Are CEOs incentivized to shelter good information?

• DOI: http://doi.org/10.1111/fire.12249
• Author: Hongrui Feng,Yuecheng Jia
• Date: 01 February 2021
• Published date: 01 February 2021

DOI: 10.1111/fire.12249

ORIGINAL ARTICLE

Are CEOs incentivized to shelter good

information?

Hongrui Feng1YuechengJia2

1Finance Department, Sam and Irene Black

School of Business, PennState Behrend, Erie,

Pennsylvania

2Chinese Academyof Finance and

Development,Central University of Finance

and Economics, Beijing, China

Correspondence

HongruiFeng, Sam and Irene Black School of

Business,Penn State Behrend, 4701 College

Erie,PA 16563.

Email:hxf51@psu.edu

Abstract

Prior theoretical studies on the agency problem hold dif-

ferent opinions from the empirical literature on two ques-

tions: (a) Are CEOs incentivized to shelter good informa-

tion? (b) Are CEOs incentivized to evenly shelter good and

bad information? This paper demonstrates that CEOs with

high pay-performanceincentives tend to successfully shelter

good information rather than bad information. Furthermore,

CEOs with high pay-performance incentives shelter good

information by using real earnings management and textual

manipulation but not accrual-based earnings management.

These asymmetric information manipulation behaviors help

to decrease corporate cash flow volatility as well as the jump

and crash risk on the stock market.

KEYWORDS

CEO Delta incentive, good information manipulation, real earnings

management, textual manipulation

JEL CLASSIFICATIONS

G34, J33

1INTRODUCTION

Prior literature documents that investors prefer firms with smooth earnings performance over those with volatile

earnings performance (e.g., Huang, 2009; Rountree, Weston, & Allayannis, 2008). The highly volatile earnings per-

formance is viewed as a signal of high probability of cash flow shortfalls and high costs of externalsources of financing

(e.g.,Minton & Schrand, 1999). Thus, to cater investors’ preference and pursue the maximization of self-interest, CEOs,

especially those with high pay-performance incentives (Delta incentives),tend to withhold information from investors

to smooth earnings performance (e.g., Bergstresser & Philippon, 2006; Benmelech, Kandel, & Veronesi, 2010;Tirole,

2010; Peng & Röell, 2008,2014).

Financial Review. 2021;56:109–132. wileyonlinelibrary.com/journal/fire ©2020 The Eastern Finance Association 109

110 FENG ANDJIA

Eventhough it is well documented that CEOs with high Delta incentives tend to hoard information from investors, it

is still an unsolved research question whether theychoose to hide good or bad news. CEOs are under asymmetric con-

straintsto hoard good and bad information. On one hand, the delayed revelation of bad news increases CEOs’ exposure

to the litigation risk, which limits the magnitude of bad news withholding (e.g., Skinner,1997; Skinner & Sloan, 2002).

On the other hand, the accounting conservatism principle sets a stricter standard to recognize profits than losses.

Thus, this principle leavesthe space for CEOs to reserve current good news to mitigate the potential bad performance

in the future.

Given the severe litigation risk to hide bad news and the easiness to reserve good news, a natural question is: Do

CEOs with high Delta incentives tend to hide good news rather than bad news? Our paper employs novel measures

on good and bad information-hoarding and finds that CEOs with high Delta incentives prefer to shelter good news

rather than bad news to smooth earnings performance. Further analysis indicates that CEOs with high Delta incen-

tives hide good news via real earnings manipulation and textual manipulation rather than the accrual-based earnings

management.

Our findings hinge on well-defined measures of good and bad news manipulation. Prior literature typically employs

accrual-based measures. However, accrual-based measures havetwo major drawbacks. First, Rountree et al. (2008)

document that investors do not value lower noncash flow–related earnings volatility. In other words, accrual-based

earnings management cannot deviate investors’ valuation on firms. Therefore, to examine the relationship between

information manipulation and investors’ reaction, one can hardly draw reliable conclusions using accrual-based mea-

sures.

The second and more severe drawback is that other than accrual-based earnings management, CEOs have many

moretools at their disposal to manipulate information. Theycan manipulate the information through CEO media expo-

sure, real manipulation, management earnings guidance, and information bundling through 10-X filings1(e.g., Cohen,

Dey,& Lys, 2008; Fudenberg & Tirole, 1995; Graham, Harvey, & Rajgopal, 2005; Hutton, Marcus, & Tehranian,2009;

Kothari,Shu & Wysocki, 2009;Tasker,1998). Thus, accrual-based measures may only partially capture the CEOs’ good

(or bad) information hoarding behaviors.

Therefore, to design a catch-all measure on corporate good and bad news hoarding, we turn to the R2literature

proposed by Roll (1988) and Morck, Yeung, and Yu(2000) among others. Following this line of studies, the R2in our

study is formed from the regression of individual stock returns on market and industry returns. A higher R2of the

market index model indicates a higher comovement with the market and less firm-specific information. Thus, high

R2indicates that less firm-specific information is observed by investors, and more information is hoarded by CEOs.

The R2measure outperforms accrual-based measures from two aspects: (a) it incorporates investors’ reaction to the

information manipulation and (b) it is a catch-all measure of CEOs’ information withholding behaviors.Thus, a positive

relationbetween R2and CEO pay-performance incentives indicates that high equity incentives drive CEOs to withhold

firm-specific information.

Togauge the type of information (either positive or negative) withholding, we employ novelmeasures—the positive

andnegative R2(R2

+andR2

)2.Specifically, a positive R2(R2

+)for a firm/year is the R2associated with the positive weekly

return sample, measuring the informativeness of a firm’s positive marketperformance. By the same token, a negative

R2for a specific firm/yeargauges the informativeness of a firm’s negative market performance.

Our main finding is that CEOs with high pay-performance incentives are associated with a higher R2

+’s but not R2

’s

in all econometric specifications. As a high R2

+is equivalent to the low informativeness of good market performance,

these findings revealthat highly incentivized CEOs shelter only the good news. CEOs’ asymmetric information manip-

ulation is consistent with the combination of two strands of theoretical studies. First, it is consistent with Fudenberg

andTirole(1995)andTirole(2010),who demonstrate that CEOs with high pay-performance incentives are motivated

110-Xreports include 10-K, 10-Q, and so on.

2The positive and negative R2’s are used in asset pricing literature such as Bris, Goetzmann, and Zhu (2007) to capture the informativeness in good and

badconditions.