ACCRUALS AND MOMENTUM
Date | 01 March 2020 |
Published date | 01 March 2020 |
DOI | http://doi.org/10.1111/jfir.12201 |
Author | Ming Gu,Yangru Wu |
The Journal of Financial Research Vol. XLIII, No. 1 Pages 63–93 Spring 2020
DOI: 10.1111/jfir.12201
ACCRUALS AND MOMENTUM
Ming Gu
Xiamen University, China
Yangru Wu
Rutgers University
Abstract
We view accruals as a natural candidate to link to momentum in the context of the
overreaction explanation. Accruals can proxy for ambiguity about the implications of new
information for a firm’s value and can vary with the business cycle. Thus, higher accruals
can lead to greater behavioral biases in the cross‐sectional and time‐series dimensions and,
hence, stronger momentum. Our results show that momentum profitability is mostly
concentrated in firms with high accruals. The previously documented cross‐sectional
characteristics of momentum and market states do not subsume the effect o f accruals on
momentum. We also find that most of the momentum returns among high‐accrual firms
are attributable to high discretionary accruals.
JEL Classification: G1, M4
I. Introduction
Jegadeesh and Titman (1993) first document that momentum strategies of buying past
winner stocks and selling past loser stocks generate statistically significant and
economically large profits. Fama and French (1996) show that their three‐factor model
(Fama and French 1993) does not explain momentum. Fama and French (2008)
highlight momentum’s pervasive effects and demonstrate that it is strong and robust in
all size groups, cross‐sectional regressions, and tests based on different sorting
methods. A variety of risk‐based explanations have been proposed to unravel this
anomaly. Several works demonstrate the significance of momentum for stocks with
certain characteristics in both cross‐sectional and time‐series analyses.
1
We appreciate helpful suggestions and valuable comments from Andres Almazan, Phil Davies, Vivian
Fang, Fangjian Fu, Rients Galema, Lixiong Guo, Armen Hovakimian, Chuan Yang Hwang, David Lesmond,
Ronald Masulis, Tavy Ronen, Rene Stulz, Sheridan Titman, Michela Verardo, Kevin Wang, Frank Zhang, the
editor, the anonymous referee, and participants at the American Finance Association meeting (Chicago), the
doctoral tutorial of the European Finance Association meeting (Stockholm), the Financial Management
Association meeting (Denver), and the Rutgers University seminar. Part of this work was completed while Wu
visited the Central University of Finance and Economics. All remaining errors are our own.
1
Risk‐based explanations include Berk, Green, and Naik (1999), Grundy and Martin (2001), Johnson
(2002), Ahn, Conrad, and Dittmar (2003), Grinblatt and Moskowitz (2004), Korajczyk and Sadka (2004),
Lesmond, Schill, and Zhou (2004), and Wang and Wu (2011). Characteristics‐based analyses include Asness
(1997), Hong, Lim, and Stein (2000), Lee and Swaminathan (2000), Chordia and Shivakumar (2002), Cooper,
Gutierrez, and Hameed (2004), Zhang (2006), Sadka (2006), Sagi and Seasholes (2007), Avramov et al. (2007),
Garlappi and Yan (2011), Antoniou, Doukas, and Subrahmanyam (2013), Wang and Xu (2015), and
Avramov, Cheng, and Hameed (2016).
63
© 2019 The Southern Finance Association and the Southwestern Finance Association
In this article, we link an important firm characteristic, accruals, to momentum
in the context of the overreaction explanation. We find that momentum profitability is
statistically significant and economically large among high‐accrual firms, but is
nonexistent or much weaker for firms with low and medium levels of accruals.
2
The
previously documented cross‐sectional characteristics of momentum and market states
do not subsume the effect of accruals on momentum. Furthermore, most of the
momentum returns among high‐accrual firms are attributable to high discretionary
accruals.
A number of studies (e.g., DeLong et al. 1990; Chan, Jegadeesh, and
Lakonishok 1996; Barberis, Shleifer, and Vishny 1998; Hong and Stein 1999;
Hirshleifer 2001; Daniel, Hirshleifer, and Subrahmanyam 1998, 2001) develop
behavioral models to explain momentum and other financial “anomalies.”For
example, Daniel, Hirshleifer, and Subrahmanyam (1998) model overconfidence as
investors’exaggeration of their information‐processing abilities. Overconfidence
is generally supported by bias in self‐attribution, which means that investors credit
their own talents and abilities for past successes and blame their failures on bad
luck. The authors that self‐attributed investors are overconfident about their stock‐
selection skills and overestimate the precision of their signals. Specifically, such
investors overweigh the value of their private signals and place inadequate weight
on public signals. With greater value ambiguity, investors trade more aggressively
on their private signals.
3
If momentum comes from investors’overconfidence and
self‐attribution, we should expect stronger momentum among firms whose
businesses are harder to value.
Several studies examine implications of the overreaction hypothesis from both
the cross‐sectional and time‐series perspectives. At the cross‐section, Jiang, Lee, and
Zhang (2005) document that momentum strategies work better among firms with
young age, high trading volume, high volatility, and long duration cash flows. Chui,
Titman, and Wei (2010) demonstrate that an individualism index featuring investor
overconfidence helps explain cross‐country differences in momentum. At the time
series, Chordia and Shivakumar (2002) and Cooper, Gutierrez, and Hameed (2004)
argue that during expansionary/bullish states, overconfidence should be stronger, thus
leading to greater momentum. We investigate how accrual‐related information
contributes to behavioral explanations for momentum. We argue that accruals work
as a natural candidate to link to momentum in the context of the overreaction
hypothesis because accruals can enhance behavioral biases in the cross‐sectional and
time‐series dimensions.
2
In this article, we study the effect of accruals on momentum in the cross‐section. Specifically, following
the literature on cross‐sectional study of momentum, we employ accruals as the sorting variable and study
momentum in each accruals group. We leave the time‐series dimension of accrual‐based momentum for future
research.
3
Jiang, Lee, and Zhang (2005) refer to value ambiguity as the degree to which a firm’s value can be
reasonably estimated by even the most knowledgeable investors at reasonable costs. Building on this concept, we
argue that the overconfidence bias is accentuated in high‐ambiguity settings.
64 The Journal of Financial Research
To continue reading
Request your trial