The Dog that Did Not Bark: Limited Price Efficiency and Strategic Nondisclosure

• Author: FRANK S. ZHOU,YUQING ZHOU
• Published date: 01 March 2020
• DOI: http://doi.org/10.1111/1475-679X.12296
• Date: 01 March 2020

DOI: 10.1111/1475-679X.12296

Journal of Accounting Research

Vol. 58 No. 1 March 2020

Printed in U.S.A.

The Dog that Did Not Bark: Limited

Price Efficiency and Strategic

Nondisclosure

FRANK S. ZHOU

∗AND YUQING ZHOU †

Received 19 January 2018; accepted 16 December 2019

ABSTRACT

Theory posits that investors can rationally infer the implications of strate-

gic nondisclosure for firm value, pressuring managers to disclose informa-

tion voluntarily. This study documents that the lack of an earnings guidance

predicts an abnormal return of −41 basis points around the subsequent quar-

terly earnings announcement, suggesting that investors do not fully incor-

porate the implications of nonguidance. Further analyses demonstrate that

limitations in price efficiency, driven by investors’ limited attention ands hort-

selling constraints, explain the mispricing of nonguidance and are associated

∗The Wharton School, University of Pennsylvania †Anderson School of Management, Uni-

versity of California at Los Angeles.

Accepted by Christian Leuz. We would like to express our special appreciation to Brett

Trueman. We appreciate the helpful comments and suggestions of two anonymous referees,

David Aboody, Chris Armstrong, Abigail Allen, Philip Berger,Judson Caskey, Ed deHaan, Paul

Fischer, Henry Friedman, Tim Gray, John Heater (discussant), Mirko Heinle, Xing Huang,

John Hughes, Brian Miller (discussant), Allison Nicoletti, Peter Nyberg (discussant), Cather-

ine Schrand, Qin Tan(discussant), Dan Taylor, Rodrigo Verdi,Robert Verrecchia, Aaron Yoon

(discussant), Frank Zhang, workshop participants at CUNY Baruch, the University of Hong

Kong, HKUST accounting symposium, the Wharton School University of Pennsylvania, the

Trans-AtlanticDoctoral Conference, the MIT–Asia Conference, the AAA Annual Meeting, and

the Helsinki Finance Summit. All errors are our own. Frank Zhou gratefully acknowledges the

Dean’s Research Fund from the Wharton School, University of Pennsylvania, for financial

support.

Correction: The SLETTEN reference incorrectly listed Econometrica upon initial publication

of this article. The reference was corrected January 30, 2020 to Review of Accounting Studies.

155

CUniversity of Chicago on behalf of the Accounting Research Center, 2020

156 F.S.ZHOU AND Y.ZHOU

with less guidance issuance. Our results collectively highlight limited price

efficiency as another friction when studying managers’ strategic disclosure

decisions.

JEL codes: G14; M41

Keywords: mispricing of nonguidance; limited attention; short-selling con-

straints; voluntary disclosure

1. Introduction

Theories of discretionary disclosure argue that firms that do not disclose

information suffer from negative capital market consequences, as investors

rationally infer that firms with bad news are more likely to be strategically

withholding the news (e.g., Grossman [1981], Milgrom [1981]).1This cap-

ital market pressure, in turn, prompts management to disclose information

voluntarily. Subsequent studies use this argument to justify why firms are

more likely to disclose information when performance improves and why

they tend not to stop providing guidance once they have begun (e.g., Miller

[2002], Houston, Lev, and Tucker [2010], Chen, Matsumoto, and Rajgopal

[2011]). The amount of pressure that the capital market exerts, however,

depends on the extent that stock prices correctly reflect the implications

of nondisclosure. Mispricing of nondisclosure can change this pressure,

which in turn affects disclosure decisions. This paper investigates whether

the market efficiently prices the implication of the lack of an earnings guid-

ance. After documenting that stock prices do not fully reflect the implica-

tion of nonguidance, we further examine the reasons driving this mispric-

ing and how the mispricing affects firms’ earnings guidance decisions.

We derive our empirical predictions by extending the classic discre-

tionary disclosure model of Grossman [1981] and Milgrom [1981] to

incorporate limitations on price efficiency. Our model shows that, when

stock prices do not fully reflect the implication of nondisclosure, nondis-

closure is interpreted as better news than it should be under rational

expectations. Market pressure to disclose then lessens, which in turn

reduces the equilibrium disclosure level.2

We use disclosure of management earnings guidance as the empirical

setting to test the model’s predictions in three steps. In the first step, we in-

vestigate whether stock prices fully impound the signal from nonguidance.

1Grossman [1981] and Milgrom [1981] demonstrate that firms disclose all information,

absent of any disclosure frictions, because investors rationally interpret nondisclosure as im-

plying the worst news, a phenomenon known as the “unraveling result.” Examples of possible

frictions include proprietary cost [Verrecchia 1983], probabilistic information endowment

[Dye 1985, Jung and Kwon 1988], and so on.

2The mechanism of our model is distinct from those of rational expectation models (e.g.,

Verrecchia [1983], Dye [1985], Jung and Kwon [1988]). In these models, the market is fully

efficient, which precludes predictable returns, whereas our prediction critically depends on

investors not correctly understanding the implication of nondisclosure, which generates pre-

dictable returns.

LIMITED PRICE EFFICIENCY AND STRATEGIC NONDISCLOSURE 157

We confirm that guidance decisions provide value-relevant information

for future firm economic fundamentals. Consistent with prior studies

(e.g., Houston, Lev, and Tucker [2010], Chen, Matsumoto, and Rajgopal

[2011], Sletten [2012]), we find that firms that do not provide manage-

ment earnings guidance (“nonguidance” hereafter) report worse earnings

than firms that do. Limited price efficiency implies that the negative im-

plication of nonguidance would not be fully incorporated into the stock

price prior to the earnings announcement. This predicts that the earnings

announcement–window abnormal return, following nonguidance, will be

negative on average, as investors initially overprice nonguidance and then

observe the earnings information from the announcement, which helps

them correct the mispricing.

Consistent with this prediction, we show that nonguidance predicts an

average size-adjusted return of −41 basis points from two days before the

subsequent earnings announcement until five days afterward. A second re-

search design uses guidance quarters as a benchmark and finds that the

difference in the average earnings announcement–window size-adjusted re-

turn between nonguidance and guidance quarters amounts to −59 basis

points.3Our results hold for factor-adjusted stock returns and are robust

to controlling for risk factors, factors that could change firm value around

earnings announcements, return momentum, and return premium before

earnings announcements that could reverse.4

In the second step, we explore two reasons for the limited price efficiency

of nonguidance quarters. The first is that limits on investors’ attention

prevent them from fully understanding the implication of nonguidance

for firm value. Inattention can simply mean that investors are not aware of

the information because of their limited capabilities to acquire and process

information (e.g., deHaan, Shevlin, and Thornock [2015], Milian [2015],

Blankespoor, deHaan, and Marinovic [2020]). Inattention could also re-

flect “limited strategic thinking.” Studies have shown that economic agents

often do not correctly understand the implications of strategic actions be-

cause inferring information from them is more difficult than interpreting

explicit disclosures (Hirshleifer and Teoh [2003], Brown, Camerer, and

Lovallo [2012]). We construct three measures of attention: the number of

Dow Jones (DJ) Newswires articles from RavenPack, the number of analysts

following a firm, and the first principal component of analyst following

and the number of news articles. Consistent with our prediction, we find

3The economic magnitude compares well to studies of market anomalies. For example,

Sloan [1996] documents that the accrual anomaly has an 11.2% annual portfolio return,

which implies around 34 basis points returns in eight trading days, assuming 252 trading days

per year. Zhang’s [2006] momentum strategy produces monthly abnormal stock returns of

63–263 basis points, which implies around 25–104 basis points returns in eight trading days,

assuming 20 trading days per month.

4Many studies document a positive return premium prior to the earnings announcement,

known as the earnings announcement premium (see, e.g., Aboody, Lehavy, and Trueman

[2010], So and Wang [2014], Chapman [2018]).