
𝑞 −4, 𝑞 −1
-0.054 0.046
[-0.79] [0.67]
ln(Mark et Equit y
𝑞 ) -1.407
∗∗∗
-1.299
∗∗∗
[-34.28] [-32.43]
Analyst Co v erage
𝑞 -0.039
∗∗∗
-0.021
∗∗∗
[-7.00] [-3.49]
EPS Guidance
𝑞 𝑞 −1]
0.629
∗∗∗
0.299
∗∗∗
[10.71] [4.99]
Beta
(𝑞 −1, 𝑞 )
0.593
∗∗∗
0.653
∗∗∗
[12.02] [13.14]
Sync hronicit y
(𝑞 −1, 𝑞 )
-0.005 -0.028
[-0.30] [-1.62]
Illiquidit y
(𝑞 −1, 𝑞 )
0.191
∗∗∗
0.134
∗∗
[3.38] [2.41]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
1.227
∗∗∗
1.080
∗∗∗
[19.89] [17.76]
𝑟 −1,−1
0.003 0.002
[0.41] [0.28]
𝑟 −20,−2
-0.020
∗∗∗
-0.019
∗∗∗
[-8.30] [-8.21]
|𝑟
0,1 𝑞 −4
| 0.107
∗∗∗
0.094
∗∗∗
[27.71] [24.69]
ln(Idiosyncratic V olatilit y
−150, −1
) 1.925
∗∗∗
1.765
∗∗∗
[28.15] [25.76]
A djusted 𝑅 2
0.001 0.031 0.030 0.031 0.158 0.166
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No Y es
N 144,847 144,847 144,847 144,847 144,847 144,847
84
T able 11 (con tin ued)
P anel B: Squared Returns
𝑟 2
0,1
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 10.500
∗∗∗
10.657
∗∗∗
11.391
∗∗∗
17.858
∗∗∗
16.527
∗∗∗
[6.38] [6.59] [7.56] [13.26] [12.34]
Disclosure Distance
𝑞 −4
0.314 -2.736
∗
-2.944
∗∗
-3.763
∗∗∗
[0.20] [-1.84] [-2.23] [-2.87]
ln(Bo ok-to-mark et
𝑞 −1
) 0.002 1.855
∗
[0.00] [1.87]
R O A
𝑞 −4, 𝑞 −1
32.370
∗∗∗
17.588
∗∗∗
[5.29] [2.83]
R&D Exp ense
𝑞 −4, 𝑞 −1
9.089 -22.562
∗∗
[0.85] [-2.04]
Financial Lev erage
𝑞 −1
16.712
∗∗∗
18.914
∗∗∗
[4.51] [4.93]
Net External Financing
𝑞 −4, 𝑞 −1
-1.920 0.301
[-0.95] [0.15]
ln(Mark et Equit y
𝑞 ) -30.519
∗∗∗
-28.681
∗∗∗
[-28.24] [-26.71]
Analyst Co v erage
𝑞 -0.820
∗∗∗
-0.403
∗∗∗
[-5.79] [-2.75]
EPS Guidance
𝑞 𝑞 −1]
11.949
∗∗∗
5.804
∗∗∗
[8.08] [3.82]
Beta
(𝑞 −1, 𝑞 )
12.752
∗∗∗
14.156
∗∗∗
[8.98] [9.84]
Sync hronicit y
(𝑞 −1, 𝑞 )
-1.321
∗∗∗
-1.668
∗∗∗
[-2.79] [-3.54]
Illiquidit y
(𝑞 −1, 𝑞 )
5.494
∗∗∗
4.968
∗∗∗
[3.58] [3.25]
A vg. T rading V olume
(𝑞 −1 , 𝑞 )
27.457
∗∗∗
25.416
∗∗∗
[16.13] [14.91]
𝑟 −1,−1
0.177 0.151
[0.47] [0.39]
𝑟 −20,−2
-0.561
∗∗∗
-0.555
∗∗∗
[-6.47] [-6.46]
𝑟 2
0,1 𝑞 −4
0.081
∗∗∗
0.075
∗∗∗
[16.19] [15.07]
ln(Idiosyncratic V olatilit y
−150, −1
) 39.567
∗∗∗
37.362
∗∗∗
[22.17] [20.68]
A djusted 𝑅 2
0.000 0.028 0.027 0.028 0.109 0.113
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No Y es
N 144,847 144,847 144,847 144,847 144,847 144,847
85
T able 12: Earnings Announcemen t Idiosyncratic V olatilit y Excluding Jumps
This table presen ts results for earnings announcemen t idiosyncratic v olatilit y for a sample excluding jumps
(as defined in Kapadia and Zekhnini ( 2019 )). P anel A (B) presen ts panel regressions for absolute (squared)
abnormal sto c k returns where abnormal returns are relativ e to the Carhart ( 1997 ) four-factor mo del. See
App e ndix A for v ariable definitions. t -v alues are in brac k ets based on standard errors clustered b y sto c k and
trading date (p < .10*, p < .05**, p < .01***). The sample spans Jan uary 2006–Decem b er 2019.
86
T able 12 (con tin ued)
P anel A: Absolute Returns
|𝑟
0,1
|
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 -0.137
∗∗∗
-0.161
∗∗∗
-0.119
∗∗∗
0.215
∗∗∗
0.204
∗∗∗
[-3.22] [-4.02] [-3.15] [7.38] [7.01]
Disclosure Distance
𝑞 −4
-0.187
∗∗∗
-0.155
∗∗∗
-0.128
∗∗∗
-0.135
∗∗∗
[-4.87] [-4.32] [-4.46] [-4.71]
ln(Bo ok-to-mark et
𝑞 −1
) 0.055
∗∗∗
0.076
∗∗∗
[2.96] [4.01]
R O A
𝑞 −4, 𝑞 −1
-0.442
∗∗∗
-0.576
∗∗∗
[-3.66] [-4.71]
R&D Exp ense
𝑞 −4, 𝑞 −1
-0.216 -0.517
∗∗
[-1.07] [-2.44]
Financial Lev erage
𝑞 −1
0.457
∗∗∗
0.421
∗∗∗
[6.35] [5.70]
Net External Financing
𝑞 −4, 𝑞 −1
0.013 0.032
[0.34] [0.83]
ln(Mark et Equit y
𝑞 ) -0.240
∗∗∗
-0.223
∗∗∗
[-11.16] [-10.19]
Analyst Co v erage
𝑞 0.008
∗∗∗
0.013
∗∗∗
[3.02] [4.72]
EPS Guidance
𝑞 𝑞 −1]
0.012 -0.050
∗
[0.46] [-1.87]
Beta
(𝑞 −1, 𝑞 )
0.132
∗∗∗
0.150
∗∗∗
[4.38] [4.89]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.038
∗∗∗
0.035
∗∗∗
[3.95] [3.61]
Illiquidit y
(𝑞 −1, 𝑞 )
0.163
∗∗∗
0.159
∗∗∗
[4.84] [4.74]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
0.354
∗∗∗
0.334
∗∗∗
[9.48] [8.93]
𝑟 −1,−1
0.022
∗∗∗
0.022
∗∗∗
[4.23] [4.20]
𝑟 −20,−2
-0.006
∗∗∗
-0.006
∗∗∗
[-3.71] [-3.70]
|𝑟
0,1 𝑞 −4
| 0.021
∗∗∗
0.019
∗∗∗
[8.90] [8.00]
ln(Idiosyncratic V olatilit y
−150, −1
) 2.248
∗∗∗
2.225
∗∗∗
[53.49] [51.68]
A djusted 𝑅 2
0.000 0.090 0.090 0.090 0.286 0.287
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No Y es
N 90,773 90,773 90,773 90,773 90,773 90,773
87
T able 12 (con tin ued)
P anel B: Squared Returns
𝑟 2
0,1
(1) (2) (3) (4) (5) (6)
Disclosure Distance
𝑞 -2.173
∗∗∗
-2.514
∗∗∗
-1.934
∗∗∗
2.124
∗∗∗
2.165
∗∗∗
[-3.43] [-4.14] [-3.33] [4.26] [4.35]
Disclosure Distance
𝑞 −4
-2.631
∗∗∗
-2.095
∗∗∗
-1.699
∗∗∗
-1.715
∗∗∗
[-4.79] [-4.08] [-3.70] [-3.70]
ln(Bo ok-to-mark et
𝑞 −1
) 1.652
∗∗∗
1.580
∗∗∗
[4.55] [4.33]
R O A
𝑞 −4, 𝑞 −1
-13.379
∗∗∗
-14.117
∗∗∗
[-5.60] [-5.84]
R&D Exp ense
𝑞 −4, 𝑞 −1
-9.658
∗∗
-11.145
∗∗∗
[-2.51] [-2.78]
Financial Lev erage
𝑞 −1
9.415
∗∗∗
8.874
∗∗∗
[6.84] [6.39]
Net External Financing
𝑞 −4, 𝑞 −1
-0.734 -0.408
[-1.08] [-0.59]
ln(Mark et Equit y
𝑞 ) -2.084
∗∗∗
-2.134
∗∗∗
[-5.20] [-5.10]
Analyst Co v erage
𝑞 0.061 0.161
∗∗∗
[1.29] [3.31]
EPS Guidance
𝑞 𝑞 − 1]
0.022 -0.468
[0.05] [-1.12]
Beta
(𝑞 −1, 𝑞 )
1.034
∗
1.301
∗∗
[1.80] [2.24]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.360
∗∗
0.334
∗
[2.06] [1.93]
Illiquidit y
(𝑞 −1, 𝑞 )
3.709
∗∗∗
3.807
∗∗∗
[5.96] [6.10]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
5.566
∗∗∗
5.663
∗∗∗
[8.12] [8.15]
𝑟 −1,−1
0.522
∗∗∗
0.518
∗∗∗
[3.53] [3.50]
𝑟 −20,−2
-0.084
∗∗
-0.086
∗∗
[-2.20] [-2.25]
𝑟 2
0,1 𝑞 −4
0.010
∗∗∗
0.010
∗∗∗
[5.32] [5.10]
ln(Idiosyncratic V olatilit y
−150, −1
) 26.719
∗∗∗
26.948
∗∗∗
[31.77] [31.32]
A djusted 𝑅 2
0.000 0.090 0.090 0.090 0.211 0.212
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No Y es
N 90,773 90,773 90,773 90,773 90,773 90,773
88
T able 13: Alternativ e Measure of Disclosure Distance: Da ys Since Disclosure
This table presen ts results for earnings announcemen t jumps and abnormal sto c k returns comp ounded
o v er ev en t da ys zero to one ( 𝑟 0,1
) using 𝐷𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 in place of 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 .
𝐷𝑎𝑦 𝑠 𝑆 𝑖𝑛𝑐 𝑒 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 is the n um b er of trading da ys since the last 8-K disclosure or the last earnings
announcemen t date for firms without in terim 8-K filings. P anel A presen ts logit regressions of jump indica-
tors on explanatory v ariables, and P anel B presen ts panel regressions of Carhart ( 1997 ) four-factor alphas
on explanatory v ariables. See App endix A for v ariable definitions. z (t) -v a lues are in brac k ets based on
standard errors clustered b y sto c k and trading date in P anel A (B) (p < .10*, p < .05**, p < .01***). The
sample spans Jan uary 2006–Decem b er 2019.
89
T able 13 (con tin ued)
P anel A: Jumps
Jump
0|1
(1) (2) (3) (4) (5) (6)
Da ys Since Disclosure
𝑞 0.006
∗∗∗
0.006
∗∗∗
0.006
∗∗∗
0.006
∗∗∗
0.005
∗∗∗
[15.50] [15.50] [16.95] [18.45] [16.66]
Da ys Since Disclosure
𝑞 −4
0.001
∗∗∗
-0.000 0.000 -0.000
[4.25] [-0.26] [1.14] [-0.79]
ln(Bo ok-to-mark et
𝑞 −1
) -0.087
∗∗∗
-0.022
∗
[-6.47] [-1.74]
R O A
𝑞 −4, 𝑞 −1
1.327
∗∗∗
0.789
∗∗∗
[15.24] [9.45]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.602
∗∗∗
-0.406
∗∗∗
[4.22] [-2.82]
Financial Lev erage
𝑞 −1
-0.069 0.043
[-1.28] [0.81]
Net External Financing
𝑞 −4, 𝑞 −1
-0.051
∗∗
-0.011
[-2.08] [-0.47]
ln(Mark et Equit y
𝑞 ) -0.448
∗∗∗
-0.387
∗∗∗
[-25.08] [-22.52]
Analyst Co v erage
𝑞 -0.027
∗∗∗
-0.019
∗∗∗
[-11.02] [-7.43]
EPS Guidance
𝑞 𝑞 −1]
0.296
∗∗∗
0.149
∗∗∗
[13.23] [6.68]
Beta
(𝑞 −1, 𝑞 )
-0.030
∗
-0.042
∗∗
[-1.72] [-2.42]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.145
∗∗∗
0.142
∗∗∗
[18.61] [19.06]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.397
∗∗∗
-0.491
∗∗∗
[-17.26] [-22.10]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.026 -0.168
∗∗∗
[-1.12] [-7.34]
𝑟 −1,−1
-0.010
∗∗∗
-0.011
∗∗∗
[-4.95] [-5.46]
𝑟 −20,−2
-0.006
∗∗∗
-0.006
∗∗∗
[-9.36] [-9.17]
Pseudo 𝑅 2
0.003 0.003 0.013 0.016 0.078 0.092
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No Y es
N 144,847 144,847 144,847 144,847 144,847 144,847
90
T able 13 (con tin ued)
P anel B: Returns
𝑟 0,1
(1) (2) (3) (4) (5) (6) (7)
Da ys Since Disclosure
𝑞 0.005
∗∗∗
0.005
∗∗∗
0.005
∗∗∗
0.003
∗∗∗
0.003
∗∗∗
-0.000
[4.21] [4.13] [4.13] [2.66] [2.60] [-0.03]
Da ys Since Disclosure
𝑞 −4
0.001 -0.000 -0.002 -0.002 0.001
[0.70] [-0.43] [-1.46] [-1.49] [0.54]
ln(Bo ok-to-mark et
𝑞 −1
) 0.087
∗∗
0.101
∗∗
0.063
[1.98] [2.27] [1.45]
R O A
𝑞 −4, 𝑞 −1
1.234
∗∗∗
1.192
∗∗∗
2.299
∗∗∗
[4.03] [3.78] [7.41]
R&D Exp ense
𝑞 −4, 𝑞 −1
1.147
∗∗
1.124
∗∗
0.364
[2.27] [2.07] [0.67]
Financial Lev erage
𝑞 −1
0.391
∗∗
0.476
∗∗∗
0.452
∗∗
[2.22] [2.61] [2.57]
Net External Financing
𝑞 −4, 𝑞 −1
-0.311
∗∗∗
-0.333
∗∗∗
-0.114
[-3.06] [-3.26] [-1.13]
ln(Mark et Equit y
𝑞 ) 0.226
∗∗∗
0.234
∗∗∗
0.211
∗∗∗
[4.26] [4.34] [3.99]
Analyst Co v erage
𝑞 0.002 0.001 0.006
[0.31] [0.18] [0.83]
EPS Guidance
𝑞 𝑞 −1 ]
0.204
∗∗∗
0.218
∗∗∗
-0.000
[2.95] [3.08] [-0.00]
Beta
(𝑞 −1, 𝑞 )
0.062 0.052 0.067
[0.79] [0.66] [0.87]
Sync hronicit y
(𝑞 −1, 𝑞 )
-0.005 -0.007 -0.023
[-0.19] [-0.29] [-0.93]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.171
∗∗
-0.180
∗∗
-0.229
∗∗∗
[-1.96] [-2.06] [-2.65]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.390
∗∗∗
-0.401
∗∗∗
-0.458
∗∗∗
[-4.07] [-4.15] [-4.78]
𝑟 −1,−1
-0.119
∗∗∗
-0.119
∗∗∗
-0.151
∗∗∗
[-6.24] [-6.24] [-7.56]
𝑟 −20,−2
-0.009
∗∗
-0.009
∗
-0.022
∗∗∗
[-1.96] [-1.96] [-4.87]
𝑟 0,1 𝑞 −4
-0.006
∗
-0.006
∗
-0.002
[-1.84] [-1.85] [-0.64]
ln(Idiosyncratic V olatilit y
−150, −1
) -0.085 -0.091 -0.027
[-0.94] [-0.98] [-0.30]
DSUE
𝑞 0.817
∗∗∗
[65.04]
A djusted 𝑅 2
0.000 0.001 0.001 0.001 0.005 0.005 0.074
Quarter FE No Y es Y es Y es Y es Y es Y es
Industry FE No No No No No Y es Y es
N 144,847 144,847 144,847 144,847 144,847 144,847 144,847
91
T able 14: Alternativ e Measure of Disclosure Distance: Rank ed Disclosure
Distance
This table presen ts results for earnings announcemen t jumps and abnormal sto c k returns comp ounded o v er
ev en t da ys zero to one ( 𝑟 0,1
) using 𝑅 𝑎𝑛𝑘𝑒𝑑 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 in place of 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 .
𝑅 𝑎𝑛𝑘𝑒𝑑 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 is the tercile rank of 𝐷𝑖𝑠𝑐 𝑙𝑜𝑠𝑢𝑟 𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐 𝑒 transformed to range from
[0,1] with ranks assigned eac h quarter. P anel A presen ts logit regressions of jump indicators on explana-
tory v ariables, and P anel B presen ts panel regressions of Carhart ( 1997 ) four-factor alphas on explanatory
v ariables. See App endix A for v ariable definitions. z (t) -v alues are in brac k ets based on standard errors clus-
tered b y sto c k and trading date in P anel A (B) (p < .10*, p < .05**, p < .01***). The sample spans Jan uary
2006–Decem b er 2019.
92
T able 14 (con tin ued)
P anel A: Jumps
Jump
0|1
(1) (2) (3) (4) (5) (6)
Rank ed Disclosure Distance
𝑞 0.291
∗∗∗
0.291
∗∗∗
0.295
∗∗∗
0.313
∗∗∗
0.278
∗∗∗
[15.55] [15.55] [16.80] [18.26] [16.35]
Rank ed Disclosure Distance
𝑞 −4
0.079
∗∗∗
0.006 0.028
∗
-0.004
[4.41] [0.35] [1.75] [-0.29]
ln(Bo ok-to-mark et
𝑞 −1
) -0.086
∗∗∗
-0.022
∗
[-6.46] [-1.74]
R O A
𝑞 −4, 𝑞 −1
1.330
∗∗∗
0.793
∗∗∗
[15.28] [9.50]
R&D Exp ense
𝑞 −4, 𝑞 −1
0.601
∗∗∗
-0.403
∗∗∗
[4.21] [-2.80]
Financial Lev erage
𝑞 −1
-0.073 0.040
[-1.34] [0.76]
Net External Financing
𝑞 −4, 𝑞 −1
-0.050
∗∗
-0.011
[-2.05] [-0.44]
ln(Mark et Equit y
𝑞 ) -0.447
∗∗∗
-0.385
∗∗∗
[-24.99] [-22.44]
Analyst Co v erage
𝑞 -0.027
∗∗∗
-0.019
∗∗∗
[-11.22] [-7.62]
EPS Guidance
𝑞 𝑞 −1]
0.301
∗∗∗
0.154
∗∗∗
[13.47] [6.87]
Beta
(𝑞 −1, 𝑞 )
-0.031
∗
-0.043
∗∗
[-1.77] [-2.48]
Sync hronicit y
(𝑞 −1, 𝑞 )
0.146
∗∗∗
0.143
∗∗∗
[18.68] [19.12]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.395
∗∗∗
-0.490
∗∗∗
[-17.19] [-22.06]
A vg. T rading V olume
(𝑞 − 1, 𝑞 )
-0.025 -0.168
∗∗∗
[-1.09] [-7.33]
𝑟 −1,−1
-0.010
∗∗∗
-0.011
∗∗∗
[-4.93] [-5.45]
𝑟 −20,−2
-0.006
∗∗∗
-0.006
∗∗∗
[-9.39] [-9.20]
Pseudo 𝑅 2
0.002 0.002 0.013 0.016 0.078 0.092
Quarter FE No Y es Y es Y es Y es Y es
Industry FE No No No No No Y es
N 144,847 144,847 144,847 144,847 144,847 144,847
93
T able 14 (con tin ued)
P anel B: Returns
𝑟 0,1
(1) (2) (3) (4) (5) (6) (7)
Rank ed Disclosure Distance
𝑞 0.291
∗∗∗
0.291
∗∗∗
0.303
∗∗∗
0.218
∗∗∗
0.215
∗∗∗
0.047
[4.95] [4.94] [5.00] [3.61] [3.56] [0.80]
Rank ed Disclosure Distance
𝑞 −4
0.025 -0.050 -0.108
∗
-0.110
∗
0.012
[0.42] [-0.81] [-1.75] [-1.78] [0.20]
ln(Bo ok-to-mark et
𝑞 −1
) 0.088
∗∗
0.101
∗∗
0.063
[2.00] [2.28] [1.45]
R O A
𝑞 −4, 𝑞 −1
1.226
∗∗∗
1.186
∗∗∗
2.293
∗∗∗
[4.01] [3.76] [7.39]
R&D Exp ense
𝑞 −4, 𝑞 −1
1.140
∗∗
1.121
∗∗
0.363
[2.26] [2.06] [0.67]
Financial Lev erage
𝑞 −1
0.393
∗∗
0.478
∗∗∗
0.454
∗∗∗
[2.24] [2.62] [2.58]
Net External Financing
𝑞 −4, 𝑞 −1
-0.311
∗∗∗
-0.333
∗∗∗
-0.114
[-3.07] [-3.27] [-1.14]
ln(Mark et Equit y
𝑞 ) 0.226
∗∗∗
0.234
∗∗∗
0.210
∗∗∗
[4.26] [4.34] [3.98]
Analyst Co v erage
𝑞 0.002 0.001 0.006
[0.32] [0.19] [0.85]
EPS Guidance
𝑞 𝑞 −1]
0.204
∗∗∗
0.219
∗∗∗
0.000
[2.95] [3.09] [0.00]
Beta
(𝑞 −1, 𝑞 )
0.061 0.051 0.067
[0.77] [0.64] [0.86]
Sync hronicit y
(𝑞 −1, 𝑞 )
-0.005 -0.007 -0.023
[-0.19] [-0.29] [-0.93]
Illiquidit y
(𝑞 −1, 𝑞 )
-0.171
∗∗
-0.180
∗∗
-0.229
∗∗∗
[-1.96] [-2.06] [-2.65]
A vg. T rading V olume
(𝑞 −1, 𝑞 )
-0.389
∗∗∗
-0.400
∗∗∗
-0.457
∗∗∗
[-4.06] [-4.13] [-4.77]
𝑟 −1,−1
-0.119
∗∗∗
-0.119
∗∗∗
-0.151
∗∗∗
[-6.24] [-6.24] [-7.56]
𝑟 −20,−2
-0.009
∗∗
-0.009
∗
-0.022
∗∗∗
[-1.96] [-1.96] [-4.87]
𝑟 0,1 𝑞 −4
-0.006
∗
-0.007
∗
-0.002
[-1.84] [-1.85] [-0.65]
ln(Idiosyncratic V olatilit y
−150, −1
) -0.081 -0.086 -0.023
[-0.90] [-0.94] [-0.26]
DSUE
𝑞 0.817
∗∗∗
[65.06]
A djusted 𝑅 2
0.000 0.001 0.001 0.001 0.005 0.005 0.074
Quarter FE No Y es Y es Y es Y es Y es Y es
Industry FE No No No No No Y es Y es
N 144,847 144,847 144,847 144,847 144,847 144,847 144,847
94

Abstract (if available)

Abstract I hypothesize that investors view earnings announcements preceded by distant disclosure as being riskier than announcements preceded by recent disclosure. Consistent with this hypothesis, distant disclosing stocks have greater idiosyncratic jump risk at earnings announcements. Distant disclosing stocks also earn positive abnormal announcement returns, suggesting that investors anticipate the greater idiosyncratic jump risk and demand a risk premium to hold these stocks at earnings announcements. Additional empirical tests drawing on the theoretical model in Merton (1987) support the investor pricing of idiosyncratic jump risk explanation.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Understanding the disclosure practices of firms affected by a natural disaster: the case of hurricanes

PDF

Does differential sensitivity to aggregate earnings shocks drive post-earnings-announcement drift?

PDF

Expectation dynamics and stock returns

PDF

Picky customers and expected returns

PDF

“What you see is all there is”: The effects of media co‐coverage on investors’ peer selection

PDF

Voluntary disclosure responses to mandated disclosure: evidence from Australian corporate tax transparency

PDF

Private disclosure contracts within supply chains and managerial learning from stock prices

PDF

Accrual quality and expected returns: the importance of controlling for cash flow shocks

PDF

Essays on the effect of cognitive constraints on financial decision-making

PDF

Dividend policy, earnings announcements, and analysts' earnings forecasts

PDF

Implied equity duration: international evidence

PDF

Essays in empirical asset pricing

PDF

Insider editing on Wikipedia

PDF

Three essays in derivatives, trading and liquidity

PDF

How large are the coinsurance benefits of mergers?

PDF

Changing fundamental analysis in the new economy: the case of DuPont analysis and STEM firms

PDF

The term structure of CAPM alphas and betas

PDF

Suicide talk as a vital sign: a theory-informed examination of individual and relational factors that influence suicidal disclosure

PDF

Essays on the firm and stakeholders relationships: evidence from mergers & acquisitions and labor negotiations

PDF

Social movements and access to credit

Asset Metadata

Creator Erhard, Ryan D. (author)

Core Title Disclosure distance and earnings announcement returns

School Marshall School of Business

Degree Doctor of Philosophy

Degree Program Business Administration

Degree Conferral Date 2022-08

Publication Date 07/08/2022

Defense Date 05/03/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag asset pricing,disclosure,earnings announcements,idiosyncratic risk,jump risk,OAI-PMH Harvest,stock returns

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Sloan, Richard (committee chair), Dechow, Patricia (committee member), Jones, Christopher (committee member), Ogneva, Maria (committee member), Soliman, Mark (committee member)

Creator Email ryan.erhard.phd@marshall.usc.edu,ryanderhard@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111369227

Unique identifier UC111369227

Legacy Identifier etd-ErhardRyan-10817

Document Type Dissertation

Format application/pdf (imt)

Rights Erhard, Ryan D.

Type texts

Source 20220708-usctheses-batch-951 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu