Description
Audit quality studies document that accruals decrease when the audit firm is large, or the
audit firm is an industry specialist, or the audit-client tenure is long. The purpose of this paper is to
posit that incentives related to highly-valued equity mitigate these results, as managers use income
increasing accruals to augment earnings.
Accounting Research Journal
Audit quality and overvalued equity
Robert Houmes Maggie Foley Richard J . Cebula
Article information:
To cite this document:
Robert Houmes Maggie Foley Richard J . Cebula, (2013),"Audit quality and overvalued equity", Accounting
Research J ournal, Vol. 26 Iss 1 pp. 56 - 74
Permanent link to this document:
http://dx.doi.org/10.1108/ARJ -08-2011-0024
Downloaded on: 24 January 2016, At: 21:17 (PT)
References: this document contains references to 44 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 1591 times since 2013*
Users who downloaded this article also downloaded:
David S. J enkins, Thomas E. Vermeer, (2013),"Audit firm rotation and audit quality: evidence from
academic research", Accounting Research J ournal, Vol. 26 Iss 1 pp. 75-84 http://dx.doi.org/10.1108/
ARJ -11-2012-0087
Victoria J . Clout, Larelle Chapple, Nilan Gandhi, (2013),"The impact of auditor independence regulations
on established and emerging firms", Accounting Research J ournal, Vol. 26 Iss 2 pp. 88-108 http://
dx.doi.org/10.1108/ARJ -Dec-2011-0045
Arifur Khan, Paul Mather, (2013),"The value of executive director share ownership and discretionary
accruals", Accounting Research J ournal, Vol. 26 Iss 1 pp. 35-55 http://dx.doi.org/10.1108/
ARJ -02-2012-0011
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Audit quality and
overvalued equity
Robert Houmes, Maggie Foley and Richard J. Cebula
Department of Accounting and Finance, Jacksonville University,
Jacksonville, Florida, USA
Abstract
Purpose – Audit quality studies document that accruals decrease when the audit ?rm is large, or the
audit ?rm is an industry specialist, or the audit-client tenure is long. The purpose of this paper is to
posit that incentives related to highly-valued equity mitigate these results, as managers use income
increasing accruals to augment earnings.
Design/methodology/approach – To test this assertion, the authors regress discretionary accruals
on: controls, a highly valued equity indicator variable equal to 1 if the client’s lagged price-to-earnings
ratio is in the highest P/E quintile, indicator variables equal to 1 for alternative measures of audit
quality, and interaction terms between the highly valued equity indicator variable and audit quality
indicator variables.
Findings – Results of tests show positive and statistically signi?cant coef?cients for each of the
highly-valued equity-audit quality interaction terms, suggesting that when a ?rm is highly valued the
accruals’ decreasing effect of high quality auditors is reduced.
Originality/value – Beginning with Jensen’s article regarding the agency costs of overvalued
equity, a stream of research examining factors associated with highly priced ?rms has developed. The
paper extends these ?ndings, as well as the considerable body of audit quality studies, by examining
the ability of a high quality auditor to attenuate this result.
Keywords Auditing, Equity capital, Price earning ratio, Audit quality, Highly valued equity,
Discretionary accruals
Paper type Research paper
Introduction
Classical agency theory asserts that alignment of management-shareholder interests
increases incentives for value creation ( Jensen and Meckling, 1976). Jensen (2005) argues,
however, that when a ?rmbecomes overvalued, i.e. the price of the ?rmbecomes greater
than its underlying economic value, managers are motivated to perpetuate
overvaluation, which is consistent in principle with arguments in Renas and Cebula
(2005). Since an overvalued ?rm, by de?nition, lacks the operational capability to achieve
performance levels re?ected in its price, managers are motivated to use aggressive
accounting policies to maximize earnings. Although numerous reporting alternatives
are available to achieve earnings management goals, accruals are an especially attractive
choice since they are a normal part of the ?nancial reporting process and their amounts
require forward looking estimates over which managers have considerable discretion.
In a recent study Houmes andSkantz (2010) assert that incentives associated with over
valued equity induce managers to support extreme valuations by using discretionary
accruals to manage earnings higher. In particular, they regress discretionary accruals
on highly valued equity proxy variables measured as companies in the highest quintile
of P/E ?rms and controls. Results show that in the year following valuation and relative
to other ?rms, highly valued ?rms report higher discretionary accruals. Among other
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1030-9616.htm
Accounting Research Journal
Vol. 26 No. 1, 2013
pp. 56-74
qEmerald Group Publishing Limited
1030-9616
DOI 10.1108/ARJ-08-2011-0024
ARJ
26,1
56
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
things they conclude that directors and audit committees should be especially sensitive to
earnings management when the ?rm is highly valued. In the spirit of these remarks, we
extend the ?ndings in Houmes and Skantz (2010) by investigating the effect that
overvaluation has on audit quality. This study provides evidence that the previously
documented inverse relation between, large audit ?rms, industry specialist audit
?rms, long tenure audit ?rms, and discretionary accruals is attenuated when the ?rm is
highly valued.
Using high price-to-earnings ?rms to proxy for highly valued equity and the level of
income increasing discretionary accruals to proxy for earnings quality, our results
show that, in accordance with prior studies, high quality auditors generally constrain
accruals. By contrast, however, accruals for highly valued clients of high quality
auditors are statistically signi?cantly higher. In particular, accruals for clients of
large Big N (Big 6, Big 5, or Big 4, depending on the period), industry specialist and
long tenure auditors increase for clients in the previous year’s highest quintile of
price-to-earnings ratios. Hence, the widely documented tendency for high quality audit
?rms to reduce accruals is mitigated when the ?rm is highly valued. We rationalize
these ?ndings within the context of traditional audit quality literature and Jensen’s
(2005) overvalued equity hypothesis by asserting that the accruals decreasing effect of
high quality auditors on earnings is reduced as informed managers of highly valued
?rms with greater incentives to perpetuate values prevail upon auditors to report
higher levels of income increasing discretionary accruals.
The value of a ?rm is a function of its prospective performance, and the greater the
performance expectations the higher the value. Since expectations are greater when a
?rm is richly priced, investors’ reactions to unexpectedly negative performance
outcomes should also be greater. For example, Skinner and Sloan (2002) document that
the market’s reaction to negative earnings surprises is greater than that for positive
earnings surprises and that this asymmetric reaction increases for high market-to-book
?rms. Given that shareholder disappointments will be greater when the market’s
expectations are the highest, the potential ?ndings of this study have implications for
audit practitioners as they conduct audits for highly valued ?rms. Prior studies report
that shareholder suits are a common source of auditor litigation (Palmrose, 1988;
Goldwasser and Eickemeyer, 2004). Extant research also documents that the likelihood
of a securities fraud class action increases with the ?rm’s market-to-book ratio (Grimm,
2009). If high valuations impair audit quality, then auditors are performing lower
quality audits on the very ?rms that historically have tended to disappoint investors
the most. Results of this study should also be relevant for boards of highly valued
?rms. If overvaluation induces lower earnings quality and by extension lower audit
quality, boards should be especially vigilant in their internal control function. This
should be particularly true for the audit committee.
From a research perspective, a substantial body of work currently exists on audit
quality and more recent studies have examined factors related to highly valued equity.
To date, however, published studies have not considered audit quality within a
valuation context. We synthesize and provide additional insight into both streams of
research by investigating the accruals constraining effect of high quality auditors
vis-a` -vis other auditors when the client is highly valued.
The remainder of this study is organized as follows. The section thereafter includes
relevant literature, along with a discussion of Jensen’s (2005) overvalued equity
Audit quality
57
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
hypothesis as it relates to earnings and audit quality. The next section includes testable
hypotheses and their motivations, followed by a description of the sample and
methodology used to conduct tests. Results of these tests are reported in the subsequent
section. The study concludes with a discussion and summary of the ?ndings.
Literature and theory
Highly valued equity
Beginning with Jensen (2005), a signi?cant stream of research has developed regarding
the overvalued equity hypothesis. Efendi et al. (2007) use a sample of 190 ?rms from a
2003 General Accounting Of?ce report to investigate factors related to overvalued
equity and document among other things that earnings restatements increase for ?rms
that are constrained by debt covenant violations or that raise new debt. They also
show that the incidence of restatements increases when CEOs hold large amounts of
in-the-money stock options.
Chi and Gupta (2009) report that overvaluation-induced earnings management as
proxied for by discretionary accruals is negatively related to following year abnormal
returns. They document that, in the following year, abnormal returns of ?rms with
high discretionary accruals are 11.88 percent lower than the returns of ?rms with lower
discretionary accruals.
A long-standing anomaly of ef?cient equity markets is the tendency for highly
valued, high price-to-earnings ?rms to earn lower returns going forward. While
explanations for the persistence of this anomaly are beyond the scope of this study,
they include the following three:
(1) economic theory predicts that competition mitigates the ability of ?rms to
achieve and sustain abnormal performance levels over time;
(2) since highly valued ?rms are less risky, lower returns re?ect lower risk; and
(3) investors (irrationally) tend to over-price high-growth-glamour ?rms.
Notwithstanding these conjectures, however, the tendency for high price-to-earnings
?rms to under perform is well documented (Basu, 1977; Chopra et al., 1992;
Lakonishok et al., 1994; Campbell and Shiller, 2001). In a more recent study, Anderson
and Brooks (2006) document that the difference in returns between value and glamour
?rms almost doubles whenP/Es are calculated using average earnings over the last eight
years. The following graphical depiction of companies from the 1991 to 2008 period
shows that ?rms in the highest quintile of price-to-earnings ratios earned lower following
year mean and median abnormal returns in every year but one (Figures 1 and 2).
Hence, from a market value perspective and relative to other companies, the very
?rms that are expected to perform the best, on average, tend to perform worse. Since
expectations are particularly high for highly valued ?rms, when managers foresee the
operational inability of their ?rms to meet expected performance targets, incentives to
manage earnings increase. An important deterrent against these incentives is the audit.
Audit quality
The value of accounting information is a function of its credibility, and a central
objective of audits is to enhance the quality of ?nancial reporting. Prior studies have
provided several empirical surrogates to measure audit quality. These include audit
?rm size, audit industry specialization, and the length of the auditor-client relationship.
ARJ
26,1
58
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Beginning with DeAngelo (1981), decades of research have shown that large audit ?rms
with greater resources and more reputation at stake perform higher quality audits
(Palmrose, 1986; Beatty, 1989; Craswell and Taylor, 1995; Lennox, 1999; Teoh and
Wong, 1993; Menon and Williams, 2004; Houmes et al., 2012, etc.). Using accruals to
Figure 1.
Annual mean CAPM
returns highly valued vs
other 1991-2008
–0.2
–0.1
0
0.1
0.2
0.3
0.4
0.5
Year 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
M
e
a
n
R
e
t
u
r
n
s
Mean Returns of Firms in the Highest Quintile of Price-to-Earnings Ratios Mean Returns Other
Figure 2.
Annual median CAPM
returns highly valued vs
other 1991-2008
–0.2
–0.15
–0.1
–0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
M
e
d
i
a
n
R
e
t
u
r
n
s
Median Returns of Firms in the Highest Quintile of Price-to-Earnings Ratios Median Return Other
Audit quality
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
proxy for earnings quality, Becker et al. (1998) show that clients of large (Big 6) audit
?rms report lower discretionary accruals. In a similar study, Francis et al. (1999)
examine a large sample of NASDAQ ?rms and ?nd that although Big 6 audit
clients have higher levels of total accruals, they have lower estimated discretionary
accruals. Krishnan (2003) provides evidence that investors ascribe higher values to
the discretionary accruals of Big 6 clients than non-Big 6 clients. Heninger (2001)
documents that the likelihood of litigation increases with levels of total and
discretionary accruals, but decreases if a Big 5 auditor. Houmes and Skantz (2010)
provide evidence that the magnitude of the inverse relation between operating cash ?ow
and accruals increases if the ?rm is highly valued and assert that incentives to manage
earnings increase when operating cash ?ow is weak. Although they do not offer any
direct tests relating to the effect of high valuation on audit quality after excluding loss
?rms, they ?nd (not tabulated) evidence large Big N auditors reduce this tendency.
Balsam et al. (2003) shows that clients of within Big 6 and Big 5 auditor industry
specialists have lower accruals and higher earnings response coef?cients. Myers et al.
(2003) document that discretionary accruals decrease with the length of audit
?rm-client relationship. Mansi et al. (2004) show that the cost of debt decreases with
tenure and Carcello and Nagy (2004) document an increase in the incidence of ?nancial
reporting fraud when audit ?rm tenure is less than three years.
Although audit opinions enhance the credibility and reliability of ?nancial reports,
they also re?ect a negotiation dimension and, within the ethical and technical con?nes
of accounting standards, a ?rm’s published ?nancial report may be perceived as a
joint statement from the manager and auditor (Antle and Nalebuff, 1991). Gibbons et al.
(2001) use a sample of 93 experienced audit partners to report that auditor-client
negotiation occurs on a regular basis. In particular, results show that negotiation is
common, with 67 percent of audit partners experiencing negotiation with 50 percent or
more of their clients. Hence, the ?nal reporting product is often the result of a
compromise between management and auditors (Ellingsen et al., 1989).
Since managers of highly valued ?rms are under increased pressure to meet
optimistic earnings forecasts, they have incentives to assume a more aggressive
negotiating stance, prevailing on auditors to, within the con?nes of existing accounting
standards, report higher earnings. Hence, income increasing negotiations that bias
earnings upward should be more prevalent for highly valued ?rms. If the auditor relents,
earnings quality will be impaired. Consequently, audit quality for highly valued ?rms
could be negatively affected. Antle and Nalebuff (1991) demonstrate that when joint
auditor-client welfare is maximized, ex post income reporting is biased upward.
Although executives have numerous opportunities to manage earnings, accruals are
particularly appealing since they are a normal, frequent, and expected component of
the ?nancial reporting process. Numerous prior studies provide evidence that earnings
are managed with discretionary accruals (Subramanyam, 1996; Bergstresser and
Philippon, 2006, etc.).
Since highly valued companies are under greater pressure to meet earnings
expectations, we expect that managers will assume a more intransigent negotiating
stance with the auditor and utilize accruals to increase earnings. Consequently, the
tendency for high quality audit ?rms to constrain accruals will be diminished for highly
valued ?rms as auditors acquiesce towards the upper bounds of accounting standards
constraints in the face of increased client pressure.
ARJ
26,1
60
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Finally, auditors may suffer, at least in some measure, from the same information
asymmetry that characterizes the relationship between managers and owners. While the
asymmetry may not be as great, managers nevertheless are closer to the ?rms they lead
than external auditors. At the very least, since managers of highly valued ?rms have
better knowledge of their company’s economic performance, they will be inclined to
accept undetected statement errors in their favor and protest only those that are adverse.
We test these assertions with the following hypotheses stated in alternative form:
H1. The magnitude of the inverse relation between the discretionary accruals of
the clients of large audit ?rms and the discretionary accruals of the clients of
other audit ?rms decreases if the clients are highly valued.
H2. The magnitude of the inverse relation between the discretionary accruals
of the clients of industry specialist audit ?rms and the discretionary accruals
of the clients of other audit ?rms decreases if the clients are highly valued.
H3. The magnitude of the inverse relation between the discretionary accruals of
audit ?rms’ clients with long tenure and the discretionary accruals of clients
with shorter tenure decreases if the clients are highly valued.
H4. The magnitude of the inverse relation between the discretionary accruals of the
clients of audit industry specialist audit ?rms with long tenure and the
discretionary accruals of other clients decreases if the clients are highly valued.
Sample and methodology
Sample
For years 1998-2009, we acquire panel data for ?nancial statement variables from the
Compustat ?les of companies. Similar to prior accruals studies, we remove all ?nancial
services (SIC codes 6000-6999) and regulated (4900-4999) ?rms. As explained below,
to control for the Big N audit quality effect on our specialists and long tenure auditor
variables, we separate the overall sample of ?rms into two groups. One group, our
Big N sample, contains Big N and non-Big N auditors. The other group, our
specialist-tenure sample contains Big N audit clients only. After deleting ?rms with
missing data for model variables, the total number of ?rm year observations in our Big
N sample is 29,405. Eliminating non-Big N audit ?rms reduces our specialist-tenure
sample to 23,937 observations. Although we also provide regression results for all
discretionary accruals, consistent with our assertion that managers of highly valued
?rms manage earnings higher, our primary tests involve income increasing
discretionary accruals only. Eliminating negative discretionary accruals reduces our
Big N (specialist-tenure) sample to 16,346 (13,172) ?rm year observations.
Discretionary accruals
Our dependent variable is discretionary accruals (DAC
it
). For all ?rms, discretionary
accruals are estimated using the cross sectional version of the modi?ed Jones model ( Jones,
1991). The modi?edJones model has beenusedina varietyof researchsettings (Becker et al.,
1998; Francis et al., 1999; Reynolds andFrancis, 2000, etc.). The model is speci?edas follows:
TAC
it
¼ b
0
þb
1
1
AT
it21
þb
2
ðDREV
it
2DAR
it
Þ þb
3
PPE
it
þ1
it
ð1Þ
Audit quality
61
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where TAC
it
is the difference between ?rm i’s year t earnings before extraordinary items
and net cash ?owfromoperations scaled by beginning of year (t 2 1) assets; AT
it21
is ?rm
i’s beginning of the year t total assets; DREV
it
is the difference inyear t andyear t 2 1 sales;
DAR
it
is the difference between year t and year t 2 1 trade account receivables; and PPE
it
is
net property plant and equipment both scaled by AT
it21
. For each year and two-digit SIC
industry group, discretionary accruals (DAC
it
) are estimated cross-sectionally as the
residuals from the above regression. Stated alternatively, coef?cients from equation (1) are
used to estimate predicted accruals for each year in each SIC group. Discretionary accruals
are the difference between each ?rm’s actual and predicted accruals, i.e.:
DAC
it
¼ TAC
it
2½b
0
þb
1
1
1=AT
it21
þb
2
ðDREV
it
2DAR
it
Þ þb
3
PPE
it
?: ð2Þ
To mitigate the potential effects of extreme observations onresults, we winsorize accruals at
the top and bottom 1 percent levels across the entire pooled sample. In our models we also
include controls that may affect the cross sectional variability in the dependent variable,
DAC
it
.
Control variables
Relative to other companies, certain industries or ?rms may tend to generate higher
accruals. In addition, it is natural that growth companies with increasing earnings and
investments in working capital are more likely to produce greater accruals, and prior
studies showthat growth ?rms report higher accruals (McNichols, 2000). To control for the
possibility that companies with greater total accruals may also have larger discretionary
accruals that our accruals model does not capture, we include total accruals (ACRL
it
) in our
multivariate tests measured as the difference between ?rm i’s year t earnings before
extraordinary items and net cash ?owfromoperations scaled by beginning of year assets.
Accruals studies typically control for size effects. Dechow and Dichev (2002) show
that larger ?rms record larger accruals. Also, larger ?rms with larger investor
following and more developed and sophisticated ?nancial reporting systems may
affect accrual levels (Becker et al., 1998; Reynolds and Francis, 2000). For each ?rm i we
include the end of ?scal year t natural log of total assets (LnASSET
it
).
Reynolds and Francis (2000) provide evidence that the tendency to manage earnings
increases with leverage. DeFond and Jiambalvo (1994) show that accruals are related to
debt covenant breeches. In addition, debt may serve as a monitoring mechanism that
constrains earnings management. To control for the effect that high debt levels may
have on accruals, we include the variable LEV
it
measured as ?rm i’s end of year t long
term debt scaled by t 2 1 total assets.
Operating cash ?ows are a component of earnings and their levels correspond
inversely with accruals. Further, the level of cash ?owmay affect the ability and/or need
to use accruals, causing ?rms with higher (lower) operating cash ?ows to report lower
(higher) discretionary accruals (Becker et al., 1998). We control for these effects by
including operating cash ?ow de?ated by the beginning of the year total assets (OCF
it
).
Kothari et al. (2005) show that discretionary accruals are impacted by ?nancial
performance. Accordingly, weinclude ROA
it
, income before extraordinary items
divided by the beginning of the year total assets. Finally, for 47 of the 48 two-digit SIC
industry codes included in our study we assign dummies equal to 1 to control for
ARJ
26,1
62
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
potential industry effects[1]. Similar to our dependent variable, DAC
it
, we winsorize all
continuously measured independent variables at the 1 percent levels.
Variables of interest: audit quality and highly valued equity
Using our initial all accruals sample, we identify highly valued clients (HV
it21
) as ?rms in
the highest quintile of PEs (i.e. P/E . 27.204) and assign an indicator variable equal to 1 if
the client is in this quintile of prior ?scal year end price-to-earnings ratios and 0 otherwise.
Our audit quality variables are as follows: BIGN
it21
equal to 1 and 0 otherwise if the
audit ?rmis one of the largest four, ?ve, or six audit ?rms (depending onthe time period);
SPEC
it21
, equal to 1 and 0 otherwise if the audit ?rm is an industry specialist; and
TEN
it21
equal to 1 and 0 otherwise if the length of the auditor-client relationship is equal
to or over ?ve years. We also include an additional audit quality indicator variable if the
auditor is both an industry specialist and has a long tenure auditor-client relationship
(SPECTEN
it21
). In accordance with prior studies (Palmrose, 1986), we de?ne an auditor
as anindustry specialist according to the auditor’s two-digit SICcode industry share and
designate the ?rm as a specialist based on the number of clients and total client sales.
In particular, an indicator variable equal to 1 and0 otherwise is assigned to auditors with
both the largest number of clients and the highest total client sales within each two-digit
SIC code. To eliminate the brand name effect of BN auditors, we restrict the sample of
specialist and tenure ?rms to Big N clients only. We eliminate SIC code industries with
fewer than 15 observations. The income increasing discretionary accruals, Big N, and
non-Big N sample consists of 16,346 ?rm year observations. The income increasing
discretionary accruals, specialist-tenure sample contains 13,172 observations.
To investigate the relation between high valuations and the tendency of high audit
quality auditors to mitigate accruals, we interact the audit quality variables with our
highly valued equity dummy. Statistically signi?cantly positive estimates for the audit
quality, highly valued equity interaction terms: BIGN
*
HV
it21
, SPEC
*
HV
it21
,
TEN
*
HV
it21
and SPECTEN
*
HV
it21
, provide support for hypotheses that incentives
associated with high valuations reduce the tendency of high quality audit ?rms to
constrain accruals. Our models are as follows:
DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
BIGN
it21
þa
7
HV
it21
þa
8
BIGN
*
HV
it21
þa
9
INDUS þ1
it
ð3Þ
DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
SPEC
it21
þa
7
HV
it21
þa
8
SPEC
*
HV
it21
þa
9
INDUS þ1
it
ð4Þ
DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
TEN
it21
þa
7
HV
it21
þa
8
TEN
*
HV
it21
þa
9
INDUS þ1
it
ð5Þ
DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
SPECTEN
it21
þa
7
HV
it21
þa
8
SPECTEN
*
HV
it21
þa
9
INDUS þ1
it
ð6Þ
Audit quality
63
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Results
Descriptive statistics
Table I de?nes variables used in models. Tables II and III provide descriptive
statistics for variables. Table II Panels A and B shows the maximum, minimum, mean,
and median values of variables in our income increasing Big N (Panel A) and
specialist-tenure (Panel B) samples. Table III provides univariate correlations. Mean
discretionary accruals (scaled) for both samples are 0.064 and 0.059, respectively[2].
In our BNauditor sample 82 percent of the observations involve large audit ?rms. In our
specialist-tenure sample, 20 percent of the observations involve specialists and
54 percent have audit-client tenures over ?ve years. Hence most Big N audit-client
relationships are greater than ?ve years. 10 percent of the observations are identi?ed as
both a specialist and long tenure audit client.
Univariate correlations in Table III reveal that size, leverage and operating cash ?ow,
are inversely related to discretionary and total accruals. ROA is positively (negatively)
related to total (discretionary) accruals. ROAand operating cash ?oware also positively
correlated with our HV
it21
indicator variable. While coef?cients are generally positive
between each of our audit quality variables and HV
it21
, only Big N is statistically
signi?cant. Overall, the signs and signi?cance of our other model variables are
consistent across all proxies for audit quality. BIGN, SPEC, TEN, and SPECTEN are
positively (negatively) correlated with size, operating cash ?owand ROA(discretionary
accruals).
Table IVprovides the differences in mean scaled discretionary accruals between each
of the high quality auditor measures and their non-high quality auditor counterparts.
DAC
it
are Firm i’s ?scal year t income increasing discretionary accruals
ACRL
it
are Firm i’s ?scal year t total accruals
LnASSET
it
are Firm i’s ?scal year t natural log of total assets
LEV
it
are Firm i’s ?scal year t long term debt divided by ?scal year t 2 1 total assets
OCF
it
are Firm i’s ?scal year t cash ?ow from operating activates divided by ?scal year
t 2 1 total assets
ROA
it
are fFrm i’s ?scal year t income before extraordinary items divided by ?scal year
t 2 1 total assets
HV
it21
are An indicator variable equal to 1 for ?rms in the highest quintile of ?scal year
t 2 1 price-to-earnings ratios
AQ
it21
are Four measures of audit quality de?ned as follows
BIGN
it21
are An indicator variable equal to 1 if at the end of ?scal year t 2 1 the auditor is a
Big N auditor
SPEC
it21
are An indicator variable equal to 1 if at the end of ?scal year t 2 1 the auditor is an
industry specialist
TEN
it21
are An indicator variable equal to 1 if at the end of ?scal year t 2 1 the auditor
tenure is greater than or equal to ?ve years
SPECTEN
it21
are An indicator variable equal to 1 if at the end of ?scal year t 2 1 the auditor is an
industry specialist and tenure is greater than or equal to ?ve years
BIGN
*
HV
it21
are An interaction term between BIGN
it21
and HV
it21
;
SPEC
*
HV
it21
are An interaction term between SPEC
it21
and HV
it21
TEN
*
HV
it21
are An interaction term between TEN
it21
and HV
it21
SPECTEN
*
HV
it21
are An interaction term between SPECTEN
it21
and HV
it21
INDUS are An indicator variable equal to1for ?rms in each two-digit SIC code
Table I.
Variable de?nitions
ARJ
26,1
64
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Results show that relative to other audit ?rms, discretionary accruals are lower for each
of the four audit quality proxy measures ( p , 0.001) with the greatest difference
occurring between Big N and non-Big N auditors (20.025).
This study predicts that incentives related to high valuations reduce the accruals
decreasing effect of high quality auditors. To initially investigate this assertion, within
each price-to-earnings group (top 20 PE and bottom 80 PE) we compare the mean
discretionary accruals between clients of high quality auditors and other audit clients.
Results show that while high quality audit clients report lower accruals vis-a` -vis other
clients in both price-to-earnings groups, for three of our audit quality measures
differences generally decrease when the client ?rm is in the highest PE quintile.
In particular, for three of the four audit quality variables, the difference in scaled
discretionary accruals between the clients of high quality auditors and the clients of
other auditors is not signi?cant when the ?rm is highly valued. For example, the
difference in mean accruals between specialist and non-specialist auditors in the top
quintile of price-to-earnings ratios is an insigni?cant 20.003 (0.349), while the difference
in mean accruals for ?rms in the bottom 80 percent is 20.006 (0.000). In addition, the
differences in accruals between the top 20 PE, TEN and SPECTENaudit clients and the
accruals of the corresponding top 20 PE, NTEN, and NSPECTENaudit quality clients is
an insigni?cant 20.002 and 20.001 ( p ¼ 0.296 and 0.788, respectively). Differences in
accruals between the bottom 80 PE, TEN, and SPECTEN are, however, signi?cantly
lower for high quality auditors. Differences in accruals between the discretionary
accruals of high quality specialists, long tenure, and specialist – long tenure audit
clients and other lower quality audit clients decrease if the client is highly valued.
The difference, however, in mean accruals for the top quintile of Big N versus NBN
auditors is signi?cant at the p , 0.001 level. Hence, the reduction in accruals between
Variables Maximum Minimum Mean Median SD
Panel A. Big N auditor-income increasing discretionary accruals (n ¼ 16,346)
DAC
it
0.331 0.000 0.064 0.044 0.066
ACRL
it
0.256 20.544 20.014 20.026 0.077
LnASSET
it
32.140 22.120 26.603 26.520 2.157
LEV
it
1.196 0.000 0.183 0.123 0.215
OCF
it
0.428 20.759 0.040 0.068 0.159
ROA
it
0.247 21.318 0.010 0.043 0.177
BIGN
it
1.000 0.000 0.820 1.000 0.388
HV
it21
1.000 0.000 0.213 1.000 0.401
Panel B. Specialists and long tenure auditors – income increasing discretionary accruals (n ¼ 13,172)
DAC
it
0.331 0.000 0.059 0.041 0.062
ACRL
it
0.269 20.544 20.017 20.028 0.073
LnASSET
it
32.360 21.990 27.021 26.978 0.890
LEV
it
1.498 0.000 0.197 0.144 0.223
OCF
it
0.498 21.708 0.048 0.073 0.168
ROA
it
0.848 22.576 0.015 0.045 0.187
SPEC
it
1.000 0.000 0.195 0.000 0.396
TEN
it
1.000 0.000 0.540 1.000 0.498
SPETENC
it
1.000 0.000 0.099 0.000 0.298
HV
it21
1.000 0.000 0.231 0.000 0.398
Note: Variables are de?ned in Table I
Table II.
Descriptive statistics
Audit quality
65
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
P
a
n
e
l
A
.
V
a
r
i
a
b
l
e
c
o
r
r
e
l
a
t
i
o
n
s
f
o
r
B
i
g
N
a
u
d
i
t
o
r
s
–
i
n
c
o
m
e
i
n
c
r
e
a
s
i
n
g
d
i
s
c
r
e
t
i
o
n
a
r
y
a
c
c
r
u
a
l
s
(
n
¼
1
6
,
3
4
6
)
V
a
r
i
a
b
l
e
D
A
C
i
t
A
C
R
L
i
t
L
n
A
S
S
E
T
i
t
L
E
V
i
t
O
C
F
i
t
R
O
A
i
t
B
I
G
N
H
V
i
t
2
1
D
A
C
i
t
1
0
.
6
2
6
(
0
.
0
0
0
)
2
0
.
2
8
9
(
0
.
0
0
0
)
2
0
.
0
4
6
(
0
.
0
5
0
)
2
0
.
3
5
3
(
0
.
0
0
0
)
2
0
.
0
7
0
(
0
.
0
0
0
)
2
0
.
1
4
6
(
0
.
0
0
0
)
2
0
.
0
2
9
(
0
.
0
0
0
)
A
C
R
L
i
t
1
2
0
.
1
6
7
(
0
.
0
0
0
)
2
0
.
0
7
2
(
0
.
0
0
0
)
2
0
.
3
3
6
(
0
.
0
0
0
)
0
.
1
1
9
(
0
.
0
0
0
)
2
0
.
0
7
3
(
0
.
0
0
0
)
2
0
.
0
0
3
(
0
.
7
4
6
)
L
n
A
S
S
E
T
i
t
1
0
.
2
7
5
(
0
.
0
0
0
)
0
.
3
7
9
(
0
.
0
0
0
)
0
.
3
3
0
(
0
.
0
0
0
)
0
.
4
0
7
(
0
.
0
0
0
)
0
.
1
0
1
(
0
.
0
0
0
)
L
E
V
i
t
1
0
.
0
3
1
(
0
.
0
0
0
)
0
.
0
2
2
(
0
.
0
0
4
)
0
.
1
1
7
(
0
.
0
0
0
)
2
0
.
0
4
3
(
0
.
0
0
0
)
O
C
F
i
t
1
0
.
8
0
1
(
0
.
0
0
0
)
0
.
1
1
8
(
0
.
0
0
0
)
0
.
1
2
2
(
0
.
0
0
0
)
R
O
A
i
t
1
0
.
0
8
5
(
0
.
0
0
0
)
0
.
1
1
6
(
0
.
0
0
0
)
B
I
G
N
i
t
1
0
.
0
3
3
(
0
.
0
0
0
)
H
V
i
t
2
1
1
P
a
n
e
l
B
.
V
a
r
i
a
b
l
e
c
o
r
r
e
l
a
t
i
o
n
s
f
o
r
s
p
e
c
i
a
l
i
s
t
s
a
n
d
l
o
n
g
t
e
n
u
r
e
a
u
d
i
t
o
r
s
–
i
n
c
o
m
e
i
n
c
r
e
a
s
i
n
g
d
i
s
c
r
e
t
i
o
n
a
r
y
a
c
c
r
u
a
l
s
(
n
¼
1
3
,
1
7
2
)
V
a
r
i
a
b
l
e
D
A
C
i
t
A
C
R
L
i
t
L
n
A
S
S
E
T
i
t
L
E
V
i
t
O
C
F
i
t
R
O
A
i
t
S
P
E
C
T
E
N
S
P
E
C
T
E
N
H
V
i
t
2
1
D
A
C
i
t
1
0
.
6
5
0
(
0
.
0
0
0
)
2
0
.
2
6
2
(
0
.
0
0
0
)
2
0
.
0
2
5
(
0
.
0
0
4
)
2
0
.
.
3
3
1
(
0
.
0
0
0
)
2
0
.
0
6
2
(
0
.
0
0
)
2
0
.
0
3
6
(
0
.
0
0
0
)
2
0
.
0
4
7
(
0
.
0
0
0
)
2
0
.
0
4
1
(
0
.
0
0
0
)
2
0
.
0
2
4
(
0
.
0
0
6
)
A
C
R
L
i
t
1
2
0
.
1
7
8
(
0
.
0
0
0
)
2
0
.
0
7
1
(
0
.
0
0
0
)
2
0
.
3
1
9
(
0
.
0
0
0
)
0
.
0
8
8
(
0
.
0
0
0
)
2
0
.
0
0
9
(
0
.
3
1
9
)
2
0
.
0
1
1
(
0
.
2
2
4
)
2
0
.
0
2
1
(
0
.
0
1
6
)
2
0
.
0
0
8
(
0
.
3
8
2
)
L
n
A
S
S
E
T
i
t
1
0
.
2
4
6
(
0
.
0
0
0
)
0
.
3
6
9
(
0
.
0
0
0
)
0
.
3
2
3
(
0
.
0
0
0
)
0
.
0
5
5
(
0
.
0
0
0
)
0
.
1
5
4
(
0
.
0
0
0
)
0
.
0
9
0
(
0
.
0
0
0
)
0
.
0
8
6
(
0
.
0
0
0
)
L
E
V
i
t
1
0
.
0
1
3
(
0
.
1
4
2
)
0
.
0
0
7
(
0
.
4
4
8
)
0
.
0
3
5
(
0
.
0
0
0
)
2
0
.
0
0
2
(
0
.
7
8
0
)
0
.
0
2
5
(
0
.
0
0
4
)
2
0
.
0
4
9
(
0
.
0
0
0
)
O
C
F
i
t
1
0
.
8
2
8
(
0
.
0
0
0
)
0
.
0
3
7
(
0
.
0
0
0
)
0
.
0
7
2
(
0
.
0
0
0
)
0
.
0
5
1
(
0
.
0
0
0
)
0
.
1
0
6
(
0
.
0
0
0
)
R
O
A
i
t
1
0
.
0
2
8
(
0
.
0
0
1
)
0
.
0
7
5
(
0
.
0
0
0
)
0
.
0
3
9
(
0
.
0
0
0
)
0
.
1
0
2
(
0
.
0
0
0
)
S
P
E
C
i
t
1
2
0
.
0
3
2
(
0
.
0
0
0
)
0
.
6
7
3
(
0
.
0
0
0
)
0
.
0
0
5
(
0
.
5
9
1
)
T
E
N
i
t
1
0
.
3
0
7
(
0
.
0
0
0
)
0
.
0
0
5
(
0
.
5
5
9
)
S
P
E
C
T
E
N
i
t
1
2
0
.
0
0
6
(
0
.
4
7
1
)
H
V
i
t
2
1
1
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
c
e
(
p
:
t
w
o
-
t
a
i
l
e
d
)
i
s
s
h
o
w
n
i
n
p
a
r
e
n
t
h
e
s
e
s
u
n
d
e
r
c
o
e
f
?
c
i
e
n
t
s
;
v
a
r
i
a
b
l
e
s
a
r
e
d
e
?
n
e
d
i
n
T
a
b
l
e
I
Table III.
ARJ
26,1
66
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
BN and NBN auditors is consistent for both PE groups. An important limitation of this
analysis is that it omits control variables that may affect accruals levels. The following
section includes results of our main tests using multivariate analysis that includes
previously described control variables (Table V).
Discretionary accruals models
Table VI shows results of main tests for our income increasing discretionary accruals
sample for models depicted in equations (3) through (6). Adjusted R
2
exceed 50 percent.
Total accruals, ACRL
it
, is signi?cant and positive. As expected and in accordance with
prior accruals studies, estimates for assets and operating cash ?ow are negative,
suggesting that greater ?rm size, and operating cash ?ow enhance accruals quality. The
coef?cients for ROA
it
are positive. Discretionary accruals increase with (scaled) income.
Audit quality
variables
Mean disc.
accruals n
Audit quality
variables
Mean disc.
accruals n
BN 0.060 13,320 TEN 0.057 7,125
NBN 0.085 3,026 NTEN 0.063 6,047
Difference 20.025 (0.000) Difference 20.006 (0.000)
SPEC 0.055 2,568 SPECTEN 0.053 1,300
NSPEC 0.061 10,604 NSPECTEN 0.061 11,872
Difference 20.006 (0.000) Difference 20.008 (0.000)
Notes: Variable de?nitions: BN – ?rm is audited by a Big N auditor; NBN – ?rm is not audited by a
Big N auditor; SPEC – auditor is an industry specialist; NSPEC – auditor is not an industry specialist;
TEN – audit-client ?rmtenure is greater than or equal to ?ve years; NTEN – audit-client ?rmtenure is
less than ?ve years; SPECTEN – audit-client ?rm tenure is greater than or equal to ?ve years and the
auditor is an industry specialist; NSPECTEN – ?rm’s auditor is not a SPECTEN auditor; n – number
of ?rm year observations
Table IV.
Difference in mean
income increasing
discretionary accruals
between high quality:
Big N, specialist and
long tenure auditors and
other auditors
BN n NBN n Difference n
Top 20 PE 0.057 2,939 0.078 543 20.021 (0.000) 3,482
Bottom 80 PE 0.060 10,381 0.086 2,483 20.026 (0.000) 12,864
SPEC n NSPEC n Difference n
Top 20 PE 0.054 603 0.057 2,450 20.003 (0.349) 3,053
Bottom 80 PE 0.056 1,965 0.062 8,154 20.006 (0.000) 10,119
TEN n NTEN n Difference n
Top 20 PE 0.056 1,695 0.058 1,358 20.002 (0.296) 3,053
Bottom 80 PE 0.057 5,430 0.064 4,689 20.007 (0.000) 10,119
SPECTEN n NSPECTEN n Difference n
Top 20 PE 0.056 297 0.057 2,756 20.001 (0.788) 3,053
Bottom 80 PE 0.052 1,003 0.062 9,116 20.010 (0.000) 10,119
Notes: Variable de?nitions: top 20 PE – ?rm is in the highest 20 percent of price-to-earnings ratios;
bottom 80 PE – ?rm is in the lowest 80 percent of price-to-earnings ratios; BN – ?rm is audited by a
Big N auditor; NBN – ?rm is not audited by a Big N auditor; SPEC – auditor is an industry specialist;
NSPEC – auditor is not an industry specialist; TEN – audit-client ?rm tenure is greater than or equal
to ?ve years; NTEN – audit-client ?rm tenure is less than ?ve years; SPECTEN – audit-client ?rm
tenure is greater than or equal to ?ve years and the auditor is an industry specialist; NSPECTEN –
?rm’s auditor is not a SPECTEN auditor; n – number of ?rm year observations
Table V.
Difference in mean
discretionary income
increasing accruals
between audit quality
variables for ?rms in
the highest 20 and
lowest 80 percent of
price-to-earnings ratios
Audit quality
67
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
HV
it21
is negative. Since our hypotheses are directional, tests for all our audit quality
variables (both main effects and interaction terms) are one-tailed. Similar to univariate
tests inthe previous sections, Table VI shows that the coef?cients for the BigN, specialist,
and long tenure-specialist audit quality measures are negative with estimates and ( p)
values of 20.006 (0.000), 20.002 (0.061), and 20.002 (0.065), respectively. However, the
estimate for our audit quality variable, TEN
it
, is an insigni?cant 20.001 (0.364).
We test hypotheses with interaction terms between our highly valued equity and audit
quality variables: BIGN
*
HV
it21
, SPEC
*
HV
it21
, TEN
*
HV
it21
, and SPECTEN
*
HV
it21
.
Respective estimates and p-values for the income increasing accruals sample are 0.004
(0.044), 0.003 (0.087), 0.003 (0.055), and 0.006 (0.003). For each of the four alternative audit
quality measures, discretionary accruals increase when the client ?rm is highly valued.
As described above, since incentives related to managing earnings higher should be
associated with higher discretionary accruals, we restrict tests in Table VI to clients
reporting income increasing discretionary accruals. Nevertheless, alternative factors may
induce negative accruals. For example, auditors with a conservative bias may provoke
income decreasingaccountingchoices (Becker et al., 1998) andthis maybe especiallylikely
if the ?rm is highly valued. To consider this possibility and other factors related to lower
accruals, we re-estimate equations (3)-(6) using signed (positive and negative)
discretionary accruals as the dependent variable. Table VII provides ?ndings from
tests of models using this all accruals sample. While results for controls are similar across
both samples, in contrast to Table VI, leverage is inversely related to discretionary
accruals. The coef?cients for the Big N, specialist, and long tenure-specialist audit quality
measures are negative with estimates and ( p) values of 20.005 (0.000), 20.002 (0.026),
BN Specialist Tenure Specialist and tenure
coef?cient
( p-value)
coef?cient
( p-value)
coef?cient
( p-value)
coef?cient
( p-value)
ACRL
it
0.546 (0.000) 0.563 (0.000) 0.561 (0.000) 0.561 (0.000)
LnASSET
it
20.005 (0.000) 20.004 (0.000) 20.004 (0.000) 20.004 (0.000)
LEV
it
0.008 (0.000) 0.009 (0.000) 0.008 (0.000) 0.008 (0.000)
OCF
it
20.042 (0.091) 20.046 (0.000) 20.046 (0.000) 20.046 (0.000)
ROA
it
0.001 (0.000) 0.011 (0.024) 0.010 (0.028) 0.010 (0.030)
HV
it21
20.006 (0.007) 20.003 (0.006) 20.004 (0.004) 20.003 (0.004)
BIGN
it21
20.006 (0.000)
BIGN
*
HV
it21
0.004 (0.044)
SPEC
it21
20.002 (0.061)
SPEC
*
HV
it21
0.003 (0.087)
TEN
it21
20.001 (0.364)
TEN
*
HV
it21
0.003 (0.055)
SPECTEN 20.002 (0.065)
SPECTEN
*
HV
it21
0.006 (0.003)
INDUS Not reported Not reported Not reported Not reported
Adjusted R
2
0.507 0.533 0.532 0.532
F( p) 0.000 0.000 0.000 0.000
n 16,346 13,172 13,172 13,172
Notes: DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
HV
it21
þ
a
7
AQ
it
þa
8
AQ
*
it
HV
it21
þa
9
INDUS þ1it; variables are de?ned in Table I
Table VI.
Results for effects
of overvaluation on the
income increasing
discretionary accruals
of Big N and audit
specialist clients
ARJ
26,1
68
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
and 2 0.002 (0.026), respectively, and once again the estimate for audit quality variable,
TEN
it2
, is not signi?cant 20.001 (0.320). Big N and specialist estimates are similar to
other audit quality – accruals studies. For example, Becker et al. (1998) use a sample of
10,881 ?rm year observations from the 1989 to 1992 period to report that non-Big
6 audit accruals are 1.5 percent higher than Big 6 accruals. In a similar study Balsamet al.
(2003) reports 20.006 estimate on an audit quality dummy from a regression of
discretionary accruals on controls and an industry leadership indicator variable.
Regarding TEN
it
, prior audit tenure studies con?ict on how differing lengths of
auditor-client tenure re?ect audit quality. For example, Johnson et al. (2002) report that
relative to medium tenures of four to six years ?nancial reporting quality as measured
by accruals decreases when the tenure is short, but ?nd no reduction in audit quality
for long tenure auditors greater than nine years. Myers et al. (2003) generally con?rm
these ?ndings. However, Davis et al. (2003) ?nd that discretionary accruals increase
over years when audit tenure is continuously measured. Although we believe our
choice of tenure equal to or greater than ?ve years re?ects the most ef?cacious measure
of audit quality from the existing literature, a possible explanation for the lack of
signi?cance may simply re?ect that documented in prior studies.
As depicted in Table VII, results of our main tests are unchanged from Table VI as
all audit quality-highly valued equity interaction terms, BIGN
*
HV
it21
, SPEC
*
HV
it21
,
TEN
*
HV
it21
, and SPECTEN
*
HV
it21
remain statistically signi?cant and positive
0.003 (0.061), 0.004 (0.004), 0.002 (0.073), and 0.004 (0.005).
These ?ndings provide evidence that relative to other ?rms, the accruals decreasing
effect of high quality auditors is reduced when the client is highly valued. Furthermore,
these results are robust across several measures of audit quality with interaction term,
BN Specialist Tenure Specialist and tenure
coef?cient
( p-value)
coef?cient
( p-value)
coef?cient
( p-value)
coef?cient
( p-value)
ACRL
it
0.798 (0.000) 0.809 (0.000) 0.809 (0.000) 0.809 (0.000)
LnASSET
it
20.003 (0.000) 20.002 (0.000) 20.002 (0.000) 20.002 (0.000)
LEV
it
20.004 (0.004) 20.003 (0.024) 20.003 (0.057) 20.003 (0.048)
OCF
it
20.023 (0.000) 20.018 (0.000) 20.019 (0.000) 20.018 (0.000)
ROA
it
0.011 (0.000) 0.010 (0.218) 0.010 (0.000) 0.010 (0.000)
HV
it21
20.006 (0.002) 20.003 (0.000) 20.004 (0.000) 20.003 (0.000)
BIGN
it21
20.005 (0.000)
BIGN
*
HV
it21
0.003 (0.061)
SPEC
it21
20.002 (0.026)
SPEC
*
HV
it21
0.004 (0.004)
TEN
it21
20.001 (0.320)
TEN
*
HV
it21
0.002 (0.073)
SPECTEN 20.002 (0.026)
SPECTEN
*
HV
it21
0.004 (0.005)
INDUS Not reported Not reported Not reported Not reported
Adjusted R
2
0.747 0.760 0.760 0.760
F( p) 0.000 0.000 0.000 0.000
N 29,405 23,937 23,937 23,937
Notes: DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
HV
it21
þ
a
7
AQ
it
þa
8
AQ
*
it
HV
it21
þa
9
INDUS þ1
it
; variables are de?ned in Table I
Table VII.
Results for effects
of overvaluation on the
discretionary accruals
of Big N and audit
specialist clients
Audit quality
69
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
SPECTEN
*
HV
it21
, showing the highest and most statistically signi?cant estimate
(0.006) from our income increasing – Big N only sample. Hence, high valuation
attenuates the accruals decreasing effect of a long tenure – industry specialist auditor
in that accruals are 0.6 percent of assets higher for clients in the highest quintile of PEs.
Unreported mean assets and common shares outstanding for SPECTEN clients are
$4,989,790,883 and 222,265,446 shares, respectively. Thus, on average net income is
increased by 0.006 times $4,989,790,883 or $29,938,745. Average earnings per share
increases by $0.135 per share ($29,938,745 divided by 222,265,446 shares). While not as
pronounced, the statistically signi?cant estimates for all other audit quality – high
valued client interaction terms similarly illustrate the potential economic impact of
increasing accruals on net income. Indeed, the unreported average diluted earnings
before extraordinary items per share for highly valued SPECTEN ?rms is $0.644,
suggesting that relatively small increases in our interaction coef?cients can have a
signi?cant economic impact on earnings per share.
Robustness
We conduct several sensitivity tests. Regarding valuation variable, HV
it21
, we
also alternatively measure price-to-earnings valuation by stratifying ?rms into
deciles (quintiles) and assign a 1 to companies in the lowest PE decile and quintile and
10 (5) to clients in the highest. Results are similar to our main ?ndings. In particular,
for both income increasing and all discretionary accruals samples, results show
signi?cantly negative (positive) coef?cients for audit quality variables (interaction
terms), BIGN, SPEC, and SPECTEN (BIGN
*
HV
it21
, SPEC
*
HV
it21
, and
SPECTEN
*
HV
it21
). TEN
*
HV
it21
is not signi?cant for either price-to-earnings
measure.
To control for factors occurring over years that our models may not capture, using
1998 as the base year, we include year dummies in each of our equations. With the
exception of TEN
*
HV
it21
( p ¼ 0.125) in our all accruals model, audit quality-highly
valued equity interaction terms are signi?cant and positive. Since ?nancing and/or
acquisition activities could impact accruals, we also assign indicator variables equal to
1 if common stock or debt increased by more than 10 percent and/or the ?rm reported
an acquisition. Results do not alter our ?ndings.
Negative prior year earnings may impact managers’ incentives to report current
year higher net income. We also run tests with a dummy equal to 1 if the ?rm reported
a loss in the prior year. In addition, to control for the impact of events not captured
in our models on our dependent variable, discretionary accruals, we further include
an abnormal returns variable measured according to the capital asset pricing model
(CAPM).
That is:
AR
it
¼ R
it
2ER
it
ð7Þ
CAPM:
ER
it
¼ R
ft
þb
it
ðR
mt
2R
ft
Þ ð8Þ
where, ER
it
is ?rm i’s expected end of ?scal year t return; R
ft
is the one month treasury
bill rate; and b
it
is the end of ?scal year t market model b calculated over the prior year
ARJ
26,1
70
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
with daily closing prices. R
mt
is the value-weighted return from the NYSE, AMEX, and
NASDAQ (from CRSP). Inclusion of these variables does not alter ?ndings as audit
quality – highly valued equity interaction terms remain generally signi?cant and
positive.
Discussion of results and conclusion
Using audit quality proxies from prior literature, this study examines the accruals
decreasing effect of high quality auditors on highly valued audit clients. We posit that
management incentives associated with highly valued equity reduce the tendency of
high quality auditors to reduce accruals. Although results generally support
hypotheses, reported increases in accruals may be affected by the particular measure of
audit quality used. While not as pronounced, the ?ndings of this study seem to provide
some evidence that high quality, Big N auditors continue to maintain their audit
quality edge, albeit marginally, over other audit ?rms. To reiterate, as per Table V, and
in contrast to our other audit quality characteristics, the difference in mean accruals
between Big N and non-Big N auditors is signi?cant even when the client is highly
valued. Hence, perhaps large audit ?rms with more clients and greater reputation at
stake are better able to mitigate the accruals increasing tendency of highly valued
?rms. Nevertheless, multivariate results with controls do reveal higher discretionary
accruals for highly valued-Big N audit clients. In addition to our main multivariate
tests, we investigate this conjecture further by eliminating long tenure and
specialist auditors from the Big N sample. For both our increasing and signed
accruals model unreported ?ndings show that while estimates for the BIGN
*
HV
it21
interaction terms are positive (0.004 and 0.005, respectively), neither are signi?cant
(0.170 and 0.206). While we do not further address this issue here, additional research
that attempts to disentangle the characteristics of large auditors, specialist auditors,
and long tenured auditors as they relate to highly valued audit clients might provide
additional insight.
Regarding other areas for further research, consideration might also be given to the
impact of corporate governance on the audit quality-highly valued equity relation.
In addition, future studies could investigate temporal aspects of audit quality and
highly valued equity both separately and in conjunction with each other. For example,
does the audit quality effect change before, during, and after periods of market tops
when prices and P/Es are at historic highs?
Notes
1. We include SIC dummies in regression models to provide an additional control for industry
effects that may not be captured by the within industry Jones model estimates.
2. As discussed above, discretionary accruals (the residuals from equation (1)) are estimated
using all (positive and negative) accruals. For the all accruals sample, the unreported mean
discretionary accruals are 0.002 for the Big N sample and 0.000 for the specialist and tenure
sample. As reported in Table II, average discretionary accruals are higher for our income
increasing accruals samples.
References
Anderson, K. and Brooks, C. (2006), “The long-term price-earnings ratio”, Journal of Business
Finance & Accounting, Vol. 7, pp. 1063-1086.
Audit quality
71
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Antle, R. and Nalebuff, B. (1991), “Conservatism and auditor-client negotiations”, Journal of
Accounting Research, Vol. 29, pp. 31-54.
Balsam, S., Krishnan, J. and Yang, J.S. (2003), “Auditor industry specialization and earnings
quality”, Auditing: A Journal of Practice & Theory, Vol. 22, pp. 71-97.
Basu, S. (1977), “Investment performance of common stocks in relation to their price
earnings ratios: a test of the ef?cient market hypothesis”, Journal of Finance, Vol. 32,
pp. 663-682.
Beatty, R. (1989), “Auditor reputation and the pricing of initial public offerings”, The Accounting
Review, Vol. 4, pp. 693-709.
Becker, C.L., Defond, M.L., Jiambalvo, J. and Subramanyam, K.R. (1998), “The effect of
audit quality on earnings management”, Contemporary Accounting Research, Vol. 15,
pp. 1-24.
Bergstresser, D. and Philippon, T. (2006), “CEO incentives and earnings management”, Journal of
Financial Economics, Vol. 80, pp. 511-529.
Campbell, J. and Shiller, R. (2001), “Valuation ratios and the long-run stock market outlook:
an update”, Cowles Foundation Discussion Paper No. 1295, Cowles Foundation for
Research in Economics, Yale University, New Haven, CT, March.
Carcello, J.V. and Nagy, A.L. (2004), “Audit ?rm tenure and fraudulent ?nancial reporting”,
Auditing: A Journal of Practice & Theory, Vol. 23, pp. 55-69.
Chi, J.D. and Gupta, M. (2009), “Overvaluation and earnings management”, Journal of Banking
& Finance, Vol. 33, pp. 1652-1663.
Chopra, N.J., Lakonishok, L. and Ritter, J.R. (1992), “Measuring abnormal performance: do stocks
overreact?”, Journal of Financial Economics, Vol. 31, pp. 235-268.
Craswell, A.J. and Taylor, S. (1995), “Auditor brand name reputations and industry
specialization and earnings quality”, Auditing: A Journal of Practice & Theory, Vol. 23,
pp. 297-322.
Davis, L.R., Soo, B.S. and Trompeter, G.M. (2003), “Auditor tenure and the ability to meet or beat
earnings forecasts”, Contemporary Accounting Research, Vol. 26, pp. 517-548.
DeAngelo, L.E. (1981), “Audit size and audit quality”, Journal of Accounting and Economics,
Vol. 3, pp. 183-199.
Dechow, P.M. and Dichev, D. (2002), “The quality of accruals and earnings: the role of accruals
estimation errors”, The Accounting Review, Vol. 77, pp. 35-59.
DeFond, M.L. and Jiambalvo, J. (1994), “Debt covenant violation and manipulation of accruals”,
Journal of Accounting and Economics, Vol. 17, pp. 145-176.
Efendi, J., Srivasta, A. and Swanson, E. (2007), “Why do corporate managers misstate ?nancial
statements? The role of option compensation and other factors”, Journal of Financial
Economics, Vol. 85, pp. 667-708.
Ellingsen, J.E., Pany, K. and Fagan, P. (1989), “SAS No. 59: how to evaluate going concern
opinions”, Journal of Accountancy, January, pp. 24-31.
Francis, J., Mayhew, E. and Sparks, H. (1999), “The role of big six auditors in
the credible reporting of accruals”, Auditing: A Journal of Practice & Theory, Vol. 18,
pp. 17-34.
Gibbons, M., Salterio, S. and Webb, A. (2001), “Evidence about auditor-client management
negotiation concerning clients’ ?nancial reporting”, Journal of Accounting Research,
Vol. 39, pp. 535-564.
ARJ
26,1
72
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Goldwasser, A. and Eickemeyer, J.H. (2004), “Introduction”, Accountant’s Liability, Chapter 1,
Practicing Law Institute, New York, NY.
Grimm, S.D. (2009), “The role of accounting quality in securities class action lawsuits”, working
paper, University of Minnesota, St Paul, MN, June.
Heninger, W.G. (2001), “The association between auditor litigation and abnormal accruals”,
The Accounting Review, Vol. 76, pp. 111-126.
Houmes, R. and Skantz, T. (2010), “Discretionary accruals and highly valued equity”, Journal of
Business Finance & Accounting, Vol. 37, pp. 60-92.
Houmes, R., Dickins, D. and O’Keefe, R. (2012), “New evidence on the incremental information
content of earnings reported using the LIFO inventory method”, Advances in Accounting,
Vol. 28, pp. 235-242.
Jensen, M. (2005), “Agency costs of highly valued equity”, Financial Management, Vol. 34,
pp. 5-20.
Jensen, M.C. and Meckling, W.H. (1976), “Theory of the ?rm: managerial behavior, agency costs
and ownership structure”, Journal of Financial Economics, Vol. 3, pp. 305-360.
Johnson, V.E., Khurana, I.K. and Reynolds, J.K. (2002), “Audit-?rm tenure and the quality of
?nancial reports”, Contemporary Accounting Review, Vol. 19, pp. 637-660.
Jones, J. (1991), “Earnings management during import relief investigations”, Journal of
Accounting Research, Vol. 29, pp. 193-228.
Kothari, S.P., Leone, A. and Wasley, C. (2005), “Performance matched discretionary accrual
measures”, Journal of Accounting and Economics, Vol. 39, pp. 163-197.
Krishnan, G.V. (2003), “Audit quality and the pricing of discretionary accruals”, Auditing:
A Journal of Practice & Theory, Vol. 22, pp. 243-263.
Lakonishok, J., Shleifer, A. and Vishny, R. (1994), “Contrarian investment extrapolation and
risk”, Journal of Finance, Vol. 49, pp. 1541-1578.
Lennox, C. (1999), “Audit quality and auditor size: an evaluation of reputation and deep
pockets hypothesis”, Journal of Business Finance & Accounting, Vol. 26, pp. 779-805.
McNichols, M. (2000), “Research design issues in earnings management studies”, Journal of
Accounting & Public Policy, Vol. 19, pp. 313-345.
Mansi, S.A., Maxwell, W.F. and Miler, D.P. (2004), “Does audit quality and tenure matter to
investors? Evidence from the bond market”, Journal of Accounting Research, Vol. 42,
pp. 755-793.
Menon, K. and Williams, D. (2004), “Former audit partners and abnormal accruals”,
The Accounting Review, Vol. 79, pp. 1095-1118.
Myers, J.N., Myers, L.A. and Omer, T.C. (2003), “Exploring the term of the auditor-client
relationship and the quality of earnings: a case for mandatory auditor rotation?”,
The Accounting Review, Vol. 78, pp. 779-799.
Palmrose, Z. (1986), “Audit fees and auditor size: further evidence”, Journal of Accounting
Research, Vol. 24, pp. 97-110.
Palmrose, Z. (1988), “An analysis of auditor litigation and audit service quality”, The Accounting
Review, Vol. 63, pp. 55-73.
Renas, S.M. and Cebula, R.J. (2005), “Enron, herding, and the deterrent effect of
disclosure of impropriety”, American Journal of Economics and Sociology, Vol. 64,
pp. 743-756.
Audit quality
73
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Reynolds, J.K. and Francis, J.R. (2000), “Does size matter? The in?uence of large clients on
of?ce-level auditor-reporting decisions”, Journal of Accounting and Economics, Vol. 30,
pp. 375-400.
Skinner, D. and Sloan, R. (2002), “Earnings surprises, growth expectations and stock returns or
don’t let an earnings torpedo sink your portfolio”, Review of Accounting Studies, Vol. 7,
pp. 289-312.
Subramanyam, K.R. (1996), “The pricing of discretionary accruals”, Journal of Accounting and
Economics, Vol. 22, pp. 249-281.
Teoh, S.H. and Wong, T.J. (1993), “Perceived auditor quality and the earnings response
coef?cient”, The Accounting Review, Vol. 68, pp. 346-367.
Corresponding author
Robert Houmes can be contacted at: [email protected]
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
ARJ
26,1
74
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
This article has been cited by:
1. Ahsan Habib, Rong Gong, Mahmud Hossain. 2013. Overvalued equities and audit fees: a research note.
Managerial Auditing Journal 28:8, 755-776. [Abstract] [Full Text] [PDF]
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_271964185.pdf
Audit quality studies document that accruals decrease when the audit firm is large, or the
audit firm is an industry specialist, or the audit-client tenure is long. The purpose of this paper is to
posit that incentives related to highly-valued equity mitigate these results, as managers use income
increasing accruals to augment earnings.
Accounting Research Journal
Audit quality and overvalued equity
Robert Houmes Maggie Foley Richard J . Cebula
Article information:
To cite this document:
Robert Houmes Maggie Foley Richard J . Cebula, (2013),"Audit quality and overvalued equity", Accounting
Research J ournal, Vol. 26 Iss 1 pp. 56 - 74
Permanent link to this document:
http://dx.doi.org/10.1108/ARJ -08-2011-0024
Downloaded on: 24 January 2016, At: 21:17 (PT)
References: this document contains references to 44 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 1591 times since 2013*
Users who downloaded this article also downloaded:
David S. J enkins, Thomas E. Vermeer, (2013),"Audit firm rotation and audit quality: evidence from
academic research", Accounting Research J ournal, Vol. 26 Iss 1 pp. 75-84 http://dx.doi.org/10.1108/
ARJ -11-2012-0087
Victoria J . Clout, Larelle Chapple, Nilan Gandhi, (2013),"The impact of auditor independence regulations
on established and emerging firms", Accounting Research J ournal, Vol. 26 Iss 2 pp. 88-108 http://
dx.doi.org/10.1108/ARJ -Dec-2011-0045
Arifur Khan, Paul Mather, (2013),"The value of executive director share ownership and discretionary
accruals", Accounting Research J ournal, Vol. 26 Iss 1 pp. 35-55 http://dx.doi.org/10.1108/
ARJ -02-2012-0011
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Audit quality and
overvalued equity
Robert Houmes, Maggie Foley and Richard J. Cebula
Department of Accounting and Finance, Jacksonville University,
Jacksonville, Florida, USA
Abstract
Purpose – Audit quality studies document that accruals decrease when the audit ?rm is large, or the
audit ?rm is an industry specialist, or the audit-client tenure is long. The purpose of this paper is to
posit that incentives related to highly-valued equity mitigate these results, as managers use income
increasing accruals to augment earnings.
Design/methodology/approach – To test this assertion, the authors regress discretionary accruals
on: controls, a highly valued equity indicator variable equal to 1 if the client’s lagged price-to-earnings
ratio is in the highest P/E quintile, indicator variables equal to 1 for alternative measures of audit
quality, and interaction terms between the highly valued equity indicator variable and audit quality
indicator variables.
Findings – Results of tests show positive and statistically signi?cant coef?cients for each of the
highly-valued equity-audit quality interaction terms, suggesting that when a ?rm is highly valued the
accruals’ decreasing effect of high quality auditors is reduced.
Originality/value – Beginning with Jensen’s article regarding the agency costs of overvalued
equity, a stream of research examining factors associated with highly priced ?rms has developed. The
paper extends these ?ndings, as well as the considerable body of audit quality studies, by examining
the ability of a high quality auditor to attenuate this result.
Keywords Auditing, Equity capital, Price earning ratio, Audit quality, Highly valued equity,
Discretionary accruals
Paper type Research paper
Introduction
Classical agency theory asserts that alignment of management-shareholder interests
increases incentives for value creation ( Jensen and Meckling, 1976). Jensen (2005) argues,
however, that when a ?rmbecomes overvalued, i.e. the price of the ?rmbecomes greater
than its underlying economic value, managers are motivated to perpetuate
overvaluation, which is consistent in principle with arguments in Renas and Cebula
(2005). Since an overvalued ?rm, by de?nition, lacks the operational capability to achieve
performance levels re?ected in its price, managers are motivated to use aggressive
accounting policies to maximize earnings. Although numerous reporting alternatives
are available to achieve earnings management goals, accruals are an especially attractive
choice since they are a normal part of the ?nancial reporting process and their amounts
require forward looking estimates over which managers have considerable discretion.
In a recent study Houmes andSkantz (2010) assert that incentives associated with over
valued equity induce managers to support extreme valuations by using discretionary
accruals to manage earnings higher. In particular, they regress discretionary accruals
on highly valued equity proxy variables measured as companies in the highest quintile
of P/E ?rms and controls. Results show that in the year following valuation and relative
to other ?rms, highly valued ?rms report higher discretionary accruals. Among other
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1030-9616.htm
Accounting Research Journal
Vol. 26 No. 1, 2013
pp. 56-74
qEmerald Group Publishing Limited
1030-9616
DOI 10.1108/ARJ-08-2011-0024
ARJ
26,1
56
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
things they conclude that directors and audit committees should be especially sensitive to
earnings management when the ?rm is highly valued. In the spirit of these remarks, we
extend the ?ndings in Houmes and Skantz (2010) by investigating the effect that
overvaluation has on audit quality. This study provides evidence that the previously
documented inverse relation between, large audit ?rms, industry specialist audit
?rms, long tenure audit ?rms, and discretionary accruals is attenuated when the ?rm is
highly valued.
Using high price-to-earnings ?rms to proxy for highly valued equity and the level of
income increasing discretionary accruals to proxy for earnings quality, our results
show that, in accordance with prior studies, high quality auditors generally constrain
accruals. By contrast, however, accruals for highly valued clients of high quality
auditors are statistically signi?cantly higher. In particular, accruals for clients of
large Big N (Big 6, Big 5, or Big 4, depending on the period), industry specialist and
long tenure auditors increase for clients in the previous year’s highest quintile of
price-to-earnings ratios. Hence, the widely documented tendency for high quality audit
?rms to reduce accruals is mitigated when the ?rm is highly valued. We rationalize
these ?ndings within the context of traditional audit quality literature and Jensen’s
(2005) overvalued equity hypothesis by asserting that the accruals decreasing effect of
high quality auditors on earnings is reduced as informed managers of highly valued
?rms with greater incentives to perpetuate values prevail upon auditors to report
higher levels of income increasing discretionary accruals.
The value of a ?rm is a function of its prospective performance, and the greater the
performance expectations the higher the value. Since expectations are greater when a
?rm is richly priced, investors’ reactions to unexpectedly negative performance
outcomes should also be greater. For example, Skinner and Sloan (2002) document that
the market’s reaction to negative earnings surprises is greater than that for positive
earnings surprises and that this asymmetric reaction increases for high market-to-book
?rms. Given that shareholder disappointments will be greater when the market’s
expectations are the highest, the potential ?ndings of this study have implications for
audit practitioners as they conduct audits for highly valued ?rms. Prior studies report
that shareholder suits are a common source of auditor litigation (Palmrose, 1988;
Goldwasser and Eickemeyer, 2004). Extant research also documents that the likelihood
of a securities fraud class action increases with the ?rm’s market-to-book ratio (Grimm,
2009). If high valuations impair audit quality, then auditors are performing lower
quality audits on the very ?rms that historically have tended to disappoint investors
the most. Results of this study should also be relevant for boards of highly valued
?rms. If overvaluation induces lower earnings quality and by extension lower audit
quality, boards should be especially vigilant in their internal control function. This
should be particularly true for the audit committee.
From a research perspective, a substantial body of work currently exists on audit
quality and more recent studies have examined factors related to highly valued equity.
To date, however, published studies have not considered audit quality within a
valuation context. We synthesize and provide additional insight into both streams of
research by investigating the accruals constraining effect of high quality auditors
vis-a` -vis other auditors when the client is highly valued.
The remainder of this study is organized as follows. The section thereafter includes
relevant literature, along with a discussion of Jensen’s (2005) overvalued equity
Audit quality
57
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
hypothesis as it relates to earnings and audit quality. The next section includes testable
hypotheses and their motivations, followed by a description of the sample and
methodology used to conduct tests. Results of these tests are reported in the subsequent
section. The study concludes with a discussion and summary of the ?ndings.
Literature and theory
Highly valued equity
Beginning with Jensen (2005), a signi?cant stream of research has developed regarding
the overvalued equity hypothesis. Efendi et al. (2007) use a sample of 190 ?rms from a
2003 General Accounting Of?ce report to investigate factors related to overvalued
equity and document among other things that earnings restatements increase for ?rms
that are constrained by debt covenant violations or that raise new debt. They also
show that the incidence of restatements increases when CEOs hold large amounts of
in-the-money stock options.
Chi and Gupta (2009) report that overvaluation-induced earnings management as
proxied for by discretionary accruals is negatively related to following year abnormal
returns. They document that, in the following year, abnormal returns of ?rms with
high discretionary accruals are 11.88 percent lower than the returns of ?rms with lower
discretionary accruals.
A long-standing anomaly of ef?cient equity markets is the tendency for highly
valued, high price-to-earnings ?rms to earn lower returns going forward. While
explanations for the persistence of this anomaly are beyond the scope of this study,
they include the following three:
(1) economic theory predicts that competition mitigates the ability of ?rms to
achieve and sustain abnormal performance levels over time;
(2) since highly valued ?rms are less risky, lower returns re?ect lower risk; and
(3) investors (irrationally) tend to over-price high-growth-glamour ?rms.
Notwithstanding these conjectures, however, the tendency for high price-to-earnings
?rms to under perform is well documented (Basu, 1977; Chopra et al., 1992;
Lakonishok et al., 1994; Campbell and Shiller, 2001). In a more recent study, Anderson
and Brooks (2006) document that the difference in returns between value and glamour
?rms almost doubles whenP/Es are calculated using average earnings over the last eight
years. The following graphical depiction of companies from the 1991 to 2008 period
shows that ?rms in the highest quintile of price-to-earnings ratios earned lower following
year mean and median abnormal returns in every year but one (Figures 1 and 2).
Hence, from a market value perspective and relative to other companies, the very
?rms that are expected to perform the best, on average, tend to perform worse. Since
expectations are particularly high for highly valued ?rms, when managers foresee the
operational inability of their ?rms to meet expected performance targets, incentives to
manage earnings increase. An important deterrent against these incentives is the audit.
Audit quality
The value of accounting information is a function of its credibility, and a central
objective of audits is to enhance the quality of ?nancial reporting. Prior studies have
provided several empirical surrogates to measure audit quality. These include audit
?rm size, audit industry specialization, and the length of the auditor-client relationship.
ARJ
26,1
58
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Beginning with DeAngelo (1981), decades of research have shown that large audit ?rms
with greater resources and more reputation at stake perform higher quality audits
(Palmrose, 1986; Beatty, 1989; Craswell and Taylor, 1995; Lennox, 1999; Teoh and
Wong, 1993; Menon and Williams, 2004; Houmes et al., 2012, etc.). Using accruals to
Figure 1.
Annual mean CAPM
returns highly valued vs
other 1991-2008
–0.2
–0.1
0
0.1
0.2
0.3
0.4
0.5
Year 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
M
e
a
n
R
e
t
u
r
n
s
Mean Returns of Firms in the Highest Quintile of Price-to-Earnings Ratios Mean Returns Other
Figure 2.
Annual median CAPM
returns highly valued vs
other 1991-2008
–0.2
–0.15
–0.1
–0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
M
e
d
i
a
n
R
e
t
u
r
n
s
Median Returns of Firms in the Highest Quintile of Price-to-Earnings Ratios Median Return Other
Audit quality
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
proxy for earnings quality, Becker et al. (1998) show that clients of large (Big 6) audit
?rms report lower discretionary accruals. In a similar study, Francis et al. (1999)
examine a large sample of NASDAQ ?rms and ?nd that although Big 6 audit
clients have higher levels of total accruals, they have lower estimated discretionary
accruals. Krishnan (2003) provides evidence that investors ascribe higher values to
the discretionary accruals of Big 6 clients than non-Big 6 clients. Heninger (2001)
documents that the likelihood of litigation increases with levels of total and
discretionary accruals, but decreases if a Big 5 auditor. Houmes and Skantz (2010)
provide evidence that the magnitude of the inverse relation between operating cash ?ow
and accruals increases if the ?rm is highly valued and assert that incentives to manage
earnings increase when operating cash ?ow is weak. Although they do not offer any
direct tests relating to the effect of high valuation on audit quality after excluding loss
?rms, they ?nd (not tabulated) evidence large Big N auditors reduce this tendency.
Balsam et al. (2003) shows that clients of within Big 6 and Big 5 auditor industry
specialists have lower accruals and higher earnings response coef?cients. Myers et al.
(2003) document that discretionary accruals decrease with the length of audit
?rm-client relationship. Mansi et al. (2004) show that the cost of debt decreases with
tenure and Carcello and Nagy (2004) document an increase in the incidence of ?nancial
reporting fraud when audit ?rm tenure is less than three years.
Although audit opinions enhance the credibility and reliability of ?nancial reports,
they also re?ect a negotiation dimension and, within the ethical and technical con?nes
of accounting standards, a ?rm’s published ?nancial report may be perceived as a
joint statement from the manager and auditor (Antle and Nalebuff, 1991). Gibbons et al.
(2001) use a sample of 93 experienced audit partners to report that auditor-client
negotiation occurs on a regular basis. In particular, results show that negotiation is
common, with 67 percent of audit partners experiencing negotiation with 50 percent or
more of their clients. Hence, the ?nal reporting product is often the result of a
compromise between management and auditors (Ellingsen et al., 1989).
Since managers of highly valued ?rms are under increased pressure to meet
optimistic earnings forecasts, they have incentives to assume a more aggressive
negotiating stance, prevailing on auditors to, within the con?nes of existing accounting
standards, report higher earnings. Hence, income increasing negotiations that bias
earnings upward should be more prevalent for highly valued ?rms. If the auditor relents,
earnings quality will be impaired. Consequently, audit quality for highly valued ?rms
could be negatively affected. Antle and Nalebuff (1991) demonstrate that when joint
auditor-client welfare is maximized, ex post income reporting is biased upward.
Although executives have numerous opportunities to manage earnings, accruals are
particularly appealing since they are a normal, frequent, and expected component of
the ?nancial reporting process. Numerous prior studies provide evidence that earnings
are managed with discretionary accruals (Subramanyam, 1996; Bergstresser and
Philippon, 2006, etc.).
Since highly valued companies are under greater pressure to meet earnings
expectations, we expect that managers will assume a more intransigent negotiating
stance with the auditor and utilize accruals to increase earnings. Consequently, the
tendency for high quality audit ?rms to constrain accruals will be diminished for highly
valued ?rms as auditors acquiesce towards the upper bounds of accounting standards
constraints in the face of increased client pressure.
ARJ
26,1
60
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Finally, auditors may suffer, at least in some measure, from the same information
asymmetry that characterizes the relationship between managers and owners. While the
asymmetry may not be as great, managers nevertheless are closer to the ?rms they lead
than external auditors. At the very least, since managers of highly valued ?rms have
better knowledge of their company’s economic performance, they will be inclined to
accept undetected statement errors in their favor and protest only those that are adverse.
We test these assertions with the following hypotheses stated in alternative form:
H1. The magnitude of the inverse relation between the discretionary accruals of
the clients of large audit ?rms and the discretionary accruals of the clients of
other audit ?rms decreases if the clients are highly valued.
H2. The magnitude of the inverse relation between the discretionary accruals
of the clients of industry specialist audit ?rms and the discretionary accruals
of the clients of other audit ?rms decreases if the clients are highly valued.
H3. The magnitude of the inverse relation between the discretionary accruals of
audit ?rms’ clients with long tenure and the discretionary accruals of clients
with shorter tenure decreases if the clients are highly valued.
H4. The magnitude of the inverse relation between the discretionary accruals of the
clients of audit industry specialist audit ?rms with long tenure and the
discretionary accruals of other clients decreases if the clients are highly valued.
Sample and methodology
Sample
For years 1998-2009, we acquire panel data for ?nancial statement variables from the
Compustat ?les of companies. Similar to prior accruals studies, we remove all ?nancial
services (SIC codes 6000-6999) and regulated (4900-4999) ?rms. As explained below,
to control for the Big N audit quality effect on our specialists and long tenure auditor
variables, we separate the overall sample of ?rms into two groups. One group, our
Big N sample, contains Big N and non-Big N auditors. The other group, our
specialist-tenure sample contains Big N audit clients only. After deleting ?rms with
missing data for model variables, the total number of ?rm year observations in our Big
N sample is 29,405. Eliminating non-Big N audit ?rms reduces our specialist-tenure
sample to 23,937 observations. Although we also provide regression results for all
discretionary accruals, consistent with our assertion that managers of highly valued
?rms manage earnings higher, our primary tests involve income increasing
discretionary accruals only. Eliminating negative discretionary accruals reduces our
Big N (specialist-tenure) sample to 16,346 (13,172) ?rm year observations.
Discretionary accruals
Our dependent variable is discretionary accruals (DAC
it
). For all ?rms, discretionary
accruals are estimated using the cross sectional version of the modi?ed Jones model ( Jones,
1991). The modi?edJones model has beenusedina varietyof researchsettings (Becker et al.,
1998; Francis et al., 1999; Reynolds andFrancis, 2000, etc.). The model is speci?edas follows:
TAC
it
¼ b
0
þb
1
1
AT
it21
þb
2
ðDREV
it
2DAR
it
Þ þb
3
PPE
it
þ1
it
ð1Þ
Audit quality
61
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where TAC
it
is the difference between ?rm i’s year t earnings before extraordinary items
and net cash ?owfromoperations scaled by beginning of year (t 2 1) assets; AT
it21
is ?rm
i’s beginning of the year t total assets; DREV
it
is the difference inyear t andyear t 2 1 sales;
DAR
it
is the difference between year t and year t 2 1 trade account receivables; and PPE
it
is
net property plant and equipment both scaled by AT
it21
. For each year and two-digit SIC
industry group, discretionary accruals (DAC
it
) are estimated cross-sectionally as the
residuals from the above regression. Stated alternatively, coef?cients from equation (1) are
used to estimate predicted accruals for each year in each SIC group. Discretionary accruals
are the difference between each ?rm’s actual and predicted accruals, i.e.:
DAC
it
¼ TAC
it
2½b
0
þb
1
1
1=AT
it21
þb
2
ðDREV
it
2DAR
it
Þ þb
3
PPE
it
?: ð2Þ
To mitigate the potential effects of extreme observations onresults, we winsorize accruals at
the top and bottom 1 percent levels across the entire pooled sample. In our models we also
include controls that may affect the cross sectional variability in the dependent variable,
DAC
it
.
Control variables
Relative to other companies, certain industries or ?rms may tend to generate higher
accruals. In addition, it is natural that growth companies with increasing earnings and
investments in working capital are more likely to produce greater accruals, and prior
studies showthat growth ?rms report higher accruals (McNichols, 2000). To control for the
possibility that companies with greater total accruals may also have larger discretionary
accruals that our accruals model does not capture, we include total accruals (ACRL
it
) in our
multivariate tests measured as the difference between ?rm i’s year t earnings before
extraordinary items and net cash ?owfromoperations scaled by beginning of year assets.
Accruals studies typically control for size effects. Dechow and Dichev (2002) show
that larger ?rms record larger accruals. Also, larger ?rms with larger investor
following and more developed and sophisticated ?nancial reporting systems may
affect accrual levels (Becker et al., 1998; Reynolds and Francis, 2000). For each ?rm i we
include the end of ?scal year t natural log of total assets (LnASSET
it
).
Reynolds and Francis (2000) provide evidence that the tendency to manage earnings
increases with leverage. DeFond and Jiambalvo (1994) show that accruals are related to
debt covenant breeches. In addition, debt may serve as a monitoring mechanism that
constrains earnings management. To control for the effect that high debt levels may
have on accruals, we include the variable LEV
it
measured as ?rm i’s end of year t long
term debt scaled by t 2 1 total assets.
Operating cash ?ows are a component of earnings and their levels correspond
inversely with accruals. Further, the level of cash ?owmay affect the ability and/or need
to use accruals, causing ?rms with higher (lower) operating cash ?ows to report lower
(higher) discretionary accruals (Becker et al., 1998). We control for these effects by
including operating cash ?ow de?ated by the beginning of the year total assets (OCF
it
).
Kothari et al. (2005) show that discretionary accruals are impacted by ?nancial
performance. Accordingly, weinclude ROA
it
, income before extraordinary items
divided by the beginning of the year total assets. Finally, for 47 of the 48 two-digit SIC
industry codes included in our study we assign dummies equal to 1 to control for
ARJ
26,1
62
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
potential industry effects[1]. Similar to our dependent variable, DAC
it
, we winsorize all
continuously measured independent variables at the 1 percent levels.
Variables of interest: audit quality and highly valued equity
Using our initial all accruals sample, we identify highly valued clients (HV
it21
) as ?rms in
the highest quintile of PEs (i.e. P/E . 27.204) and assign an indicator variable equal to 1 if
the client is in this quintile of prior ?scal year end price-to-earnings ratios and 0 otherwise.
Our audit quality variables are as follows: BIGN
it21
equal to 1 and 0 otherwise if the
audit ?rmis one of the largest four, ?ve, or six audit ?rms (depending onthe time period);
SPEC
it21
, equal to 1 and 0 otherwise if the audit ?rm is an industry specialist; and
TEN
it21
equal to 1 and 0 otherwise if the length of the auditor-client relationship is equal
to or over ?ve years. We also include an additional audit quality indicator variable if the
auditor is both an industry specialist and has a long tenure auditor-client relationship
(SPECTEN
it21
). In accordance with prior studies (Palmrose, 1986), we de?ne an auditor
as anindustry specialist according to the auditor’s two-digit SICcode industry share and
designate the ?rm as a specialist based on the number of clients and total client sales.
In particular, an indicator variable equal to 1 and0 otherwise is assigned to auditors with
both the largest number of clients and the highest total client sales within each two-digit
SIC code. To eliminate the brand name effect of BN auditors, we restrict the sample of
specialist and tenure ?rms to Big N clients only. We eliminate SIC code industries with
fewer than 15 observations. The income increasing discretionary accruals, Big N, and
non-Big N sample consists of 16,346 ?rm year observations. The income increasing
discretionary accruals, specialist-tenure sample contains 13,172 observations.
To investigate the relation between high valuations and the tendency of high audit
quality auditors to mitigate accruals, we interact the audit quality variables with our
highly valued equity dummy. Statistically signi?cantly positive estimates for the audit
quality, highly valued equity interaction terms: BIGN
*
HV
it21
, SPEC
*
HV
it21
,
TEN
*
HV
it21
and SPECTEN
*
HV
it21
, provide support for hypotheses that incentives
associated with high valuations reduce the tendency of high quality audit ?rms to
constrain accruals. Our models are as follows:
DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
BIGN
it21
þa
7
HV
it21
þa
8
BIGN
*
HV
it21
þa
9
INDUS þ1
it
ð3Þ
DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
SPEC
it21
þa
7
HV
it21
þa
8
SPEC
*
HV
it21
þa
9
INDUS þ1
it
ð4Þ
DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
TEN
it21
þa
7
HV
it21
þa
8
TEN
*
HV
it21
þa
9
INDUS þ1
it
ð5Þ
DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
SPECTEN
it21
þa
7
HV
it21
þa
8
SPECTEN
*
HV
it21
þa
9
INDUS þ1
it
ð6Þ
Audit quality
63
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Results
Descriptive statistics
Table I de?nes variables used in models. Tables II and III provide descriptive
statistics for variables. Table II Panels A and B shows the maximum, minimum, mean,
and median values of variables in our income increasing Big N (Panel A) and
specialist-tenure (Panel B) samples. Table III provides univariate correlations. Mean
discretionary accruals (scaled) for both samples are 0.064 and 0.059, respectively[2].
In our BNauditor sample 82 percent of the observations involve large audit ?rms. In our
specialist-tenure sample, 20 percent of the observations involve specialists and
54 percent have audit-client tenures over ?ve years. Hence most Big N audit-client
relationships are greater than ?ve years. 10 percent of the observations are identi?ed as
both a specialist and long tenure audit client.
Univariate correlations in Table III reveal that size, leverage and operating cash ?ow,
are inversely related to discretionary and total accruals. ROA is positively (negatively)
related to total (discretionary) accruals. ROAand operating cash ?oware also positively
correlated with our HV
it21
indicator variable. While coef?cients are generally positive
between each of our audit quality variables and HV
it21
, only Big N is statistically
signi?cant. Overall, the signs and signi?cance of our other model variables are
consistent across all proxies for audit quality. BIGN, SPEC, TEN, and SPECTEN are
positively (negatively) correlated with size, operating cash ?owand ROA(discretionary
accruals).
Table IVprovides the differences in mean scaled discretionary accruals between each
of the high quality auditor measures and their non-high quality auditor counterparts.
DAC
it
are Firm i’s ?scal year t income increasing discretionary accruals
ACRL
it
are Firm i’s ?scal year t total accruals
LnASSET
it
are Firm i’s ?scal year t natural log of total assets
LEV
it
are Firm i’s ?scal year t long term debt divided by ?scal year t 2 1 total assets
OCF
it
are Firm i’s ?scal year t cash ?ow from operating activates divided by ?scal year
t 2 1 total assets
ROA
it
are fFrm i’s ?scal year t income before extraordinary items divided by ?scal year
t 2 1 total assets
HV
it21
are An indicator variable equal to 1 for ?rms in the highest quintile of ?scal year
t 2 1 price-to-earnings ratios
AQ
it21
are Four measures of audit quality de?ned as follows
BIGN
it21
are An indicator variable equal to 1 if at the end of ?scal year t 2 1 the auditor is a
Big N auditor
SPEC
it21
are An indicator variable equal to 1 if at the end of ?scal year t 2 1 the auditor is an
industry specialist
TEN
it21
are An indicator variable equal to 1 if at the end of ?scal year t 2 1 the auditor
tenure is greater than or equal to ?ve years
SPECTEN
it21
are An indicator variable equal to 1 if at the end of ?scal year t 2 1 the auditor is an
industry specialist and tenure is greater than or equal to ?ve years
BIGN
*
HV
it21
are An interaction term between BIGN
it21
and HV
it21
;
SPEC
*
HV
it21
are An interaction term between SPEC
it21
and HV
it21
TEN
*
HV
it21
are An interaction term between TEN
it21
and HV
it21
SPECTEN
*
HV
it21
are An interaction term between SPECTEN
it21
and HV
it21
INDUS are An indicator variable equal to1for ?rms in each two-digit SIC code
Table I.
Variable de?nitions
ARJ
26,1
64
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Results show that relative to other audit ?rms, discretionary accruals are lower for each
of the four audit quality proxy measures ( p , 0.001) with the greatest difference
occurring between Big N and non-Big N auditors (20.025).
This study predicts that incentives related to high valuations reduce the accruals
decreasing effect of high quality auditors. To initially investigate this assertion, within
each price-to-earnings group (top 20 PE and bottom 80 PE) we compare the mean
discretionary accruals between clients of high quality auditors and other audit clients.
Results show that while high quality audit clients report lower accruals vis-a` -vis other
clients in both price-to-earnings groups, for three of our audit quality measures
differences generally decrease when the client ?rm is in the highest PE quintile.
In particular, for three of the four audit quality variables, the difference in scaled
discretionary accruals between the clients of high quality auditors and the clients of
other auditors is not signi?cant when the ?rm is highly valued. For example, the
difference in mean accruals between specialist and non-specialist auditors in the top
quintile of price-to-earnings ratios is an insigni?cant 20.003 (0.349), while the difference
in mean accruals for ?rms in the bottom 80 percent is 20.006 (0.000). In addition, the
differences in accruals between the top 20 PE, TEN and SPECTENaudit clients and the
accruals of the corresponding top 20 PE, NTEN, and NSPECTENaudit quality clients is
an insigni?cant 20.002 and 20.001 ( p ¼ 0.296 and 0.788, respectively). Differences in
accruals between the bottom 80 PE, TEN, and SPECTEN are, however, signi?cantly
lower for high quality auditors. Differences in accruals between the discretionary
accruals of high quality specialists, long tenure, and specialist – long tenure audit
clients and other lower quality audit clients decrease if the client is highly valued.
The difference, however, in mean accruals for the top quintile of Big N versus NBN
auditors is signi?cant at the p , 0.001 level. Hence, the reduction in accruals between
Variables Maximum Minimum Mean Median SD
Panel A. Big N auditor-income increasing discretionary accruals (n ¼ 16,346)
DAC
it
0.331 0.000 0.064 0.044 0.066
ACRL
it
0.256 20.544 20.014 20.026 0.077
LnASSET
it
32.140 22.120 26.603 26.520 2.157
LEV
it
1.196 0.000 0.183 0.123 0.215
OCF
it
0.428 20.759 0.040 0.068 0.159
ROA
it
0.247 21.318 0.010 0.043 0.177
BIGN
it
1.000 0.000 0.820 1.000 0.388
HV
it21
1.000 0.000 0.213 1.000 0.401
Panel B. Specialists and long tenure auditors – income increasing discretionary accruals (n ¼ 13,172)
DAC
it
0.331 0.000 0.059 0.041 0.062
ACRL
it
0.269 20.544 20.017 20.028 0.073
LnASSET
it
32.360 21.990 27.021 26.978 0.890
LEV
it
1.498 0.000 0.197 0.144 0.223
OCF
it
0.498 21.708 0.048 0.073 0.168
ROA
it
0.848 22.576 0.015 0.045 0.187
SPEC
it
1.000 0.000 0.195 0.000 0.396
TEN
it
1.000 0.000 0.540 1.000 0.498
SPETENC
it
1.000 0.000 0.099 0.000 0.298
HV
it21
1.000 0.000 0.231 0.000 0.398
Note: Variables are de?ned in Table I
Table II.
Descriptive statistics
Audit quality
65
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
P
a
n
e
l
A
.
V
a
r
i
a
b
l
e
c
o
r
r
e
l
a
t
i
o
n
s
f
o
r
B
i
g
N
a
u
d
i
t
o
r
s
–
i
n
c
o
m
e
i
n
c
r
e
a
s
i
n
g
d
i
s
c
r
e
t
i
o
n
a
r
y
a
c
c
r
u
a
l
s
(
n
¼
1
6
,
3
4
6
)
V
a
r
i
a
b
l
e
D
A
C
i
t
A
C
R
L
i
t
L
n
A
S
S
E
T
i
t
L
E
V
i
t
O
C
F
i
t
R
O
A
i
t
B
I
G
N
H
V
i
t
2
1
D
A
C
i
t
1
0
.
6
2
6
(
0
.
0
0
0
)
2
0
.
2
8
9
(
0
.
0
0
0
)
2
0
.
0
4
6
(
0
.
0
5
0
)
2
0
.
3
5
3
(
0
.
0
0
0
)
2
0
.
0
7
0
(
0
.
0
0
0
)
2
0
.
1
4
6
(
0
.
0
0
0
)
2
0
.
0
2
9
(
0
.
0
0
0
)
A
C
R
L
i
t
1
2
0
.
1
6
7
(
0
.
0
0
0
)
2
0
.
0
7
2
(
0
.
0
0
0
)
2
0
.
3
3
6
(
0
.
0
0
0
)
0
.
1
1
9
(
0
.
0
0
0
)
2
0
.
0
7
3
(
0
.
0
0
0
)
2
0
.
0
0
3
(
0
.
7
4
6
)
L
n
A
S
S
E
T
i
t
1
0
.
2
7
5
(
0
.
0
0
0
)
0
.
3
7
9
(
0
.
0
0
0
)
0
.
3
3
0
(
0
.
0
0
0
)
0
.
4
0
7
(
0
.
0
0
0
)
0
.
1
0
1
(
0
.
0
0
0
)
L
E
V
i
t
1
0
.
0
3
1
(
0
.
0
0
0
)
0
.
0
2
2
(
0
.
0
0
4
)
0
.
1
1
7
(
0
.
0
0
0
)
2
0
.
0
4
3
(
0
.
0
0
0
)
O
C
F
i
t
1
0
.
8
0
1
(
0
.
0
0
0
)
0
.
1
1
8
(
0
.
0
0
0
)
0
.
1
2
2
(
0
.
0
0
0
)
R
O
A
i
t
1
0
.
0
8
5
(
0
.
0
0
0
)
0
.
1
1
6
(
0
.
0
0
0
)
B
I
G
N
i
t
1
0
.
0
3
3
(
0
.
0
0
0
)
H
V
i
t
2
1
1
P
a
n
e
l
B
.
V
a
r
i
a
b
l
e
c
o
r
r
e
l
a
t
i
o
n
s
f
o
r
s
p
e
c
i
a
l
i
s
t
s
a
n
d
l
o
n
g
t
e
n
u
r
e
a
u
d
i
t
o
r
s
–
i
n
c
o
m
e
i
n
c
r
e
a
s
i
n
g
d
i
s
c
r
e
t
i
o
n
a
r
y
a
c
c
r
u
a
l
s
(
n
¼
1
3
,
1
7
2
)
V
a
r
i
a
b
l
e
D
A
C
i
t
A
C
R
L
i
t
L
n
A
S
S
E
T
i
t
L
E
V
i
t
O
C
F
i
t
R
O
A
i
t
S
P
E
C
T
E
N
S
P
E
C
T
E
N
H
V
i
t
2
1
D
A
C
i
t
1
0
.
6
5
0
(
0
.
0
0
0
)
2
0
.
2
6
2
(
0
.
0
0
0
)
2
0
.
0
2
5
(
0
.
0
0
4
)
2
0
.
.
3
3
1
(
0
.
0
0
0
)
2
0
.
0
6
2
(
0
.
0
0
)
2
0
.
0
3
6
(
0
.
0
0
0
)
2
0
.
0
4
7
(
0
.
0
0
0
)
2
0
.
0
4
1
(
0
.
0
0
0
)
2
0
.
0
2
4
(
0
.
0
0
6
)
A
C
R
L
i
t
1
2
0
.
1
7
8
(
0
.
0
0
0
)
2
0
.
0
7
1
(
0
.
0
0
0
)
2
0
.
3
1
9
(
0
.
0
0
0
)
0
.
0
8
8
(
0
.
0
0
0
)
2
0
.
0
0
9
(
0
.
3
1
9
)
2
0
.
0
1
1
(
0
.
2
2
4
)
2
0
.
0
2
1
(
0
.
0
1
6
)
2
0
.
0
0
8
(
0
.
3
8
2
)
L
n
A
S
S
E
T
i
t
1
0
.
2
4
6
(
0
.
0
0
0
)
0
.
3
6
9
(
0
.
0
0
0
)
0
.
3
2
3
(
0
.
0
0
0
)
0
.
0
5
5
(
0
.
0
0
0
)
0
.
1
5
4
(
0
.
0
0
0
)
0
.
0
9
0
(
0
.
0
0
0
)
0
.
0
8
6
(
0
.
0
0
0
)
L
E
V
i
t
1
0
.
0
1
3
(
0
.
1
4
2
)
0
.
0
0
7
(
0
.
4
4
8
)
0
.
0
3
5
(
0
.
0
0
0
)
2
0
.
0
0
2
(
0
.
7
8
0
)
0
.
0
2
5
(
0
.
0
0
4
)
2
0
.
0
4
9
(
0
.
0
0
0
)
O
C
F
i
t
1
0
.
8
2
8
(
0
.
0
0
0
)
0
.
0
3
7
(
0
.
0
0
0
)
0
.
0
7
2
(
0
.
0
0
0
)
0
.
0
5
1
(
0
.
0
0
0
)
0
.
1
0
6
(
0
.
0
0
0
)
R
O
A
i
t
1
0
.
0
2
8
(
0
.
0
0
1
)
0
.
0
7
5
(
0
.
0
0
0
)
0
.
0
3
9
(
0
.
0
0
0
)
0
.
1
0
2
(
0
.
0
0
0
)
S
P
E
C
i
t
1
2
0
.
0
3
2
(
0
.
0
0
0
)
0
.
6
7
3
(
0
.
0
0
0
)
0
.
0
0
5
(
0
.
5
9
1
)
T
E
N
i
t
1
0
.
3
0
7
(
0
.
0
0
0
)
0
.
0
0
5
(
0
.
5
5
9
)
S
P
E
C
T
E
N
i
t
1
2
0
.
0
0
6
(
0
.
4
7
1
)
H
V
i
t
2
1
1
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
c
e
(
p
:
t
w
o
-
t
a
i
l
e
d
)
i
s
s
h
o
w
n
i
n
p
a
r
e
n
t
h
e
s
e
s
u
n
d
e
r
c
o
e
f
?
c
i
e
n
t
s
;
v
a
r
i
a
b
l
e
s
a
r
e
d
e
?
n
e
d
i
n
T
a
b
l
e
I
Table III.
ARJ
26,1
66
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
BN and NBN auditors is consistent for both PE groups. An important limitation of this
analysis is that it omits control variables that may affect accruals levels. The following
section includes results of our main tests using multivariate analysis that includes
previously described control variables (Table V).
Discretionary accruals models
Table VI shows results of main tests for our income increasing discretionary accruals
sample for models depicted in equations (3) through (6). Adjusted R
2
exceed 50 percent.
Total accruals, ACRL
it
, is signi?cant and positive. As expected and in accordance with
prior accruals studies, estimates for assets and operating cash ?ow are negative,
suggesting that greater ?rm size, and operating cash ?ow enhance accruals quality. The
coef?cients for ROA
it
are positive. Discretionary accruals increase with (scaled) income.
Audit quality
variables
Mean disc.
accruals n
Audit quality
variables
Mean disc.
accruals n
BN 0.060 13,320 TEN 0.057 7,125
NBN 0.085 3,026 NTEN 0.063 6,047
Difference 20.025 (0.000) Difference 20.006 (0.000)
SPEC 0.055 2,568 SPECTEN 0.053 1,300
NSPEC 0.061 10,604 NSPECTEN 0.061 11,872
Difference 20.006 (0.000) Difference 20.008 (0.000)
Notes: Variable de?nitions: BN – ?rm is audited by a Big N auditor; NBN – ?rm is not audited by a
Big N auditor; SPEC – auditor is an industry specialist; NSPEC – auditor is not an industry specialist;
TEN – audit-client ?rmtenure is greater than or equal to ?ve years; NTEN – audit-client ?rmtenure is
less than ?ve years; SPECTEN – audit-client ?rm tenure is greater than or equal to ?ve years and the
auditor is an industry specialist; NSPECTEN – ?rm’s auditor is not a SPECTEN auditor; n – number
of ?rm year observations
Table IV.
Difference in mean
income increasing
discretionary accruals
between high quality:
Big N, specialist and
long tenure auditors and
other auditors
BN n NBN n Difference n
Top 20 PE 0.057 2,939 0.078 543 20.021 (0.000) 3,482
Bottom 80 PE 0.060 10,381 0.086 2,483 20.026 (0.000) 12,864
SPEC n NSPEC n Difference n
Top 20 PE 0.054 603 0.057 2,450 20.003 (0.349) 3,053
Bottom 80 PE 0.056 1,965 0.062 8,154 20.006 (0.000) 10,119
TEN n NTEN n Difference n
Top 20 PE 0.056 1,695 0.058 1,358 20.002 (0.296) 3,053
Bottom 80 PE 0.057 5,430 0.064 4,689 20.007 (0.000) 10,119
SPECTEN n NSPECTEN n Difference n
Top 20 PE 0.056 297 0.057 2,756 20.001 (0.788) 3,053
Bottom 80 PE 0.052 1,003 0.062 9,116 20.010 (0.000) 10,119
Notes: Variable de?nitions: top 20 PE – ?rm is in the highest 20 percent of price-to-earnings ratios;
bottom 80 PE – ?rm is in the lowest 80 percent of price-to-earnings ratios; BN – ?rm is audited by a
Big N auditor; NBN – ?rm is not audited by a Big N auditor; SPEC – auditor is an industry specialist;
NSPEC – auditor is not an industry specialist; TEN – audit-client ?rm tenure is greater than or equal
to ?ve years; NTEN – audit-client ?rm tenure is less than ?ve years; SPECTEN – audit-client ?rm
tenure is greater than or equal to ?ve years and the auditor is an industry specialist; NSPECTEN –
?rm’s auditor is not a SPECTEN auditor; n – number of ?rm year observations
Table V.
Difference in mean
discretionary income
increasing accruals
between audit quality
variables for ?rms in
the highest 20 and
lowest 80 percent of
price-to-earnings ratios
Audit quality
67
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
HV
it21
is negative. Since our hypotheses are directional, tests for all our audit quality
variables (both main effects and interaction terms) are one-tailed. Similar to univariate
tests inthe previous sections, Table VI shows that the coef?cients for the BigN, specialist,
and long tenure-specialist audit quality measures are negative with estimates and ( p)
values of 20.006 (0.000), 20.002 (0.061), and 20.002 (0.065), respectively. However, the
estimate for our audit quality variable, TEN
it
, is an insigni?cant 20.001 (0.364).
We test hypotheses with interaction terms between our highly valued equity and audit
quality variables: BIGN
*
HV
it21
, SPEC
*
HV
it21
, TEN
*
HV
it21
, and SPECTEN
*
HV
it21
.
Respective estimates and p-values for the income increasing accruals sample are 0.004
(0.044), 0.003 (0.087), 0.003 (0.055), and 0.006 (0.003). For each of the four alternative audit
quality measures, discretionary accruals increase when the client ?rm is highly valued.
As described above, since incentives related to managing earnings higher should be
associated with higher discretionary accruals, we restrict tests in Table VI to clients
reporting income increasing discretionary accruals. Nevertheless, alternative factors may
induce negative accruals. For example, auditors with a conservative bias may provoke
income decreasingaccountingchoices (Becker et al., 1998) andthis maybe especiallylikely
if the ?rm is highly valued. To consider this possibility and other factors related to lower
accruals, we re-estimate equations (3)-(6) using signed (positive and negative)
discretionary accruals as the dependent variable. Table VII provides ?ndings from
tests of models using this all accruals sample. While results for controls are similar across
both samples, in contrast to Table VI, leverage is inversely related to discretionary
accruals. The coef?cients for the Big N, specialist, and long tenure-specialist audit quality
measures are negative with estimates and ( p) values of 20.005 (0.000), 20.002 (0.026),
BN Specialist Tenure Specialist and tenure
coef?cient
( p-value)
coef?cient
( p-value)
coef?cient
( p-value)
coef?cient
( p-value)
ACRL
it
0.546 (0.000) 0.563 (0.000) 0.561 (0.000) 0.561 (0.000)
LnASSET
it
20.005 (0.000) 20.004 (0.000) 20.004 (0.000) 20.004 (0.000)
LEV
it
0.008 (0.000) 0.009 (0.000) 0.008 (0.000) 0.008 (0.000)
OCF
it
20.042 (0.091) 20.046 (0.000) 20.046 (0.000) 20.046 (0.000)
ROA
it
0.001 (0.000) 0.011 (0.024) 0.010 (0.028) 0.010 (0.030)
HV
it21
20.006 (0.007) 20.003 (0.006) 20.004 (0.004) 20.003 (0.004)
BIGN
it21
20.006 (0.000)
BIGN
*
HV
it21
0.004 (0.044)
SPEC
it21
20.002 (0.061)
SPEC
*
HV
it21
0.003 (0.087)
TEN
it21
20.001 (0.364)
TEN
*
HV
it21
0.003 (0.055)
SPECTEN 20.002 (0.065)
SPECTEN
*
HV
it21
0.006 (0.003)
INDUS Not reported Not reported Not reported Not reported
Adjusted R
2
0.507 0.533 0.532 0.532
F( p) 0.000 0.000 0.000 0.000
n 16,346 13,172 13,172 13,172
Notes: DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
HV
it21
þ
a
7
AQ
it
þa
8
AQ
*
it
HV
it21
þa
9
INDUS þ1it; variables are de?ned in Table I
Table VI.
Results for effects
of overvaluation on the
income increasing
discretionary accruals
of Big N and audit
specialist clients
ARJ
26,1
68
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
and 2 0.002 (0.026), respectively, and once again the estimate for audit quality variable,
TEN
it2
, is not signi?cant 20.001 (0.320). Big N and specialist estimates are similar to
other audit quality – accruals studies. For example, Becker et al. (1998) use a sample of
10,881 ?rm year observations from the 1989 to 1992 period to report that non-Big
6 audit accruals are 1.5 percent higher than Big 6 accruals. In a similar study Balsamet al.
(2003) reports 20.006 estimate on an audit quality dummy from a regression of
discretionary accruals on controls and an industry leadership indicator variable.
Regarding TEN
it
, prior audit tenure studies con?ict on how differing lengths of
auditor-client tenure re?ect audit quality. For example, Johnson et al. (2002) report that
relative to medium tenures of four to six years ?nancial reporting quality as measured
by accruals decreases when the tenure is short, but ?nd no reduction in audit quality
for long tenure auditors greater than nine years. Myers et al. (2003) generally con?rm
these ?ndings. However, Davis et al. (2003) ?nd that discretionary accruals increase
over years when audit tenure is continuously measured. Although we believe our
choice of tenure equal to or greater than ?ve years re?ects the most ef?cacious measure
of audit quality from the existing literature, a possible explanation for the lack of
signi?cance may simply re?ect that documented in prior studies.
As depicted in Table VII, results of our main tests are unchanged from Table VI as
all audit quality-highly valued equity interaction terms, BIGN
*
HV
it21
, SPEC
*
HV
it21
,
TEN
*
HV
it21
, and SPECTEN
*
HV
it21
remain statistically signi?cant and positive
0.003 (0.061), 0.004 (0.004), 0.002 (0.073), and 0.004 (0.005).
These ?ndings provide evidence that relative to other ?rms, the accruals decreasing
effect of high quality auditors is reduced when the client is highly valued. Furthermore,
these results are robust across several measures of audit quality with interaction term,
BN Specialist Tenure Specialist and tenure
coef?cient
( p-value)
coef?cient
( p-value)
coef?cient
( p-value)
coef?cient
( p-value)
ACRL
it
0.798 (0.000) 0.809 (0.000) 0.809 (0.000) 0.809 (0.000)
LnASSET
it
20.003 (0.000) 20.002 (0.000) 20.002 (0.000) 20.002 (0.000)
LEV
it
20.004 (0.004) 20.003 (0.024) 20.003 (0.057) 20.003 (0.048)
OCF
it
20.023 (0.000) 20.018 (0.000) 20.019 (0.000) 20.018 (0.000)
ROA
it
0.011 (0.000) 0.010 (0.218) 0.010 (0.000) 0.010 (0.000)
HV
it21
20.006 (0.002) 20.003 (0.000) 20.004 (0.000) 20.003 (0.000)
BIGN
it21
20.005 (0.000)
BIGN
*
HV
it21
0.003 (0.061)
SPEC
it21
20.002 (0.026)
SPEC
*
HV
it21
0.004 (0.004)
TEN
it21
20.001 (0.320)
TEN
*
HV
it21
0.002 (0.073)
SPECTEN 20.002 (0.026)
SPECTEN
*
HV
it21
0.004 (0.005)
INDUS Not reported Not reported Not reported Not reported
Adjusted R
2
0.747 0.760 0.760 0.760
F( p) 0.000 0.000 0.000 0.000
N 29,405 23,937 23,937 23,937
Notes: DAC
it
¼ a
0
þa
1
ACRL
it
þa
2
LnASSET
it
þa
3
LEV
it
þa
4
OCF
it
þa
5
ROA
it
þa
6
HV
it21
þ
a
7
AQ
it
þa
8
AQ
*
it
HV
it21
þa
9
INDUS þ1
it
; variables are de?ned in Table I
Table VII.
Results for effects
of overvaluation on the
discretionary accruals
of Big N and audit
specialist clients
Audit quality
69
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
SPECTEN
*
HV
it21
, showing the highest and most statistically signi?cant estimate
(0.006) from our income increasing – Big N only sample. Hence, high valuation
attenuates the accruals decreasing effect of a long tenure – industry specialist auditor
in that accruals are 0.6 percent of assets higher for clients in the highest quintile of PEs.
Unreported mean assets and common shares outstanding for SPECTEN clients are
$4,989,790,883 and 222,265,446 shares, respectively. Thus, on average net income is
increased by 0.006 times $4,989,790,883 or $29,938,745. Average earnings per share
increases by $0.135 per share ($29,938,745 divided by 222,265,446 shares). While not as
pronounced, the statistically signi?cant estimates for all other audit quality – high
valued client interaction terms similarly illustrate the potential economic impact of
increasing accruals on net income. Indeed, the unreported average diluted earnings
before extraordinary items per share for highly valued SPECTEN ?rms is $0.644,
suggesting that relatively small increases in our interaction coef?cients can have a
signi?cant economic impact on earnings per share.
Robustness
We conduct several sensitivity tests. Regarding valuation variable, HV
it21
, we
also alternatively measure price-to-earnings valuation by stratifying ?rms into
deciles (quintiles) and assign a 1 to companies in the lowest PE decile and quintile and
10 (5) to clients in the highest. Results are similar to our main ?ndings. In particular,
for both income increasing and all discretionary accruals samples, results show
signi?cantly negative (positive) coef?cients for audit quality variables (interaction
terms), BIGN, SPEC, and SPECTEN (BIGN
*
HV
it21
, SPEC
*
HV
it21
, and
SPECTEN
*
HV
it21
). TEN
*
HV
it21
is not signi?cant for either price-to-earnings
measure.
To control for factors occurring over years that our models may not capture, using
1998 as the base year, we include year dummies in each of our equations. With the
exception of TEN
*
HV
it21
( p ¼ 0.125) in our all accruals model, audit quality-highly
valued equity interaction terms are signi?cant and positive. Since ?nancing and/or
acquisition activities could impact accruals, we also assign indicator variables equal to
1 if common stock or debt increased by more than 10 percent and/or the ?rm reported
an acquisition. Results do not alter our ?ndings.
Negative prior year earnings may impact managers’ incentives to report current
year higher net income. We also run tests with a dummy equal to 1 if the ?rm reported
a loss in the prior year. In addition, to control for the impact of events not captured
in our models on our dependent variable, discretionary accruals, we further include
an abnormal returns variable measured according to the capital asset pricing model
(CAPM).
That is:
AR
it
¼ R
it
2ER
it
ð7Þ
CAPM:
ER
it
¼ R
ft
þb
it
ðR
mt
2R
ft
Þ ð8Þ
where, ER
it
is ?rm i’s expected end of ?scal year t return; R
ft
is the one month treasury
bill rate; and b
it
is the end of ?scal year t market model b calculated over the prior year
ARJ
26,1
70
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
with daily closing prices. R
mt
is the value-weighted return from the NYSE, AMEX, and
NASDAQ (from CRSP). Inclusion of these variables does not alter ?ndings as audit
quality – highly valued equity interaction terms remain generally signi?cant and
positive.
Discussion of results and conclusion
Using audit quality proxies from prior literature, this study examines the accruals
decreasing effect of high quality auditors on highly valued audit clients. We posit that
management incentives associated with highly valued equity reduce the tendency of
high quality auditors to reduce accruals. Although results generally support
hypotheses, reported increases in accruals may be affected by the particular measure of
audit quality used. While not as pronounced, the ?ndings of this study seem to provide
some evidence that high quality, Big N auditors continue to maintain their audit
quality edge, albeit marginally, over other audit ?rms. To reiterate, as per Table V, and
in contrast to our other audit quality characteristics, the difference in mean accruals
between Big N and non-Big N auditors is signi?cant even when the client is highly
valued. Hence, perhaps large audit ?rms with more clients and greater reputation at
stake are better able to mitigate the accruals increasing tendency of highly valued
?rms. Nevertheless, multivariate results with controls do reveal higher discretionary
accruals for highly valued-Big N audit clients. In addition to our main multivariate
tests, we investigate this conjecture further by eliminating long tenure and
specialist auditors from the Big N sample. For both our increasing and signed
accruals model unreported ?ndings show that while estimates for the BIGN
*
HV
it21
interaction terms are positive (0.004 and 0.005, respectively), neither are signi?cant
(0.170 and 0.206). While we do not further address this issue here, additional research
that attempts to disentangle the characteristics of large auditors, specialist auditors,
and long tenured auditors as they relate to highly valued audit clients might provide
additional insight.
Regarding other areas for further research, consideration might also be given to the
impact of corporate governance on the audit quality-highly valued equity relation.
In addition, future studies could investigate temporal aspects of audit quality and
highly valued equity both separately and in conjunction with each other. For example,
does the audit quality effect change before, during, and after periods of market tops
when prices and P/Es are at historic highs?
Notes
1. We include SIC dummies in regression models to provide an additional control for industry
effects that may not be captured by the within industry Jones model estimates.
2. As discussed above, discretionary accruals (the residuals from equation (1)) are estimated
using all (positive and negative) accruals. For the all accruals sample, the unreported mean
discretionary accruals are 0.002 for the Big N sample and 0.000 for the specialist and tenure
sample. As reported in Table II, average discretionary accruals are higher for our income
increasing accruals samples.
References
Anderson, K. and Brooks, C. (2006), “The long-term price-earnings ratio”, Journal of Business
Finance & Accounting, Vol. 7, pp. 1063-1086.
Audit quality
71
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Antle, R. and Nalebuff, B. (1991), “Conservatism and auditor-client negotiations”, Journal of
Accounting Research, Vol. 29, pp. 31-54.
Balsam, S., Krishnan, J. and Yang, J.S. (2003), “Auditor industry specialization and earnings
quality”, Auditing: A Journal of Practice & Theory, Vol. 22, pp. 71-97.
Basu, S. (1977), “Investment performance of common stocks in relation to their price
earnings ratios: a test of the ef?cient market hypothesis”, Journal of Finance, Vol. 32,
pp. 663-682.
Beatty, R. (1989), “Auditor reputation and the pricing of initial public offerings”, The Accounting
Review, Vol. 4, pp. 693-709.
Becker, C.L., Defond, M.L., Jiambalvo, J. and Subramanyam, K.R. (1998), “The effect of
audit quality on earnings management”, Contemporary Accounting Research, Vol. 15,
pp. 1-24.
Bergstresser, D. and Philippon, T. (2006), “CEO incentives and earnings management”, Journal of
Financial Economics, Vol. 80, pp. 511-529.
Campbell, J. and Shiller, R. (2001), “Valuation ratios and the long-run stock market outlook:
an update”, Cowles Foundation Discussion Paper No. 1295, Cowles Foundation for
Research in Economics, Yale University, New Haven, CT, March.
Carcello, J.V. and Nagy, A.L. (2004), “Audit ?rm tenure and fraudulent ?nancial reporting”,
Auditing: A Journal of Practice & Theory, Vol. 23, pp. 55-69.
Chi, J.D. and Gupta, M. (2009), “Overvaluation and earnings management”, Journal of Banking
& Finance, Vol. 33, pp. 1652-1663.
Chopra, N.J., Lakonishok, L. and Ritter, J.R. (1992), “Measuring abnormal performance: do stocks
overreact?”, Journal of Financial Economics, Vol. 31, pp. 235-268.
Craswell, A.J. and Taylor, S. (1995), “Auditor brand name reputations and industry
specialization and earnings quality”, Auditing: A Journal of Practice & Theory, Vol. 23,
pp. 297-322.
Davis, L.R., Soo, B.S. and Trompeter, G.M. (2003), “Auditor tenure and the ability to meet or beat
earnings forecasts”, Contemporary Accounting Research, Vol. 26, pp. 517-548.
DeAngelo, L.E. (1981), “Audit size and audit quality”, Journal of Accounting and Economics,
Vol. 3, pp. 183-199.
Dechow, P.M. and Dichev, D. (2002), “The quality of accruals and earnings: the role of accruals
estimation errors”, The Accounting Review, Vol. 77, pp. 35-59.
DeFond, M.L. and Jiambalvo, J. (1994), “Debt covenant violation and manipulation of accruals”,
Journal of Accounting and Economics, Vol. 17, pp. 145-176.
Efendi, J., Srivasta, A. and Swanson, E. (2007), “Why do corporate managers misstate ?nancial
statements? The role of option compensation and other factors”, Journal of Financial
Economics, Vol. 85, pp. 667-708.
Ellingsen, J.E., Pany, K. and Fagan, P. (1989), “SAS No. 59: how to evaluate going concern
opinions”, Journal of Accountancy, January, pp. 24-31.
Francis, J., Mayhew, E. and Sparks, H. (1999), “The role of big six auditors in
the credible reporting of accruals”, Auditing: A Journal of Practice & Theory, Vol. 18,
pp. 17-34.
Gibbons, M., Salterio, S. and Webb, A. (2001), “Evidence about auditor-client management
negotiation concerning clients’ ?nancial reporting”, Journal of Accounting Research,
Vol. 39, pp. 535-564.
ARJ
26,1
72
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Goldwasser, A. and Eickemeyer, J.H. (2004), “Introduction”, Accountant’s Liability, Chapter 1,
Practicing Law Institute, New York, NY.
Grimm, S.D. (2009), “The role of accounting quality in securities class action lawsuits”, working
paper, University of Minnesota, St Paul, MN, June.
Heninger, W.G. (2001), “The association between auditor litigation and abnormal accruals”,
The Accounting Review, Vol. 76, pp. 111-126.
Houmes, R. and Skantz, T. (2010), “Discretionary accruals and highly valued equity”, Journal of
Business Finance & Accounting, Vol. 37, pp. 60-92.
Houmes, R., Dickins, D. and O’Keefe, R. (2012), “New evidence on the incremental information
content of earnings reported using the LIFO inventory method”, Advances in Accounting,
Vol. 28, pp. 235-242.
Jensen, M. (2005), “Agency costs of highly valued equity”, Financial Management, Vol. 34,
pp. 5-20.
Jensen, M.C. and Meckling, W.H. (1976), “Theory of the ?rm: managerial behavior, agency costs
and ownership structure”, Journal of Financial Economics, Vol. 3, pp. 305-360.
Johnson, V.E., Khurana, I.K. and Reynolds, J.K. (2002), “Audit-?rm tenure and the quality of
?nancial reports”, Contemporary Accounting Review, Vol. 19, pp. 637-660.
Jones, J. (1991), “Earnings management during import relief investigations”, Journal of
Accounting Research, Vol. 29, pp. 193-228.
Kothari, S.P., Leone, A. and Wasley, C. (2005), “Performance matched discretionary accrual
measures”, Journal of Accounting and Economics, Vol. 39, pp. 163-197.
Krishnan, G.V. (2003), “Audit quality and the pricing of discretionary accruals”, Auditing:
A Journal of Practice & Theory, Vol. 22, pp. 243-263.
Lakonishok, J., Shleifer, A. and Vishny, R. (1994), “Contrarian investment extrapolation and
risk”, Journal of Finance, Vol. 49, pp. 1541-1578.
Lennox, C. (1999), “Audit quality and auditor size: an evaluation of reputation and deep
pockets hypothesis”, Journal of Business Finance & Accounting, Vol. 26, pp. 779-805.
McNichols, M. (2000), “Research design issues in earnings management studies”, Journal of
Accounting & Public Policy, Vol. 19, pp. 313-345.
Mansi, S.A., Maxwell, W.F. and Miler, D.P. (2004), “Does audit quality and tenure matter to
investors? Evidence from the bond market”, Journal of Accounting Research, Vol. 42,
pp. 755-793.
Menon, K. and Williams, D. (2004), “Former audit partners and abnormal accruals”,
The Accounting Review, Vol. 79, pp. 1095-1118.
Myers, J.N., Myers, L.A. and Omer, T.C. (2003), “Exploring the term of the auditor-client
relationship and the quality of earnings: a case for mandatory auditor rotation?”,
The Accounting Review, Vol. 78, pp. 779-799.
Palmrose, Z. (1986), “Audit fees and auditor size: further evidence”, Journal of Accounting
Research, Vol. 24, pp. 97-110.
Palmrose, Z. (1988), “An analysis of auditor litigation and audit service quality”, The Accounting
Review, Vol. 63, pp. 55-73.
Renas, S.M. and Cebula, R.J. (2005), “Enron, herding, and the deterrent effect of
disclosure of impropriety”, American Journal of Economics and Sociology, Vol. 64,
pp. 743-756.
Audit quality
73
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Reynolds, J.K. and Francis, J.R. (2000), “Does size matter? The in?uence of large clients on
of?ce-level auditor-reporting decisions”, Journal of Accounting and Economics, Vol. 30,
pp. 375-400.
Skinner, D. and Sloan, R. (2002), “Earnings surprises, growth expectations and stock returns or
don’t let an earnings torpedo sink your portfolio”, Review of Accounting Studies, Vol. 7,
pp. 289-312.
Subramanyam, K.R. (1996), “The pricing of discretionary accruals”, Journal of Accounting and
Economics, Vol. 22, pp. 249-281.
Teoh, S.H. and Wong, T.J. (1993), “Perceived auditor quality and the earnings response
coef?cient”, The Accounting Review, Vol. 68, pp. 346-367.
Corresponding author
Robert Houmes can be contacted at: [email protected]
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
ARJ
26,1
74
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
This article has been cited by:
1. Ahsan Habib, Rong Gong, Mahmud Hossain. 2013. Overvalued equities and audit fees: a research note.
Managerial Auditing Journal 28:8, 755-776. [Abstract] [Full Text] [PDF]
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
:
1
7
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_271964185.pdf