You are on page 1of 55

Accepted Manuscript

Title: The Accrual Anomaly in the U.K. Stock Market:


Implications of Growth and Accounting Distortions

Author: Leonidas C. Doukakis Georgios A.


Papanastasopoulos

PII: S1042-4431(14)00080-8
DOI: http://dx.doi.org/doi:10.1016/j.intfin.2014.06.006
Reference: INTFIN 705

To appear in: Int. Fin. Markets, Inst. and Money

Received date: 19-9-2013


Revised date: 22-6-2014
Accepted date: 23-6-2014

Please cite this article as: Doukakis, L.C., Papanastasopoulos, G.A.,The Accrual
Anomaly in the U.K. Stock Market: Implications of Growth and Accounting
Distortions, Journal of International Financial Markets, Institutions and Money (2014),
http://dx.doi.org/10.1016/j.intfin.2014.06.006

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
*Highlights (for review)

Highlights

 We examine the accrual anomaly in the U.K. using total operating accruals
 We decompose total operating accruals into a component capturing growth in
output and a component capturing accounting distortions

t
 The accrual effect on future earnings is driven only by the component attributable

ip
to accounting distortions
 The accrual effect on future stock returns is driven by both the growth and the

cr
accounting distortions components
 The two components act as complements in driving the accrual effect on future
stock returns in the U.K.

us
an
M
ed
pt
ce
Ac

Page 1 of 54
The Accruals Anomaly in the U.K. Stock Market:
Implications of Growth and Accounting Distortions*

t
ip
cr
us
Leonidas C. Doukakis
Faculty of Business and Economics (HEC Lausanne)

an
University of Lausanne
email: leonidas.doukakis@unil.ch
M
d
Georgios A. Papanastasopoulos #
Department of Business Administration
te

University of Piraeus
email: papanast@unipi.gr
p
ce

Revised Draft: June 2014


Ac

*
The authors appreciate helpful comments and suggestions from an anonymous reviewer, Dimitrios
Thomakos, seminar participants at HEC Lausanne and conference participants at the 37th EAA Annual
Congress in Tallinn (2014), at the World Finance Conference in Venice (2014) and at the Annual
Conference of the Hellenic Finance and Accounting Association (2013). The usual disclaimer applies.
# Corresponding author.

Page 2 of 54
The Accrual Anomaly in the U.K. Stock Market:

Implications of Growth and Accounting Distortions

Abstract: On the basis of an accrual decomposition into two components capturing output

growth and accounting distortions, this paper analyzes the effects of accounting accruals on

t
firms’ future performance in the U.K. stock market. Findings reveal a strong negative

ip
association of accruals with future profitability and stock returns. The effect of accruals on

cr
future earnings performance is driven only by the component attributable to accounting

us
distortions, and the accrual effect on stock price performance is driven by both the component

attributable to accounting distortions and the component attributable to growth. These two
an
components complement each other in driving the accrual effect on stock returns.
M
Keywords: accruals, growth, accounting distortions, profitability, stock returns

JEL Classification: M41


ed
pt
ce
Ac

1
Page 3 of 54
1. Introduction

Extensive accounting and finance literature documents the negative association of accounting

accruals with future earnings and stock price performance. This negative relationship, first

demonstrated in the U.S. capital market by Sloan (1996), is called the accrual anomaly, and

t
suggests that firms with high (low) reported accruals in a fiscal period tend to have low (high)

ip
future profitability and stock returns.

cr
Anomalies by their nature challenge existing theory and offer a promising avenue for

us
empirical research. The accrual anomaly directly tests capital market efficiency with respect

to publicly available accounting information. The potential for generalizability and the
an
reasons for accounting anomalies, as well as their implications for market efficiency, continue

to attract researchers’ interest (Chudek et al., 2010; Pincus et al., 2007; Titman et al., 2013;
M
Watanabe et al., 2013). The vast majority of prior literature has explored various aspects of

the accrual anomaly (e.g., earnings persistence, return predictability, economic exploitation,
ed

underlying driving forces) within the U.S. stock market.1 In this paper, we seek to provide

insights into the accrual anomaly by analyzing a sample of firms listed on the London Stock
pt

Exchange, the world’s oldest stock exchange and one of the top three exchanges in total
ce

capitalization.

While researchers agree that the accrual anomaly exists, consensus on its interpretation
Ac

has been elusive. Research is divided into two broad streams: the stream that attributes the

anomaly to accounting distortions (Dechow and Dichev, 2002; Richardson et al., 2005; Xie,

2001) and the stream that attributes the anomaly to growth-related factors (Fairfield et al.,

2003a; Fairfield et al., 2003b; Zhang, 2007; Wu et al., 2010).

1
A partial list includes Barth and Hutton (2004), Beneish and Vargus (2002), Chan et al. (2006), Collins and
Hribar (2002), Core et al. (2008), Dechow et al. (2008), De Fond and Park (2001), Hirshleifer et al. (2012), Lev
and Nissim (2006), and Thomas and Zhang (2002).

2
Page 4 of 54
The first international investigation in the literature shows that the negative

relationship of accounting accruals with future earnings and stock returns can be generalized

to other capital markets and documents the occurrence of the accrual anomaly in the U.K.

(Pincus et al., 2007). However, additional research within the U.K. capital market produces

conflicting results regarding the accrual effect on stock returns. LaFond (2005) and Chan et

t
al. (2006) show that accruals could predict abnormal returns. Soares and Stark (2009, 2011)

ip
document evidence partially consistent with return predictability and profitable exploitation of

cr
the accrual anomaly. Leippold and Lohre (2012) report significant raw returns on accrual

us
trading strategies, but insignificant abnormal returns.

Thus, whether accruals are negatively related with future returns within the U.K. stock
an
market is still under debate. At the same time, none of the above-mentioned studies examines

the implications of accruals for earnings performance.2 More importantly, the contributing
M
role of accounting-based and growth-based factors on the subsequent drift in earnings and

returns attributable to accruals remains unresolved. Despite the several similarities between
ed

the U.S. and the U.K. stock markets (e.g., common law tradition, extensive use of accruals,

lower ownership concentration etc.), several differences exist that could potentially impact the
pt

existence and/or the intensity of the accrual anomaly in the U.K. Differences in the
ce

accounting standards followed, in the emphasis of the stock market on accounting information

(value relevance), in the education and training backgrounds of the market participants, in
Ac

institutional conditions, and in earnings management practices can have an impact on the

accrual anomaly in the U.K. as compared to the accrual anomaly in the U.S. These differences

motivate our attempt to provide additional insights into the accrual anomaly in the U.K. stock

market.

2
Pincus et al. (2007) offer the only exception, providing evidence consistent with the lower persistence of
accruals.

3
Page 5 of 54
The objective of our research is twofold: (1) to examine the possible generalizability

of the accrual anomaly with respect to both future profitability and stock returns, and (2) to

investigate whether the accrual anomaly is driven by accounting-related factors or by growth-

related factors, or both. In accordance with the first objective, we contribute to the literature

by following Richardson et al. (2005, 2006) and considering in our analysis total operating

t
accruals (i.e., change in net operating assets); this is regarded as a more comprehensive

ip
accrual measure than the working capital accruals utilized by prior research into the accrual

cr
anomaly in the U.K. Richardson et al. (2005, 2006) argue that the latter measure omits

us
accruals related to changes in noncurrent operating assets/liabilities, which are also

manifestations of the accrual accounting process (i.e., long-term accruals). As Richardson et


an
al. (2005, 2006) show, this extended measure of total accruals leads to lower earnings

persistence and greater predictable returns relative to working capital accruals in the U.S.
M
stock market.

In accordance with the second objective, we extend prior research by investigating


ed

Richardson et al.’s (2006) accrual decomposition with respect to future earnings and stock

returns. The authors decompose total operating accruals into a growth component and an
pt

efficiency component and examine the components’ implications on earnings persistence in


ce

the U.S. capital market. The growth component captures accruals that are attributable to

growth in output and the efficiency component captures accruals attributable to either
Ac

accounting distortions or less efficient use of existing capital (Jansen et al., 2012). Richardson

et al. (2006) show that both components contribute to the lower persistence of accruals.

In considering Richardson et al.’s (2006) accrual decomposition, we are interested in

assessing the importance of accounting- and growth-based factors underlying the possible

occurrence of the accrual anomaly in the U.K. stock market. To the best of our knowledge,

4
Page 6 of 54
this study is the first systematic attempt to investigate the predictive ability of Richardson et

al.’s (2006) accrual components for future returns.

Our findings can be outlined as follows. The accrual anomaly, first documented by

Sloan (1996), occurs in the U.K. stock market: accruals are negatively related to future

operating profitability and stock returns. The economic significance of this relationship is

t
summarized by the earnings and stock price performance of decile portfolios formed on the

ip
magnitude of total accruals.3 An accrual hedge portfolio (i.e., taking a long position on firms

cr
within the lowest accrual portfolio and a short position on firms within the highest accrual

us
portfolio) consists of firms that experience an increase in their future operating profitability.

In a similar vein, we show that the accrual hedge portfolio earns large positive and highly

significant raw and abnormal returns in the future. an


In contrast to Richardson et al. (2006), we show that only the efficiency component
M
leads to lower earnings persistence. The growth component is not significantly related to

future operating profitability. Similar insights emerge from the performance of portfolios
ed

formed on the magnitude of accrual components. On the one hand, a hedge portfolio based on

accruals attributable to accounting distortions (i.e., long on firms with low accounting
pt

distortions and short on firms with high accounting distortions) exhibits positive growth in
ce

operating profitability in the subsequent period. On the other hand, a hedge portfolio based on

accruals attributable to growth (i.e., long on firms with low growth and short on firms with
Ac

high growth) is characterized by an insignificant change in future operating profitability.

Thus, only accounting distortions play an important role in explaining the lower persistence of

accruals within the U.K. stock market.

At the same time, we find that both accrual components contribute to the negative

relationship between accounting accruals and stock returns. Hedge portfolios on accruals

3
Richardson et al. (2005, p. 478) and Dechow et al. (2008, p. 558), among others, argue that the economic
significance of the accrual anomaly can be summarized by the performance of ranked portfolios on accruals,
especially from the resulting hedge accrual portfolio.

5
Page 7 of 54
attributable to both growth and accounting distortions generate large positive and highly

significant raw and abnormal stock returns. The results are robust to different econometric

techniques as well as across different sub-periods marked by structural changes in the U.K.

stock market.

We also show that neither component subsumes or dominates the other in predicting

t
future returns. In particular, the effect of the growth component on stock returns is more

ip
severe after controlling for accounting distortions and vice-versa. Thus, accounting- and

cr
growth-based factors appear to complement each other in driving the accrual effect on stock

us
returns within the U.K. stock market.

The remainder of the paper is organized as follows. In section 2 we present our


an
hypotheses and research design, and in section 3 we present our data, sample formation and

variable measurement. In section 4 we critically discuss our results. Finally, in section 5, we


M
offer concluding remarks.
ed

2. Hypothesis development and research design


pt

2.1. Hypothesis development


ce

The accrual anomaly was first documented by Sloan (1996) in the U.S. equity market and
Ac

refers to the finding that firms with high (low) accruals experience low (high) earnings

performance and stock returns in the future. Subsequent research, based on U.S. data,

documents supporting evidence of its robustness and raises the question of whether the

accrual effect also occurs in other countries. Pincus et al. (2007) show that the accrual

anomaly can be generalized to other capital markets: accruals have low persistence in 13 of 19

counties (other than the U.S.) examined and generate superior abnormal returns in 10 of 19

6
Page 8 of 54
countries (other than the U.S.) examined. Evidence associated with the generalizability of the

accrual effect on stock returns has also been reported (LaFond, 2005; Leippold and Lohre,

2012). Recognizing that the accrual anomaly is not merely a freak occurrence in the U.S.,

researchers have attempted to examine its implications by focusing more closely on major

stock markets in terms of market capitalization and especially on that of the United

t
Kingdom.4

ip
Regarding the implications of accruals on future earnings performance, Pincus et al.

cr
(2007) provide evidence consistent with the lower persistence of accruals within the U.K.

us
stock market. However, we know of no other study relying on U.K. data to examine the

accrual effect on future earnings performance. At the same time, related evidence is
an
ambiguous regarding the existence of the accrual effect on stock returns.

LaFond (2005) and Chan et al. (2006) document that accrual hedge portfolios (i.e.,
M
long position on the lowest accrual portfolio and short position on the highest accrual

portfolio) earn significant size-adjusted returns and risk-adjusted alphas from the Fama-
ed

French (1993) three-factor model. Soares and Stark (2009) document that when transaction

costs are taken into account, profitable exploitation of accrual hedge portfolios is not
pt

necessarily possible. Leippold and Lohre (2012) show that when risk-adjusted alphas from the
ce

Carhart (1997) four-factor model are considered, the performance of accrual hedge portfolios

is insignificant. In follow-up research, Soares and Stark (2011) report evidence against the
Ac

existence of the accrual anomaly after controlling for risk and other variables (book-to-market

ratio, earnings yield, research and development expense to market-value ratio, growth rate in

capital expenditures, leverage, and cash flow from operations).

4
For the occurrence of the accrual anomaly in Japan, the other major stock market, Pincus et al. (2007) report
lower accrual persistence and insignificant size-adjusted returns from accrual hedge trading strategies. LaFond
(2005) and Leippold and Lohre (2012) document significant risk-adjusted alphas from the Fama-French (2003)
three-factor model and the Carhart (1997) four-factor model from accrual hedge trading strategies.

7
Page 9 of 54
The motivation for this study is a desire to extend prior investigations on the accrual

anomaly in the U.K. context that find little evidence on the accrual effect on future

profitability and conflicting results on the accrual effect on future stock returns. Despite the

several similarities between the U.S. and the U.K. stock markets (e.g. common law tradition,

extensive use of accruals, lower ownership concentration etc.) that, according to prior

t
literature, are related to the accrual anomaly in an international context, a number of

ip
arguments suggest that focusing on the U.K. could enhance our understanding of the accrual

cr
anomaly.

us
First, accruals that drive the accrual anomaly are independent of the market but are

determined by the accounting system and, in particular, by the application of the measurement
an
rules prescribed under the accounting standards. Although U.K. and U.S. GAAP are not

fundamentally different, Weetman et al. (1998) identify a number of de jure and de facto
M
differences, leading them to conclude that there is increasing disharmony between the two

sets of standards. Similarly, Weetman and Gray (1991) analyzing a sample of UK firms that
ed

filed the Form 20-F with the SEC find that U.K. GAAP are significantly less conservative

than U.S. GAAP. Kaserer and Klinger (2008) find evidence consistent with the accrual
pt

anomaly for German firms that voluntarily adopted IFRS, but no evidence of such an anomaly
ce

for firms that continued using German GAAP. Chan et al. (2009) report a decrease in the

accrual anomaly following the application of the FRS 3 in the U.K. The above evidence
Ac

stresses the role of the accounting standards and measurement methods as potential

determinants of the accrual anomaly and underlines the importance of examining the

existence of the accrual anomaly in the U.K. context.

Second, even if one finds support for the existence of the accrual anomaly in the U.K.,

this evidence does not necessarily imply that the accrual anomaly is driven by the same

underlying factors as it is in the U.S. Third, differences in the value relevance of earnings and

8
Page 10 of 54
book values across countries potentially due to the differential emphasis of the market on

accounting information provides another motivation to examine the presence of the accrual

anomaly outside the U.S.

Fourth, as Soares and Stark (2011) suggest, educational and training backgrounds

might vary among investors operating in different countries and stock markets. Different

t
backgrounds might give rise to different forms of irrationalities. Taking into account that the

ip
accrual anomaly was initially regarded as a form of irrationality in the U.S., it seems

cr
reasonable to examine its existence in the U.K. stock market.

us
Finally, although the U.S. and U.K. have similar legal environments, there are a

number of institutional differences that could have an impact on the existence and/or the
an
intensity of the accrual anomaly. Hofstede (2001) points to a number of organizational

differences. Other studies refer to a number of corporate governance-related differences such


M
as the composition of boards of directors (Monks and Minow, 2004), the levels of executive

compensation (Coffee, 2005) and the appointment of the auditors (Turnbull, 2005) etc.
ed

Similarly, prior evidence suggest different degrees of earnings manipulation practices

between the U.S. and the U.K. Brown and Higgins (2001) report that U.S. managers
pt

manipulate earnings surprises significantly more compared to the U.K. Wright et al. (2006)
ce

find that managers of U.S. management buyout firms manage earnings downward

significantly more compared to their U.K. counterparts.


Ac

The above arguments underline the importance of examining the existence of the

accrual anomaly in the U.K. context. Despite several similarities between the two countries,

significant differences exist. Those differences suggest that the existence of the accrual

anomaly in the U.K. is an open empirical question. In order to investigate this issue, we

extend prior research by considering in our analysis total operating accruals. Notably, the

current conflicting evidence on the occurrence of the accrual anomaly within the U.K. is

9
Page 11 of 54
based on studies that consider only working capital accruals as a measure of accounting

accruals. As argued by Richardson et al. (2005), due to several accounting issues regarding

the measurement of accounting accruals, total accruals constitute a more comprehensive

accrual measure relative to working capital accruals. The first hypothesis of the study (in

alternative form) is as follows:

t
ip
Hypothesis 1 (H1): The accrual anomaly occurs in the U.K. capital market.

cr
us
Further, we investigate, for the first time in the literature concerning the U.K. stock

market, the importance of contributing factors suggested by existing explanations of the


an
accrual anomaly. Prior research falls into three broad streams that put forward accounting-

based factors and growth-based factors to interpret the anomaly. For this purpose, we use
M
Richardson et al.’s (2006) decomposition of accruals into those attributable to growth and

those attributable to accounting distortions.


ed

The first stream of research adopts the viewpoint that the accrual anomaly can be

explained by earnings fixation. Sloan (1996) hypothesizes that investors naively fixate on
pt

bottom-line earnings and fail to understand the lower persistence of the accrual component of
ce

earnings. Consequently, investors overvalue (undervalue) firms with high (low) accruals.

Sloan (1996) conjectures that the lower persistence of the accrual component of
Ac

earnings is driven by the greater subjectivity involved in the estimation of accruals. Xie

(2001) decomposes accruals into a nondiscretionary (i.e., normal) and a discretionary (i.e.,

abnormal) component and shows that Sloan’s findings are primarily attributable to the latter.

Chan et al. (2006) report similar evidence. Both studies interpret the accrual anomaly as a

consequence of managerial opportunism, with the implicit assumption that managers exploit

discretionary accruals to manipulate earnings. Dechow and Dichev (2002) provide evidence

10
Page 12 of 54
that firms with low accrual quality have less persistent earnings. Richardson et al. (2005) rate

accounting accruals according to their reliability, and document that less reliable accruals lead

to lower earnings persistence and that investors do not fully anticipate the low earnings

persistence, leading to significant predictable stock returns.

The second stream of research adopts the viewpoint that the accrual anomaly can be

t
explained as a more general growth anomaly. Fairfield et al. (2003a) point out that accruals

ip
are components not only of current profitability but also of growth. Working capital accruals,

cr
long-term accruals, and total accruals represent growth in net current operating assets, net

us
noncurrent operating assets, and net operating assets, respectively.5 Fairfield et al. (2003a)

find no statistical differences in the negative relationship between working capital accruals
an
and long-term accruals with future earnings performance and stock returns.

On the basis of these findings, Fairfield et al. (2003a) conjecture that the lower
M
persistence of accruals arises from the interaction between firm growth and the lower rates of

economic profit associated with diminishing marginal returns on increased investment. At the
ed

same time, they claim that investors misunderstand the incremental implications of growth for

future earnings performance, leading to significant security mispricing. In follow-up research,


pt

Fairfield et al. (2003b) argue that the lower persistence of accruals relative to cash flows is
ce

driven by growth in investment base that is not matched by growth in income. Zhang (2007)

shows that the magnitude of the accrual anomaly is greater when accruals are more likely to
Ac

measure growth.

In sum, evidence from the existing research is inconclusive on the underlying driving

sources of the accrual anomaly. The first stream attributes the lower persistence of accruals to

accounting distortions, while the second stream attributes it to growth. Both streams assume

some form of irrationality to interpret the accrual effect on stock returns: naive investors

5
Fairfield et al. (2003a) label long-term accruals as a generic form of growth.

11
Page 13 of 54
ignore the implications of accounting- and growth-based factors for future earnings

performance, leading to significant security mispricing.

However, we cannot discard the possibility that the accrual anomaly can be driven by

both accounting- and growth-based factors. Regarding the lower persistence of accruals,

Dechow et al. (2008) argue that diminishing returns to scale could coexist with susceptibility

t
of accruals to measurement error, since generally accepted accounting principles require the

ip
immediate impairment of most unprofitable investments. Indeed, Richardson et al. (2006)

cr
show that both factors can explain the lower accrual persistence of accruals within the U.S.

us
stock market.

Regarding the accrual effect on stock returns from a mispricing perspective, we follow
an
the argument of Shi and Zhang (2012) that the growth anomaly may to some degree be

consistent with the earnings fixation. The interpretation of the growth anomaly is currently
M
under debate in the literature. Dechow and Sloan (1997), Lakonishok et al. (1994), La Porta

(1996), and La Porta et al. (1997) attribute the growth anomaly to investor error in
ed

expectations: firms with low (high) growth are undervalued (overvalued) because investors

extrapolate their weak (strong) past performance to form pessimistic (optimistic) estimates
pt

about their future performance.


ce

Indeed, earnings fixation can be associated with investors’ upward- (downward-)

biased expectations about the future earnings performance of firms with high (low) accruals. In
Ac

this vein, Sloan (1996) shows that negative (positive) abnormal returns of firms with high

(low) accruals are concentrated around future earnings announcements. Bradshaw et al.

(2001) show that financial analysts’ forecasts are relatively optimistic (pessimistic) for firms

with high (low) accruals. Thus, one could argue that the growth anomaly could also arise from

investors’ fixation on earnings. Conceptually, in such a case accounting distortions and

growth could complement each other in driving the accrual effect on both future earnings and

12
Page 14 of 54
stock returns. Such a possibility could also arise from motives for earnings management.

Executives of firms with increased investment (or related overinvestment) may manipulate

earnings to render diminishing returns to scale less noticeable to investors (Polk and Sapienza,

2009; Wei and Xie, 2008). These motives may be stronger insofar as investors and analysts

extrapolate trends in past performance to form expectations about future performance (Chan

t
et al., 2006).

ip
The third stream of research suggests the accrual anomaly may reflect rational risk

cr
premia. Khan (2008) finds that the accrual anomaly can be explained by a four-factor model

us
motivated by the Intertemporal Capital Asset Pricing Model (ICAPM) and based on Campbell

and Vuolteenaho (2004) and Fama and French (1993). The four risk factors are news about
an
future expected dividends on the market portfolio, news about future expected returns on the

market portfolio, size, and book to market, respectively. Khan (2008) concludes that firms
M
with high accruals better enable investors to smooth intertemporal consumption patterns,

while firms with low accruals expose investors to distress risk.


ed

In follow-up research, Wu et al. (2010) find that after adding an investment-based

return factor into the CAPM and the Fama-French model (1993), the accrual anomaly is
pt

substantially reduced. They treat accruals as capital investment and argue, as suggested by the
ce

neoclassical q theory with real investment frictions (Cochrane, 1991), that firms adjust their

investment in response to discount rate changes.


Ac

Importantly, a rational explanation places growth-based factors in a prominent role in

explaining the accrual effect on stock returns. Empirical evidence suggests that firms with

low (high) sales growth are more likely to have low (high) accruals (Dechow et al., 1998).

Opler and Titman (1994) show that firms with high financial distress experience low sales

growth owing to the aggressive behavior of competitors and risk aversion of customers. In a

similar vein, Fama and French (1992, 1993, 1996) argue that value firms with low growth

13
Page 15 of 54
rates in sales are fundamentally riskier and thus earn lower expected returns than firms with

high growth rates in sales. The type of risk behind this argument is distress risk.

According to q theory, investment growth is a rational response to reductions in the

discount rate. As Dechow et al. (2008) and others argue, accruals measure growth in working

capital (i.e., working capital accruals) and growth in long-term capital (i.e., long-term

t
accruals). When the discount rate falls (rises), more (fewer) investment projects become more

ip
(less) profitable, and thus accruals will increase (decrease) as firm executives will optimally

cr
adjust their investment expenditures upwards (downwards). In turn, firms with high (low)

us
accruals are associated with lower (higher) risk and lower (higher) future returns.

We stress that a rational explanation has two important implications. First, it is silent
an
about the effect of accruals on future profitability. On the one hand, from the viewpoint that

accruals contain distress risk information, no theory or empirical evidence suggests higher
M
(lower) future profitability for high (low) distressed firms with low (high) accruals. On the

other hand, from the viewpoint that the level of accruals arises from optimal investment
ed

growth, one might expect stronger long-term earnings growth for high accrual firms, but these

effects in the short run should be less pronounced owing to the confounding effects of
pt

transitory earnings in the current period (Zhang 2007).6 Second, one cannot make risk-based
ce

assessments for a potential association of accounting distortions with future returns. A risk-

based explanation holds that firms with high (low) accounting distortions should be
Ac

negatively (positively) related to future returns owing to lower (higher) risk. However, the

current evidence from the literature suggests that firms with high accounting distortions are

riskier than firms with low accounting distortions (Francis et al., 2004; Kim and Qi, 2010).

However, mispricing associated with accounting distortions could co-exist with

optimal investment as underlying sources of the accrual effect on stock returns. Wu et al.

6
Note that, as Wu et al. (2008) suggest, investment in working capital and long-term capital as a rational
response to negative changes in the discount rate does not require the presence of decreasing marginal returns to
increased investment (see proposition 2, p. 184).

14
Page 16 of 54
(2008) show that accrual reliability is negatively correlated with real investment growth.

Indeed, overpricing has the power to reduce financing constraints and therefore to lead to

optimal investment by firm executives.

According to Lam and Wei (2012) and Li and Zhang (2010), mispriced firms whose

stocks are more costly to arbitrage face higher financing frictions. Birru (2012) shows that in

t
overvalued countries only firms that are financially constrained invest more and raise more

ip
external finance. Linck et al. (2013) provide evidence that constrained firms with high

cr
discretionary accruals experience higher earnings-announcement returns, obtain more equity

us
and debt financing, and invest more optimally than constrained firms with low accruals. These

findings are consistent with overpricing relaxing financing frictions and leading to value-

enhancing investment.7 an
The above discussion on the role of accounting distortions and growth in driving the
M
accrual effect on future profitability and stock returns leads to the following testable

hypotheses:
ed

Hypothesis 2 (H2): The lower persistence of the accrual component of earnings is attributable
pt

to accounting distortions and/or growth.


ce

Hypothesis 3 (H3): The accrual effect of stock returns is attributable to accounting distortions

and/or growth.
Ac

2.2. Research design

Prior studies on the possible occurrence of the accrual anomaly within the U.K. stock

market have used working capital accruals as the basic accrual measure. Working capital

7
See Baker et al. (2003) and Campello and Graham (2007) for further evidence on the positive relationship
between mispricing and value-enhancing investment.

15
Page 17 of 54
accruals are defined as the change in net working capital (i.e., net current operating assets)

less depreciation expense (Healy, 1985). However, this measure is narrow, since it ignores

accruals relating to net noncurrent operating assets (e.g., capitalized software development

costs, capitalized expenditures, long-term receivables) and only incorporates the reversal of a

subset of long-term accruals through subtraction of depreciation expense.

t
To address this issue, Richardson et al. (2005, 2006) extend the definition of accruals

ip
to include changes in noncurrent asset/liability accounts (i.e., long-term accruals). According

cr
to this definition, total accruals (ACC) represent the percentage change in net operating assets

us
(NOA). NOA are equal to the difference between operating assets (OA) and operating

liabilities (OL). Following Hirshleifer et al. (2004), operating assets are calculated as the
an
residual amount from total assets (TA) after subtracting cash and cash equivalents (C), and

operating liabilities are calculated as the residual amount from total assets after subtracting
M
minority interest (MINT), ordinary and preferred stock (OPS), and total debt. Therefore, NOA

and ACC are equal to:


ed

NOAt  TAt  Ct   TAt  MINTt  OPS t  TDt  (1)

ACC t  NOAt NOAt 1


pt

(2)

The current evidence from the U.S. capital market suggests that this extended measure
ce

of accruals is associated with lower earnings persistence and greater predictable stock return
Ac

relative to both working capital accruals and long-term accruals. Richardson et al. (2005) find

that conditional on current profitability, working capital accruals and long-term accruals are

negatively related to future profitability, with negative coefficients of about -0.116 and -

0.047, respectively. In follow-up research, Richardson et al. (2006) show that total accruals

contribute negatively to the future accounting rate of return with a coefficient of about -0.131.

Further, on the basis of these studies, the hedge size-adjusted returns for working capital

accruals, long-term accruals, and total accruals are 12.8%, 16.5%, and 18%, respectively.

16
Page 18 of 54
Recognizing that total accruals are more properly defined and could have more severe

implications on future earnings and returns relative to working capital accruals, we consider

total accruals to be the basic accrual measure in our study.

To advance the understanding of the underlying causes of the accrual anomaly with

respect to both future profitability and stock returns within the U.K. capital market, we rely on

t
Richardson et al.’s (2006) decomposition of total accruals into a growth and an efficiency

ip
component. This decomposition of accruals allows discrimination between the role of growth

cr
and accounting distortions as factors contributing to the accrual anomaly.8 The decomposition

us
is based on the idea that accruals are positively related to growth rate in sales and negatively

related to reductions in NOA efficiency, as captured by the NOA turnover ratio (ratio of sales

to NOA noted by AT): an


ACCt  Sales t Sales t 1  ATt ATt  Sales t 1 Sales t   (ATt ATt ) (3)
M
According to this decomposition, the first term is the growth component of accruals

(SG), the second term is the efficiency component of accruals (ΔAT), and the third term is the
ed

interaction between the growth and efficiency components of accruals (INT). Richardson et
pt

al. (2006) argue that if accruals reflect increased investment, then growth will lead to higher

sales, whereas if accruals increase with no change in sales, then the accrual increase is likely
ce

due to declines in efficiency, either because of accounting distortions or because of the less

efficient use of capital.


Ac

Thus, the growth component is more likely to capture the effects of investment

growth. In contrast, firms that have the largest decrease in the efficiency component are those

that have the greatest concentration of asset-related accounting distortions. In this respect, any

measurement problems associated with the accrual accounting process, such as the

8
As Richardson et al. (2006) point out, this decomposition has two advantages over statistically-oriented
decomposition of accruals into discretionary and nondiscretionary components commonly adopted in the
accounting literature (Chan et al., 2006; Jones, 1991): it controls for non-linearities, and it is an algebraic identity
and thus does not require estimation of any parameters.

17
Page 19 of 54
subjectivity of the accrual measure or earnings management, are more likely to be captured by

the efficiency component than by the growth component of accruals. The interaction term

captures correlated changes between growth and accounting distortions.

Richardson et al. (2006) find that the growth and efficiency components both

contribute to the lower persistence accruals, although the contribution of the efficiency

t
component is stronger. Further, the authors show a strong relationship between the efficiency

ip
component and SEC enforcement actions for alleged earnings manipulation, suggesting the

cr
existence of opportunistic managerial discretion to generate accounting distortions that

us
temporarily inflate earnings. Richardson et al. (2006) conclude that temporary accounting

distortions play the most important role in the lower persistence of the accrual component of
an
earnings, although they cannot rule out a supplementary role for diminishing marginal returns

in increased investment.
M
Richardson et al. (2006) do not extend their analysis to the prediction of future stock

returns, raising a question of how the growth and efficiency components of accruals could be
ed

related to future returns. Starting with the growth component, we follow the existing literature

originated by Graham and Dodd (1934) that predicts a negative relationship between growth
pt

in fundamentals (sales, earnings, cash flows) and future returns: firms with low (high) sales
ce

growth are more likely to have high (low) future returns. This prediction is also consistent

with Richardson et al.’s (2006) decomposition, where accruals are positively related with
Ac

sales growth. If NOA efficiency remains unchanged, increases (decreases) in sales lead to

increases (decreases) in accruals.

According to Richardson et al.’s (2006) decomposition, accruals are negatively related

with the efficiency component. In the absence of sales growth, decreases (increases) in the

efficiency component lead to proportional increases (decreases) in accruals. Thus, the

18
Page 20 of 54
efficiency is expected to be positively related to future returns: firms with low (high)

efficiency component are expected to earn low (high) future returns.

Notably, we cannot make risk-based assessments for this component for two reasons.

First, to the extent that this component captures accounting distortions, the literature suggests

a positive relationship between accounting distortions and risk (Francis et al., 2004; Kim and

t
Qi, 2010) and thus a negative relationship between the efficiency component and future

ip
returns. Second, to the degree that this component reflects less efficient use of existing capital,

cr
we should expect a negative relationship between efficiency and risk and expected returns.

us
Nevertheless, based on Richardson et al.’s (2006) decomposition, the efficiency component of

accruals is expected to be negatively related to accruals, and thus positively related to future

returns. an
As Richardson et al. (2006) point out, a potential limitation of the decomposition is
M
that the growth component may also pick up accounting distortions, such as overstatement of

accounts receivable. At the same time, the properties of the efficiency component are not
ed

attributable to growth. Thus, the ability of the growth component to explain cross-sectional

variation in future profitability and stock returns could be overstated, while the ability of the
pt

efficiency component could be understated to the extent that it is contaminated by the growth
ce

component.

We organize our empirical analysis along three dimensions. First, we investigate the
Ac

possible occurrence of the accrual anomaly by estimating cross-sectional regressions of both

future profitability and future stock returns on total accruals, conditional on current operating

profitability. We also examine the change between future and current profitability and the

level of future stock returns of portfolios formed on the magnitude of total accruals. Second,

we assess the contributing role of growth and accounting distortions by repeating cross-

sectional regression tests and portfolio tests after replacing total accruals with accrual

19
Page 21 of 54
components. Third, in the event that we find that both components could explain the accrual

anomaly, we investigate the possibility that one subsumes or dominates the other.

3. Data, sample formation, and variable measurement

t
3.1. Data and sample formation

ip
cr
Our sample covers all common stocks listed on the London Stock Exchange with available

financial statement and market data (returns, market capitalization) on Worldscope and

us
Datastream International for the period 1980–2009. Closed-end funds, trusts, ADRs, REITs,

an
units of beneficial interest, other financial institutions, and foreign firms are excluded from

the sample. Following Wu and Li (2011), we perform all initial data screenings for basic
M
coding errors via the methods outlined in Ince and Porter (2006). Further, we eliminate firm-

year observations with negative book value of equity, negative value of net operating assets, 9
ed

and insufficient data to compute total accruals, accrual components, current and one-year-

ahead operating profitability (i.e., return on net operating assets), market capitalization, book-
pt

to-market ratio, and one-year-ahead raw and abnormal returns. These criteria yield a final
ce

sample size of 20,675 firm-year observations with nonmissing accounting-based and market-

based variables. All accounting variables are winsorized at the top and bottom 1% of their
Ac

distributions to mitigate the influence of outliers.

3.2. Measurement of accounting variables

We use the indirect (balance sheet) method to measure total accruals (ACC) as the

percentage change in net operating assets (NOA):

9
The criterion associated with the negative value of net operating assets is imposed by Richardson et al. (2006).

20
Page 22 of 54
NOAt  TAt  Ct   TAt  MINTt  OPS t  TDt 

ACC t  NOAt NOAt 1

where:

TAt = Total assets (Worldscope data item 02999).

t
C = Cash and cash equivalents (item 02001).

ip
t

MINTt = Minority interest (item 03426).

cr
OPS t = Ordinary and preferred shares (item 03995).

us
TDt = Total debt (item 03255).

For the growth versus accounting distortions decomposition, sales growth (SG) is
an
measured as the percentage change in sales (data item 01001) and change in NOA turnover
M
ratio (ΔAT) as shown in equation (3):

SGt  Sales t  Sales t 1  Sales t 1


ed

ATt ATt  Salest NOAt   Salest 1 NOAt 1  Salest NOAt 

In accordance with Richardson et al. (2006), we use current and one-year-ahead


pt

operating profitability (i.e., return on net operating assets ~ RNOA) that is measured as
ce

operating profit (data item 01250) deflated by lagged net operating assets. To eliminate the

influence of outliers, all accounting variables (ACC, SG, ΔAT, RNOA, ΔRNOA) are then
Ac

winsorized at the top and bottom 1% of their distribution.

3.3. Measurement of stock returns

The calculation of raw stock returns starts six months after the financial year-end, since

this is the period within which financial statements are required to be published in the U.K. Stock

returns are calculated inclusive of dividends using the return index provided by Datastream

21
Page 23 of 54
(item RI), which is defined as the theoretical growth in the value of a share-holding unit of

equity at the closing price applicable on the ex-dividend date. The raw equity return for a firm

RI j 1
at month j is calculated as: r j  1. 10
Once we get firm-monthly returns, we calculate
RI j

one-year-ahead annual raw stock return (RETt+1) using compounded 12-monthly buy-and-

t
hold returns.

ip
For the measurement of abnormal returns, we follow the characteristic-based

cr
benchmark approach, since recent studies suggest that researchers need to exercise caution

us
when relying on alphas from existing asset-pricing models (Murtazashvilia and

Vozlyublennaiab, 2013). Further, in recognition that in the U.K. stock market adjusting for

an
more than two characteristics is difficult owing to the limited number of listed firms

(compared for example with the U.S. stock market), we choose to calculate size- and book-to-
M
market-adjusted returns (SBMRETt+1). In this way, we control for the size effect and the book-

to-market effect.
ed

To calculate size- and book-to-market-adjusted returns, for each year we sort firms

into four equally weighted portfolios (i.e., quartiles) by market capitalization (MV), and in
pt

each of the resulting portfolios we further sort firms into another four equally weighted
ce

portfolios by the book-to-market ratio (BM). MV is measured as the natural logarithm of

market value of equity (data item 08001), six months after financial year-end. BM is
Ac

calculated as the natural logarithm of the ratio of the book value of equity (data item 03501)

to MV. This procedure results in 16 benchmark portfolios and the matching return is the

annual one-year-ahead weighted average return of all firms in the benchmark portfolio.

10
Following Ince and Porter (2006) and McLean et al. (2009), to avoid the generation of extreme outliers we
delete all the zero returns from the last observation to the first observation with non-zero return, set the returns of
two consecutive months as missing if we observe an increase of over 300% a month and a decrease of more than
50% and delete monthly returns at the top and bottom 1% of their distribution.

22
Page 24 of 54
Thus, the size- and BM-adjusted return (SBMRETt+1) for a firm is the difference

between the raw return (RETt+1) and the matching return of the benchmark portfolio to which

the firm belongs. Note that for the calculation of abnormal returns, if a firm delists during the

period, then the last available return index (RI) before delisting is used to calculate the

delisting return and the proceeds are reinvested into the benchmark portfolio.

t
ip
4. Main empirical results

cr
us
In this section, we present the empirical results of the analysis. We start by examining the

presence or not of the accrual anomaly in the U.K. stock market. This analysis includes
an
regression- and portfolio-based results on both future profitability and future stock returns.

We then examine the importance of accounting- and growth-based factors behind the possible
M
occurrence of the accrual anomaly in the U.K. Again, the conclusions are based on both

regression- and portfolio-based analysis. Finally, we provide evidence on the relative roles of
ed

the growth and accounting distortions in driving the accrual anomaly in the U.K. stock

market.
pt

Table 1 presents the characteristics of decile portfolios based on total accruals (ACC).
ce

Six months after financial year-end, stocks are allocated into decile portfolios and equally

weighted characteristics are computed. ACC_(L_H) is the difference between the lowest and
Ac

highest portfolios. The findings suggest that the most extreme ACC deciles also exhibit the

most extreme growth (SG) and change in efficiency (ΔAT). In other words, both SG and ΔAT

closely track the patterns in accruals. SG monotonically increases with ACC whereas an

inverse relationship occurs between ACC and ΔAT. The differences (ACC_L-H) between the

lowest and the highest portfolios are always statistically significant at below the 5% level.

23
Page 25 of 54
Insert Table 1 about here

Table 2 presents pair-wise correlations between total accruals, accrual components,

and profitability variables. The growth and efficiency components are strongly positively

correlated, suggesting that growing firms experience increases in asset turnover. According to

t
Richardson et al. (2006), this positive correlation indicates a potentially important role for

ip
both components in explaining the accrual anomaly.

cr
us
Insert Table 2 about here

an
We then seek to provide evidence on the lower persistence of accruals in the U.K.

stock market. Following prior literature (Richardson et al., 2005, 2006; Fairfield et al., 2003)
M
we regress future RNOA on current RNOA and ACC (Table 3, Panel A). We estimate annual

cross-sectional regressions over the period 1980–2009 and report the time-series averages of
ed

the parameter coefficients. Reported t-statistics are based on the time-series variation of the

annual regression coefficient estimates. The coefficient on ACC represents the difference
pt

between the accrual and the cash component of RNOA. Consistent with prior literature, the
ce

coefficient γ2 is negative, confirming the lower persistence of accruals in the U.K.

Going a step further, we examine the association between future (one-year-ahead) raw
Ac

returns and current RNOA and total accruals (ACC). Panel B, Table 3 shows a negative and

significant relationship between ACC and future stock returns (RETt+1). The negative

coefficient γ2 (-0.049) confirms the predictive power of total accruals on future stock price

performance. Similar results are reported in Panel C, where raw returns are replaced by the

size- and book-to-market-adjusted returns (SBMRETt+1). These findings provide support for

H1 concerning the occurrence of the accrual anomaly in the U.K. stock market.

24
Page 26 of 54
Insert Table 3 about here

We then investigate the future performance of portfolios based on the magnitude of

total accruals (Table 4). In doing so, we focus on future changes in operating profitability

t
(ΔRNOAt+1), future raw returns (RETt+1), and future abnormal returns (SBMRETt+1) of decile

ip
portfolios on total accruals. The lowest (highest) accrual-based portfolios experience large

cr
positive (negative) changes in future operating profitability. The difference between the

us
lowest and highest portfolios is significantly positive at the 5% level. Similarly, raw returns

are positive for the lowest accruals deciles and decline monotonically as we move from low to
an
high accruals deciles. The difference in average returns between the lowest and highest

accruals decile is 10.9% and is statistically significant at the 1% level. We find similar
M
patterns for the size- and book-to-market-adjusted returns (SBMRETt+1). However, for the

highest accruals portfolios, the abnormal returns are now negative. The mean abnormal
ed

portfolio returns differ by 11.6% between the lowest and highest accruals portfolios in the

first post-formation year. The differences in raw and abnormal returns across the extreme
pt

decile portfolios are consistent with the regression results presented in Table 3, Panels B and
ce

C. These findings underline the economic significance of the accrual anomaly in the U.K.

capital market with respect to both future earnings and stock returns.
Ac

Insert Table 4 about here

To gain insight into the reasons for the occurrence of the accrual anomaly in the U.K.

stock market, we decompose total accruals into the growth component (SG), the efficiency

component (ΔAT), and the interaction term (SG*ΔAT). We estimate annual cross-sectional

25
Page 27 of 54
regressions over the period 1980–2009 and report the time-series averages of the parameter

coefficients. Reported t-statistics are based on the time-series variation of the annual

regression coefficient estimates. Following Richardson et al. (2006), we build the model

gradually. The regression results in Panel A of Table 5 show that the growth component is not

significantly related to future profitability (γ2 = -0,015, t-stat = -0.895). On the other hand, the

t
significantly positive coefficient on –ΔAT is consistent with lower efficiency and accounting

ip
distortions explaining the lower persistence of accruals. Richardson et al. (2006) provide

cr
evidence suggesting a positive relationship between changes in assets turnover (ΔAT) and

us
severe cases of earnings management that result in SEC enforcement actions. Taken as a

whole, these findings attribute a significant role only to earnings management considerations
an
in explaining the negative relationship between accruals and future profitability.

We then extend Richardson et al. (2006) by investigating accrual decomposition with


M
respect to future returns. Panel B of Table 5 presents the results for the future (one-year-

ahead) raw returns. Taking into account that the growth component of accruals does not
ed

contribute to lower earnings persistence, if the negative relationship of accruals with future

returns is driven solely by the growth component, then investors’ misunderstanding of


pt

diminishing marginal returns to increased investment cannot be put forward as an explanation


ce

for the accrual anomaly. As we have explained in developing the hypotheses, such a

possibility seems to be more consistent with an interpretation suggesting optimal investment


Ac

by firm executives in response to discount rate reductions as the driving force of the negative

relationship between accruals and stock returns. On the other hand, if the efficiency

component of accruals (ΔAT) reflects earnings management considerations and ΔAT is the

only driver of the negative relationship of accruals with stock returns, then the earnings

fixation may explain the accrual anomaly in the U.K. At the same time, the possibility

remains that both components could drive the accrual effect on stock returns. Such a case

26
Page 28 of 54
requires examination of whether one component subsumes or dominates the other in

predicting future returns. This approach calls for recognition that a potential limitation of

Richardson et al.’s (2006) decomposition is that the growth component of accruals may also

capture accounting distortions related to sales. At the same time, Wu et al. (2008) show that

accounting distortions are positively correlated with real investment growth.

t
ip
Insert Table 5 about here

cr
us
Table 5, Panel B presents the results for future raw (RETt+1) and abnormal

(SBMRETt+1) returns. In the first row, we include the growth component of accruals. The
an
negative coefficient of SG (-0.051) suggests an important role for this component as a

contributing factor to the accrual anomaly with respect to future returns. However, in the
M
second row, the significance of the coefficient γ3 (0.042) supports an important role for the

efficiency component as well. Findings for the full model (fourth row) suggest that both
ed

components have implications for future stock returns. In other words, the growth and

efficiency components of accruals both explain one-year-ahead stock returns. Similar results
pt

are reported for the size- and book-to-market-adjusted returns. These findings provide support
ce

for H2 and H3.

The subsequent portfolio analyses confirm the above-reported results. Table 6 presents
Ac

future changes in operating profitability and future raw and abnormal returns of portfolios

formed on the growth component of accruals. The lowest growth-based portfolios experience

positive changes in future operating profitability, whereas the highest portfolios generally

exhibit negative changes in the ΔRNOAt+1. However, the difference between the lowest and

highest portfolios is not statistically significant, confirming the insignificant role of growth in

explaining the lower persistence of accruals. Firms in the lowest SG portfolio experience raw

27
Page 29 of 54
returns of 11.3%, while firms in the highest portfolio earn only 4%. The difference between

the lowest and the highest portfolios generates statistically significant returns of 7.3%. Similar

patterns are observed for the size- and book-to-market-adjusted returns. In particular, hedge

portfolios based on the growth component of accruals experience significant abnormal returns

of 9.6%.

t
ip
Insert Table 6 about here

cr
us
Table 7 presents changes in future operating profitability along with raw and abnormal

returns for portfolios formed on the basis of the efficiency component of accruals. Confirming
an
a significant role for accounting distortions in explaining the lower persistence of accruals, the

results show a statistically significant difference between the highest and lowest portfolio for
M
ΔRNOAt+1. Similarly, the hedge portfolio based on the efficiency component of accruals

generates significant raw and abnormal returns of 6.5% and 5% respectively.


ed

Insert Table 7 about here


pt
ce

Taken together, the results suggest that accounting distortions are responsible for the

lower earnings persistence, while growth is not. However, both factors contribute to the
Ac

accrual effect on stock returns.

As we have stressed, a question that naturally arises is whether the two components of

accruals act as substitute or complementary mechanisms in driving the accrual anomaly in the

U.K. To this end, we form portfolios based on the magnitude of the growth component of

accruals conditional on the efficiency component of accruals. In particular, each year, six

months after financial year-end, stocks are first allocated into equally weighted decile

28
Page 30 of 54
portfolios based on the efficiency component of accruals, and subsequently allocated into

equally weighted decile portfolios based on the growth component of accruals. We then

combine all sub-portfolios on the growth component of accruals of each rank within each

decile and report time-series averages of one-year-ahead raw and abnormal returns for each of

these sub-deciles. Table 8 presents the empirical results. The hedge portfolio generates

t
positive and statistically significant raw and abnormal returns. Interestingly, after controlling

ip
for accounting distortions the effect of the growth component on stock returns is even more

cr
important (9.2% as compared to 7.3% for raw returns and 11.7% as compared to 9.6% for

us
abnormal returns). We perform a similar analysis for the efficiency component conditional on

the growth component of accruals. The results reported in Table 9 confirm a significant role
an
for accounting distortions in explaining the accrual anomaly, even after controlling for the

effect of the growth component.


M

Insert Tables 8 and 9 about here


ed

Altogether, the results clearly suggest that the growth and efficiency components act
pt

as complements in driving the accrual effect on future stock returns in the U.K. In other
ce

words, both growth- and accounting-based explanations seem to contribute to the inverse

relationship between current accruals and future stock returns. However, accounting
Ac

distortions seem to be the only driving factor in explaining the lower persistence of accruals

with regard to future earnings. The growth component of accruals does not seem to contribute

to the lower persistence of accruals.

5. Robustness checks

29
Page 31 of 54
In this section, we present results from robustness checks. First, we extend our analysis by

examining whether the accrual anomaly persists over time in the U.K. stock market.11 In

doing so, we seek not only to gain additional insights on one of the most puzzling findings in

accounting research but also to explore the value relevance of accounting figures, such as

accruals, for users of financial information. Such a question has also been investigated by

t
researchers in U.S. capital market setting. Lev and Nissim (2006) present evidence on the

ip
pervasiveness of the anomaly: accrual trading portfolios generate positive returns in 80s, 90s

cr
and up to early 2000s. Green et al. (2011) show that after the 2003 endpoint of the study by

us
Lev and Nissim (2006) the superior performance of these portfolios has been substantially

attenuated to the point that they no longer generate positive returns.


an
Our tests on the persistence of accrual anomaly are conducted conditional on major

events that have affected the structure of the U.K. equity market. First, we divide our sample
M
period into two parts according to the deregulation of the market known as ―Big Bang‖ in

1986 (see also Chambers and Dimson, 2009). The Big Bang refers to the liberalization of
ed

fixed brokerage commissions and the termination of membership restrictions on the London

Stock Exchange.12 Based on this cut-off point, we get two subperiods: 1980 to 1986 (i.e., the
pt

pre-Big Bang period) and 1987 to 2009 (i.e., the post-Big Bang period). In order to capture
ce

the effects from the introduction of the Alternative Investment Market in 1995, the launch of a

new trading system in 1997 and the demutualization of the market in 2000, we further divide
Ac

the post-Big Bang period into additional sub-periods (1987-1995 and 1996-2009).

Other things being equal, we would expect such events to improve market efficiency

(for instance by reducing barriers to arbitrage) and, thus, to moderate the level of any

mispricing pattern. This issue motivates us to investigate the stock price performance of

11
We are grateful to the reviewer for suggesting this additional analysis.
12
The process of deregulation of the world's securities markets began in the U.S. on "Mayday" (May 1, 1975)
with the abolition of fixed commission rates by the Securities and Exchange Commission.

30
Page 32 of 54
portfolios formed on total accruals, after taking into account major structural changes in the

U.K. capital market setting over time. The results are presented in Table 10. The mean raw

return differs by 14.7% between the lowest and highest accrual portfolio in the pre-Big Bang

period. The difference is lower in the post-Big Bang period, but is still highly positive and

significant. It is equal to 9.8% in 1987-2009, 9.2% in 1987-1995 and 10.2% in 1996-2008.

t
The spread in abnormal returns between the lowest and highest accrual deciles is around

ip
11.5% and statistically significant at the 1% level before and after Big Bang. From 1987 to

cr
1995 the spread is equal to 10.6% and from 1996 to 2008 equal to 12.4%. Overall, results are

us
consistent across the different sub-periods and clearly suggest that the accrual anomaly

persists over time in the U.K., even after controlling for major changes that affect the

structure of the stock market.13 an


M
Insert Table 10 about here
ed

Further, following Petersen (2009) we investigate the sensitivity of our regression

results to the use of the clustered standard errors approach (instead of the Fama-MacBeth
pt

standard errors used in the main analysis). In particular, we estimate regressions with the
ce

Ordinary Least Squares (OLS) approach clustered at firm and year level using the pooled

sample over the period 1980–2009. An advantage of the two-way clustering is that it uses all
Ac

available information in one go (by data pooling), thus improving estimation precision and, in

addition, computes standard errors by clustering observations based on their firm-year

characteristics.

Table 11 presents results from regressions of future operating profitability and raw and

abnormal returns on total accruals, after controlling for current operating profitability.

13
We derive consistent results when we divide our sample period into the three respective decades: 80s (i.e.,
1980-1989), 90s (i.e., 1990-1999) and 2000s (i.e., 2000-2009).

31
Page 33 of 54
Table 11
Regressions of future operating profitability and returns on total accruals

We investigate cross-sectional regressions of future (one-year-ahead) operating profitability and raw


and abnormal returns on total accruals, after controlling for current operating profitability. We
estimate regressions with the Ordinary Least Squares (OLS) approach clustered by firm and year using
the pooled sample over the period 1980–2009. We report the resulted parameter coefficients along
with their associated t-statistics (in italics). Bold t-statistics (in italics) indicate significance of
coefficients at below the 5% level (2-tailed). Panel A presents results for future operating profitability,
Panel B for future raw returns and Panel C for future abnormal returns. Sample formation and

t
definition of total accruals are provided in Table 1, definition of current and future (one-year-ahead)

ip
operating profitability is provided in Table 2, and definition of returns is provided in Table 3.

Panel A: RNOAt 1  γ0  γ1 RNOAt  γ2 ACCt  υt 1

cr
0 1 2 R2

us
0.068 0.547 -0.033
0.47
5.115 11.461 -3.229
Panel B: RETt 1  γ0  γ1 RNOAt  γ2 ACCt  υt 1

0.086
2.219
0.015
2.452
an
-0.019
-3.811
Panel C: SBMRETt 1  γ0  γ1 RNOAt  γ2 ACCt  υt 1
0.01
M
-0.014 0.006 -0.012 0.01
-1.64 1.145 -3.962
ed
pt
ce
Ac

52
Page 34 of 54
Empirical results are almost identical to the respective results based on the Fama-MacBeth

approach on Table 3. Accruals exhibit a strong negative relation with future profitability and

stock returns (-0.033, t-stat: -3.229 for RNOAt+1; -0.019, t-stat: -3.811 for RETt+1; -0.012, t-

stat: -3.962 for SBMRETt+1). Thus, these results confirm the existence of the accrual anomaly

in the U.K.

t
ip
Insert Table 11 about here

cr
us
Untabulated results from regressions of accrual decomposition with respect to future

operating profitability suggest that only the efficiency component of accruals has a significant
an
relation to future earnings performance (0.066, t-stat: 5.206). However, consistent with the

results under the Fama-MacBeth approach, both the growth and the efficiency component of
M
accruals have a significant relation to future stock price performance based on both raw and

abnormal returns (growth component: -0.042, t-stat: -5.928 for RETt+1 and -0.036, t-stat: -
ed

4.808 for SBMRETt+1; efficiency component: 0.030, t-stat: 3.676 for RETt+1 and 0.016, t-stat:

4.701 for SBMRETt+1). Thus, the lower persistence of accruals in the U.K. is attributable only
pt

to accounting distortions, while the accrual effect on stock returns is attributable to both
ce

accounting distortions and growth.


Ac

5. Conclusion

This paper examines the existence of the accrual anomaly in the U.K. stock market and its

driving factors. We base our inferences on the total accruals measure, incorporating long-term

accruals that seem to contribute to the accruals puzzle. Following Richardson et al.’s (2006)

32
Page 35 of 54
decomposition of total accruals, we are able to distinguish between the contributing role of

accounting-based and that of growth-based factors.

The empirical findings confirm the existence of the accrual anomaly in the U.K. stock

market. Total accruals are negatively related to future profitability and stock returns, and the

earnings and stock price performance of accrual hedge portfolios provide a meaningful

t
economic summary of this relationship. Further, we show that accounting distortions are

ip
responsible for the lower earnings persistence, while growth is not. However, both factors

cr
contribute to the accrual effect on stock returns. We also show that accounting distortions and

us
growth complement each other in explaining the accrual effect on stock returns.

Our findings are consistent with mispricing of the impact of accounting distortions on
an
future profitability. Indeed, they suggest investor overpricing with respect to that part of

accruals that is attributable to accounting distortions. However, we cannot put forward a


M
similar mispricing hypothesis for the part of accruals attributable to growth, and the predictive

ability of the growth component of accruals seems to be more consistent with a rational
ed

explanation based on the q theory of investment. Indeed, the overpricing on the part of

accounting distortions could reduce financing constraints, leading to value-enhancing


pt

investment by firm executives as a rational response to the reduction in the cost of capital.
ce

Our evidence suggests that an eclectic interpretation may be more appropriate for the

accrual anomaly in the U.K. stock market. We thus echo and extend the view of Soares and
Ac

Stark (2009, p. 321): ―Overall, we conclude that, whilst there is evidence of mispricing

consistent with the accrual anomaly, the profitable exploitation of the anomaly is not

necessarily possible when transactions costs are taken into account. Thus, the accrual anomaly

is not so egregious in the UK as to challenge the semi-strong version efficient markets

hypothesis.‖

33
Page 36 of 54
The world is gray and complex. Market inefficiencies could be responsible for

efficient investment. Overall, we conclude that optimal investment growth motivated by

mispricing on accounting distortions appears to be a consistent interpretation of the accrual

anomaly within the U.K. stock market.

t
ip
cr
us
an
M
ed
pt
ce
Ac

34
Page 37 of 54
References

Baker, M., Stein, J., Wurgler, J. 2003. When does the market matter? stock prices and the

investment of equity-dependent firms. Quarterly Journal of Economics, 118, 969-1005.

Barth., M., Hutton., A.2004. Analyst forecast revisions and the pricing of accruals. Review of

Accounting Studies, 9, 59-96.

t
Beneish, M. D., Vargus, M.E. 2002. Insider trading, earnings quality and accrual mispricing.

ip
The Accounting Review, 77, 755-791.

cr
Birru, J. 2012. Inefficient markets, efficient investment? Working Paper, University of Ohio.

us
Bradshaw, M., Richardson, S., Sloan, R. 2001. Do analysts and auditors use information in

accruals? Journal of Accounting Research, 39, 45-74.


an
Brown, L., Higgins, H., 2001. Managing earnings surprises in the U.S. versus 12 other

countries. Journal of Accounting and Public Policy 20, 373–398.


M
Campbell, J., Vuolteenaho, T. 2004. Bad beta, good beta. American Economic Review, 94,

1249-1275.
ed

Campello, M., Graham, J. 2013. Do stock prices influence corporate decisions? Evidence

from the technology bubble. Journal of Financial Economics, 107, 189–110.


pt

Carhart, M. 1997. On persistence in mutual fund performance. Journal of Finance, 52, 57-82.
ce

Chan, K., Chan, L., Jegadeesh, N., Lakonishok, J. 2006. Earnings quality and stock returns.

Journal of Business, 79, 1041-1082.


Ac

Chan, A. L. C., Lee, E., Lin, S. 2009. The impact of accounting information quality on the

mispricing of accruals: The case of FRS3 in the UK. Journal of Accounting and Public

Policy 28, 189-206.

Chudek, M., Truong C., Veeraraghavan, M. 2011. Is trading on earnings surprises a profitable

strategy? Canadian evidence. Journal of International Financial Markets, Institutions

and Money, 21, 832-850.

35
Page 38 of 54
Cochrane, J. 1991. Production-based asset pricing and the link between stock returns and

economic fluctuations. Journal of Finance, 46, 209-237.

Coffee, J. 2005. A theory of corporate scandals: Why the U.S. and Europe differ. Oxford

Review Economic Policy, 21, 198-211.

Collins, D., Hribar, P. 2002. Earnings-based and accrual-based market anomalies: one effect

t
or two? Journal of Accounting and Economics, 29, 101-123.

ip
Core, J., Guay, W., Verdi, R. 2008. Is accruals quality a priced risk factor? Journal of

cr
Accounting and Economics, 46, 2-22.

us
Dechow, P., Sloan, R. 1997. Returns to contrarian investment strategies: Tests of naïve

expectation hypotheses. Journal of Financial Economics, 43, 3-27.


an
Dechow, P., Kothari, S., Watts, R. 1998. The relation between earnings and cash flows.

Journal of Accounting and Economics, 25, 133-168.


M
Dechow, P., Dichev, I. 2002. The quality of accruals and earnings: The role of accrual

estimation errors. The Accounting Review, 77, 35-59.


ed

Dechow, P., Richardson, S., Sloan, R. 2008. The persistence and pricing of the cash

component of earnings. Journal of Accounting Research, 46, 537-566.


pt

Fairfield, P., Whisenant, J., Yohn, T. 2003a. Accrued earnings and growth: implications for
ce

current profitability and market mispricing. The Accounting Review, 78, 353–371.

Fairfield, P., Whisenant, J., Yohn, T. 2003b. The differential persistence of accruals and cash
Ac

flows for future operating income versus future profitability. Review of Accounting

Studies, 8, 221–243.

Fama, E., MacBeth, J. 1973. Risk, return, and equilibrium: empirical tests. Journal of Political

Economy, 81, 607-636.

Fama, E., French, K. 1992. The cross-section of expected stock returns. Journal of Finance,

47, 427-465.

36
Page 39 of 54
Fama, E., French, K. 1993. Common risk factors in the returns on stocks and bonds. Journal

of Financial Economics, 33, 3–56.

Fama, E., French, K. 1996. Multifactor explanations of asset pricing anomalies. Journal of

Finance, 51, 3–56.

Francis, J., LaFond, R., Olsson, P., Schipper, K. 2004. Costs of equity and earnings attributes,

t
The Accounting Review, 79, pp. 967–1010.

ip
Healy, P. 1985. The effect of bonus schemes on accounting decisions. Journal of Accounting

cr
and Economics, 7, 85-107.

us
Hirshleifer, D., Hou, K., Teoh, S. 2012. The accrual anomaly: risk or mispricing?

Management Science, 58, 320-355.


an
Hirshleifer, D., Hou, K., Teoh, S., Zhang, Y. 2004. Do investors overvalue firms with bloated

balance sheets? Journal of Accounting and Economics, 38, 297-331.


M
Hofstede, G. 2001. Culture’s consequences: Comparing values, behaviors, institutions, and

organizations across nations. 2nd Edition. Thousand Oaks, CA: Sage Publications.
ed

Ince, O., Porter, R.B. 2006. Individual equity return data from Thomson Datastream: handle

with care. Journal of Financial Research, 29, 463-479.


pt

Jansen, I., Ramnath, S., Yohn, T. 2012. A diagnostic for earnings management using changes
ce

in asset turnover and profit margin. Contemporary Accounting Research, 29, 221-251.

Jones, J. 1991. Earnings management during import relief investigations. Journal of


Ac

Accounting Research, 29, 193–228.

Kaserer, C., Klinger, C., 2008. The accrual anomaly under different accounting standards –

lessons learned from the German experiment. Journal of Business Finance and

Accounting 35, 837-859.

Khan, M. 2008. Are accruals really mispriced? Evidence from tests of an intertemporal capital

asset pricing model. Journal of Accounting and Economics, 45, 55-77.

37
Page 40 of 54
Kim, D. Qi, Y. 2010. Accruals quality, stock returns, and macroeconomic conditions. The

Accounting Review, 85, pp. 937–78.

LaFond, R., 2005. Is the accrual anomaly a global anomaly? Working paper, MIT Sloan

School of Management.

Lakonishok, J., Shleifer, A., Vishny, R. 1994. Contrarian investment, extrapolation, and risk.

t
Journal of Finance, 49, 1541–1578.

ip
La Porta, R. 1996. Expectations and the cross section of stock returns. Journal of Finance, 51,

cr
1715-1742.

us
La Porta, R., Lakonishok, J., Shleifer, A., Vishny, R. 1997. Good news for value stocks:

further evidence on market efficiency. Journal of Finance, 52, 859-874.


an
Leippold, M., Lohre, H. 2012. Data snooping and the global accrual anomaly. Applied

Financial Economics, 7, 509-535.


M
Lev, B., Nissim, D. 2006. The persistence of the accruals anomaly. Contemporary Accounting

Research, 23, 193-226.


ed

Lam, E., Wei, J. 2011. Limits to arbitrage, investment frictions, and the asset growth anomaly.

Journal of Financial Economics, 102, 127-149.


pt

Li, D., Zhang, L. 2010. Does q-theory with investment frictions explain anomalies in the
ce

cross-section of returns? Journal of Financial Economics, 98, 297-314.

Linck, J., Netter, J., Shu, T. 2013. Can managers use discretionary accruals to ease financial
Ac

constraints? Evidence from discretionary accruals prior to investment. The Accounting

Review, 88, 2117-2143.

McLean, D., Pontiff, J., Watanabe, A. 2009. Share issuance and cross-sectional returns:

international evidence, Journal of Financial Economics, 94, pp. 1-17.

Monks, A., and N. Minow. 2004. Corporate Governance. 3rd Edition. Malden, MA:

Blackwell Publishing.

38
Page 41 of 54
Murtazashvili, I., Vozlyublennaia, N. 2013. When do characteristics-sorted factors

mechanically explain returns? Journal of International Financial Markets, Institutions

and Money, 21, 119-143.

Opler, T., Titman, S. 1994. Financial distress and corporate performance, Journal of Finance,

49, 1015-1040.

t
Pincus, M., Rajgopal, S., Venkatachalam, M. 2007. The accrual anomaly: international

ip
evidence. The Accounting Review, 82, 169-203.

cr
Polk, C., Sapienza, P. 2009. The stock market and corporate investment: a test of catering

us
theory. Review of Financial Studies, 22, 187-217.

Richardson, S., Sloan, R., Soliman, M., Tuna, I. 2005. Accrual reliability, earnings persistence
an
and stock prices. Journal of Accounting and Economics, 39, 437-485.

Richardson, S., R. Sloan, M. Soliman, and I. Tuna, 2006. The implications of firm growth and
M
accounting distortions for accruals and profitability. The Accounting Review 81, 713-

743.
ed

Shi, L., Zhang, H. 2012. Can the earnings fixation hypothesis explain the accrual anomaly?

Review of Accounting Studies, 17, 1-21.


pt

Sloan, R. 1996. Do stock prices fully reflect information in accruals and cash flows about
ce

future earnings? The Accounting Review, 71, 289-315.

Soares, N., Stark., A. 2009. The accruals anomaly - can implementable portfolios strategies be
Ac

developed that are profitable in the UK? Accounting and Business Research, 39, 321-

345.

Soares, N., Stark., A. 2011. Is there an accruals or a cash flow anomaly in UK stock returns?

Working paper, University of Porto.

Thomas, J., Zhang, H. 2002. Inventory changes and future returns. Review of Accounting

Studies, 7, 163-187.

39
Page 42 of 54
Titman, S., Wei, J., Xie, F. 2013. Market development and the asset growth effect:

international evidence. Journal of Financial and Quantitative Analysis, 48, 1405-1432.

Turnbull, S., 2005. How US and UK auditing practices became muddled to muddle corporate

governance principles. Available at SSRN: http://ssrn.com/abstract=608241

Watanabe, A., Xu, Y., Yao, T., Yu, T. 2013. The asset growth effect: insights from

t
international equity markets. Journal of Financial Economics, 108, 529-563.

ip
Wei, J., Xie, F. 2008. Accruals, capital investments, and stock returns. Financial Analysts

cr
Journal, 64, 34-44.

us
Weetman, P., Gray, S. J., 1991. A comparative international analysis of the impact of

accounting principles on profits: the USA versus the UK, Sweden and The Netherlands.
an
Accounting and Business Research, 2 l, 363-319.

Weetman, P., Jones, E. A. E., Adams, C. A., Gray, S. J. 1998. Profit measurement and UK
M
accounting standards: A case of increasing disharmony in relation to US GAAP and

IASs. Accounting and Business Research, 28, 189-208.


ed

Wright, C. J., Riley Shaw J., Guan, L., 2006. Corporate governance and investor protection:

Earnings management in the U.K. and U.S.. Journal of International Accounting


pt

Research, 5, 25-40.
ce

Wu, Y., Li, Y. 2011. Long-term return reversals—Value and growth or tax? UK evidence.

Journal of International Financial Markets, Institutions and Money, 21, 347-368.


Ac

Wu. J., Zhang, L., Zhang, F. 2010. The q-theory approach to understanding the accrual

anomaly. Journal of Accounting Research, 48, 177-223.

Xie, H. 2001. The mispricing of abnormal accruals. The Accounting Review, 76, 357-373.

Zhang, X. 2007. Accruals, investment, and the accrual anomaly. The Accounting Review, 82,

1333-1363.

40
Page 43 of 54
Table 1
Summary statistics

We examine characteristics (total accruals and accrual components) of portfolios formed on the
magnitude of total accruals over the period 1980–2009. Each year, six months after financial year-end,
stocks are allocated into equally weighted decile portfolios based on total accruals, and time-series
averages of characteristics are computed. Bold t-statistics (in italics) indicate significant differences, at
below the 5% level (2-tailed), between the lowest and the highest accrual portfolio (i.e., the hedge
portfolio on accruals). Sample formation and variable definition are provided in the note below this
table. 

t
Sort on ATt

ip
ACC t SGt
ACC t
ACC t _L -0.461 0.115 0.385

cr
ACC t _2 -0.140 0.078 0.109
ACC t _3 -0.049 0.092 0.062

us
ACC t _4 0.006 0.121 0.031
ACC t _5 0.057 0.129 0.005
ACC t _6
ACC t _7
0.114
0.188
an 0.169
0.207
-0.036
-0.068
ACC t _8 0.308 0.261 -0.134
M
ACC t _9 0.575 0.413 -0.325
ACC t _H 2.783 0.879 -1.437
ed

ACC t _(L-H) -3.244 -0.764 1.822


(-5.961) (-7.386) (6.676)
pt
ce
Ac

Sample Formation
The sample consists of non-financial firms listed on the London Stock Exchange with coverage on Worldscope
and Datastream files over the period 1980– 2009. We eliminate firm-year observations with insufficient data to
compute the primary financial statement variables. These criteria yield a final sample size of 20,675 annual firm-
year observations.
Variable Measurement
ACC t is total accruals, calculated as the percentage change in NOA . NOA is net operating assets, measured
as the difference between operating assets (data item 02999 - data item 02001) and operating liabilities (data
item 02999 - data item 03995 - data item 03255 - data item 03426). SGt is sales growth, calculated as the
percentage change in sales (data item 01001).ATt is change in (net operating) asset turnover, measured as the
change in NOA turnover (ratio of sales (data item 01001) to NOA ) deflated by current NOA turnover ratio.

41
Page 44 of 54
Table 2
Pair-wise correlations

We examine pair-wise correlations between total accruals, accrual components and current and future
operating profitability (i.e., return on net operating assets). Bold numbers indicate significance of pair-
wise correlations at below the 5% level (2-tailed). Definition of current and future operating
profitability is provided in the note below this table. Sample formation and definition of other
variables are provided in Table 1. 

ACC t SGt ATt RNOAt RNOAt 1

t
ACC t 1

ip
0.358 -0.744 0.038 -0.015
SGt 1
0.123 -0.111 -0.091
ATt 1

cr
-0.006 0.043
RNOAt 1
0.541
RNOAt 1 1

us
an
M
ed
pt
ce
Ac

Variable Measurement
RNOAt is return on net operating assets, calculated as operating profit (data item 01250) deflated by lagged net
operating assets ( NOA ). RNOAt 1 is future (one-year ahead) return on net operating assets.

42
Page 45 of 54
Table 3
Regressions of future operating profitability and returns on total accruals

We investigate cross-sectional regressions of future (one-year-ahead) operating profitability and raw


and abnormal returns on total accruals, after controlling for current operating profitability. We
estimate annual cross-sectional regressions over the period 1980–2009 and report the time-series
averages of the parameter coefficients. Reported t-statistics (in italics) are based on the time-series
variation of the annual regression coefficient estimates. Bold t-statistics indicate significance of
coefficients at below the 5% level (2-tailed). Panel A presents results for future operating profitability,
Panel B for future raw returns and Panel C for future abnormal returns. Raw and abnormal returns are
defined in the note below. Sample formation and definition of total accruals are provided in Table 1,

t
while definition of current and future (one-year-ahead) operating profitability is provided in Table 2. 

ip
Panel A: RNOAt 1  γ0  γ1 RNOAt  γ2 ACCt  υt 1

0 1 2

cr
R2
0.059 0.543 -0.062
0.37

us
3.133 9.318 -2.277
Panel B: RETt 1  γ0  γ1 RNOAt  γ2 ACCt  υt 1

0.133 -0.014 -0.049


0.02
2.858 -0.612 -4.668
Panel C: SBMRETt 1  γ0  γ1 RNOAt  γ2 ACCt  υt 1
an
-0.000 -0.034 -0.047
0.01
M
-0.011 -1.954 -4.684
ed
pt
ce

Variable Measurement
Ac

RETt 1 is the one-year-ahead raw return for a firm. The raw equity return at month j is first calculated using the
return index provided in Datastream (item RI), as r j  RI j 1 RI j  1 . RI j is the theoretical growth in the
value of a share holding unit of equity at the closing price applicable on the ex-dividend date. Then, RETt 1 is
calculated using compounded 12-monthly buy-and-hold returns. The 12-month return cumulation period begins
six months after the financial year-end.
SBMRETt 1 is the one-year ahead abnormal return adjusted for size and book to market. Six months after each
financial year-end, firms are first sorted into four equally weighted portfolios (i.e., quartiles) by market
capitalization (data item 08001) and in each of the resulting portfolios firms are further sorted into four equally
weighted portfolios by book-to-market ratio (ratio of data item 03501 to data item 08001). This procedure results
in 16 benchmark portfolios and the matching return is the annual one-year ahead weighted average return for
each benchmark portfolio. SBMRETt 1 is the difference between the RETt 1 and the matching return of the
benchmark portfolio to which the firm belongs.

43
Page 46 of 54
Table 4
Operating and stock price performance of portfolios on total accruals

We examine the change between future (one-year-ahead) operating profitability and current operating
profitability, as well as future (one-year-ahead) raw and abnormal returns of portfolios formed on the
magnitude of total accruals over the period 1980–2009. Each year, six months after financial year-end,
stocks are allocated into equally weighted decile portfolios based on total accruals and time-series
averages of characteristics of profitability changes and returns are computed. Bold t-statistics (in
italics) indicate significant differences, at below the 5% level (2-tailed), between the lowest accrual
portfolio and the highest accrual portfolio (i.e., the hedge portfolio on accruals). Sample formation and
definition of total accruals are provided in Table 1, definition of current and future (one-year-ahead)

t
operating profitability is provided in Table 2, and definition of returns is provided in Table 3.

ip
Sort on RNOAt 1 RETt 1 SBMRETt 1

cr
ACC t
ACC t _L 0.109 0.149 -0.003
ACC t _2

us
0.049 0.155 0.020
ACC t _3 0.033 0.156 0.032
ACC t _4 0.014 0.150 0.021
ACC t _5
ACC t _6
0.009
-0.007
an 0.152
0.137
0.017
-0.001
ACC t _7
M
-0.010 0.114 -0.018
ACC t _8 -0.038 0.104 -0.032
ACC t _9 0.000 0.070 -0.070
ed

ACC t _H -0.137 0.040 -0.119


ACC t _(L-H) 0.246 0.109 0.116
(2.182) (5.599) (5.668)
pt
ce
Ac

44
Page 47 of 54
Table 5
Regressions of future operating profitability and returns on accruals attributable to growth and
accounting distortions

We investigate cross-sectional regressions of future (one-year-ahead) operating profitability, as well as


raw and abnormal returns on accrual components, after controlling for current operating profitability.
We estimate annual cross-sectional regressions over the period 1980–2009 and report the time-series
averages of the parameter coefficients. Reported t-statistics (in italics) are based on the time-series
variation of the annual regression coefficient estimates. Bold t-statistics (in italics) indicate
significance of coefficients at below the 5% level (2-tailed). Panel A presents results for future
operating profitability, Panel B for future raw returns, and Panel C for future abnormal returns. Sample

t
formation and definition of accrual components is provided in Table 1, definition of current and future

ip
(one-year-ahead) operating profitability is provided in Table 2, and definition of returns is provided in
Table 3.

cr
Panel A: RNOAt 1  γ0  γ1 RNOAt  γ2 SGt  γ3 ATt  γ4 SGt * ATt   υt 1

us
0 1 2 3 4 R2

0.066 0.496 -0.015 0.34


4.145 9.381 -0.895
0.062
3.682
0.060
0.518
9.866
0.503
an 0.074
2.641
0.013
0.36

0.34
3.773 9.742 0.299
M
0.065 0.531 -0.040 0.089 0.033
0.38
3.897 9.197 -1.787 2.863 0.908

Panel B: RETt 1  γ0  γ1 RNOAt  γ2 SGt  γ3 ATt  γ4 SGt * ATt   υt 1


ed

0.139 -0.030 -0.051 0.02


2.949 1.173 -2.990
0.134 -0.032 0.042
0.01
pt

2.855 -1.165 3.581


0.133 -0.035 -0.036
0.01
2.808 -1.287 -1.250
ce

0.140 -0.016 -0.081 0.044 0.015


0.02
2.999 -0.682 -4.001 3.360 0.850

Panel C: SBMRETt 1  γ0  γ1 RNOAt  γ2 SGt  γ3 ATt  γ4 SGt * ATt   υt 1


Ac

0.008 -0.048 -0.051 0.01


0.923 -2.301 -4.253
0.001 -0.050 0.036 0.01
0.161 -2.464 2.943
0.000 -0.054 -0.035
0.01
0.106 -2.602 -1.465
0.009 -0.036 -0.077 0.044 0.020
0.02
1.124 -1.983 -5.721 3.181 1.126

45
Page 48 of 54
Table 6
Operating and stock price performance of portfolios on accruals attributable to growth

We examine the change between future (one-year-ahead) operating profitability and current operating
profitability, as well as future (one-year-ahead) raw and abnormal returns of portfolios formed on the
magnitude of the growth component of accruals over the period 1980–2009. Each year, six months
after financial year-end, stocks are allocated into equally weighted decile portfolios based on the
growth component of accruals and time-series averages of profitability changes and returns are
computed. Bold t-statistics (in italics) indicate significant differences, at below the 5% level (2-tailed),
between the lowest portfolio and the highest portfolio (i.e., the hedge portfolio on the growth
component of accruals). Sample formation and definition of the growth component of accruals are

t
provided in Table 1, definition of current and future (one-year-ahead) operating profitability is

ip
provided in Table 2, and definition of returns is provided in Table 3.

RNOAt 1

cr
Sort on SGt RETt 1 SBMRETt 1
SGt _L 0.105 0.113 -0.014

us
SGt _2 0.011 0.152 0.027
SGt _3 0.020 0.165 0.041
SGt _4
SGt _5
0.007
-0.012
an 0.153
0.166
0.018
0.023
SGt _6 -0.043 0.146 0.005
M
SGt _7 -0.065 0.134 -0.007
SGt _8 -0.116 0.105 -0.045
SGt _9
ed

-0.039 0.059 -0.092


SGt _H 0.152 0.040 -0.110
SGt _(L-H) -0.047 0.073 0.096
pt

(-0.410) (3.035) (4.034)


ce
Ac

46
Page 49 of 54
Table 7
Operating and stock price performance of portfolios on accruals attributable to accounting distortions

We examine the change between future (one-year-ahead) and current operating profitability, as well as
future (one-year-ahead) raw and abnormal returns of decile portfolios formed on the magnitude of the
efficiency component of accruals over the period 1980–2009. Each year, six months after financial
year-end, stocks are allocated into equally weighted decile portfolios based on the efficiency
component of accruals and time-series averages of profitability changes and returns are computed.
Bold t-statistics (in italics) indicate significant differences, at below the 5% level (2-tailed), between
the highest portfolio and the lowest portfolio (i.e., the hedge portfolio on the efficiency component of
accruals). Sample formation and definition of the efficiency component of accruals are provided in

t
Table 1, definition of current and future (one-year-ahead) operating profitability is provided in Table

ip
2, and definition of returns is provided in Table 3.

Sort on ATt RNOAt 1

cr
RETt 1 SBMRETt 1
ATt _L -0.160 0.061 -0.084

us
ATt _2 -0.064 0.103 -0.027
ATt _3 -0.030 0.109 -0.024
ATt _4 -0.005 an 0.142 0.009
ATt _5 -0.008 0.121 -0.015
ATt _6 -0.001 0.143 0.006
M
ATt _7 0.017 0.157 0.021
ATt _8 0.043 0.155 0.014
ATt _9
ed

0.083 0.116 -0.022


ATt _H 0.141 0.126 -0.033
ATt _(H-L) 0.301 0.065 0.050
pt

(2.899) (3.859) (3.116)


ce
Ac

47
Page 50 of 54
Table 8
Stock price performance of portfolios on accruals attributable to growth, conditional on accruals
attributable to accounting distortions

We examine the change between future (one-year-ahead) raw and abnormal returns of portfolios
formed on the magnitude of the growth component of accruals, conditional on the efficiency
component of accruals, over the period 1980–2009. Each year, six months after financial year-end,
stocks are first allocated into equally weighted decile portfolios based on the efficiency component of
accruals and subsequently allocated into equally weighted decile portfolios based on the growth
component of accruals. We then combine all subportfolios on the growth component of accruals of
decile rank 1, rank 2, etc., and report time-series averages of one-year-ahead raw and abnormal returns

t
for each of these subdeciles. Bold t-statistics (in italics) indicate significant differences, at below the

ip
5% level (2-tailed), between the lowest portfolio and the highest portfolio (i.e., the hedge subdecile-
portfolio on the growth component of accruals). Sample formation and definition of accrual
components are provided in Table 1, while definition of returns is provided in Table 3.

cr
us
RETt 1 SBMRETt 1
Sort on SGt conditional on ATt
SGt _L 0.134 0.004
SGt _2
SGt _3
an
0.140
0.167
0.019
0.032
SGt _4 0.168 0.034
M
SGt _5 0.151 0.015
SGt _6 0.129 -0.015
ed

SGt _7 0.130 -0.018


SGt _8 0.104 -0.038
SGt _9 0.074 -0.070
pt

SGt _H 0.042 -0.113


SGt _(L-H) 0.092 0.117
ce

(3.441) (5.416)
Ac

48
Page 51 of 54
Table 9
Stock price performance of portfolios on accruals attributable to accounting distortions, conditional on
accruals attributable to growth

We examine the change between future (one-year-ahead) raw and abnormal returns of portfolios
formed on the magnitude of the efficiency component of accruals, conditional on the growth
component of accruals, over the period 1980–2009. Each year, six months after financial year-end,
stocks are first allocated into equally weighted decile portfolios based on the growth component of
accruals and subsequently allocated into equally weighted decile portfolios based on the efficiency
component of accruals. We then combine all subportfolios on the efficiency component of accruals of
decile rank 1, rank 2, etc., and report time-series averages of one-year-ahead raw and abnormal returns

t
for each of these subdeciles. Bold t-statistics (in italics) indicate significant differences, at below the

ip
5% level (2-tailed), between the highest portfolio and the lowest portfolio (i.e., the hedge subdecile-
portfolio on the efficiency component of accruals). Sample formation and definition of accrual
components are provided in Table 1, while definition of returns is provided in Table 3.

cr
RETt 1 SBMRETt 1
Sort on ATt conditional on

us
SGt
ATt _L -0.078 0.069
ATt _2 an
-0.042 0.096
ATt _3 -0.035 0.101
ATt _4 -0.018 0.121
M
ATt _5 -0.020 0.115
ATt _6 -0.005 0.134
ATt _7
ed

0.013 0.142
ATt _8 0.011 0.144
ATt _9 0.009 0.147
ATt _H
pt

-0.005 0.151
ATt _(H-L) 0.073 0.082
ce

(3.589) (4.114)
Ac

49
Page 52 of 54
Table 10

Stock price performance of portfolios on total accruals, conditional on major events that affect the
structure of the U.K. equity market.

We examine the change between future (one-year-ahead) raw and abnormal returns of portfolios
formed on the magnitude of total accruals over the period 1980–2009, conditional on major events that
affect the structure of the U.K. equity market. First we divide our sample period into two sub-periods
conditional on the deregulation of the market known as ―Big Bang‖ in 1986: 1980 to 1986 (i.e., the
pre-Big Bang period) and 1987 to 2009 (i.e., the post-Big Bang period). Subsequently, we also divide
the post-Big Bang period into two parts: 1987 to 1995 and 1996-2009 (to capture the effects from

t
events such as the introduction of the Alternative Investment Market in 1995, the launch of a new

ip
trading system in 1997 and the demutualization of the market in 2000). Then, each year, six months
after financial year-end, stocks are allocated into equally weighted decile portfolios based on total
accruals and time-series averages of characteristics of profitability changes and returns are computed.

cr
Bold t-statistics (in italics) indicate significant differences, at below the 5% level (2-tailed), between
the lowest accrual portfolio and the highest accrual portfolio (i.e., the hedge portfolio on accruals).
Sample formation and definition of total accruals are provided in Table 1, and definition of returns is

us
provided in Table 3.

Panel A : Raw Returns of Portfolios on Total Accruals

1980-1986
an
1987-2009 1987-1995 1996-2009

Sort on RETt 1 RETt 1 RETt 1 RETt 1


M
ACC t
ACC t _L 0.421 0.074 0.115 0.046
ACC t _2 0.388 0.091 0.149 0.051
ed

ACC t _3 0.362 0.1 0.123 0.084


ACC t _4 0.329 0.101 0.106 0.097
pt

ACC t _5 0.385 0.089 0.109 0.075


ACC t _6 0.318 0.088 0.118 0.067
ce

ACC t _7 0.276 0.07 0.076 0.066


ACC t _8 0.306 0.049 0.072 0.033
ACC t _9 0.227 0.027 0.035 0.022
Ac

ACC t _H 0.274 -0.024 0.023 -0.056


ACC t _(L-H) 0.147 0.098 0.092 0.102
(2.615) (5.099) (3.219) (3.824)

50
Page 53 of 54
Panel B : Abnormal Returns of Portfolios on Total Accruals

1980-1986 1987-2009 1987-1995 1996-2009

Sort on SBMRETt 1 SBMRETt 1 SBMRETt 1 SBMRETt 1


ACC t
ACC t _L 0.065 -0.021 -0.007 -0.032
ACC t _2 0.039 0.014 0.044 -0.006

t
ACC t _3

ip
0.05 0.027 0.014 0.036
ACC t _4 -0.018 0.031 0.002 0.052

cr
ACC t _5 0.064 0.005 -0.002 0.01
ACC t _6 -0.018 0.004 0.001 0.007

us
ACC t _7 -0.034 -0.014 -0.034 0.001
ACC t _8 -0.011 -0.038 -0.043 -0.035
ACC t _9
ACC t _H
ACC t _(L-H)
-0.09
-0.048
0.113
an
-0.064
-0.138
0.117
-0.075
-0.113
0.106
-0.057
-0.156
0.124
(2.678) (5.912) (4.268) (4.241)
M
ed
pt
ce
Ac

51
Page 54 of 54

You might also like