Masculinity, Testosterone, and Financial Misreporting

• Author: YACHANG ZENG,LAURENCE VAN LENT,YUPING JIA
• DOI: http://doi.org/10.1111/1475-679X.12065
• Published date: 01 December 2014
• Date: 01 December 2014

DOI: 10.1111/1475-679X.12065

Journal of Accounting Research

Vol. 52 No. 5 December 2014

Printed in U.S.A.

Masculinity, Testosterone,

and Financial Misreporting

YUPING JIA,

∗LAURENCE VAN LENT,

†AND YACHANG ZENG

†

Received 15 May 2013; accepted 6 September 2014

ABSTRACT

We examine the relation between a measure of male CEOs’ facial masculin-

ity and financial misreporting. Facial masculinity is associated with a complex

of masculine behaviors (including aggression, egocentrism, riskseeking, and

maintenance of social status) in males. One possible mechanism for this rela-

tion is that the hormone testosterone influences both behavior and the de-

velopment of the face shape. We document a positive association between

CEO facial masculinity and various misreporting proxies in a broad sample

of S&P1500 firms during 1996–2010. We complement this evidence by docu-

menting that a CEO’s facial masculinity predicts his firm’s likelihood of being

subject to an SEC enforcement action. We also show that an executive’s facial

∗Frankfurt School of Finance & Management; †Center and Department of Accountancy,

Tilburg University.

Accepted by Christian Leuz. We thank Wenjiao (Amanda) Cao, Sujie Chen, Michael

Geulen, Anne Koning, Moritz Kostrzewa, Gita Lieuw, Willem Mobach, Mitzi Perez Padilla,

Tetiana Shevchenko, Violeta Shtereva, and Fei Wang for their help in collecting the data.

We thank the Center for Accounting Research and Education at Notre Dame for provid-

ing us with the data on SEC enforcement releases. We benefited from feedback from Shane

Dikolli, Igor Goncharov, Kelvin Law, Max Mueller, Bill Mayew, Cheryl McCormick, Valeri

Nikolaev, Per Olsson, Thorsten Sellhorn, as well as from workshop participants at Duke, ES-

SEC Singapore, LMU, Tilburg, the Frankfurt School of Finance and Management, the Uni-

versity of Melbourne, University of Queensland, VU University Amsterdam, WHU, and the

2012 European Accounting Association Annual Meeting. We are also grateful for the con-

structive comments by three anonymous reviewers. This study was approved by the Frankfurt

School of Finance and Management Research Ethics Committee. An Online Appendix to this

paper can be downloaded at http://research.chicagobooth.edu/arc/journal-of-accounting-

research/online-supplements.

1195

Copyright C, University of Chicago on behalf of the Accounting Research Center,2014

1196 Y.JIA,L.VAN LENT,AND Y.ZENG

masculinity is associated with the likelihood of the SEC naming him as a per-

petrator. We find that facial masculinity is not a measure of overconfidence.

Finally, we demonstrate that facial masculinity also predicts the incidence of

insider trading and option backdating.

JEL codes: M12; M20; M41; M50

Keywords: misreporting; testosterone; facial masculinity; insider trading;

option backdating

1. Introduction

Interdisciplinary collaborations between social scientists and biologists

open up novel avenues for research. In particular, researchers interested

in financial decision making have used biological measures to help explain

(economic) behavior.1Even relatively simple metrics such as a person’s

height or body mass index appear to have explanatory value (see, e.g., Bo-

denhorn, Moehling, and Price [2012], Case and Paxson [2008], Korniotis

and Kumar [2014]). We focus on a relatively recent line of biological re-

search, which documents that an individual’s facial structure is associated

with a complex of related “masculine” behaviors including aggression, ego-

centrism, risk seeking, and a desire to maintain social status (Carr´

e and Mc-

Cormick [2008], Haselhuhn and Wong [2012], Stirrat and Perrett [2010]).

While in the strict neoclassical model, the personal characteristics of

managers do not matter for corporate decision making, an emerging

stream of work in accounting and finance recognizes that financial report-

ing practices vary predictably with specific individual traits of executives

(Davidson, Dey, and Smith [2013], Dikolli, Mayew, and Steffen [2012],

Schrand and Zechman [2012]). We argue that facial structures meaning-

fully capture variation in those CEO personal characteristics that are asso-

ciated with financial reporting. We focus on the risk of materially misstated

financial statements and test the hypothesis that it varies systematically with

the degree to which the CEO is disposed to masculine behavior.

It is not obvious why the physical traits of individuals should be associated

with their behavior, and much ongoing research attempts to unravel the

mechanisms involved. The available evidence currently allows only specula-

tions. One conjecture is that testosterone, a steroid hormone, underlies the

association between facial structures and behavior (Lefevre et al. [2013]).

Testosterone is thought to influence behavior because it shapes an individ-

ual’s neural circuitry early in life and because the brain responds to changes

in current testosterone levels. Testosterone not only affects brain devel-

opment but is also associated with face shape. Some clinical studies have

documented that testosterone administration causes craniofacial growth

1See, for example, Harlow and Brown [1990], Kuhnen and Knutson [2005], and Cesarini

et al. [2009, 2010].

MASCULINITY,TESTOSTERONE,AND FINANCIAL MISREPORTING 1197

in adolescents (Verdonck et al. [1999]).2Other studies show positive

correlations between more “masculine” facial features and testosterone

(Lefevre et al. [2013]) but can only speculate about the exact mechanisms

involved.

While the mechanisms linking facial features to masculine behavior are

not yet fully understood and remain subject to debate in the neurosciences,

the literature proposes that an individual’s facial features can be used as

a biological marker of masculine behavior (Stirrat and Perrett [2012]).3

Future research might find evidence of a (causal) role of testosterone, but

it is also possible that testosterone is not implicated at all. Indeed, behavior

is a very complex and “plastic” expression of a multitude of factors, only

partly hormonal or even biological.4

Nevertheless, the evidence that documents an association between facial

“masculinity” and (masculine) behavior is strong, and we rely on it for our

empirical strategy to examine the association between a CEO’s facial shape

and misreporting. Our strategy is related to the approach taken by Wong,

Ormiston, and Haselhuhn [2011], who document a positive association be-

tween CEO facial masculinity and financial performance.5

We collect a sample of pictures of 1,136 CEOs from S&P1500 companies

in 2009. We compute a facial structure metric, the facial width-to-height ra-

tio (fWHR), by measuring the distance between cheekbones and the height

of the upper face of each CEO from his picture. We then trace the full em-

ployment history of each CEO to construct a panel data set of 3,909 firm-

years during the 1996–2010 period. We use a composite measure of the

likelihood of accounting manipulation developed by Dechow et al. [2011]

to identify firms that are likely to have materially misstated accounting re-

ports. In our main set of analyses, we examine the relation between a given

CEO’s fWHR and the likelihood of having material accounting misstate-

ments. We find that the likelihood for a substantial risk of misreporting is

2For more evidence from clinical studies that testosterone alters craniofacial morphology,

see, for example, Lindberg et al. [2005], Thornhill and Gangestad [1999], Thornhill and

Moller [1997], and Vanderschuerenand Boullion [1995]. These studiesallow one to speculate

that there might be a relation between a sexual dimorphism (i.e., difference between men and

women) in the face structure and testosterone exposure in adolescence. Lefevre et al. [2013],

however, note that the evidence on sexual dimorphism is weak. These authors propose that

differences in testosterone may be more directly related to facial bone size within men (while

other (hormonal) factors influence the face shape of women).

3Our sample comprises male CEOs only. Thus, throughout the paper, we only consider

men and use the pronoun he.

4In particular, sociocultural practices can reshape behaviors, which highlights the need to

include a range of environmental control variables in the empirical tests (Jablonka, Lamb, and

Zeligowski [2005]). What’s more, hormones are also “social” and are often studied as outcome

variables to examine how they respond to social context.

5Some other authors have also used facial features in economic and/or corporate settings

(Graham, Harvey, and Puri [2010], Rule and Ambady [2008]), but these studies pertain to

perceptions of how competent a CEO looks rather than using the face to infer masculine

behavior.