Accrual or Cash Flow Anomaly Evidence from New Zealand

Description
This paper investigates the presence of the
accrual and the cash flow anomalies in the New
Zealand stock market for the period of 1987 to
2003. We observe insignificant evidence of the
accrual anomaly but find strong evidence of the
presence of the cash flow anomaly.

Accounting Research Journal
Accrual or Cash Flow Anomaly? Evidence from New Zealand
Hardjo Koerniadi Alireza Tourani-Rad
Article information:
To cite this document:
Hardjo Koerniadi Alireza Tourani-Rad, (2007),"Accrual or Cash Flow Anomaly? Evidence from New Zealand", Accounting
Research J ournal, Vol. 20 Iss 1 pp. 21 - 36
Permanent link to this document:http://dx.doi.org/10.1108/10309610780000687
Downloaded on: 24 January 2016, At: 21:01 (PT)
References: this document contains references to 0 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 518 times since 2007*
Users who downloaded this article also downloaded:
Harun Harun, Karen Van Peursem, Ian Eggleton, (2012),"Institutionalization of accrual accounting in the
Indonesian public sector", J ournal of Accounting & Organizational Change, Vol. 8 Iss 3 pp. 257-285 http://
dx.doi.org/10.1108/18325911211258308
Ibrahim El-Sayed Ebaid, (2011),"Accruals and the prediction of future cash flows: Empirical evidence from an emerging
market", Management Research Review, Vol. 34 Iss 7 pp. 838-853http://dx.doi.org/10.1108/01409171111146715
Alan Gregory, Yuan-Hsin Wang, (2013),"Cash acquirers: Can free cash flow, debt and institutional ownership explain long-
run performance?", Review of Behavioural Finance, Vol. 5 Iss 1 pp. 35-57http://dx.doi.org/10.1108/RBF-02-2013-0010
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please visit
www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication
Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Accrual or Cash Flow Anomaly? Evidence from New Zealand

21

Accrual or Cash Flow Anomaly?
Evidence from New Zealand
Hardjo Koerniadi and Alireza Tourani-Rad
Department of Finance
Business School
Auckland University of Technology

Abstract

This paper investigates the presence of the
accrual and the cash flow anomalies in the New
Zealand stock market for the period of 1987 to
2003. We observe insignificant evidence of the
accrual anomaly but find strong evidence of the
presence of the cash flow anomaly. However,
from 1987 to 1992 — a period before the
introduction of the Companies and the Financial
Reporting Acts 1993 — the presence of the
accrual anomaly was statistically significant
suggesting that the introduction of the FRA had
a significant impact on the occurrence of the
anomaly. We observe further that firms with
high discretionary accruals experience
significant negative future stock returns. This
evidence is consistent with the notion that
managers of these firms engage in earnings
management.
1. Introduction
The accounting and finance literature provides
extensive evidence that the magnitude of
accruals (cash flows) component in current
earnings is negatively (positively) correlated
with future stock return. This anomaly
apparently occurs because market participants
use current reported earnings to forecast future
earnings but seem to be uninformed of the
difference in persistence between the accruals
and cash flows components of current earnings

Key Words: Accruals, Cash flows, Anomalies
JEL classification: G14, M41
Acknowledgments: We would like to thank the editor, an
anonymous referee of the Journal and the participants
at the Australasian Finance and Banking Conference,
Sidney, December 2005, and at the New Zealand Finance
Colloquium, Otago, January 2006, for their helpful
comments and suggestions. All remaining errors are ours.
into future earnings. Accruals are less persistent
than cash flows (Bradshaw, Richardson and
Sloan (2001) and Barth and Hutton (2004)).
Consequently, when current earnings are
accompanied by high accruals (cash flows), the
persistence of current earnings is low (high)
which results in lower (higher) than expected
future earnings. When future earnings are
lower (higher) than expected, investors react
negatively (positively) to the earnings
announcements. Thus, the market tends to
overprice (underprice) high accrual (cash flow)
stocks and underprice (overprice) low accrual
(cash flow) stocks. This market fixation on
earnings provides an opportunity to profit from
an arbitrage investment strategy. A hedge
trading strategy, taking a short (long) position in
a high accrual (cash flow) firms and a long
(short) position in a low accrual (cash flow)
firms, would generate a positive and significant
abnormal investment return.
The accrual anomaly was first documented
by Sloan (1996). Sloan finds that the
predictability of stock returns is correlated to the
different persistence of the accruals and cash
flows components of current earnings. Accruals
show mean reversion quicker than cash flows
and are negatively correlated with future stock
returns. He shows that low (high) accrual stocks
generate positive (negative) abnormal future
returns and a hedge strategy that exploits this
anomaly generates a significant annual
abnormal return of 10.4%. Because accruals and
cash flows are negatively correlated, Sloan
argues that a trading strategy of simultaneously
buying high cash flows stocks and selling low
cash flows stocks will also generate a positive
abnormal return. He postulates that the cash
flow anomaly coexists with the accrual
anomaly.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 20 NO 1 (2007) 

22

Collins and Hribar (2000) and Houge and
Lougran (2000) provide further evidence of the
coexistence of the accrual and the cash flow
anomaly. Collins and Hribar (2000) report that
these two anomalies are robust using quarterly
data instead of annual data and are distinct from
the post-earnings announcement drift anomaly.
Houge and Loughran (2000) show that these
two anomalies are robust when applying the
three factor model of Fama and French (1993).
They report, however, that the characteristics of
accrual stocks are different from those of cash
flow stocks, and that the accrual anomaly arises
primarily from the poor performance of high
accrual stocks.
Xie (2001) contends that the accruals
mispricing reported by Sloan (1996) can be
attributed to the discretionary part of accruals.
Xie reports that the market overprices the
discretionary part of accruals more than the
nondiscretionary ones. Discretionary accruals
are used synonymously with earnings
management in the literature (Kothari (2001)).
The mispricing of discretionary accruals (Xie
(2001)) combined with the lower persistence of
accruals on future earnings (Sloan (1996)) and
the poor performance of high accrual stocks
(Houge and Loughran (2000)) indicate that the
accrual anomaly may arise from earnings
management.
Kothari, Sabino and Zach (2005) and Kraft,
Leone and Wasley (2005), however, find that
prior studies on the accrual anomaly suffer from
sample selection bias. Kraft et al. (2005) show
that the accrual anomaly and the cash flow
anomaly are attributed to firms with buy and
hold annual returns of more than 200%. After
eliminating these outliers, which account for
less than 1% of total observations, they find that
both low and high accrual portfolios generate
negative abnormal returns. Further, the
magnitude of the abnormal return of the accrual
strategy is reduced to 1.7%. They also report
that the high cash flow portfolio abnormal
return is reduced from 3.3% to 1.1%. The
abnormal return of the hedge strategy based on
cash flows, however, is still positive at 23%.
In addition to the sample selection problem,
the accrual anomaly documented in the US is
not a global phenomenon. Pincus, Rajgopal, and
Venkatachalam (2005) examine the presence of
the accrual and the cash flow anomalies in 20
countries. They find that the presence of one of
these anomalies does not imply the coexistence
of the other anomaly. They report that the
accrual anomaly, but not the cash flow
anomaly, occurs in certain countries (the
U.S., the U.K., Canada and Australia), while
the opposite is true in 8 other countries.
Furthermore, Pincus et al. (2005) report that the
accrual anomaly tends to occur in countries with
certain institutional and accounting structures.
They find that the occurrence of the accrual
anomaly is correlated with extensive use of
accruals accounting, with a common law
tradition, with weak shareholder protections and
with low share-ownership concentration.
The present study is motivated by two
reasons. First, as discussed by Pincus et al.
(2005) the occurrence of the accrual anomaly
is not a global phenomenon and seems to
be related with a country’s legal system
and corporate governance. New Zealand’s
institutional and accounting structures provide a
setting in which the accrual anomaly is likely to
occur. New Zealand adopts a common law legal
system, allows an extensive use of accruals
accounting (Hung (2001)) and at the same time
it has a rather weak shareholders protection
apparatus in place as measured by large
shareholders blocks (La Porta, Lopez-de-Silanes
and Shleifer (1998), and Walker (2003))
1
.
Particularly, prior to 1993, New Zealand had
poor corporate governance due to inadequacies
of the then existing legislation (Seebold (1993)
and Quigg and Land (1994)). These
shortcomings include insufficient disclosure
requirements, lack of protection for minority
shareholders, weak insider trading regulation
and frequent abuse of controlling shareholdings
(Quigg and Land (1994), p. 40). Claims of poor
compliance with the NZ accounting standards
had also been frequently reported (Bradbury and
Van Zijl (2005)). Although Statements of
Standard Accounting Practice (SSAPs) were in
place, prior to 1993 there was not sufficient legal
backing to ensure its implementation. These

1 Although the antidirector index for New Zealand is
relatively high at 4 based on La porta et al. the securities
regulation regime is notoriously weak. According to
Kusnadi and Wei (2006), the New Zealand public
enforcement index which measures the power of the
capital market supervisory agency in regulating and
enforcing the securities laws, is relatively low at 0.40
compared to Australia at 0.90 and below the average
public enforcement index for Asia Pacific region at 0.59.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Accrual or Cash Flow Anomaly? Evidence from New Zealand

23

problems urged New Zealand to review and
change its corporate laws. The Companies Act
1993 enhances directors’ duties and increases
directors’ responsibilities (Seebold (1993) and
Quigg and Land (1994)) and the Financial
Reporting Act 1993 (FRA93) was introduced to
provide a legal backing to ensure that financial
reports are made in compliance with the
accounting standards. We would, as a spin-off of
our investigation, test the impact of these
regulations on the presence of accrual anomaly
in New Zealand. To the best of our knowledge,
this is the first study investigating the presence
of the accrual anomaly in New Zealand. As the
accrual anomaly is a contradiction to the widely
believed efficient market hypothesis, evidence
from a different country that confirms the
existence or non-existence of this anomaly
would contribute to the existing literature and
could be of interest to New Zealand investors.
Second, Pincus et al. (2005) also report
primarily descriptive evidence that the
occurrence of accruals mispricing does not
imply cash flows mispricing to occur, or vice
versa. The lack of cross-country evidence on the
coexistence of the accrual and the cash flows
anomalies casts doubt on the coexistence
hypothesis of the two anomalies. As the
characteristics of the accruals-based portfolios
are different from those of cash flows-based
portfolios (Houge and Loughran (2000)), this
evidence suggests that the two anomalies,
although accruals and cash flows are negatively
correlated, may not be due to exact same
reasons. Therefore evidence on the (non-)
coexistence of these anomalies would indicate
whether both of these anomalies arise from the
same cause, or each from a different cause.
This paper employs data from 1987 to 2003
and applies a data-selection procedure similar to
that suggested by Kraft et al. (2005). Contrary to
prior studies, we find that, on average, accruals
are not associated with future returns. The
abnormal return based on the accrual strategy,
although positive at 2.56%, is not statistically
significant. The abnormal return of high accrual
firms is significantly negative at 4.13% while the
abnormal return of low accrual firms is negative
but statistically insignificant at 1.57%. Thus, the
positive abnormal return from the accrual
strategy arises mostly from high accrual firms.
The significantly negative abnormal return of the
high accrual stocks indicates that investors
overvalue accruals in high earnings firms.
Further, we find a similar abnormal return
pattern when we sort firms based on two
discretionary accrual models. As discretionary
accruals are positively correlated with firms’
earnings, the negative stock return of the high
accrual firms gives support to the earnings
management hypothesis.
Sorting firms based on the magnitude of cash
flows, however, presents a different picture.
Cash flows are positively and significantly
related to future returns. The average abnormal
return of high (low) cash flow firms is
significantly positive (negative). A hedge
strategy, simultaneously taking a long position
in the high cash flow portfolio and a short
position in the low cash flow portfolio,
generates a significantly positive abnormal
return of 16%. It is further observed that the
characteristics of cash flow-sorted portfolios are
different from those based on accruals. Both
extreme, high and low, accrual portfolios
consist of small firms while only the low cash
flow firms consist of small firms.
The rest of the paper is organised as follows.
In section 2 we formulate the hypotheses to be
tested, describe the sample selection process
and describe the research method. The results
are reported in section 3. We conclude the paper
in section 4.
2. Research Design
2.1 Hypothesis
Prior studies on the accrual anomaly report that
market participants do not take into account the
difference between the persistence of accruals
and the persistence of cash flows in current
earnings when predicting future earnings (Sloan
(1996) and Bradshaw et al. (2001)). Instead, they
focus only on current earnings and are
“surprised” when future earnings performance is
lower (higher) than expected. Sloan (1996) and
Bradshaw et al. (2001) examine the relation
between future earnings and the components of
current earnings. They find that both coefficients
of accruals and cash flows are statistically
significant and lie between 0 and 1 which mean
that the two components contribute to the mean
reversion of earnings. The coefficient of
accruals, however, is smaller than that of cash
flows indicating that the mean reversion of
accruals is faster than for cash flows.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 20 NO 1 (2007) 

24

Current earnings performance, when
accompanied by high accruals, therefore sees a
quicker mean reversion than when accompanied
by high cash flows. As a result, firms with high
earnings attributed to high accruals (cash
flows), ceteris paribus, will end up with lower
(higher) future earnings. The accrual anomaly
arises because investors do not price the
different persistence of accruals and cash flows.
Therefore, to examine the presence of the
accrual anomaly, our first hypothesis is:
H1: The performance of current earnings that
is mainly attributed to accruals is less
persistent than when it is mainly attributed
to the cash flows component of earnings.
The accrual anomaly arises because the
market incorrectly prices accruals and cash
flows as if they have the same persistence on
future earnings. As accruals are less persistent
than cash flows, the market seems to overprice
(underprice) accruals (cash flows). Therefore
our second hypothesis is:
H2: The market overprices (underprices) the
accruals (cash flows) component.
When future earnings are unexpectedly
lower (higher) the market reacts negatively
(positively) to the earnings announcement. The
higher the accrual component in current
earnings, the bigger the earnings surprise and
the more negative is the market’s reaction to the
earnings surprise.
Sloan (1996) suggests that, because accruals
and cash flows are negatively correlated, the
accrual strategy can also be expressed in terms
of the magnitude of cash flow. Houge and
Loughran (2000) and Collins and Hribar (2000)
find that the magnitude of cash flows (accruals)
are positively (negatively) correlated with future
stock return. The predictive association between
accruals (and cash flows) and future stock
return then creates an arbitrage investment
opportunity and leads us to our third and fourth
hypotheses:
H3: A trading strategy that takes a long position
in the portfolio of low accruals firms and a
short position in the portfolio of high
accruals firms generates a positive
abnormal return.
H4: A trading strategy that takes a long position
in the portfolio of high cash flows firms and
a short position in the portfolio of low cash
flows firms generates a positive abnormal
return.
Pincus et al. (2005), however, find that the
occurrence of the accrual anomaly in a country
does not always imply that the cash flow
anomaly coexists, or vice versa. They find the
accrual anomaly is present in the US, the UK,
Canada and Australia, but find no evidence of
the presence of the cash flow anomaly in these
countries. On the other hand, they do not find
the accrual anomaly in other countries in their
sample but instead find the presence of the cash
flow anomaly. This evidence shows that the
cash flow anomaly is more pervasive across
different countries. Their results also indicate
that the two anomalies may not coexist. Thus
our fifth hypothesis is:
H5: The accrual anomaly coexists with the cash
flows anomaly.
2.2. Methodology
Hribar and Collins (2002) report that computing
accruals directly from statements of cash flows
is a more precise measure of accruals and
avoids measurement errors in estimating
accruals using the balance sheet approach. This
approach has been acknowledged and employed
extensively in the literature
2
. Accordingly, we
use the cash flow approach to measure accruals.
Accruals are calculated as the difference
between earnings and operating cash flows.
Operating cash flows data are obtained from the
statements of cash flows. We measure earnings
as operating income after depreciation but
before interest expense, taxes and special items.
All the three variables (earnings, cash flows and
accruals) are standardized by the average of the
beginning and end of the fiscal year book value
of total assets.
Firm statements of cash flows prior to 1991
are not available from Datex
3
. Therefore, for
periods 1987 to 1991, this study applies a
balance sheet approach (as employed in Sloan

2 See for example, Subramanyam (1996), Teoh, Welch
and Wong (1998a and 1998b), Collins and Hribar
(2000), Klein (2002), Desai, Rajgopal and
Venkatachalam (2004), Chan, Chan, Jegadeesh and
Lakonishok (2005), Pincus et al. (2005) and Coulton,
Taylor and Taylor (2005).
3 Although SSAP 10 explicitly requires firms to report this
statement, poor legal backing results with poor
compliance with the accounting standard (Bradbury and
Van Zijl (2005)). This problem was resolved with the
introduction of the Financial Reporting Act 1993 that
requires firms to include a cash flow statement in their
financial reports.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Accrual or Cash Flow Anomaly? Evidence from New Zealand

25

(1996), Houge and Loughran (2000) and Desai
et al. (2004)) in computing total accruals
4
:
( ) ( ) Dep TP STD CL Cash CA Accruals ? ? ? ? ? ? ? ? ? ? =
(1)
?CA is the change in current assets. ?Cash is
the change in cash or cash equivalent. ?CL is
the change in current liabilities. ?STD is the
change in debt included in current liabilities.
?TP is the change in tax payables, and Dep is
the depreciation and amortization expense.
Following Sloan (1996) we use a model that
estimates the average persistence of current
earnings on future earnings and another model
that does not restrict the accrual and the cash
flow components of current earnings to be equal
to examine the different persistence of accrual
and cash flow components of current earnings.
1 2 1 + +
+ + =
t t t t
Earnings Earnings ? ? ?
(2)
1 2 1 1 1 + +
+ + + =
t it it t
Accruals Cashflows Earnings ? ? ? ?

(3)
Model (2) estimates the average persistence
of current earnings on future earnings. The
accrual anomaly arises from the different
persistence of accrual and cash flow
components of earnings. Model (3) breaks
current earnings into accrual and cash flow
components of earnings. The smaller the
component from the other, the faster it is to
mean revert, indicating the less persistence of
the component.
To test the market’s pricing on accruals and
cash flows we employ the Mishkin (1983) test
and the hedge portfolio test. These tests have
been frequently used in studies on the accrual
anomaly (Sloan (1996), Collins and Hribar
(2000), Xie (2001) and Pincus et al. (2005)) to
examine whether the market efficiently prices
the accrual and the cash flow components of
earnings.
2.2.1. Mishkin test
Mishkin (1983) provides a framework to test for
the existence of the accrual anomaly. As in prior
studies, we estimate the following:
1 2 1 1 1 + +
+ + + =
t it it t
Accruals Cashflows Earnings ? ? ? ?

(4a)

4 The results, not reported but available from the authors,
are similar when we deleted data prior to 1991 and redid
the analysis based only on the accruals estimated from
the cash flow approach.
1
2
*
1
*
1 1 1 1
) (
+ + +
+ ? ? ? =
t it it t t
Accruals Cashflows Earnings AR ? ? ? ? ?
(4b)
AR is a stock’s abnormal return defined as
the difference between the stock return and the
size matched portfolio return. The idea is to
figure out if investors assign a higher valuation
coefficient to accruals than the one expected in
the association between accruals and future
earnings. If markets are efficient, we should
expect the two coefficients not to be statistically
different from each other. Accruals (cash flows)
mispricing is observed if the market assigns a
significantly larger or smaller coefficient than
implied in the association between accruals
(cash flows) and future earnings.
The Mishkin (1983) test is carried out first by
estimating the regressions jointly using an
iterative weighted nonlinear least squares
method to obtain the coefficient estimates. Then
the joint regressions are re-estimated by
imposing the constraints ?
p
= ?*
p
. We test this
by using a likelihood statistic ratio which is
asymptotically ?
2
(q) distributed:
2*N*Ln
?
?
?
?
?
?
?
?
u
c
SSR
SSR
(5)
N = number of observation
q = number of restrictions
SSR
c
= sum of squared residuals of the
constrained regression
SSR
u
= sum of squared residuals of the
unconstrained regression
2.2.2. Hedge portfolio test
We group stocks into five categories based on
the magnitude of accruals and cash flows. Stock
returns are computed as the buy and hold
returns that are measured beginning from four
months after the end of the firms’ fiscal years.
Prior studies find that although more than one
year ahead abnormal stock returns are positive,
these returns are not significantly different from
zero. Furthermore, the inclusion of more than
one year ahead stock returns will decrease the
sample size of this study. Therefore, future
stock return is examined only as a one
year ahead stock return. These portfolios are
rebalanced every year. To generate the
benchmark portfolio returns, five equally
weighted portfolios are constructed based on the
size or market value of the firms. The buy and
hold returns of these portfolios are calculated
within each group. Following similar studies on
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 20 NO 1 (2007) 

26

the accrual anomaly, the abnormal stock return
is defined as the difference between the stock
return and the size matched portfolio return:
pt it it
R R AR ? =
(6)
AR
it
is the size adjusted returns of stock i, R
it
is
the raw return of the individual stock and R
pt
is
the size matched portfolio return.
2.3. Data
This study is conducted using all non financial
firms listed on the New Zealand Stock
Exchange with available data in the Datastream
and the Datex financial company report files.
We delete firm-year observations that have
insufficient data for the calculation of accruals
as defined below as well as firms that change
their fiscal year ends. The sample period is from
1987 to 2003. This process results in a sample
of 1,202 firm year observations with the
required financial statement and share price
data. In order to avoid any data errors and the
effects of outliers as in prior studies, we delete
from the sample those stocks with annual buy
and hold returns of more than ±100%. The final
sample is 189 firms with 1,127 firm-year
observations.
3. Empirical Results
3.1. The accrual and the cash flow
anomalies
As reported in Panel A of Table 1, the
coefficient of current earnings in model (2a) is
between 0 and 1 indicating that current earnings
is mean reverting. Pincus et al. (2005) report the
cross-country range of mean reversion of
earnings is between 0.6 and 0.8. The mean
reversion of NZ firms’ earnings is within the
mean range (?
2
= 0.71). Results in panel B of
Table 1 show that both accrual and cash flow
components of current earnings significantly
explain future earnings. The coefficient of
accruals (0.54) is however smaller than the
coefficient of cash flows (0.94) and less than
unity which means that accruals are mean
reverting faster than cash flows. An F test
confirms that the coefficient of accruals is
smaller than the coefficient of cash flows. This
evidence supports the hypothesis that accruals
are less persistent than cash flows in shaping
future earnings.
Table 1
The Persistence of Accrual and Cash Components of Earnings
The dependent variable is one-year ahead earnings, the explanatory variables are cash flows and accruals.
Earnings are measured as operating income after depreciation but before interest expense, taxes and special
items. Cash Flows are operating cash flows. Accruals are the difference between earnings and cash flows.
All variables are deflated by average total assets. Sample period is from 1987 to 2003. Unavailability of
one-year ahead future earnings data for some firms reduces the sample size to 956 firm year observations.
Two-tail t statistics are in parentheses.
Panel A.
1 2 1 1 + +
+ + =
t t t
Earnings Earnings ? ? ?

?
?
?
2
Adj. R
2
%
0.00 0.71 30.76
(0.49) (20.62)
***

Panel B.
1 2 1 1 1 + +
+ + + =
t t t t
Accruals Cashflows Earnings ? ? ? ?

?
?
?
?
?
2
Adj. R
2
%
-0.01 0.95 0.54 34.25
(-1.78) (20.26)
***
(13.23)
***

F test: ?
?
= ?
2
47.71
p-value 0.000
***

*** significant at 1%

D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Accrual or Cash Flow Anomaly? Evidence from New Zealand

27

The results from the Mishkin test reported in
Panel A of Table 2 indicate that on average the
NZ market underprices current earnings
(?
1
> ?*
1
). This underpricing of current earnings
is attributed to the underpricing of both accruals
and cash flows components of earnings (Panel
B). The underpricing of both accruals and cash
flows is not unique to New Zealand, but as
reported by Pincus et al., does occur in other
countries as well.
Table 3 provides statistics of five portfolios
of stocks sorted by the magnitude of accruals.
Earnings are positively correlated with total
accruals but cash flows are negatively correlated
with total accruals. The average annual
correlation between accruals and cash flows,
however, is weak, only -0.22. The magnitude of
this correlation is much lower than that reported
in prior studies which is typically more than
-0.5.
The inverted “U” shape pattern in the market
value of the firms sorted by accruals shows that
the two extreme portfolios consist of small
stocks. A hedge portfolio strategy taking a long
position in the low accrual portfolio and a short
position in the high accrual portfolio should
therefore eliminate the size-risk factor of the
strategy.
The average abnormal return from the hedge
strategy during the sample period is 2.56% per
year but insignificant. The positive hedge return
is derived mainly from the negative return of the
high accrual portfolio.
Table 2
Results from the Iterative Weighted Non-linear Least Squares Regressions
of the Stock Price Reaction to Information in Earnings and the
Components of Current Earnings
Earnings are measured as operating income after depreciation but before interest expense, taxes and special
items. Cash Flows are operating cash flows. Accruals are the difference between earnings and cash flows.
Abnormal return is computed as the stock’s buy and hold annual raw return minus the size-matched buy
and hold annual portfolio return. Sample period is from 1987 to 2003. Unavailability of one-year ahead
future earnings data for some firms reduces the sample size to 956 firm year observations.
L = 2*n*ln(SSR
c
/SSR
u
). p-values are in parentheses.
Panel A. Test on the pricing of current earnings
1 1 1 1 + +
+ + =
t t t
Earnings Earnings ? ? ?

1
*
1 1 1 1 1
) (
+ + +
+ ? ? =
t it t t
Earnings Earnings AR ? ? ? ?

?
1
?*
1

0.71 0.21
L: ?
1 =
?*
1
19.38
(0.000)
***

Panel B. Test on the pricing of the components of current earnings
1 2 1 1 1 + +
+ + + =
t it it t
Accruals Cashflows Earnings ? ? ? ?

1
2
*
1
*
1 1 1 1
) (
+ + +
+ ? ? ? =
t it it t t
Accruals Cashflows Earnings AR ? ? ? ? ?

?
1
?*
1
?
2
?*
2

0.94 0.14 0.54 0.30
L: ?
1 =
?*
1
27.30
(0.000)
***

L: ?
2 =
?*
2
2.75
(0.0973)
*

*** significant at 1%
* significant at 10%
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 20 NO 1 (2007) 

28

Table 3
Average of Firm Variables Sorted by Accruals
Earnings are measured as operating income after depreciation but before interest expense, taxes and special
items. Cash Flows are operating cash flows. Accruals are the difference between earnings and cash flows.
All variables are deflated by average total assets. Size is market value of firms’ equity and B/M is the book
to market ratio of firms’ equity. Book equity is total asset minus total liabilities. Raw return is defined as
the buy and hold return calculated from 4 months after the end of the firm fiscal year. Abnormal return
(AR) is the size-adjusted return measured as the difference between the stock return minus the size-
matched portfolio return. The hedge portfolio consists of a long position in portfolio one (the lowest total
accruals) and a short position in portfolio five (the highest total accruals). Sample consists of 1,127 firm
years observations from 1987 to 2003. Two-tail t statistics are in parentheses.
Portfolio 1 2 3 4 5
Earnings -0.10 0.06 0.08 0.11 0.14
Cash Flows 0.12 0.10 0.08 0.07 -0.02
Accrual -0.22 -0.04 0.01 0.04 0.17
Size 85.16 206.12 183.38 183.32 112.99
B/M 1.14 1.28 1.45 1.13 0.99
Raw Return -2.15% 5.21% 2.82% 3.25% 4.90%
Abnormal Return -1.57% 3.77% 0.85% 1.03% -4.13%
(-0.63) (1.71)
*
(0.44) (0.53) (-1.90)
*

Hedge Abnormal Return 2.56%
(0.55)
N 227 228 226 227 219
* significant at 10%

The abnormal return of the high accrual
portfolio is -4.13% and statistically significant,
while the abnormal return of the low portfolio is
-1.57% and statistically insignificant. This
evidence confirms the results in prior studies
that the positive abnormal return of the accrual
strategy is mainly due to the poor performance
of firms reporting high accruals (Houge and
Loughran (2000)) and that the abnormal returns
of both extreme accrual portfolios after
excluding the outliers are negative (Kraft et al.
(2004)).
Figure 1 shows that the accrual strategy
generates positive abnormal returns in only
9 (53%) of 17 years during the sample period.
The highest positive return is 53.15% in 1991
and the lowest abnormal return is -16.15% in
1989. This evidence shows that the extensive
use of accruals in an accounting system and a
country’s legal tradition may not always be
indicative of the possibility of occurrence of the
accrual anomaly in a particular country as
suggested by Pincus et al. (2005).
The significant abnormal return of the high
accrual portfolio indicates that investors
overvalue high accrual stocks. Indeed, when we
apply the Mishkin test on high accrual stocks,
we find that the market significantly overprices
accruals in the high accrual portfolio
5
. Table 3
shows that firms with high accruals are also
firms with high earnings. As the investors seem
to overvalue high accrual stocks, the poor
performance of the high accrual portfolio
provides a preliminary indication that when
high earnings are accompanied by high
accruals, managers of these firms engage in
income increasing accruals.

5 Results are not reported but available from the authors
and they are significant at the conventional level of 5%.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Accrual or Cash Flow Anomaly? Evidence from New Zealand

29

Figure 1
Abnormal Returns of the Trading Strategy Based on Accruals by
Calendar Year
Abnormal returns are firms’ size adjusted returns. The strategy’s abnormal returns are based on going long
on the lowest accrual portfolio and short on the highest accrual portfolio. Accruals are the difference
between earnings and cash flows. All variables are deflated by average total assets. Sample consists of
1,127 firm years observations from 1987 to 2003.

To investigate the possibility of income
increasing management, we sort portfolios into
nondiscretionary and discretionary accruals.
Nondiscretionary accruals are accruals that arise
from normal business activity and discretionary
accruals are accruals that arise from managers’
discretion. We employ two measures of
discretionary accruals. Our first measure is the
cross sectional modified Jones model (Dechow,
Sloan and Sweeney (1995)):
t
t
t
t
t t
t
A
PPE
A
REC REV
A
TA ? ? ? ? +
?
?
?
?
?
?
?
?
+
?
?
?
?
?
?
?
? ? ? ?
+
?
?
?
?
?
?
?
?
=
? ? ? 1
3
1
2
1
1
1
(7)
TA is accruals, A is total assets, ?REV is the
change in revenues, ?REC is the change in
account receivables and PPE is property plant
and equipment. The nondiscretionary accruals
(NDA) are the fitted values and the
discretionary accruals (DA) are the residuals of
the model.
The cross-sectional modified Jones model is
chosen instead of the time series Jones model
because the parameter estimates obtained from
the cross sectional version of the modified Jones
model are specified better and do not suffer
from the “look ahead” bias as in the time series
version (Subramanyam (1996) and Bartov, Gul
and Tsui (2000)). In addition, few NZ firms
have a long series of historical data. Hence the
cross sectional Jones model generates a larger
sample and the power of the tests in this
study. Although the modified Jones model is
not perfect in partitioning discretionary and
nondiscretionary accruals from total accruals,
this model is widely used in the earnings
management literature (Dechow, Sloan and
Sweeney (1995) and Guay, Kothari and Watts
(1996)).
The modified Jones model, however, does
not control for the effect of firm performance on
accruals. Firms with high growth in
performance are likely to report high accruals
not attributed to earnings management.
Consequently, the commonly used discretionary
accrual models are too often biased toward
rejecting the null hypothesis of no earnings
management. Kothari, Leone and Wasley
(2005) report that the performance-matched
discretionary accrual models and the Jones
model with current Return on Assets (ROA
t
)
included as an additional regressor enhance the
reliability of inferences from earnings
management research. While Kothari et al.
report that the performance-matched Jones or
modified Jones models tend to be the best
measure of discretionary accruals, these
methods are best applied to examine earnings
management related to an event (e.g., an IPO or
a SEO) where control firms are not expected to
report the same degree of discretionary accruals.
As our study does not examine a specific
earnings management event but uses all firms
available, employing the ROA-adjusted Jones
model as our second measure of discretionary
accruals is the best alternative to control the
effect of firm performance in detecting earnings
management.
HEDGE RETURN
-12.63%
7.08%
-16.15%
20.83%
53.15%
19.46%
-8.93%
-14.23%
3.85%
-0.41%
12.89%
0.56% 0.59%
-8.25% -9.18%
16.85%
-0.14%
3.84%
-20%
-10%
0%
10%
20%
30%
40%
50%
60%
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 AVG
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 20 NO 1 (2007) 

30

t t
t
t
t
t
t
ROA
A
PPE
A
REV
A
TA ? ? ? ? ? + +
?
?
?
?
?
?
?
?
+
?
?
?
?
?
?
?
? ?
+
?
?
?
?
?
?
?
?
=
? ? ?
4
1
3
1
2
1
1
1
(8)
Return on assets is measured as earnings
before interest and taxes divided by total
assets (Bodie, Kane and Marcus (2002)). The
nondiscretionary accruals (NA) are the fitted
values and the discretionary accruals (ABA) are
the residuals of the model.
Table 4 presents the descriptive statistics of
the coefficients and the variables used in the
modified Jones and the ROA-adjusted Jones
models. The means and the statistical
significance of the coefficients are calculated
following Fama-MacBeth (1973). The means of
the regression coefficients are calculated as
the average (across years) of the regression
coefficients and the t-statistics are defined as the
mean divided by its standard error.
In Panel A of Table 4, the mean coefficient
of the change in revenues minus the change in
receivables is statistically significant. The
average adjusted R
2
of the modified Jones
model is 15%. Panel B of Table 4 shows
Table 4
Descriptive Statistics of the Discretionary Accrual Models
Accruals are defined as the difference between operating earnings and operating cash flows. Operating
earnings are defined as the net income after depreciation. ?REV is the change in total revenues. ?REC is
the change in total receivables. PPE is property, plant and equipment. ROA is net profit before interest and
taxes. All variables are scaled by lagged total assets. Nondiscretionary (discretionary) accruals are the fitted
values (residuals) of the modified Jones and the ROA-adjusted Jones models. Sample period is from 1987
to 2003. The final sample consists of 1,127 firm year observations. The means of the year-by-year
regression coefficients are calculated as the average of the each-year regression coefficients. t-statistics for
the means, t(Mn), is defined as the mean of the coefficient divided by its standard error (time series
standard deviation of the coefficient divided by (17)
1/2
. The table also shows the average (across years) of
the means and standard deviation (SD) of the regression variables.
Panel A. The Modified Jones model
1. Mean and t-Statistics for the Means of the Year-by-Year Regression Coefficients
Intercept ?REV-?REC PPE Adj. R
2

Mean -274.4596 0.0720 0.0003 14.53%
t(Mn) (-1.64) (2.39)
***
(0.01) (2.75)
***

2. Means and Standards Deviation of the Regression Variables
Accruals Intercept ?REV ?REC PPE
Mean -0.0125 0.0012 0.1238 0.0326 0.3521
SD 0.2481 0.0017 0.1743 0.0307 0.3872
Panel B. The ROA-adjusted Jones model
1. Mean and t-Statistics for the Means of the Year-by-Year Regression Coefficients
Intercept ?REV PPE ROA Adj. R
2

Mean -104.6360 0.0299 -0.0563 0.4412 36.37%
t(Mn) (-1.11) (1.11) (-2.13)
**
(4.56)
***
(5.78)
***

2. Means and Standards Deviation of the Regression Variables
Accruals Intercept ?REV PPE ROA
Mean -0.0125 0.0012 0.1238 0.3521 0.0600
SD 0.2481 0.0088 0.1743 0.3872 0.2157
** siginificant at 5%
*** significant at 1%.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Accrual or Cash Flow Anomaly? Evidence from New Zealand

31

average coefficients from the ROA-adjusted
Jones model. The coefficient of the change in
revenues is positive but not significant.
Consistent with prior studies in earnings
management, the coefficient of PPE is negative
and significant. The positive and significant
coefficient on ROA is consistent with the
hypothesis that ROA captures the performance
effect on accruals. The performance of the
ROA-adjusted Jones model, in terms of the
adjusted R
2
, is better than that of the modified
Jones model. The adjusted R
2
of the ROA-
adjusted Jones model is 36% and strongly
significant.
Table 5 reports results of portfolio abnormal
returns sorted on the nondiscretionary and
the discretionary components of accruals. For
portfolios sorted on the nondiscretionary
component of accruals (NDA and NA), the
patterns of abnormal returns across the quintile
portfolios are inconsistent with the accrual
anomaly. The abnormal returns of portfolio one
(five) are both negative (positive). For portfolios
sorted on the discretionary component of
accruals (DA and ABA), the patterns of the
abnormal returns are consistent with the accrual
anomaly. The abnormal returns of portfolio
one are both positive. On the other hand, the
abnormal returns of portfolio five are
significantly negative. The positive hedge
abnormal return, however, is not significant for
the modified Jones model and is slightly
significant for the performance-adjusted Jones
model. The significantly negative abnormal
return of the high discretionary accruals firms
therefore support the hypothesis that managers
of high accrual firms engage in earnings
management.
As discussed earlier, the Companies and the
Financial Reporting Act 1993 (FRA93) were
introduced to provide a legal backing to ensure
that financial reports are made in compliance
with the accounting standards. Based on visual
inspection of Figure 1, after 1992, the accrual
anomaly hardly exists during the sample period.
To test the effects of these regulations, we
Table 5
Portfolio Abnormal Returns Sorted on the Discretionary and
Nondiscretionary Components of Accruals
Abnormal return (AR) is the size-adjusted return measured as the difference between the buy and hold
stock return calculated from 4 months after the end of the firm fiscal year minus the size-matched portfolio
return. The hedge portfolio consists of a long position in portfolio one (the lowest component of accruals)
and a short position in portfolio five (the highest component of accruals). NDA (DA) is the fitted (residual)
value of the modified Jones model. NA (ABA) is the fitted (residual) value of the Jones model with current
ROA as an additional regressor. Sample consists of 1,127 firm years observations from 1987 to 2003. Two-
tail t statistics are in parentheses.
Portfolio 1 2 3 4 5 Hedge
NDA -2.69% 2.29% -0.15% -1.75% 2.22% -4.91%
(-1.06) (1.14) (-0.07) (-0.82) (1.09) (-1.07)
NA -7.29% 1.43% 4.75% 0.25% 0.64% -7.93%
(-2.98)
***
(0.68) (2.12)
**
(0.14) (0.29) (-1.70)
*

DA 0.91% 4.36% -0.39% -1.00% -3.85% 4.76%
(0.37) (2.01)
**
(-0.19) (-0.54) (-1.73)
*
(1.01)
ABA 2.86% 1.80% 1.83% -1.22% -5.22% 8.08%
(1.14) (0.89) (0.87) (-0.58) (2.36)
***
(1.72)
*

* significant at 10%
** significant at 5%
*** significant at 1%

D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 20 NO 1 (2007) 

32

separate the sample into two sample periods
before and after the introduction of these Acts
and repeat our analysis on the accrual anomaly
from 1987 to 1992 and from 1993 to 2003.
Table 6 reports results of the hedge abnormal
return of portfolios sorted on the accruals and
the discretionary component of accruals during
the pre-acts from 1987 to 1992 and the post-acts
from 1993 to 2003. For the period prior to the
introduction of the Companies and the Financial
Reporting Acts (from 1987 to 1992), the hedge
abnormal returns of portfolios sorted on both
measures of discretionary accruals are positive
at around 12% and statistically significant.
However, the abnormal return of the hedge
strategy for the period after the Acts were
introduced (1993 to 2003) is not only smaller in
magnitude but also statistically insignificant for
both measures of discretionary accruals.
Table 6
The Hedge Abnormal Return of Portfolios Sorted on Total Accruals and
Discretionary Accruals
Portfolios of stocks are sorted based on total accruals and discretionary accruals. Total accruals are the
difference between earnings and cash flows. Earnings are measured as operating income after depreciation
but before interest expense, taxes and special items. Cash Flows are operating cash flows. All variables are
deflated by lagged total assets. Abnormal return is the size-adjusted return measured as the difference
between the buy and hold stock return calculated from 4 months after the end of the firm fiscal year minus
the size-matched portfolio return. The hedge portfolio consists of a long position in the lowest total accrual
or discretionary accrual portfolio and a short position in the highest total accrual or discretionary accrual
portfolio. DA is the residual value of the modified Jones model. ABA is the residual value of the Jones
model with current ROA as an additional regressor. Sample consists of 1,127 firm years observations from
1987 to 2003. Two-tail t statistics are in parentheses.
Hedge Abnormal Return
Period Total Accruals DA ABA N
87-03 2.56% 4.76% 8.08% 1,127
(0.55) (1.42) (1.72)*
87-92 10.34% 12.11% 12.58% 310
(1.48) (1.71) (1.77)*
93-03 -0.38% 2.06% 6.41% 817
(0.07) (0.39) (1.22)
* significant at 10%

Prior to 1993, claims of poor compliance
with the NZ accounting standards had been
frequently reported (Bradbury and Van Zijl
(2005)). The Financial Reporting Act provides a
sufficient legal backing to the existing
accounting standards but the Companies Act
1993 clearly states that the responsibility to
comply with the standards is placed on the
company directors. The absence of the accrual
anomaly for the period after the introduction of
these Acts suggests that the occurrence of this
anomaly may also be related with corporate
governance. Klein (2002) finds that the
magnitude of discretionary accruals is positively
correlated with poor corporate governance.
Examining the relation between the occurrence
of the accrual anomaly and corporate
governance is beyond the scope of this study
and is left for future research.
Addressing the cash flow anomaly, Table 7
presents summary statistics of firms sorted by
the magnitude of their cash flows. Earnings are
positively (negatively) correlated with cash
flows (accruals). Firms in the low (high) cash
flow portfolio exhibit the lowest (highest)
performance in future returns. Consistent with
Houge and Loughran (2002), we find that the
profile of portfolios based on cash flows is
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Accrual or Cash Flow Anomaly? Evidence from New Zealand

33

Table 7
Average of Firm Variables Sorted by Cash Flows
Earnings are measured as operating income after depreciation but before interest expense, taxes and special
items. Cash Flows are operating cash flows. Accruals are the difference between earnings and cash flows.
All variables are deflated by average total assets. Size is market value of firms’ equity and B/M is the book
to market ratio of firms’ equity. Book equity is total asset minus total liabilities. Raw return is defined as
the buy and hold return calculated from 4 months after the end of the firm fiscal year. Abnormal return
(AR) is the size-adjusted return measured as the difference between the stock return minus the size-
matched portfolio return. The hedge portfolio consists of a long position in portfolio five and a short
position in portfolio one. Sample consists of 1,127 firm years observations from 1987 to 2003. Two-tail t
statistics are in parentheses
Portfolio 1 2 3 4 5
Earnings -0.12 0.05 0.08 0.11 0.17
CF -0.15 0.03 0.08 0.12 0.24
Accrual 0.03 0.02 0.01 -0.01 -0.07
Size 48.72 142.74 195.91 235.35 147.25
B/M 1.26 1.50 1.42 0.98 0.83
Raw Return -11.62% -1.84% 5.76% 3.18% 8.60%
Abnormal Return -8.82% -1.70% 3.01% 0.86% 6.35%
(-3.77)
***
(-0.71) (1.60) (0.45) (2.99)
***

Hedge Abnormal Return 15.84%
(3.55)
***

N 227 228 226 227 219
*** significant at 1%

Figure 2
Abnormal Returns of the Trading Strategy Based on Cash Flows by
Calendar Year
Abnormal returns are firms’ size adjusted returns. The strategy’s abnormal returns are based on going long
on the highest cash flow portfolio and short on the lowest cash flow portfolio. Cash flows are the operating
cash flows obtained from the statements of cash flows. Sample consists of 1,127 firm years observations
from 1987 to 2003.

Hedge Return
-10.24%
20.76%
1.64%
42.68%
33.67%
12.97%
-4.56%
15.42%
7.01%
21.18%
9.94%
-0.56%
24.13%
44.38%
11.97%
17.10%17.76%
15.60%
-20%
-10%
0%
10%
20%
30%
40%
50%
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 AVG
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 20 NO 1 (2007) 

34

different from that based on accruals. The low
cash flow portfolio consists of small stocks
while the high cash flow portfolio consists of
big stocks. The average abnormal return in
the low (high) cash flow portfolio is negative
(positive) at -9% (6%) and statistically
significant. The average return of the hedge
strategy is around 16% and statistically
significant. Furthermore, the relation between
the magnitude of cash flows and the abnormal
returns is more stable. The average abnormal
return is generally increasing according to the
order of the quintile portfolios.
The abnormal returns of the cash flow
strategy are positive in 14 (82%) of 17 years
during the sample period (Figure 2). The
highest return is 44.38% in 2000 and the
lowest return is -10.24% in 1987 which may
be attributed to the stock market crash in
October 1987.
3.2. Robustness tests
The positive abnormal returns of the cash flow
portfolios in almost all of the calendar years
across the sample period (Figure 3) suggest that
investors underweight the persistence of the
cash flows component of current earnings.
However, these positive abnormal returns may
also reflect other unidentified risk factors.
Fama and French (1992 and 1993) report that
beta, size and the book to market ratio explain
most of the cross sectional variation in portfolio
returns. They argue that their asset pricing
model captures the cross sectional returns
attributed to systematic, size and book to market
ratio risk factors.
Table 8
Monthly Time Series Regressions of Buy and Hold Returns of
Equally Weighted Cash Flow Portfolios on Market Risk, Size and
Book to Market Ratio
Stocks are ranked based on the magnitude of operating cash flows scaled by average total assets. Equally
weighted cash flow portfolios are formed on July of year t until June of year t+1. The sample period is from
1990 to 2003. A Fama and French 3 factor model is conducted for each quintile portfolio. R
pt
is stock
return of portfolio p in month t. R
ft
is 1-month bank bill rate. R
mt
is the market (NZSX All) return in month
t. SMB is size factor (small minus big) in month t. HML is book to market (high minus low) factor in
month t. Two-tail t statistics are in parentheses.
r
pt
– r
ft
= ?
0
+ ?
1
(r
mt
– r
ft
) + ?
2
SMB + ?
3
HML + ?
pt

Portfolio ?
0
?
1
?
2
?
3
Adj. R
2
%
Low -0.01 0.98 0.31 0.13 43.31
(-2.24)
**
(9.84)
***
(3.70)
***
(1.24)
2 0.00 0.79 -0.06 0.11 54.05
(-0.26) (13.98)
***
(-1.28) (1.97)
**

3 0.00 0.99 -0.16 0.14 62.26
(1.59) (16.59)
***
(-3.21)
***
(2.25)
**

4 0.00 0.87 -0.17 0.14 65.28
(1.67)
*
(17.64)
***
(-4.23)
***
(2.70)
***

High 0.01 0.89 -0.12 0.04 62.19
(3.48)
***
(16.66)
***
(-2.64)
***
(0.76)
Hedge Return 0.02 -0.09 -0.42 -0.08 11.97
(3.52)
***
(-0.73) (-4.38)
***
(-0.71)
* significant at 10%
** significant at 5%
*** significant at 1%

D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Accrual or Cash Flow Anomaly? Evidence from New Zealand

35

The Fama and French three-factor model is:
r
pt
– r
ft
= ?
0
+ ?
1
(r
mt
– r
ft
) + ?
2
SMB
+ ?
3
HML + ?
pt
(9)
r
pt
is stock return of portfolio p in month t. r
ft
is
the risk free rate in month t. r
mt
is the market
return in month t. SMB is the size factor (small
minus big) in month t. HML is book to market
(high minus low) factor in month t. The
intercept, ?
0
, measures average monthly
abnormal return of the portfolio in year t+1. To
get the annualized abnormal return, ?
0
is
multiplied by 12.
The significant abnormal returns of the high
and low cash flow portfolio may be attributed to
these three risk factors. To test this hypothesis,
we construct equally weighted monthly time
series cash flow portfolios beginning from July
in year t and held until June in year t+1. The
median size of NZSE firms is used to split
stocks into small and big portfolios. We also
sort firms based on their book to market ratios
and classify the bottom (top) 30% as the low
(high) book to market portfolio. The 1-month
bank bill rate is employed as the risk free rate.
We then run the three factor model for each
quintile of cash flow portfolio. As the market
index, the NZSE All, is available only from
1990, the sample period for this test is from
1990 to 2003.
Table 8 shows the results of the three-factor
model for the cash flows-based portfolios.
Similar to previous results, the abnormal returns
of cash flow portfolios 2 to 4 are not statistically
significant. Beta, size and the book to market
ratio significantly explain the cross sectional
variation of these portfolio returns.
However the abnormal returns of the two
extreme cash flow portfolios are still robust
after controlling for these three risk factors. The
monthly average abnormal return of the low
(high) cash flow portfolio is -0.99% (0.83%) or
-11.85% (9.94%) annually and statistically
significant. Buying high and selling low cash
flow portfolio strategy during the sample
period generates a significant average monthly
abnormal return of 1.82% or 21.79% annually
6
.

6 Carhart (1997) argues that adding a factor representing
one-year momentum in stock returns factor into the
Fama and French three factor model better explains the
variation in stock returns. We, therefore, repeat the
analysis using the Carhart four-factor model. The results
are similar to the results using the three-factor model.
4. Summary
During the total sample period of 1987 to 2003,
we do not find a significant accrual anomaly in
New Zealand. However, we find that, from
1987 to 1992 - a period before the introduction
of the Companies and the Financial Reporting
Acts 1993 - the presence of the accrual anomaly
is statistically significant suggesting that these
Acts have a significant impact on the
occurrence of the anomaly.
In contrast, the presence of the cash flow
anomaly during the sample period is significant.
This evidence is consistent with a recent
international study that the accrual anomaly
does not always coexist with the cash flow
anomaly. The abnormal return of the low cash
flow portfolio is negative and significant,
while the abnormal return of the high cash
flow portfolio is significantly positive. A
corresponding cash-flow based trading strategy
generates positive returns in 14 (82%) of the 17
years period.
There are several reasons, however, that New
Zealand investors may not be able to fully
benefit from exploiting this anomaly. First, the
prohibition of short selling in New Zealand
prevents the use of the hedge strategy and as a
result reduces the abnormal return of the
strategy. Second, even though buying only high
cash flows stocks still generates a positive and
significant average abnormal return of 6.35%,
firms in the sample have different fiscal periods.
As a result, the hedge strategy requires
portfolios to be constructed more than once in a
given year. The information acquisition and the
processing costs to implement this strategy
would limit the benefit of the trading strategy.
5. References
Barth, M. and Hutton, A. (2004), “Analyst Earnings Forecast
Revisions and the Pricing of Accruals”, Review of
Accounting Studies, vol. 9, pp. 59-96.
Bartov, E., Gul, F. and Tsui, J. (2000), “Discretionary
Accrual Models and Audit Qualifications”, Journal of
Accounting and Economics, vol. 30, pp. 421-452.
Beaver, W. (1998), Financial Reporting: An Accounting
Revolution, Upper Saddle, NJ: Prentice Hall.
Block, S.B. (1999), “A Study of Financial Analysts: Practice
and Theory”, Financial Analyst Journal, vol. 55, pp. 86-
92.
Bradbury, M. and Van Zijl, T. (2005), The New Zealand
Financial Reporting Framework, John Wiley and Sons
Australia Ltd.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 20 NO 1 (2007) 

36

Bradshaw, M., Richardson, S. and Sloan, R. (2001), “Do
Analysts and Auditors Use Information in Accruals”,
Journal of Accounting Research, vol. 39, pp. 45-74.
Bushman, R. and Smith, A. (2001), “Financial Accounting
Information and Corporate Governance”, Journal of
Accounting and Economics, vol. 32, pp. 234-333.
Callen, J. and Segal, D. (2004), “Do Accruals Drive Firm-
Level Stock Returns? A Variance Decomposition
Analysis”, Journal of Accounting Research, vol. 42,
pp. 527-560.
Chan, K., Jegadeesh, N. and Sougiannis, T. (2004), “The
Accrual Effect on Future Earnings”, Review of
Quantitative Finance and Accounting, vol. 22,
pp. 97-121.
Collins, D. and Hribar, P. (2000), “Earnings-Based and
Accrual-Based Anomalies: One Effect or Two?”
Journal of Accounting and Economics, vol. 29,
pp. 101-123.
Coulton, J., Taylor, S. and Taylor, S. (2005), “Is ‘Benchmark
Beating’ by Australian Firms Evidence of Earnings
Management?” Accounting and Finance, vol. 45,
pp. 553-576.
Dechow, P., Sloan, R. and Sweeney, A. (1995), “Detecting
Earnings Management”, The Accounting Review,
vol. 70, pp. 193-225.
Fama, E. and French, K. (1992), “The Cross-Section of
Expected Stock Returns”, Journal of Finance, vol. 47,
pp. 427-465.
Fama, E. and French, K. (1993), “Common Risk Factors in
the Returns on Stocks and Bonds”, Journal of Financial
Economics, vol. 33, pp. 3-56.
Finger, C.A. (1994), “The Ability of Earnings to Predict
Future Earnings and Cash Flow”, Journal of Accounting
Research, vol. 32, pp. 210-223.
Firth, M. (1997), “Takeovers in New Zealand: Motives,
Stockholder Returns, and Executive Share Ownership”,
Pacific-Basin Finance Journal, vol. 5, pp. 419-440.
Foster, G., Olsen, C. and Shevlin, T. (1984), “Earnings
Releases, Anomalies, and the Behavior of Security
Returns”, The Accounting Review, vol. 59, pp. 574-603.
Glazebrook, S., Jones, A., James, J. and Cole, G. (1999), The
New Zealand Accrual Regime, 2
nd
ed, CCH New
Zealand Ltd.
Guay, W., Kothari, S.P. and Watts, R. (1996), “A Market-
Based Evaluation of Discretionary Accrual Models”,
Journal of Accounting Research, vol. 34, pp. 83-105.
Hand, J. (1990), “A Test of the Extended Functional
Fixation Hypothesis”, The Accounting Review, vol. 65,
pp. 740-763.
Houge, T. and Loughran, T. (2000), “Cash Flow is King:
Cognitive Errors by Investors”, Journal of Psychology
and Financial Markets, vol 2, pp. 161-175.
Hung, M. (2001), ‘Accounting Standards and Value
Relevance of Financial Statements: An International
Analysis”, Journal of Accounting and Economics,
vol. 30, pp. 401-420.
Hribar, P. and Collins, D. (2002), “Errors in Estimating
Accruals: Implications for Empirical Research”,
Journal of Accounting Research, vol. 40, pp. 105-134.
Klein, A. (2002), “Audit Committee, Board of Director
Characteristics, and Earnings Management”, Journal of
Accounting and Economics, vol. 33, pp. 375-400.
Kothari, S.P. (2001), “Capital Markets Research in
Accounting”, Journal of Accounting and Economics,
vol. 31, pp. 105-231.
Kothari, S.P., Sabino, J.S. and Zach, T. (2005),
“Implications of Survival and Data Trimming for Tests
of Market Efficiency”, Journal of Accounting and
Economics, vol. 39, pp. 129-161.
Kothari. S.P., Leone, A.J. and Wasley, C.E. (2005),
“Performance Matched Discretionary Accrual
Measures”, Journal of Accounting and Economics,
vol. 39, pp. 163-197.
Kraft, A., Leone, A. and Wasley, C. (2004), “Research
Design Issues and Related Inference Problems
Underlying Tests of the Market Pricing of Accounting
Information”, Working Paper, University of Rochester.
Kusnadi, Y. and Wei, K.C. (2006), “Legal Protection, Equity
Dependence and Corporate Investment: Evidence from
around the World”, Working Paper, Hong Kong
University of Science and Technology.
La Porta, R., Lopez-de-Silanes, F. and Shleifer, R. W.
(1998), “Law and Finance”, Journal of Political
Economy, vol. 106, pp. 1113-1155.
Levine, R. and Zervos, S. (1993), “What We Have Learned
about Policy and Growth from Cross-Country
Regressions?” The American Economic Review, vol. 83,
pp. 426–443.
Mackenzie, R., Kelleher, B. and Vos, E. (2000), “New
Zealand Takeover Evidence: Firm Characteristics and
Payment Method”, Managerial Finance, vol. 26,
pp. 55-65.
Pincus, M., Rajgopal, S. and Venkatachalam, M. (2005),
“The Accrual Anomaly: International Evidence”,
Working Paper, The University of Iowa.
Quigg, D. and Land, J. (1994), “NZ’s Directors to Carry the
Can”, International Corporate Law, vol. 34, pp. 40-42.
Seebold, N. (1993), “Kiwis Take Lead in Company Law
Reform”, International Financial Law Review, vol. 12,
pp. 37-38.
Sloan, R. (1996), “Do Stock Prices Fully Reflect
Information in Accruals and Cash Flows about Future
Earnings?” The Accounting Review, vol. 71,
pp. 289-315.
Subramanyam, K. (1996), “The Pricing of Discretionary
Accruals”, Journal of Accounting and Economics,
vol. 22, pp. 249-281.
Teoh, S., Welch, I. and Wong, T. (1998a), “Earnings
Management and the Long-Run Underperformance of
Seasoned Equity Offerings”, Journal of Financial
Economics, vol. 50, pp. 63-100.
Teoh, S., Welch, I. and Wong, T. (1998b), “Earnings
Management and the Long-Run Underperformance of
Initial Public Equity Offerings”, Journal of Finance,
vol. 53, pp. 1935-1974.
Walker, G. (2003), “Securities Regulation Reform in New
Zealand: Australian Rules OK?”, C&SLJ, vol. 21,
pp. 533-537.
Xie, H. (2001), “The Mispricing of Abnormal Accruals”,
The Accounting Review, vol. 76, pp. 357-373.

D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
This article has been cited by:
1. Young Jun Kim, Jung Hoon Kim, Sewon Kwon, Su Jeong Lee. 2015. Percent accruals and the accrual anomaly: Korean
evidence. Pacific-Basin Finance Journal 35, 340-366. [CrossRef]
2. Nasif Ozkan, Mustafa Mesut Kayali. 2015. The accrual anomaly: Evidence from Borsa Istanbul. Borsa Istanbul Review .
[CrossRef]
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
0
1

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)

doc_277496651.pdf