The Asymmetric Effect of Reporting Flexibility on Priced Risk

• Published date: 01 June 2021
• Author: MATTHEW J. BLOOMFIELD
• Date: 01 June 2021
• DOI: http://doi.org/10.1111/1475-679X.12346

DOI: 10.1111/1475-679X.12346

Journal of Accounting Research

Vol. 59 No. 3 June 2021

Printed in U.S.A.

The Asymmetric Effect

of Reporting Flexibility on Priced

Risk

MATTHEW J. BLOOMFIELD∗

Received 15 August 2017; accepted 16 December 2020

Abstract

Most firms covary more positively with downmarkets than upmarkets—a phe-

nomenon I refer to as “risk asymmetry.” I predict and find that risk asymmetry

is caused, at least in part, by a firm’s ability to selectively obfuscate poor per-

formance. Risk asymmetry decreases significantly when firms are required to

adhere to the more stringent auditing standards mandated under Section

404 of the Sarbanes-Oxley Act, however this decrease is more muted for firms

with weak internal controls. Consistent with my predictions, these patterns

are stronger for more market-sensitive firms and weaker for firms that include

relative performance evaluation in their CEOs’ pay packages. Taken together

with prior literature (which documents that risk asymmetry is priced), my

∗The Wharton School of the University of Pennsylvania

Accepted by Mark Lang. I am grateful to Chris Armstrong, Ray Ball, Phil Berger,

Robert Bloomfield, Judson Caskey, Ben Charoenwong, Jung-Ho Choi, Eugene Fama, Joseph

Gerakos, Pat Hopkins, SP Kothari, Christian Leuz (co-advisor), Steven Levitt, Valeri Nikolaev

(co-advisor), Haresh Sapra, Doug Skinner, Kevin Smith, Dan Taylor, Sorabh Tomar, and an

anonymous referee for their many helpful comments. I would also like to thank all participants

in Phil Berger’s, Eugene Fama’s, and Haresh Sapra’s PhD research seminars, as well as work-

shop participants at the University of Chicago, European Accounting Association 2016 Annual

Congress, and American Accounting Association 2016 Annual Meeting for their contributions.

Lastly, I wish to express my gratitude to Peter Iliev for graciously allowing me access to his

hand-collected public float data. An online appendix to this paper can be downloaded at http:

//research.chicagobooth.edu/arc/journal-of-accounting- research/online-supplements.

867

© 2021 The Chookaszian Accounting Research Center at the University of Chicago Booth School of

Business

868 m. j. bloomfield

results suggest that a firm can lower its cost of capital by credibly reducing its

ability to obfuscate value-relevant information.

JEL codes: G14, M41, M42, M49

Keywords: reporting quality; obfuscation; asymmetric risk; Sarbanes-Oxley;

internal controls

1. Introduction

In this study, I develop and test predictions regarding the link between

a manager’s ability to obfuscate value-relevant information, hereafter “re-

porting flexibility,” and systematic risk. In particular, I examine how report-

ing flexibility creates asymmetric systematic risk exposure, whereby the firm

covaries more positively with downmarkets than with upmarkets—a phe-

nomenon I refer to as “risk asymmetry.”

Motivating my predictions are the notions that: (1) poor information

quality causes assessed systematic risk exposure to amplify (e.g., Yee [2006],

Lambert, Leuz, and Verrecchia [2007]; and (2) managers prefer to be less

transparent with bad news than with good news (e.g., Verrecchia [1983],

Dye [1985], Jung and Kwon [1988], Lang and Lundholm [1993], Bloom-

field [2002], Miller [2002], Li [2008], Kothari, Shu, and Wysocki [2009],

Bertomeu, Ma, and Marinovic [2015]). Coupled with (1), managers’ asym-

metric desire to obfuscate will result in greater systematic risk in bad states

than in good states—but only insofar as obfuscation is feasible. If obfusca-

tion is sufficiently costly (e.g., because of stringent audit practices), then

management’s asymmetric incentives to obfuscate will not translate into

asymmetric information quality. Thus, ceteris paribus, I expect firms with

more (less) reporting flexibility, to exhibit greater (lower) risk asymmetry.1

Risk asymmetry is a feature of asset return comovements and cannot be

diversified away in large economies, and is therefore of potential hedging

concern to investors. Prior work by Ang, Chen, and Xing [2006], hereafter

“ACX,” demonstrates that risk asymmetry is positively priced in American

equity markets, even after controlling for common determinants of the cost

of capital (β, size, book-to-market, momentum, liquidity, and coskewness

risk). Thus, empirical support for my predictions would suggest that report-

ing flexibility contributes to a firm’s cost of capital beyond what is currently

captured by common multifactor asset pricing models.

I measure risk asymmetry as the difference between a firm’s downmar-

ket βand its upmarket β.2For my primary measure of reporting flexibil-

ity, I exploit variation in treatment by Section 404 of the Sarbanes-Oxley

Act (“SOX 404”), which requires firms to have their internal controls as-

sessed and reported on by their auditors. I find that firms that are man-

dated to comply with SOX 404 requirements (“SOX 404 reporters”) exhibit

1I provide an analytical derivation of my predictions in the online appendix.

2A detailed description of the construction of this measure can be found in section 3.

the asymmetric effect of reporting flexibility 869

significantly less risk asymmetry than do firms that are not mandated to

comply with SOX 404 requirements (“SOX 404 non-reporters”).

To justify a causal interpretation, I follow Iliev [2010] and use a regres-

sion discontinuity design (“RDD”) to implement an event study based on

plausibly exogenous variation in firms’ exposures to the SOX 404 man-

date. For fiscal years ending after November 15, 2004, domestic firms with

public floats in excess of $75 million (“accelerated filers”) are subject to

more stringent auditing requirements, whereas firms with smaller public

floats (“non-accelerated filers”) are left unaffected. For firms within a nar-

row bandwidth of the $75 million treatment threshold, the assumption

that treatment is “as good as random” allows the estimated difference-in-

differences to be attributed to SOX 404.3I find that, compared to other-

wise similar firms, being an accelerated filer is associated with a roughly 0.3

(0.27 standard deviation) decrease in the firm’s risk asymmetry—but only

after SOX 404 comes into effect. I caveat that the estimated magnitude

should be interpreted as a “local average treatment effect” and is plausibly

larger than for the average public firm in the economy.

Consistent with my predictions, I further find that the effect of SOX 404

is greater for more market-sensitive firms (i.e., firms whose market βsare

large in magnitude). The estimated post–SOX 404 decrease in risk asymme-

try is roughly five times greater for firms of above-median market sensitivity,

compared to firms of below-median market-sensitivity.

Many of the firms in the RDD sample switch from being non-accelerated

filers to accelerated filers, after the initial SOX 404 adoption date. By de-

sign, the RDD specifications disregard these endogenous switches, relying

instead on instrumented filer status based on 2002 public floats. However,

in subsequent analyses, I exploit this variation and analyze changes in firms’

risk asymmetry as they transition from being SOX 404 non-reporters to

SOX 404 reporters. I find that the switch to being a SOX 404 reporter coin-

cides with a substantial drop in risk asymmetry relative to comparably sized

nontransitioning firms.

Lastly, I explore variation in reporting flexibility and risk asymmetry

among SOX 404 reporters. I winnow the sample to include only the firm-

year observations for which an auditor reported an opinion on the firm’s

internal controls, and exploit variation in the auditors’ opinions. I docu-

ment that risk asymmetry is greater for firm-year observations in which the

auditor assessed the internal controls to be weak. With respect to the mag-

nitude of the effect, weak internal controls appear to undo most of the ben-

efits from SOX 404 reporting, vis-à-vis reducing risk asymmetry. On average,

SOX 404 reporters with internal control weaknesses exhibit risk asymmetry

on par with SOX 404 non-reporters. The same is true of a within-firm com-

parison; on average SOX 404 reporters with internal control weaknesses

3Prior work by Gao, Wu, and Zimmerman [2009] and Iliev [2010] suggests that some firms

manage their 2004 public floats to avoid treatment—a violation of random assignment. To

address this concern, I use the IV approach from Iliev [2010].