The effect of mandatory disclosure requirements and disclosure types of auditor fees

Description
The Taiwanese authorities issued and revised the disclosure regulations with regard to auditor fees three
times during the period 2002e2012. The most important change in disclosure regulation is that firms
have to disclose their auditor fees and have been able to disclose fees in the form of individual amount or
fee range since 2009. This study extends the perspective of auditor independence to explore the effect of
mandatory disclosure requirements and disclosure types of auditor fees on earnings management of
listed corporations. The results show that the enhanced information transparency induced by the
mandatory disclosure requirements of auditor fees is useful to reduce both positive accruals-based
earnings management and real earnings management. Furthermore, firms that disclose their auditor
fees in the form of individual amount have lower positive accruals-based earnings management than
those in the form of fee range. The overall findings are consistent with the notion that the enhanced
information transparency related to auditor fees is associated with enhanced auditor independence

The effect of mandatory disclosure requirements and disclosure types of auditor
fees on earnings management: Evidence from Taiwan
Chieh-Shuo Chen
*
Department of Accounting, National Changhua University of Education, Taiwan, ROC
a r t i c l e i n f o
Article history:
Received 19 August 2015
Accepted 8 December 2015
Available online xxx
Keywords:
Auditor fees
Mandatory disclosure requirements
Disclosure type
Discretionary accruals
a b s t r a c t
The Taiwanese authorities issued and revised the disclosure regulations with regard to auditor fees three
times during the period 2002e2012. The most important change in disclosure regulation is that ?rms
have to disclose their auditor fees and have been able to disclose fees in the form of individual amount or
fee range since 2009. This study extends the perspective of auditor independence to explore the effect of
mandatory disclosure requirements and disclosure types of auditor fees on earnings management of
listed corporations. The results show that the enhanced information transparency induced by the
mandatory disclosure requirements of auditor fees is useful to reduce both positive accruals-based
earnings management and real earnings management. Furthermore, ?rms that disclose their auditor
fees in the form of individual amount have lower positive accruals-based earnings management than
those in the form of fee range. The overall ?ndings are consistent with the notion that the enhanced
information transparency related to auditor fees is associated with enhanced auditor independence and
thus support Dye's (1991) theory.
© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier
Taiwan LLC. All rights reserved.
1. Introduction
A number of major ?nancial scandals, like Enron and Worldcom
in the US, or Paci?c Electric Wire & Cable, Procomp, Infodisc
Technology, and Simmit Technology in Taiwan, arose in the early
2000s, and exposed the ethical and quality problems that had
existed among auditors for decades. Issues related to the pricing of
auditor fees and the impacts of these on audit and earnings quality
have received signi?cant attention from regulators and stake-
holders in many countries. This is because an audit of ?nancial
reports by a third party, such as auditors, can enhance the quality of
the ?nancial information reported by management, and improve
the quality of information that stakeholders have about the value of
traded securities. Auditors are commissioned and their auditor fees
are paid by management. An important concern about ?nancial-
reporting quality is thus whether auditors can resist the eco-
nomic dependence on their clients or management pressures, such
as the threat of changes in auditor or reductions in non-audit ser-
vices, and charge fair audit fees to maintain audit quality.
1
The in-
formation transparency of auditor fees is thus useful not only to
understand supply and demand for audit and non-audit services,
but also to investigate the relationship between auditor fees and
audit quality.
2
In order to respond to concerns over the possible loss of auditor
independence and improve the information transparency of
auditor fees, the US Securities and Exchange Commission (US SEC)
has required companies to disclose audit and non-audit fees in
their proxy statements ?led on or after February 5, 2001. Moreover,
listed companies in China have been required by the China
* No. 2, Shi-Da Road, Changhua City 50074, Taiwan, ROC Tel.: þ886 4
7232105x7521; fax: þ886 4 7211294.
E-mail address: [email protected].
Peer review under responsibility of College of Management, National Cheng
Kung University.
1
Nichols and Price (1976) and Wu, Chiou, and Cheng (2011) claimed that the cost
of changing auditor is low for a ?rm, and thus the auditor will not wish to increase
the possibility of being dismissed by entering into con?icts with a client. This
economic dependence consequently compromises the auditor's independence. In
addition, Francis and Wang (2005) showed that clients gained bargaining power
over auditors after US SEC mandated public disclosure of audit fees.
2
Watkins, Hillison, and Morecroft (2004) showed that audit fees play an
important role in audit quality, and the pricing of audit services continues to be an
important issue. Huang, Lin, and Tien (2012) indicated that audit quality will
become worse when the audit fee decreases, regardless of whether there is an
auditor switch.
HOSTED BY
Contents lists available at ScienceDirect
Asia Paci?c Management Review
j ournal homepage: www. el sevi er. com/ l ocat e/ apmrvhttp://dx.doi.org/10.1016/j.apmrv.2015.12.002
1029-3132/© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier Taiwan LLC. All rights reserved.
Asia Paci?c Management Review xxx (2016) 1e15
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
Securities Regulatory Commission (CSRC) to publicly disclose audit
fees since their 2001 annual reports. However, Taiwanese regula-
tors required listed ?rms to disclose the amount of audit or non-
audit fees before 2009 only when non-audit fees are excessive,
audit fees decrease with an audit ?rm switch, or when audit fees
decrease substantially. On the other hand, in order to encourage
public ?rms to improve their information disclosure and further
enhance the information transparency of capital markets, the
Taiwanese information disclosure and evaluation system adopted
since 2002, and corporate governance assessment system in use
since 2005, both regard voluntary disclosure of auditor fees as a
bonus point. Nevertheless, less than 40% of listed companies dis-
closed auditor fees information from 2002 to 2008. Since 2009,
Taiwanese regulators have required all public ?rms to disclose
auditor fees information by two alternative disclosure types (i.e.,
fee range and individual amount), except for three speci?c condi-
tions for which the amount must be disclosed.
3
This thus motivated
the current study to explore whether the enhanced information
transparency induced by the mandatory regulation related to
alternative disclosure types of auditor fees is associated with
enhanced auditor independence, and thus better audit quality.
This study divides the sample into two main sub-periods based
on the mandatory regulation with regard to alternative disclosure
types of auditor fees (more transparent information period of
auditor fees for 2009e2012 versus less transparent information
period of auditor fees for 2002e2008). Therefore, this study ana-
lyzes two following research issues: (1) whether earnings man-
agement is lower in the more transparent information period than
the less transparent information period, and (2) whether earnings
management in the more transparent information period is lower
when the auditor fees are disclosed by individual amount than by
fee range.
This study makes some contributions. Firstly, prior literature
suggested that the legal environment, regulatory regime, or
disclosure requirement affects audit pricing (Choi, Kim, Liu, &
Simunic, 2008; Taylor & Simon, 1999). Taiwanese regulators have
required listed ?rms to disclose auditor fees in the form of amount
or range since 2009. This study thus has some academic value in
relation to Taiwan. Secondly, in contrast to the amount type of
auditor fees disclosure required by other countries, two alternative
disclosure types are allowed in Taiwan, and this can not only
mitigate the concerns of auditors about price competition,
4
but also
provide this study with another opportunity to compare the audit
quality in relation to different disclosure types. The ?ndings of this
study are thus valuable for Taiwanese regulators if they consider
revising the related regulations in the future. Finally, there is a clear
controversy in the literature concerning whether the disclosure of
audit fees affects auditor independence (DeAngelo, 1981a; Dye,
1991), and no studies have yet examined the relationship be-
tween earnings management and the disclosure of auditor fees.
This study ?lls this gap in the literature by investigating whether
earnings management decreases after mandatory disclosure re-
quirements of auditor fees, implying that auditor independence
increases.
The next section of this study brie?y introduces the changes in
disclosure regulations related to auditor fees in Taiwan, reviews the
literature, and develops the hypotheses. The empirical methodol-
ogy is then discussed, along with the sample selection process. The
fourth section presents the empirical results, with the conclusions
are in the ?nal section.
2. Changes in disclosure regulations, literature review, and
hypotheses development
2.1. Changes in disclosure regulations related to auditor fees in
Taiwan
Taiwanese disclosure regulations with regard to auditor fees
were revised three times during the period 2002e2012. Taiwanese
regulators had required public ?rms to disclose auditor fees since
2002 under three speci?c conditions. Firstly, the listed ?rms should
disclose audit and non-audit fees along with service scope if the
non-audit fees are more than 500 thousand Taiwan dollars or more
than 25% of audit fees. Secondly, the listed ?rms should disclose the
reduced amount, percentage, and cause of audit fees if the audit
?rm is changed and the audit fees are lower than the prior fees.
Thirdly, the listed ?rms should disclose the reduced amount, per-
centage, and cause of audit fees if the audit fees decrease by more
than 15% of prior fees. The disclosure regulations were revised in
2006 when Taiwanese regulators removed the condition in which
the non-audit fees are more than 500 thousand Taiwan dollars.
Taiwanese regulators further revised the regulations in 2009 to
require all public ?rms to disclose auditor fees by individual
amount or fee range, except for three speci?c conditions that
require individual amount disclosure.
5
According to Table 1, the
number of listed ?rms that did not disclose auditor fees is about
90%, 65%, and 4% during the period of 2002e2005, 2006e2008, and
2009e2012, respectively. After 2009, most ?rms disclosed auditor
fees by individual amount.
2.2. Literature review and hypotheses development
Taiwanese list ?rms were only required to disclose auditor fees
conditionally before 2009, which limited the related research and
analyses. Most prior studies explored audit pricing in Taiwan using
only limited database or questionnaire (Chang & Tsao, 2005; Chen,
Liu, &Lin, 2003; Chen &Wu, 2004; Lee, Liao, &Huang, 2010; Shiue,
Chang, & Kao, 2008; Su, 2000). Moreover, Francis and Wang (2005)
showed that there is greater downward fee adjustment than up-
ward adjustment that suggests that clients have bargaining power
over auditors after the US SEC's mandated public disclosure of audit
fees. They did not further investigated whether the bargaining
power after the requirement of disclosure is likely associated with
weaker audit quality due to the less audit resource. The aspect of
economic dependence suggests the lower audit fees will reduce the
economic bonding between the auditor and their client, which is
less likely to impair auditor independence (DeFond, Raghunandan,
& Subramanyam, 2002), whereas the aspect of input resources
suggests that the lower audit fees will induce less audit effort and
then impair audit quality (Mitra, Deis, & Hossain, 2009). Recently,
Liao, Wang, and Chi (2012) used the sample period of 2002e2012
and showed the average audit fee is higher after than before the
disclosure requirement of Taiwanese regulators (2009). They also
indicated that abnormal audit fees does not impair audit quality
(abnormal accruals or earnings threshold) before and after the
disclosure requirement of auditor fees, consistent with DeFond
3
Liao et al. (2012) used samples of audit fees disclosed from 2002 to 2008 as the
selected sample period, and those from 2009 to 2010 as the complete sample
period.
4
Generally, each auditor regards auditor fees, including audit and non-audit fees,
as secret, and is thus unwilling to disclose the fees they charge (Lee et al., 2003).
5
According to the new regulations revised in 2009, auditor fees can been dis-
closed by the following fee ranges: (1) less than 2000 thousand Taiwan dollars, (2)
2000e4000 thousand Taiwan dollars, (3) 4000e6000 thousand Taiwan dollars, (4)
6000e8000 thousand Taiwan dollars, (5) 8000e10,000 thousand Taiwan dollars,
and (6) more than 10,000 thousand Taiwan dollars.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 2
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
et al. (2002). However, a controversial and important relationship
between mandatory disclosure requirements of auditor fees and
audit quality has not been investigated in prior studies, especially
for the most important change with regard to allowing two
different disclosure types.
DeAngelo (1981a) suggested that the disclosure of audit fees
could not affect the transaction costs that give rise to quasi-rents,
resulting in no relationship between the disclosure of audit fees
and audit opinion used to proxy for auditor independence. How-
ever, Dye (1991) suggested that the disclosure of audit fees would
induce future quasi-rents to be zero and improve auditor inde-
pendence, leading the issued audit opinions to be more credible.
Francis and Ke (2006) found that the market's perception of auditor
independence and earnings quality (earnings response coef?cient)
was signi?cantly lower for ?rms with high levels of non-audit fees
than for ?rms with low levels of such fees following the US SEC
mandated disclosure of auditor fees. In contrast, there was no
reduction in earnings response coef?cients for ?rms with high
levels of non-audit fees in the year prior to the new fee disclosures.
Their ?ndings suggested that the disclosure of auditor fees provides
market participants with new information to appraise auditor in-
dependence and earnings quality. Lai (2009) showed that the
?nancial statements issued after the disclosure requirement of US
SEC (2000) are more likely to be associated with a going concern
opinion than those issued before the disclosure requirement, sug-
gesting the disclosure of auditor fees can enhance auditor inde-
pendence. In all the literature concerning Taiwan, most prior
studies used Taiwanese information disclosure and evaluation
system as the proxy for information transparency. Chang and Fang
(2006) showed that earnings management decreased after the
implementation of the information disclosure and evaluation sys-
tem. However, the difference between ?rms with more trans-
parency and those with less was not statistically signi?cant.
In addition, Singhvi and Desai (1971) suggested that the quality
of corporate disclosure will affect the quality of investment de-
cisions made by investors. Hence, disclosure quality refers to the
completeness, accuracy and reliability of disclosure in annual re-
ports. Alternative ?nancial-statement formats will affect the pos-
sibility of earnings manipulation being found and investors'
judgments (Hirst & Hopkins, 1998; Jones & Sharma, 2001; Maines
& McDaniel, 2000). Bloom?eld (2002) and Hirshleifer and Teoh
(2003) indicated that investors care not only that information is
made publicly available, but also about the form in which it is
revealed, even when the information content of the alternative
formats is identical.
The most important change in disclosure regulation related to
auditor fees in Taiwan is that ?rms have to disclose their auditor
fees and have been able to disclose fees in the form of individual
amount or fee range since 2009. Based on Dye's (1991) argument
and other studies' empirical evidence (Chang & Fang, 2006; Francis
& Ke, 2006; Lai, 2009), this study argues that during the more
transparent information period of auditor fees, due to the imple-
mentation of the newdisclosure regulations, auditors will not have
quasi-rents from clients and could maintain or improve
Table 1
Summary of auditor fees disclosure.
Year Disclosure and type Number (proportion) Number (proportion) of
mandatory disclosure
Number (proportion) of
voluntary disclosure
2002 No 626 (92.74%) e e
Yes Amount 49 (7.26%) 49 (7.26%) 0 (0.00%)
2003 No 714 (91.19%) e e
Yes Amount 69 (8.81%) 64 (8.17%) 5 (0.64%)
2004 No 776 (92.27%) e e
Yes Amount 65 (7.73%) 57 (6.78%) 8 (0.95%)
2005 No 771 (88.01%) e e
Yes Amount 105 (11.99%) 92 (10.50%) 13 (1.49%)
2006 No 591 (64.66%) e e
Yes Amount 323 (35.34%) 98 (10.72%) 225 (24.62%)
2007 No 582 (63.89%) e e
Yes Amount 329 (36.11%) 118 (12.95%) 211 (23.16%)
2008 No 574 (65.98%) e e
Yes Amount 296 (34.02%) 110 (12.64%) 186 (21.38%)
2009 No 17 (1.78%) e e
Yes Range 466 (48.74%) 466 (48.74%) e
Amount 473 (49.48%) 473 (49.48%) e
2010 No 16 (1.62%) e e
Yes Range 410 (41.54%) 410 (41.54%) e
Amount 561 (56.84%) 561 (56.84%) e
2011 No 37 (3.66%) e e
Yes Range 330 (32.67%) 330 (32.67%) e
Amount 643 (63.67%) 643 (63.67%) e
2012 No 36 (3.48%) e e
Yes Range 379 (36.69%) 379 (36.69%) e
Amount 618 (59.83%) 618 (59.83%) e
2002e2012 No 4740 (48.09%) e e
Yes Range 1585 (16.08%) 1585 (16.08%) e
Amount 3531 (35.83%) 2883 (29.25%) 648 (6.58%)
Note: Yes means that ?rms disclose the complete information of audit and non-audit fees. Firms that disclose audit and non-audit fees in the formof full or partial fee range are
classi?ed into the Range group, while those that disclose audit and non-audit fees in the form of individual amount are classi?ed into the Amount group.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 3
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
independence, further enhancing audit quality, as proxied by
accrual-based earnings management. Moreover, the enhanced
audit quality should be better for ?rms that disclose the individual
amount of auditor fees than those that disclose the fee range. This
study thus proposes the alternative hypotheses as follows.
H1. Ceteris paribus, ?rms have lower accrual-based earnings
management after mandatory disclosure requirements than before
the requirements.
H2. Ceteris paribus, after mandatory disclosure requirements,
?rms that disclose the individual amount of auditor fees have lower
accrual-based earnings management than those that disclose the
fee range.
3. Empirical methodology
3.1. Regression models and variable de?nitions
In order to explore the effect of mandatory disclosure re-
quirements of auditor fees on earnings management, this study
?rst tests H1 by estimating the following accruals-based earnings
management model
6
:
DA
it
¼ a
0
þ a
1
After
it
þa
2
Asset
it
þ a
3
Lev
it
þ a
4
Turnover
it
þ a
5
CFO
it
þa
6
D_loss
it
þ a
7
ROA
it
þ a
8
MB
it
þ a
9
Growth
it
þa
10
Age
it
þ a
11
D_big4
it
þ a
12
D_expert
it
þ a
13
D_newaudit
it
þa
14
Tenure
it
þa
15
import
it
þ
X
uIndustry
it
þy
it
ð1Þ
In the regression, the variables for ?rm i in year t are de?ned as
follows:
DA ¼ performance-adjusted abnormal accruals, measured by
jDAj, DAþ, DAÀ, respectively;
After ¼ 1 if the observation relates to a ?nancial year-end after
2009 (i.e., 2009e2012), and 0 if the observation relates to a
?nancial year-end before 2009 (i.e., 2002e2008);
Asset ¼ natural logarithm of total assets;
Lev ¼ leverage, measured by total liabilities divided by total
assets;
Turnover ¼ total assets turnover;
CFO ¼ ratio of cash ?ows from operation to lagged total assets;
D_loss ¼ 1 if the ?rm has reported a loss in year t or year t À 1,
and 0 otherwise;
ROA ¼ return on total assets, measured by net income divided
by average total assets;
MB ¼ market to book ratio, measured by total market capitali-
zation to book value of equity;
Growth ¼ sales growth rate;
Age ¼ the number of years the ?rm has been listed;
D_big4 ¼ 1 if the ?rm's auditor is a Big 4 auditor, and
0 otherwise;
D_expert ¼ 1 if the ?rm's auditor is an industry specialist (i.e.,
the audit ?rm is the largest supplier in terms of total sales of all
clients in the industry), and 0 otherwise;
D_newaudit ¼ 1 if the ?rm changes the auditor, and
0 otherwise;
Tenure ¼ the number of years the ?rm is audited by the same
auditor;
Import ¼ ratio of sales of the ?rm to total sales of all clients of
the auditor;
Industry ¼ a set of dummy variables representing industry.
Because audit quality can enhance earnings quality, many pre-
vious studies employed discretionary accruals (DA) to measure
audit or earnings quality (Chen, Lin, &Lin, 2008; Chi, Huang, Liao, &
Xie, 2009; Choi, Kim, & Zang, 2010; Hsu, Chen, & Chen, 2013; Liao
et al., 2012; Tseng, Chang, & Shiue, 2012). This study uses DA as a
proxy for audit quality since it captures the degree of negotiations
related to the audit adjustment decisions.
7
The DA measure is
estimated by the performance-adjusted abnormal accrual (Kothari,
Leone, & Wasley, 2005), as follows.
8
TAC
it
TA
itÀ1
¼ g
0
1
TA
itÀ1
þ g
1
ðDREV
it
ÀDREC
it
Þ
TA
itÀ1
þ g
2
PPE
it
TA
itÀ1
þ g
3
ROA
itÀ1
þl
it
(2)
Total accrual (TAC) is measured by the earnings before extraordi-
nary items and discontinued operations minus the operating cash
?ows. DREV is the change in revenues from the preceding year.
DREC is the change in accounts receivables fromthe preceding year.
PPE is the gross value of property, plant, and equipment. ROA is net
income divided by total assets. TA is total assets in the beginning of
year t. This study estimates model (2) for each year and industry
through the cross-sectional regression, and the discretionary ac-
cruals are the residuals of this model.
This study uses absolute discretionary accruals as a common
proxy for earnings management (Chen et al., 2008; Chi et al., 2009;
Choi et al., 2010; Liao et al., 2012; Liao, Sang, & Lin, 2013).
Furthermore, the prior literature suggests that managing earnings
upward to meet or beat benchmarks is more common, and may be
of more concern to auditors, i.e., that the auditor's legal exposure
primarily emanates from ?rms overstating income (Boone,
Khurana, & Raman, 2010; Heninger, 2001; Kinney & Martin, 1994;
Lee, Hsu, & Chen, 2003; Trompeter, 1994). This study also sepa-
rately examines positive (i.e., income increasing) and negative (i.e.,
income decreasing) performance-adjusted DA, respectively (Chen
et al., 2008; Chi et al., 2009; Lee & Lin, 2005; Liao et al., 2012,
2013; Mitra et al., 2009).
6
In testing the statistical signi?cance of the coef?cient estimations, this study
compute the t-statistics using the two-way robust standard errors with ?rm-level
and year-level clustering to adjust standard errors for cross-sectional and time-
series dependence (Cameron & Miller, 2011; Clinton, White, & Woidtke, 2014;
Gow, Ormazabal, & Taylor, 2010; Liao et al., 2012, 2013; Thompson, 2011; Tseng
et al., 2012; Young et al., 2012).
7
Financial statements should be a joint statement from the auditor and manager.
The auditor and client begin negotiations in which the auditor may offer some
adjustments in the revised statement. On average, the overall effect of the ad-
justments is an aggregate decrease in recorded earnings and assets. The client may
threaten to dismiss the auditor and ?nd one more accepting of its views. The
auditor should have better audit quality if he can persist in the audit adjustment
decisions for better earnings quality. This study thus uses the accrual-based earn-
ings management as the proxy for audit quality.
8
Kothari et al. (2005) argued that the DA as estimated by both Jones and
modi?ed Jones models may induce severe measurement error when these models
do not control for the prior performance of the ?rm. Two alternative techniques
were thus suggested to control for prior performance, by adding a prior perfor-
mance measure, like the return on assets of the previous year as an additional
regressor variable to the cross-sectional model, or by adjusting DA through the
average DA of a portfolio matched on industry and previous performance. Liao and
Hung (2010) and Liao et al. (2013) stated that not all research can adopt portfolio
matching. Most recent studies in Taiwan calculated performance-adjusted
abnormal accruals for DA (Hsu et al., 2013; Liao et al., 2013; Liao et al., 2012;
Tseng et al., 2012). In addition, this study also uses the modi?ed Jones model of
DA proposed by Dechow et al. (1995) and real earnings management employed by
Cohen et al. (2008) for additional tests.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 4
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
The dependent variable is thus performance-adjusted DA, and
the variable for testing H1 is AFTER, which takes a value of 1 if the
observations are after mandatory disclosure requirements, and
0 otherwise.
9
If auditor independence is improved and further
enhances audit quality when almost all the listed ?rms disclose
auditor fees, the coef?cient of AFTER is expected to be negative.
10
Other variables are included as controls in model (1). Becker,
DeFond, Jiambalvo, and Subramanyam (1998) suggested that ?rm
size may surrogate for numerous omitted variables. In addition,
?rms with larger ?rm size and longer age have lower discretionary
accruals due to their more stable business. This study thus includes
‘Asset’ and ‘Age’ to control for ?rm size and business life cycle
(Anthony & Ramesh, 1992; Chen et al., 2008; Chi et al., 2009; Lee &
Chen, 2004; Liao et al., 2012, 2013). This study also includes Lev as a
control variable because prior research found that managers of
highly leveraged ?rms have incentives to manage earnings upward
to avoid debt covenant violations (Dechow, Sloan, & Sweeney,
1996; Lee & Chen, 2004; Press & Weintrop, 1990; Watts &
Zimmerman, 1978) or downward to engage in contractual re-
negotiations (DeAngelo, DeAngelo, & Skinner, 1994; Lee et al.,
2003; Lee & Lin, 2005). This study also includes Turnover as a
control variable for business operation capability (Hsu et al., 2013).
Because ?rms with lower cash ?ow or worse pro?tability are more
likely to manage earnings, CFO, D_loss, and ROA are added as
control variables (Choi et al., 2010; Dechow, Sloan, & Sweeney,
1995; Frankel, Johnson, & Nelson, 2002; Healy, 1985; Liao et al.,
2012, 2013; Tseng et al., 2012). This study controls for ?rm
growth by including sales growth (Growth) and market-to-book
ratio (MB) as additional control variables, because DA are higher
for growth ?rms (Carcello &Nagy, 2004; Chi, Lisic, &Pevzner, 2011;
Ghosh & Moon, 2005; Liao et al., 2012, 2013; Su, 2005; Tseng et al.,
2012; Young &Wu, 2003). This study adds Big 4 auditor (D_big4) as
a control variable, because larger auditors are more likely to offer
higher audit quality, inducing higher earnings quality (DeAngelo,
1981b; Francis, Maydew, & Sparks, 1999; Lee & Lin, 2013; Liao
et al., 2012, 2013; Tseng et al., 2012; Wang, Chang, & Lin, 2012).
In addition, prior research found that audit ?rms with industry
specialization can more effectively restrict earnings management
(Balsam, Krishnan, & Yang, 2003; Jiang & Yang, 2005; Krishnan,
2003), and this study thus includes D_expert as a control vari-
able. This study also includes auditor tenure (Tenure) to control for
the potential effects of this even if the literature is inconsistent
(Carey &Simnett, 2006; Ghosh &Moon, 2005; Lee &Lin, 2005; Liao
& Hung, 2010; Liao et al., 2012; Myers, Myers, & Omer, 2003).
Becker et al. (1998) and DeFond and Subramanyam (1998) indi-
cated that DA tends to change when ?rms change auditor. This
study thus adds D_newaudit as a control variable (Liao et al., 2012;
Tseng et al., 2012). The variable Import is included to control for the
importance of the clients to the auditor (Fan, Chen, &Wu, 2007; Lee
& Chen, 2004; Liao et al., 2013).
Moreover, this study explores whether the degree of earnings
management is different between ?rms that disclose the individual
amount of auditor fees and those that disclose a fee range after
mandatory disclosure requirements. The H2 is thus tested by esti-
mating the following model:
DA
it
¼ b
0
þ b
1
D_amount
it
þ b
2
Asset
it
þ b
3
Lev
it
þ b
4
Turnover
it
þ b
5
CFO
it
þb
6
D_loss
it
þ b
7
ROA
it
þ b
8
MB
it
þ b
9
Growth
it
þ b
10
Age
it
þ b
11
D_big4
it
þ b
12
D_expert
it
þ b
13
D_newaudit
it
þ b
14
Tenure
it
þ b
15
Import
it
þ
X
uIndustry
it
þ
X
jYear
it
þ ?
it
ð3Þ
In the regression, the variable for testing H2 is D_amount, which
takes the value of 1 if the ?rm discloses the individual amount of
audit and non-audit fees, and 0 if the ?rm discloses auditor fees
fully or partially in the form of fee range. The de?nitions of other
variables are same as those in model (1).
3.2. Sample selection and data sources
The sample period ranges from 2002 to 2012 for Taiwanese
listed ?rms, due to data availability. Taking into account the
external validity, the delisted ?rms are also included in the sample
to avoid the potential survivorship bias. Financial ?rms are
excluded due to their speci?c characteristics, which are dissimilar
to those of other ?rms. The US SEC has required companies to
disclose audit and non-audit fees since 2001. However, many listed
?rms inTaiwan only disclose information about either audit or non-
audit fees. This study refers to the disclosure requirements of the
US SEC (2000), and classi?es the ?rms with complete disclosure of
audit and non-audit fees as those with fees disclosure. Those ?rms
that only disclose audit or non-audit fees, or that do not disclose
any such fees are classi?ed as those without fees disclosure.
11
Financial data, audit ?rm data, and market value of equity are ob-
tained fromthe Taiwan Economic Journal (TEJ) database. This study
excludes ?rms with negative value or missing data on total assets,
sales or market value of equity. Finally, this study also excludes
observations with values of continuous variables outside the 1st
and 99th percentiles of the sample ?rms (Young, Tsai, Chen, & Liao,
2012). The selection procedure yields 9856 ?rm-year observations
for testing H1. After excluding the sample period of 2002e2008 and
the ?rms without fees disclosure, the total number of ?rm-year
observations for testing H2 is 3880.
4. Empirical results and analyses
4.1. Descriptive statistics and correlation analysis
Table 2 reports the descriptive statistics of the research variables
in the analysis. The mean, maximum, and minimum of absolute
discretionary accruals (jDAj) are 0.0631, 0.3337, and 0, respectively,
similar to Liao et al. (2012) and Tseng et al. (2012). The observations
after 2009 (After) are about 40.44% of the total sample. With
respect to ?rm characteristics, sample ?rms have an average of
15.2033 ?rm size (Asset), an average of 0.3751 leverage (Lev), and
an average of 0.0622 operating cash ?ow scaled by lagged total
assets (CFO), respectively. About 29.98% of sample ?rms have a loss
in the current or preceding year (D_loss). On average, the total
assets turnover (Turnover), return on total assets (ROA), market-to-
book ratio (MB), sales growth rate (Growth), and listed year (Age)
are 0.8488, 0.0441, 1.4224, 0.0621, and 10.5825, respectively.
Regarding auditor characteristics, about 84.81% of sample ?rms are
audited by one of the Big 4 audit ?rms (D_big4). Moreover, 27.82%
9
Several prior studies used dummy variables, such as AFTER, TIME, Enron, or
SOR, to test the difference between two different periods (Chang & Fang, 2006;
Chang, Shen, & Fang, 2009; Clinton et al., 2014; Guan, Chien, & Hsu, 2008; Lai,
2009; Raghunandan & Rama, 1995).
10
The US SEC has required companies to disclose audit and non-audit fees in their
proxy statements since 2001. Taiwanese regulators also have required public ?rms
to disclose audit and non-audit fees in their annual reports since 2009. Thus, a
common research limitation of this study and Lai (2009) is that the mandatory
disclosure effect of non-audit fees cannot be discriminated from audit fees.
11
Taiwan Economic Journal database (TEJ) and Liao et al. (2012) both regard the
?rms with only disclosure of audit fees but not non-audit fees as those with
samples with disclosure, will leads to the overestimation of Taiwanese listed ?rms
with complete disclosure of auditor fees.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 5
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
of sample ?rms are audited by an industry-specialist auditor
(D_expert), and 6.07% of sample ?rms change the auditor
(D_newaudit). The average auditor tenure (Tenure) and average
client importance (Import) are 8.4692 and 2.3144%, respectively.
Table 3 represents the Pearson (below diagonal) and Spearman
(above diagonal) correlations among the variables used in this
study. There are signi?cant and negative correlations between
absolute discretionary accruals (jDAj) and Asset, Age, CFO, and
Tenure, indicating that ?rms with larger size, longer listed period,
more operating cash ?ow or longer auditor tenure are less likely to
manipulate discretionary accruals. The absolute discretionary
accruals (jDAj) are signi?cantly and positively correlated with Lev,
Turnover, D_loss, MB, and Growth, indicating that ?rms with more
leverage, higher operation capability, lower pro?tability, and
greater growth are more likely to manipulate discretionary ac-
cruals. Most correlations are similar to those reported in prior
research (Hsu et al., 2013; Tseng et al., 2012). Moreover, the abso-
lute discretionary accruals (jDAj) decrease signi?cantly after 2009.
Table 3 also shows that there are signi?cant correlations among
control variables. Because these correlation coef?cients are lower
than 0.8, and the untabulated variance in?ation factors (VIF) are
lower than 10, this study ?nds no signs of any problem with mul-
ticollinearity (Chang, Chen, & Su, 2010; Gujarati, 1995; Peng &
Young, 2013).
Moreover, the mean and median tests for the variables used in
the regression are shown in Table 4. Firms in the more transparent
information period with regard to auditor fees (i.e., after mandatory
disclosure requirements, along with two alternative disclosure
types) have signi?cantly lower absolute and positive discretionary
accruals (jDAj and DAþ) than those in the less transparent period
(i.e., before mandatory disclosure requirements), as predicted.
There is no signi?cant difference in negative discretionary accruals
(DAÀ) between ?rms in the more transparent information period
and those in the less transparent period. The remaining variables
also indicate several signi?cant differences in the ?rm and auditor
Table 2
Descriptive statistics.
Variable Mean Standard deviation Min Median Max
jDAj 0.0631 0.0565 0.0000 0.0476 0.3337
After 0.4044 0.4908 0.0000 0.0000 1.0000
Asset 15.2033 1.1971 12.8023 15.0543 19.2275
Lev 0.3751 0.1614 0.0555 0.3689 0.8780
Turnover 0.8488 0.5413 0.0300 0.7400 3.3800
CFO 0.0622 0.0837 À0.2347 0.0574 0.3383
D_loss 0.2998 0.4582 0.0000 0.0000 1.0000
ROA 0.0441 0.0827 À0.3177 0.0465 0.2791
MB 1.4224 0.8468 0.2955 1.1937 6.1923
Growth 0.0621 0.2791 À0.6885 0.0371 1.7638
Age 10.5825 7.1036 2.0000 9.0000 44.0000
D_big4 0.8481 0.3586 0.0000 1.0000 1.0000
D_expert 0.2782 0.4481 0.0000 0.0000 1.0000
D_newaudit 0.0607 0.2387 0.0000 0.0000 1.0000
Tenure 8.4692 5.1987 1.0000 7.0000 25.0000
Import 2.3144 8.3897 0.0000 0.0900 85.1400
Note: jDAj is the absolute value of performance-adjusted abnormal accruals pro-
posed by Kothari et al. (2005). After equals one if the observation is after the new
disclosure regulation for alternative disclosure types, and zero otherwise. Asset is
the nature logarithm of total assets. Lev is measured by total liabilities divided by
total assets. Turnover is total assets turnover. CFO is the ratio of cash ?ows from
operation to lagged total assets. D_loss equals one if the ?rm has reported loss in
year t or year t À1, and 0 otherwise. ROA is return on total assets. MB is measured by
total market capitalization to book value of equity. Growth is sales growth rate. Age
is the number of years the ?rm has been listed. D_big4 equals one if the ?rm's
auditor is a Big 4 auditor, and 0 otherwise. D_expert equals one if the ?rm's auditor
is an industry specialist, and zero otherwise. D_newaudit equals one if the ?rm
change the auditor, and 0 otherwise. Tenure is the number of years the ?rm is
audited by the same auditor. Import is the ratio of sales of the ?rmto total sales of all
clients of the auditor.
T
a
b
l
e
3
C
o
r
r
e
l
a
t
i
o
n
m
a
t
r
i
x
.
V
a
r
i
a
b
l
e
1
.
2
.
3
.
4
.
5
.
6
.
7
.
8
.
9
.
1
0
.
1
1
.
1
2
.
1
3
.
1
4
.
1
5
.
1
6
.
1
.
j
D
A
j
À
0
.
0
4
9
a
À
0
.
0
4
5
a
0
.
0
8
1
a
0
.
0
7
8
a
À
0
.
0
6
9
a
0
.
0
4
5
a
0
.
0
3
4
a
0
.
0
7
6
a
0
.
0
2
0
b
À
0
.
1
6
3
a
0
.
0
1
4
0
.
0
0
9
0
.
0
0
2
À
0
.
1
3
3
a
À
0
.
0
0
2
2
.
A
f
t
e
r
À
0
.
0
6
1
a
À
0
.
0
0
6
À
0
.
1
3
5
a
À
0
.
0
7
4
a
À
0
.
0
2
3
b
0
.
0
1
8
c
À
0
.
0
3
1
a
0
.
1
2
6
a
À
0
.
1
3
7
a
0
.
3
3
2
a
0
.
0
3
1
a
0
.
0
3
0
a
À
0
.
0
9
4
a
0
.
2
6
8
a
À
0
.
1
5
2
a
3
.
A
s
s
e
t
À
0
.
0
3
9
a
À
0
.
0
0
6
0
.
1
8
0
a
À
0
.
1
0
3
a
0
.
0
3
1
a
À
0
.
1
3
9
a
0
.
1
3
0
a
0
.
0
1
4
0
.
1
0
5
a
0
.
4
0
9
a
0
.
0
6
4
a
0
.
0
7
1
a
À
0
.
0
1
6
0
.
3
5
9
a
0
.
6
0
1
a
4
.
L
e
v
0
.
1
0
1
a
À
0
.
1
3
1
a
0
.
1
7
2
a
0
.
1
9
2
a
À
0
.
2
7
7
a
0
.
1
9
4
a
À
0
.
2
5
1
a
À
0
.
0
9
8
a
0
.
1
0
1
a
À
0
.
0
2
2
b
À
0
.
0
5
7
a
À
0
.
0
1
4
0
.
0
4
9
a
À
0
.
0
4
8
a
0
.
2
4
1
a
5
.
T
u
r
n
o
v
e
r
0
.
0
8
5
a
À
0
.
0
6
4
a
À
0
.
0
6
4
a
0
.
2
1
4
a
0
.
1
1
1
a
À
0
.
1
9
2
a
0
.
2
5
3
a
0
.
1
8
5
a
0
.
2
4
3
a
À
0
.
2
1
1
a
0
.
0
5
4
a
0
.
0
1
3
À
0
.
0
3
9
a
À
0
.
0
9
8
a
0
.
1
4
6
a
6
.
C
F
O
À
0
.
0
8
7
a
À
0
.
0
2
4
b
0
.
0
4
6
a
À
0
.
2
9
0
a
0
.
0
5
8
a
À
0
.
2
7
5
a
0
.
4
2
9
a
0
.
2
5
3
a
0
.
0
0
8
À
0
.
1
1
6
a
0
.
1
0
1
a
0
.
0
4
0
a
À
0
.
0
3
6
a
À
0
.
0
4
0
a
À
0
.
0
2
4
b
7
.
D
_
l
o
s
s
0
.
0
5
9
a
0
.
0
1
8
c
À
0
.
1
2
4
a
0
.
2
1
1
a
À
0
.
1
6
0
a
À
0
.
2
6
2
a
À
0
.
6
7
8
a
À
0
.
3
0
8
a
À
0
.
2
0
4
a
0
.
0
5
8
a
À
0
.
0
4
9
a
À
0
.
0
4
5
a
0
.
0
5
3
a
À
0
.
0
1
0
À
0
.
0
8
7
a
8
.
R
O
A
À
0
.
0
2
9
a
À
0
.
0
2
8
a
0
.
1
2
6
a
À
0
.
2
6
5
a
0
.
1
8
9
a
0
.
4
2
6
a
À
0
.
6
5
6
a
0
.
5
8
7
a
0
.
3
7
1
a
À
0
.
1
4
0
a
0
.
0
8
9
a
0
.
0
5
9
a
À
0
.
0
4
1
a
À
0
.
0
6
5
a
0
.
0
8
8
a
9
.
M
B
0
.
0
8
9
a
0
.
0
9
9
a
0
.
0
1
6
À
0
.
1
0
7
a
0
.
1
4
9
a
0
.
2
5
3
a
À
0
.
2
5
4
a
0
.
4
9
5
a
0
.
2
5
9
a
À
0
.
2
0
9
a
0
.
1
0
8
a
0
.
0
3
9
a
0
.
0
0
3
À
0
.
1
5
4
a
À
0
.
0
2
3
b
1
0
.
G
r
o
w
t
h
0
.
0
5
5
a
À
0
.
1
0
5
a
0
.
1
0
2
a
0
.
0
9
8
a
0
.
1
9
6
a
À
0
.
0
1
5
À
0
.
1
5
7
a
0
.
3
4
1
a
0
.
2
2
1
a
À
0
.
0
9
0
a
0
.
0
0
3
0
.
0
1
3
0
.
0
3
2
a
À
0
.
0
7
2
a
0
.
1
3
6
a
1
1
.
A
g
e
À
0
.
1
6
3
a
0
.
1
8
2
a
0
.
3
9
0
a
À
0
.
0
0
8
À
0
.
2
0
8
a
À
0
.
1
1
2
a
0
.
0
4
6
a
À
0
.
0
9
2
a
À
0
.
1
8
6
a
À
0
.
0
6
1
a
À
0
.
1
0
3
a
0
.
0
3
6
a
À
0
.
0
5
5
a
0
.
7
3
5
a
0
.
2
1
0
a
1
2
.
D
_
b
i
g
4
0
.
0
0
7
0
.
0
3
1
a
0
.
0
7
8
a
À
0
.
0
6
2
a
0
.
0
6
9
a
0
.
0
9
7
a
À
0
.
0
4
9
a
0
.
0
7
1
a
0
.
1
0
2
a
0
.
0
0
5
À
0
.
0
9
7
a
0
.
2
6
2
a
À
0
.
0
5
5
a
À
0
.
0
1
3
À
0
.
5
9
0
a
1
3
.
D
_
e
x
p
e
r
t
À
0
.
0
0
3
0
.
0
3
0
a
0
.
0
9
0
a
À
0
.
0
1
4
0
.
0
3
0
a
0
.
0
4
3
a
À
0
.
0
4
5
a
0
.
0
5
4
a
0
.
0
4
4
a
0
.
0
1
5
0
.
0
4
5
a
0
.
2
6
2
a
0
.
0
4
5
a
0
.
0
4
0
a
À
0
.
1
5
6
a
1
4
.
D
_
n
e
w
a
u
d
i
t
À
0
.
0
0
7
À
0
.
0
9
4
a
À
0
.
0
1
5
0
.
0
5
1
a
À
0
.
0
3
9
a
À
0
.
0
3
2
a
0
.
0
5
3
a
À
0
.
0
4
8
a
0
.
0
0
1
0
.
0
3
3
a
À
0
.
0
2
6
b
À
0
.
0
5
5
a
0
.
0
4
5
a
À
0
.
2
5
5
a
0
.
0
3
1
a
1
5
.
T
e
n
u
r
e
À
0
.
1
4
5
a
0
.
2
3
0
a
0
.
3
7
0
a
À
0
.
0
4
2
a
À
0
.
1
1
1
a
À
0
.
0
5
3
a
À
0
.
0
0
4
À
0
.
0
4
3
a
À
0
.
1
5
7
a
À
0
.
0
6
7
a
0
.
6
6
7
a
À
0
.
0
3
1
a
0
.
0
3
3
a
À
0
.
2
1
4
a
0
.
1
5
2
a
1
6
.
I
m
p
o
r
t
0
.
0
1
7
c
À
0
.
0
3
3
a
0
.
0
8
3
a
0
.
0
6
9
a
À
0
.
0
1
1
À
0
.
0
4
3
a
À
0
.
0
1
9
c
0
.
0
1
4
À
0
.
0
3
8
a
0
.
0
3
5
a
0
.
0
9
3
a
À
0
.
5
8
4
a
À
0
.
1
4
8
a
0
.
0
1
8
c
0
.
0
4
1
a
N
o
t
e
:
T
h
e
P
e
a
r
s
o
n
c
o
r
r
e
l
a
t
i
o
n
s
a
r
e
i
n
t
h
e
b
e
l
o
w
d
i
a
g
o
n
a
l
a
n
d
S
p
e
a
r
m
a
n
c
o
r
r
e
l
a
t
i
o
n
s
a
r
e
i
n
t
h
e
a
b
o
v
e
d
i
a
g
o
n
a
l
.
V
a
r
i
a
b
l
e
d
e
?
n
i
t
i
o
n
s
a
r
e
s
a
m
e
a
s
t
h
o
s
e
i
n
T
a
b
l
e
2
.
a
,
b
,
c
d
e
n
o
t
e
t
w
o
-
t
a
i
l
e
d
s
i
g
n
i
?
c
a
n
c
e
a
t
t
h
e
0
.
0
1
,
0
.
0
5
,
a
n
d
0
.
1
0
l
e
v
e
l
s
,
r
e
s
p
e
c
t
i
v
e
l
y
.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 6
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
characteristics in the two sub-periods. The mean and median of
Leverage (Lev), operating cash ?ow (CFO), total assets turnover
(Turnover), return on total assets (ROA), or sales growth (Growth)
in the more transparent information period are signi?cantly
smaller than those in the less transparent period. A higher pro-
portion of ?rms in the more transparent information period re-
ported a loss (D_loss) than those in the less transparent period.
Firms in the more transparent information period have a higher
market-to-book ratio (MB) and longer listed period (Age) than
those in the less transparent period. The mean and median of larger
auditor (D_big4), industry-specialist auditor (D_expert), and audit
tenure (Tenure) in the more transparent information period are
signi?cantly higher than those in the less transparent period.
Finally, the mean and median of auditor switch (D_newaudit) and
client importance (Import) are signi?cantly smaller for the more
transparent information period.
4.2. Multivariate analyses
This section explores the effects of mandatory disclosure re-
quirements of auditor fees on reducing the degree of discretionary
accruals, and the differences in discretionary accruals between
?rms reporting the individual amount of auditor fees and those
reporting the fee range. After controlling for multiple other vari-
ables from the regression analyses, Tables 5 and 6 show empirical
evidence of H1 and H2, respectively.
Table 5 shows that the results from the regression analysis of
discretionary accruals on the dummy variable representing
mandatory disclosure requirements, as well as the control vari-
ables. As in Chang, Fang, and Tseng (2007), this study reports three
models presenting absolute, positive, and negative discretionary
accruals for comparative purposes. The main variable of interest is
‘After’ (for observations after mandatory disclosure requirements)
in model (1). As shown in Table 5, ‘After’ is signi?cantly and
negatively related to absolute discretionary accruals
(coef. ¼ À0.0065, signi?cant at the 1% level) and positive discre-
tionary accruals (coef. ¼ À0.0135, signi?cant at the 1% level),
respectively. In addition, the coef?cient of ‘After’ is not signi?cantly
negative at the 10% level (coef. ¼ À0.0008). For these three
regressions, the adjusted R
2
shows the independent variables
explain approximately 10.18e48.44% of the variation in accrual-
based earnings management. The DurbineWatson (DW) statistic
in the three regressions ranges from 1.828 to 1.964, which is be-
tween 1.5 and 2.5, indicating that there is no autocorrelation in the
residual value. Therefore, after mandatory disclosure requirements
of auditor fees, ?rms are less more likely to manipulate accruals-
based earnings management, especially upward earnings man-
agement activity, which tends to be consistent with Dye (1991) and
Lai (2009) for better auditor independence. The evidence thus
suggests that H1 is supported.
Next, this study considers the difference in earnings manage-
ment after mandatory disclosure requirements of auditor fees if
Table 4
The mean and median tests between before and after mandatory disclosure requirements of auditor fees.
Variable Before mandatory disclosure
requirements for 2002e2008
(N ¼ 5870)
After mandatory disclosure
requirements for 2009e2012
(N ¼ 3986)
Mean diff. (t value) Median diff. (z value)
Mean Median Mean Median
jDAj 0.0659 0.0497 0.0590 0.0448 6.14
a
4.85
a
DAþ
*
0.0695 0.0521 0.0577 0.0424 7.21
a
6.33
a
DAÀ
*
À0.0620 À0.0468 À0.0603 À0.0475 À1.12 0.37
Asset 15.2096 15.0644 15.1941 15.0333 0.63 0.56
Lev 0.3925 0.3915 0.3495 0.3385 13.15
a
13.35
a
Turnover 0.8774 0.7600 0.8067 0.7000 6.42
a
7.36
a
CFO 0.0639 0.0585 0.0598 0.0549 2.39
b
2.25
b
D_loss 0.2928 0.0000 0.3101 0.0000 À1.83
c
À1.83
c
ROA 0.0460 0.0475 0.0413 0.0450 2.85
a
3.06
a
MB 1.3532 1.1237 1.5242 1.2966 À9.82
a
À12.50
a
Growth 0.0862 0.0599 0.0266 À0.0021 10.29
a
13.63
a
Age 9.5198 7.0000 12.1475 11.0000 À19.17
a
À33.00
a
D_big4 0.8392 1.0000 0.8620 1.0000 À3.14
a
À3.10
a
D_expert 0.2670 0.0000 0.2948 0.0000 À3.01
a
À3.03
a
D_newaudit 0.0792 0.0000 0.0334 0.0000 10.12
a
9.36
a
Tenure 7.4823 6.0000 9.9225 9.0000 À23.29
a
À26.57
a
Import 2.5399 0.1200 1.9824 0.0700 3.32
a
15.07
a
Note: DAþ(DAÀ) is the positive (negative) value of performance-adjusted abnormal accruals proposed by Kothari et al. (2005), and other variable de?nitions are same as those
in Table 2.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
* denotes that the observations before and after mandatory disclosure requirements for DAþ (DAÀ) are 3078 and 2013 (2792, and 1973), respectively.
Table 5
The regression model of mandatory disclosure requirements of auditor fees and
control variables on discretionary accruals during 2002e2012.
Variable jDAj DAþ DAÀ
Intercept 0.0429
a
(4.70) 0.0247
b
(2.26) À0.0364
a
(À3.44)
After À0.0065
a
(À5.63) À0.0135
a
(À10.70) À0.0008 (À0.65)
Asset 0.0005 (0.91) 0.0035
a
(5.07) 0.0022
a
(3.56)
Lev 0.0009 (0.19) 0.0019 (0.37) 0.0051 (1.09)
Turnover 0.0073
a
(5.75) À0.0023 (À1.59) À0.0083
a
(À5.45)
CFO À0.0548
a
(À4.48) À0.6523
a
(À40.80) À0.5390
a
(À39.10)
D_loss 0.0067
a
(3.91) 0.0185
a
(9.92) 0.0061
a
(3.43)
ROA À0.0289
c
(À1.85) 0.5101
a
(27.75) 0.5005
a
(33.33)
MB 0.0075
a
(8.59) À0.0007 (À0.71) À0.0109
a
(À12.21)
Growth 0.0037 (1.30) À0.0271
a
(À8.52) À0.0362
a
(À10.90)
Age À0.0005
a
(À5.35) À0.0008
a
(À6.83) À0.0001 (À0.13)
D_big4 0.0006 (0.29) À0.0001 (À0.04) À0.0001 (À0.01)
D_expert 0.0012 (0.93) 0.0003 (0.21) À0.0001 (À0.01)
D_newaudit À0.0058
a
(À2.62) À0.0081
a
(À2.84) 0.0002 (0.10)
Tenure À0.0004
a
(À2.89) À0.0004
a
(À2.78) 0.0002 (1.54)
Import 0.0002
b
(1.96) 0.0001 (0.94) À0.0001 (À0.65)
Industry dummies included
N 9856 5091 4765
F value 40.01
a
82.63
a
75.23
a
Adj R
2
0.1018 0.4813 0.4844
Note: Variable de?nitions are same as those in Tables 2 and 4. The t-statistic (in
parentheses) is based on the two-way robust standard errors clustered at the ?rm
and year levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 7
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
such fees are disclosed by individual amount rather than by range.
According to Dye (1991), the disclosure of audit fees would induce
future quasi-rents to be zero and improve auditor independence.
The disclosure of audit fees in the US in Dye's (1991) study is carried
out in the form of individual amount (i.e., more transparent
disclosure). In contrast, the mandatory disclosure requirements of
auditor fees in Taiwan allowtwo alternative types of disclosures, by
individual amount and fee range, with the former being clearer and
more precise. The main variable of interest is D_amount (for ?rms
reporting the individual amount of auditor fees after mandatory
disclosure requirements) in model (3). Table 6 does not show any
signi?cant relationship between D_amount and absolute discre-
tionary accruals (jDAj). After dividing the sample into two sub-
samples for positive and negative discretionary accruals (DAþ and
DAÀ), Table 6 shows that D_amount is signi?cantly and negatively
related to positive discretionary accruals (coef. ¼ À0.0028, signi?-
cant at the 10% level) and negative discretionary accruals
(coef. ¼ À0.0034, signi?cant at the 5% level), respectively. There-
fore, after mandatory disclosure requirements, ?rms that reveal the
amount of auditor fees have lower upward and downward earnings
management activity than those that disclose the fee range, offer-
ing support for H2. In addition, the adjusted R
2
shows the inde-
pendent variables explain approximately 10.28e63.95% of the
variation in accrual-based earnings management. The DW statistic
in the three regressions ranges from 1.788 to 1.945, which is be-
tween 1.5 and 2.5, indicating that there is no autocorrelation in the
residual value.
4.3. Additional tests
To enhance the internal validity, this study performs additional
analyses to test two hypotheses, by using year-to-year dummy
variables after mandatory disclosure requirements, the restrained
sample for auditor fees disclosure, controlling for current accruals
and prior discretionary accruals, self-selection of auditor, employ-
ing the truncated regression, modi?ed Jones model, examining real
earnings management, considering changes in disclosure types,
and carrying out other sensitivity analyses.
4.3.1. Year-to-year dummy variables after mandatory disclosure
requirements
This study switches the main variable in model (1) from a single
dummy variable (i.e., After) to year-to-year dummy variables (i.e.,
Y2009eY2012), and thus explores whether the decrease in the
accrual-based earnings management by ?rms is different for each
year after mandatory disclosure requirements of auditor fees.
Table 7 shows that the coef?cients of the year-to-year dummy
variables in the model of absolute discretionary accruals are
À0.0046, À0.0074, À0.004, À0.01, which all are signi?cant at the 5%
or 1% levels. Furthermore, the coef?cients of Y2009eY2012 in the
model of positive discretionary accruals are also all negative, and
most are signi?cant at the 1% level, indicating that upward earnings
management activity decreases signi?cantly, especially since the
second year after mandatory disclosure requirements. On the other
hand, the coef?cient signs of Y2009eY2012 are signi?cantly
different for negative discretionary accruals as the dependent var-
iable, indicating that the effect of mandatory disclosure re-
quirements on downward earnings management may be
inconsistent for eachyear. These results are consistent with those in
Table 5.
4.3.2. The restrained sample for auditor fees disclosure
The total sample includes observations without disclosure of
audit or non-audit fees, regardless of whether they are after
mandatory disclosure requirements of auditor fees. This study ex-
tracts observations with auditor fees disclosure from the total
sample, and excludes those without it. Both stricter and looser
criteria are applied. Under the stricter criteria for auditor fees
disclosure the sample includes only ?rms that disclose their audit
and non-audit fees, regardless of whether by individual amount or
Table 6
The regression model of the disclosure type of auditor fees and control variables on
discretionary accruals during 2009e2012.
Variable jDAj DAþ DAÀ
Intercept 0.0689
a
(5.03) 0.0802
a
(5.37) À0.0459
a
(À2.62)
D_amount 0.0021 (1.29) À0.0028
c
(À1.83) À0.0034
b
(À2.15)
Asset À0.0016
c
(À1.82) 0.0002 (0.21) 0.0038
a
(4.54)
Lev 0.0020 (0.30) 0.0179
a
(2.85) 0.0141
b
(2.10)
Turnover 0.0061
a
(3.05) À0.0025 (À1.27) À0.0105
a
(À5.21)
CFO À0.0296 (À1.45) À0.7283
a
(À28.95) À0.5999
a
(À27.84)
D_loss 0.0059
b
(2.19) 0.0161
a
(6.96) 0.0045
c
(1.88)
ROA À0.0539
b
(À2.10) 0.6199
a
(23.00) 0.5777
a
(27.27)
MB 0.0077
a
(5.65) 0.0017 (1.35) À0.0087
a
(À7.13)
Growth 0.0112
b
(2.55) À0.0018 (À0.40) À0.0187
a
(À4.39)
Age À0.0009
a
(À4.77) À0.0005
a
(À2.69) 0.0002 (0.85)
D_big4 0.0018 (0.57) À0.0025 (À0.79) 0.0018 (0.54)
D_expert 0.0024 (1.35) À0.0006 (À0.37) À0.0013 (À0.75)
D_newaudit 0.0036 (0.79) À0.0066 (À1.12) À0.0089 (À1.54)
Tenure 0.0003
c
(1.70) À0.0001 (À0.62) À0.0001 (À0.42)
Import 0.0003
c
(1.75) 0.0001 (1.07) À0.0001 (À0.09)
Industry and year dummies included
N 3880 1960 1920
F value 24.02
a
39.35
a
42.45
a
Adj R
2
0.1028 0.6395 0.5816
Note: D_amount equals one if the ?rmdiscloses audit and non-audit fees in the form
of individual amount, and zero if the ?rm discloses audit and non-audit fees in the
form of full or partial fee range. The other variable de?nitions are same as those in
Tables 2 and 4. The t-statistic (in parentheses) is based on the two-way robust
standard errors clustered at the ?rm and year levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
Table 7
The regression model of mandatory disclosure requirements of auditor fees and
control variables on discretionary accruals during 2002e2012: year-to-year dummy
variables after mandatory disclosure requirements.
Variable jDAj DAþ DAÀ
Intercept 0.0434
a
(4.76) 0.0267
b
(2.45) À0.0323
a
(À3.04)
Y2009 À0.0046
b
(À2.29) À0.0012 (À0.59) 0.0102
a
(5.14)
Y2010 À0.0074
a
(À3.78) À0.0235
a
(À11.68) À0.0105
a
(À5.04)
Y2011 À0.0040
b
(À2.27) À0.0167
a
(À9.70) À0.0087
a
(À4.66)
Y2012 À0.0100
a
(À6.02) À0.0113
a
(À6.68) 0.0066
a
(4.12)
Asset 0.0005 (0.83) 0.0034
a
(4.96) 0.0020
a
(3.30)
Lev 0.0008 (0.19) 0.0016 (0.32) 0.0039 (0.84)
Turnover 0.0073
a
(5.75) À0.0021 (À1.46) À0.0083
a
(À5.44)
CFO À0.0550
a
(À4.46) À0.6603
a
(À40.90) À0.5508
a
(À39.88)
D_loss 0.0067
a
(3.89) 0.0182
a
(9.83) 0.0057
a
(3.25)
ROA À0.0295
c
(À1.88) 0.5180
a
(27.95) 0.5048
a
(34.01)
MB 0.0075
a
(8.34) À0.0015 (À1.53) À0.0116
a
(À12.92)
Growth 0.0043 (1.44) À0.0224
a
(À6.75) À0.0310
a
(À8.90)
Age À0.0005
a
(À5.28) À0.0007
a
(À6.67) 0.0001 (0.14)
D_big4 0.0006 (0.31) À0.0001 (À0.02) À0.0001 (À0.02)
D_expert 0.0012 (0.96) 0.0003 (0.23) 0.0002 (0.13)
D_newaudit À0.0056
b
(À2.53) À0.0085
a
(À2.98) À0.0002 (À0.08)
Tenure À0.0004
a
(À2.76) À0.0004
a
(À2.64) 0.0002 (1.52)
Import 0.0002
c
(1.95) 0.0001 (0.89) À0.0001 (À0.60)
Industry dummies included
N 9856 5091 4765
F value 32.19
a
135.48
a
130.76
a
Adj R
2
0.1023 0.4875 0.4951
Note: Y2009eY2012 is the dummy variable for 2009, 2010, 2011, and 2012,
respectively. The other variable de?nitions are same as those in Tables 2 and 4. The
t-statistic (in parentheses) is based on the two-way robust standard errors clustered
at the ?rm and year levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 8
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
range. Under the looser criteria, in addition to the observations
from the stricter criteria, the sample also includes those with only
disclosure of audit fees under the second or third speci?c
requirement conditions (i.e., audit fees decrease with audit ?rm
switch, or audit fees decrease substantially). This study summarizes
the results in Table 8. As shown in Table 8, the coef?cient of
mandatory disclosure requirements of auditor fees (i.e., After) are
still signi?cantly and negatively related to absolute and positive
discretionary accruals (i.e., jDAj and DAþ), regardless of whether
the sample is formed under the stricter criteria or not, while the
coef?cients of ‘After’ are still not signi?cant for negative discre-
tionary accruals as the dependent variable, consistent with Table 5.
4.3.3. Controlling for current accruals and prior discretionary
accruals
Becker et al. (1998), Francis et al. (1999) and Frankel et al. (2002)
indicated that ?rms with larger total accruals have relatively large
discretionary accruals. This study thus includes total current ac-
cruals as an additional control variable (Lee & Chen, 2004; Wang,
Shiue, & Tseng, 2011). In addition, Barton and Simko (2002) sug-
gested that accrual-based earnings management has a reversal
characteristic, and thus this study also includes a variable for lagged
discretionaryaccruals tocapture this (Kim, Chung, &Firth, 2003; Lee
et al., 2003; Liao &Hung, 2010; Tseng et al., 2012; Wang et al., 2011;
Young & Wu, 2003). Table 9 shows that total current accruals (i.e.,
TAC) and prior discretionary accruals (i.e., LagDA) are signi?cantly
associated with discretionary accruals in models (1) and (3). After
controlling for TAC and LagDA, the coef?cients of ‘After’ in model (1)
are still signi?cant and negative for jDAj and DAþ as the dependent
variables, consistent with Table 5. In addition, the coef?cients of
D_amount inmodel (3) are still signi?cant andnegative for DAþand
DAÀ as the dependent variables, consistent with Table 6.
4.3.4. Heckman's correction procedure for self-selection of auditors
The prior literature indicated that auditor selection is
endogenously determined, and suggested a correction for the
self-selection of auditor bias (Chaney, Jeter, & Shivakumar, 2004;
Liao et al., 2012; Shiue et al., 2008). This study employs Heck-
man's two-step correction procedure (Heckman, 1979) to control
this, and reexamines the two hypotheses. In the ?rst step of
Heckman's method, this study estimates a probit model used by
Liao et al. (2012), in which the dependent variable D_big4 is
whether the ?rm is audited by a Big 4 auditor or not, and the
independent variables includes the natural logarithm of total
assets, return on total assets, leverage, current ratio, reported
loss, total assets turnover, listed period, ratio of cash ?ows from
operation to lagged total assets, year and industry dummy vari-
ables. The ?rst-step estimator of Heckman's method is used to
determine the inverse Mill's ratio (i.e., Invmills) as the condi-
tional expectation of the error term of the regression model. The
second step involves the OLS estimation of the regression model
similar to this study's models (1) and (3). As shown in Table 10,
the relationships between ‘After’ and discretionary accruals (i.e.,
jDAj and DAþ) are still signi?cantly negative, and D_amount is
still signi?cantly and negatively associated with discretionary
accruals (i.e., DAþ and DAÀ). The results are consistent with
Tables 5 and 6.
4.3.5. Truncated regression
Myers et al. (2003) suggested that OLS estimates of a truncated
sample are generally biased towards zero. This study re-estimates
models (1) and (3) using the signed value of accruals as the
dependent variable for the subsample of client ?rms with either
positive (DAþ) or negative (DAÀ) discretionary accruals.
12
Because
both subsamples are truncated, this study ?ts maximum likelihood
truncated regression models for these two subsamples following
prior research (Boone et al., 2010; Liao et al., 2012, 2013; Myers
et al., 2003). The results in Table 11 are still consistent with
Tables 5 and 6.
Table 8
The regression model of mandatory disclosure requirements of auditor fees and control variables on discretionary accruals during 2002e2012: the restrained sample for
auditor fees disclosure.
Variable Under the stricter criteria Under the looser criteria
jDAj DAþ DAÀ jDAj DAþ DAÀ
Intercept 0.0634
a
(5.17) 0.0606
a
(4.29) À0.0620
a
(À4.39) 0.0616
a
(5.05) 0.0604
a
(4.27) À0.0587
a
(À4.17)
After À0.0089
a
(À4.70) À0.0120
a
(À6.29) 0.0031 (1.50) À0.0087
a
(À4.68) À0.0120
a
(À6.35) 0.0025 (1.22)
Asset À0.0010 (À1.28) 0.0008 (0.99) 0.0035
a
(4.42) À0.0009 (À1.18) 0.0008 (0.99) 0.0034
a
(4.31)
Lev 0.0030 (0.49) 0.0181
a
(2.82) 0.0143
b
(2.23) 0.0030 (0.49) 0.0177
a
(2.79) 0.0133
b
(2.09)
Turnover 0.0057
a
(3.26) À0.0035
c
(À1.89) À0.0096
a
(À4.53) 0.0058
a
(3.33) À0.0034
c
(À1.81) À0.0097
a
(À4.57)
CFO À0.0373
b
(À2.18) À0.6946
a
(À34.98) À0.5476
a
(À28.84) À0.0384
b
(À2.25) À0.6974
a
(À35.10) À0.5449
a
(À28.76)
D_loss 0.0074
a
(3.14) 0.0154
a
(7.06) 0.0025 (1.12) 0.0071
a
(3.02) 0.0153
a
(7.01) 0.0029 (1.26)
ROA À0.0303 (À1.40) 0.5814
a
(25.32) 0.5225
a
(25.34) À0.0333 (À1.55) 0.5795
a
(25.21) 0.5196
a
(24.97)
MB 0.0081
a
(7.15) 0.0013 (1.17) À0.0093
a
(À8.59) 0.0083
a
(7.35) 0.0015 (1.33) À0.0093
a
(À8.52)
Growth 0.0049 (1.36) À0.0241
a
(À6.10) À0.0366
a
(À9.09) 0.0044 (1.21) À0.0247
a
(À6.23) À0.0357
a
(À8.87)
Age À0.0007
a
(À4.74) À0.0004
b
(À2.37) 0.0002 (0.99) À0.0008
a
(À4.80) À0.0004
b
(À2.38) 0.0002 (1.10)
D_big4 À0.0017 (À0.62) À0.0015 (À0.54) 0.0042 (1.40) À0.0011 (À0.40) À0.0013 (À0.45) 0.0024 (0.79)
D_expert 0.0014 (0.86) À0.0018 (À1.12) À0.0036
b
(À2.13) 0.0013 (0.80) À0.0019 (À1.19) À0.0035
b
(À2.07)
D_newaudit 0.0010 (0.25) À0.0025 (À0.50) À0.0048 (À1.10) 0.0016 (0.41) À0.0017 (À0.35) À0.0041 (À0.94)
Tenure 0.0001 (0.25) À0.0003
c
(À1.80) À0.0001 (À0.07) 0.0001 (0.26) À0.0003
c
(À1.76) À0.0001 (À0.08)
Import 0.0002 (1.64) 0.0001 (0.72) 0.0001 (0.15) 0.0003
c
(1.93) 0.0001 (0.76) À0.0001 (À0.40)
Industry and year dummies included
N 5116 2608 2508 5148 2620 2528
F value 18.66
a
107.87
a
80.97
a
18.94
a
108.47
a
80.51
a
Adj R
2
0.1023 0.5750 0.5128 0.1031 0.5752 0.5094
Note: Under the stricter criteria for auditor fees disclosure, the sample includes only ?rms that disclose audit and non-audit fees. Under the looser criteria, the sample also
includes those that only disclose audit fees under the second or third speci?c requirement condition (i.e., audit fees decrease with audit ?rm switch, or audit fees decrease
substantially). Variable de?nitions are same as those in Tables 2 and 4. The t-statistic (in parentheses) is based on the two-way robust standard errors clustered at the ?rmand
year levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
12
In other words, the model is estimated using a truncated regression approach
since only observations with a positive (negative) value of the dependent variable
DA are included.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 9
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
4.3.6. Modi?ed Jones model for estimation of discretionary accruals
This study also estimates the discretionary accruals using the
cross-sectional modi?ed Jones model proposed by Dechow et al.
(1995) as a robust test (Lee & Chen, 2004; Lee et al., 2003; Wang
et al., 2011; Young & Wu, 2003). The results in Table 12 are
consistent with those in Tables 5 and 6.
4.3.7. Real earnings management as a proxy for earnings
management
Following Schipper (1989), earnings management can be clas-
si?ed into two categories: accrual-based earnings management and
Table 9
The regression model of mandatory disclosure requirements of auditor fees, the disclosure type and control variables on discretionary accruals: controlling for current accruals
and prior discretionary accruals.
Variable During 2002e2012 During 2009e2012
jDAj DAþ DAÀ jDAj DAþ DAÀ
Intercept 0.0365
a
(3.91) 0.0545
a
(5.05) À0.0390
a
(À3.67) 0.0497
a
(3.66) 0.0648
a
(4.14) À0.0352
c
(À1.93)
After À0.0035
a
(À3.01) À0.0077
a
(À6.61) À0.0006 (À0.46)
D_amount 0.0027
c
(1.70) À0.0025
c
(À1.76) À0.0028
c
(À1.82)
Asset À0.0001 (À0.05) 0.0003 (0.53) 0.0025
a
(4.10) À0.0009 (À1.01) 0.0005 (0.62) 0.0035
a
(4.25)
Lev 0.0159
a
(3.41) 0.0187
a
(4.16) 0.0172
a
(3.54) 0.0137
b
(2.03) 0.0248
a
(4.14) 0.0201
a
(3.16)
Turnover 0.0028
b
(2.06) À0.0033
b
(À2.34) À0.0085
a
(À5.78) 0.0020 (0.99) À0.0067
a
(À3.55) À0.0108
a
(À5.24)
CFO 0.8373
a
(11.91) 0.0885 (0.94) À0.0932 (À1.51) 1.0220
a
(7.57) 0.1033 (0.67) À0.1913
c
(À1.86)
D_loss 0.0053
a
(3.12) 0.0162
a
(9.84) 0.0064
a
(3.69) 0.0056
b
(2.22) 0.0160
a
(7.20) 0.0052
b
(2.29)
ROA À0.8505
a
(À12.46) À0.1640
c
(À1.80) 0.0790 (1.29) À1.0328
a
(À7.92) À0.1559 (À1.05) 0.1882
c
(1.80)
MB 0.0070
a
(7.37) À0.0011 (À1.16) À0.0085
a
(À9.49) 0.0070
a
(4.85) 0.0002 (0.15) À0.0071
a
(À6.24)
Growth 0.0074
b
(2.37) À0.0364
a
(À10.69) À0.0339
a
(À9.70) 0.0133
a
(3.09) À0.0049 (À1.09) À0.0116
b
(À2.34)
Age À0.0003
a
(À3.46) À0.0002
b
(À2.05) À0.0001 (À0.33) À0.0007
a
(À4.09) À0.0005
b
(À2.55) 0.0002 (0.80)
D_big4 0.0010 (0.49) 0.0001 (0.08) À0.0011 (À0.51) 0.0030 (0.98) À0.0012 (À0.35) À0.0007 (À0.22)
D_expert 0.0004 (0.33) 0.0005 (0.43) 0.0002 (0.13) 0.0018 (1.01) 0.0001 (0.00) À0.0012 (À0.72)
D_newaudit À0.0060
a
(À2.72) À0.0067
a
(À2.90) 0.0002 (0.09) 0.0080 (1.62) À0.0027 (À0.63) À0.0044 (À0.93)
Tenure À0.0003
a
(À2.65) À0.0001 (À0.90) 0.0002 (1.53) 0.0002 (0.86) À0.0002 (À1.18) À0.0001 (À0.39)
Import 0.0001 (0.90) 0.0001 (0.46) 0.0001 (0.77) 0.0002 (0.86) À0.0001 (À0.20) 0.0001 (0.71)
LagDA 0.1034
a
(8.75) À0.0233
a
(À3.04) À0.0274
a
(À3.58) 0.1220
a
(6.10) À0.0113 (À0.90) À0.0411
a
(À3.30)
TAC 0.8320
a
(11.78) 0.7265
a
(8.35) 0.5192
a
(7.65) 1.0034
a
(7.31) 0.7772
a
(5.26) 0.4695
a
(4.05)
Industry and year dummies included
N 8041 4066 3975 3438 1730 1708
F value 44.39
a
183.26
a
142.78
a
20.44
a
99.97
a
75.79
a
Adj R
2
0.1589 0.6108 0.5553 0.1769 0.6851 0.6247
Note: LagDA is the prior discretionary accruals. TAC is total current accruals. Variable de?nitions are same as those in Tables 2, 4, and 6. The t-statistic (in parentheses) is based
on the two-way robust standard errors clustered at the ?rm and year levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
Table 10
The regression model of mandatory disclosure requirements of auditor fees, the disclosure type and control variables on discretionary accruals: Heckman's correction pro-
cedure for self-selection of auditors.
Variable During 2002e2012 During 2009e2012
jDAj DAþ DAÀ jDAj DAþ DAÀ
Intercept 0.0414
b
(2.13) 0.0612
a
(2.83) À0.0240 (À1.46) 0.0687
b
(2.23) 0.0619
b
(2.05) À0.0692
a
(À2.72)
After À0.0065
a
(À4.84) À0.0147
a
(À10.68) À0.0014 (À1.06)
D_amount 0.0021 (1.30) À0.0028
c
(À1.83) À0.0034
b
(À2.17)
Asset 0.0006 (0.54) 0.0016 (1.29) 0.0014 (1.34) À0.0016 (À0.82) 0.0013 (0.71) 0.0052
a
(3.66)
Lev 0.0006 (0.12) 0.0047 (0.90) 0.0082 (1.35) 0.0020 (0.27) 0.0164
b
(2.52) 0.0099 (1.28)
Turnover 0.0074
a
(4.31) À0.0043
b
(À2.48) À0.0094
a
(À4.53) 0.0062
a
(2.90) À0.0025 (À1.28) À0.0093
a
(À3.90)
CFO À0.0544
a
(À4.04) À0.6652
a
(À38.48) À0.5426
a
(À39.75) À0.0295 (À1.36) À0.7271
a
(À29.14) À0.5897
a
(À27.33)
D_loss 0.0068
a
(3.94) 0.0178
a
(9.52) 0.0063
a
(3.56) 0.0059
b
(2.17) 0.0163
a
(7.15) 0.0038 (1.63)
ROA À0.0288
c
(À1.85) 0.5058
a
(27.56) 0.5008
a
(33.53) À0.0539
b
(À2.12) 0.6244
a
(22.82) 0.5771
a
(27.62)
MB 0.0075
a
(8.52) À0.0004 (À0.44) À0.0108
a
(À12.15) 0.0077
a
(5.64) 0.0017 (1.33) À0.0088
a
(À7.35)
Growth 0.0037 (1.30) À0.0271
a
(À8.55) À0.0362
a
(À10.97) 0.0112
b
(2.56) À0.0018 (À0.40) À0.0188
a
(À4.43)
Age À0.0005
a
(À3.46) À0.0005
a
(À3.41) 0.0001 (0.53) À0.0009
a
(À2.67) À0.0007
b
(À2.18) À0.0001 (À0.08)
D_big4 0.0006 (0.29) À0.0005 (À0.22) À0.0002 (À0.09) 0.0018 (0.58) À0.0023 (À0.73) 0.0022 (0.67)
D_expert 0.0012 (0.94) 0.0003 (0.25) 0.0001 (0.00) 0.0024 (1.36) À0.0007 (À0.44) À0.0014 (À0.82)
D_newaudit À0.0058
a
(À2.62) À0.0086
a
(À3.00) 0.0002 (0.07) 0.0036 (0.79) À0.0065 (À1.10) À0.0089 (À1.53)
Tenure À0.0004
a
(À2.82) À0.0005
a
(À3.20) 0.0002 (1.30) 0.0003
c
(1.69) À0.0001 (À0.40) À0.0001 (À0.20)
Import 0.0002
c
(1.96) 0.0001 (0.79) À0.0001 (À0.71) 0.0003
c
(1.75) 0.0002 (1.14) À0.0001 (À0.00)
Invmills 0.0015
a
(0.09) À0.0329
b
(À2.02) À0.0134 (À0.92) 0.0002 (0.01) 0.0133 (0.64) 0.0231 (1.15)
Industry and year dummies included
N 9856 5091 4765 3880 1960 1920
F value 33.83
a
140.05
a
132.64
a
12.98
a
94.89
a
73.15
a
Adj R
2
0.1017 0.4815 0.4844 0.1026 0.6394 0.5818
Note: Invmills is Inverse Mills ratio from the Probit model of Heckman's (1979) method, and other variable de?nitions are same as those in Tables 2, 4, and 6. The t-statistic (in
parentheses) is based on the two-way robust standard errors clustered at the ?rm and year levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 10
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
real earnings management (REM).
13
Accrual-based earnings man-
agement refers to the manipulation of earnings through the
exploitation of the accounting discretion available in GAAP,
whereas real earnings management involves management's at-
tempts to change reported earnings by undertaking actions that
deviate from the ?rst best decisions (i.e., making suboptimal de-
cisions) on the timing and scale of the underlying business activ-
ities (Taylor and Xu, 2010). In other words, accrual-based earnings
management involves with-GAAP choices. In contrast, REM in-
volves real activities manipulation (Chi et al., 2011; Cohen, Dey, &
Lys, 2008; Cohen & Zarowin, 2010; Gupta, Pevzner, &
Seethamraju, 2010; Hsu et al., 2013; Roychowdhury, 2006; Tseng
et al., 2012; Young et al., 2012; Zang, 2012). Roychowdhury
(2006) argued that managers may manage earnings by sales
manipulation, overproduction, and reduction of discretionary ex-
penditures. This study thus also uses REM as a proxy for earnings
quality for robustness, as in Tseng et al. (2012), and captures the
degree of real activities manipulation by managers from abnormal
cash ?ows (Abn_CFO), abnormal production costs (Abn_Prod), and
abnormal discretionary expenses (Abn_discexp) following prior
Table 11
The regression model of mandatory disclosure requirements of auditor fees, the disclosure type and control variables on discretionary accruals: truncated regression.
Variable During 2002e2012 During 2009e2012
DA+ DAÀ DA+ DAÀ
Intercept À0.0731
a
(À3.35) 0.0466
b
(2.51) 0.0899
a
(4.55) 0.0025 (0.11)
After À0.0265
a
(À9.59) À0.0033 (À1.41)
D_amount À0.0051
b
(À2.23) À0.0055
b
(À2.12)
Asset 0.0079
a
(6.06) 0.0027
a
(2.64) 0.0002 (0.15) 0.0052
a
(4.11)
Lev 0.0041 (0.44) 0.0066 (0.84) 0.0342
a
(3.85) 0.0205
b
(2.17)
Turnover À0.0104
a
(À4.32) À0.0134
a
(À5.82) À0.0072
a
(À3.18) À0.0154
a
(À5.49)
CFO À1.1950
a
(À42.94) À0.8709
a
(À41.52) À1.0560
a
(À42.69) À0.8448
a
(À33.13)
D_loss 0.0338
a
(9.75) 0.0093
a
(2.83) 0.0204
a
(6.50) 0.0041 (1.06)
ROA 0.9568
a
(30.21) 0.7826
a
(34.26) 0.9109
a
(31.51) 0.7878
a
(28.33)
MB À0.0049
a
(À2.81) À0.0175
a
(À12.47) 0.0005 (0.29) À0.0118
a
(À7.07)
Growth À0.0451
a
(À10.18) À0.0572
a
(À13.79) 0.0021 (0.52) À0.0255
a
(À4.77)
Age À0.0015
a
(À5.14) 0.0004 (1.54) À0.0009
a
(À2.79) 0.0007
c
(1.80)
D_big4 À0.0021 (À0.47) À0.0012 (À0.29) À0.0066 (À1.55) 0.0033 (0.66)
D_expert 0.0019 (0.69) 0.0006 (0.26) À0.0005 (À0.21) À0.0024 (À0.83)
D_newaudit À0.0144
a
(À2.61) 0.0028 (0.61) À0.0117 (À1.63) À0.0140
b
(À2.06)
Tenure À0.0006
c
(À1.66) 0.0006
c
(1.75) 0.0001 (0.21) À0.0001 (À0.24)
Import 0.0001 (0.29) À0.0001 (À0.41) 0.0001 (0.28) 0.0001 (0.43)
Industry and year dummies included
N 5091 4765 1960 1920
Wald value 2149.5
a
2190.1
a
2173.8
a
1482.0
a
Note: Variable de?nitions are same as those in Tables 2, 4, and 6. The t-statistic (in parentheses) is based on the two-way robust standard errors clustered at the ?rm and year
levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
Table 12
The regression model of mandatory disclosure requirements of auditor fees, the disclosure type and control variables on discretionary accruals: modi?ed Jones model for
estimation of DA.
Variable During 2002e2012 During 2009e2012
jDAj DAþ DAÀ jDAj DAþ DAÀ
Intercept 0.0450
a
(4.78) 0.0107 (0.98) À0.0451
a
(À4.25) 0.0525
a
(3.72) 0.0463
a
(3.51) À0.0639
a
(À3.80)
After À0.0074
a
(À6.19) À0.0135
a
(À11.13) À0.0003 (À0.30)
D_amount 0.0029
c
(1.73) À0.0031
b
(À2.17) À0.0028
c
(À1.79)
Asset 0.0006 (1.05) 0.0044
a
(6.32) 0.0028
a
(4.75) À0.0008 (À0.91) 0.0008 (1.01) 0.0037
a
(4.64)
Lev 0.0018 (0.40) À0.0016 (À0.31) 0.0051 (À1.15) 0.0021 (0.31) 0.0165
b
(2.46) 0.0090 (1.48)
Turnover 0.0059
a
(4.65) À0.0045
a
(À3.03) À0.0085
a
(À5.69) 0.0047
b
(2.36) À0.0055
a
(À2.80) À0.0102
a
(À4.91)
CFO À0.0666
a
(À5.29) À0.6929
a
(À44.72) À0.5553
a
(À42.50) À0.0380
c
(À1.88) À0.7406
a
(À34.76) À0.5908
a
(À31.31)
D_loss 0.0037
b
(2.18) 0.0095
a
(5.02) 0.0039
b
(2.43) 0.0042 (1.59) 0.0056
b
(2.40) 0.0048
b
(2.23)
ROA À0.0803
a
(À4.85) 0.6137
a
(33.76) 0.6100
a
(42.96) À0.0987
a
(À3.73) 0.6949
a
(28.75) 0.6953
a
(36.93)
MB 0.0099
a
(10.87) 0.0007 (0.80) À0.0080
a
(À9.34) 0.0097
a
(6.97) 0.0050
a
(4.19) À0.0057
a
(À5.12)
Growth 0.0062
b
(2.12) À0.0370
a
(-11.37) À0.0503
a
(À15.65) 0.0103
b
(2.51) À0.0222
a
(À5.03) À0.0445
a
(À11.14)
Age À0.0006
a
(À5.74) À0.0010
a
(À8.85) À0.0001 (À1.24) À0.0009
a
(À4.73) À0.0008
a
(À4.10) À0.0001 (À0.37)
D_big4 0.0017 (0.84) 0.0007 (0.31) 0.0001 (0.07) 0.0029 (0.92) À0.0014 (À0.45) 0.0010 (0.32)
D_expert 0.0011 (0.87) 0.0003 (0.22) 0.0002 (0.18) 0.0016 (0.89) À0.0005 (À0.30) À0.0009 (À0.52)
D_newaudit À0.0045
b
(À2.07) À0.0093
a
(À3.38) À0.0008 (À0.35) 0.0016 (0.37) À0.0069 (À1.43) À0.0021 (À0.48)
Tenure À0.0003
b
(À2.28) À0.0004
a
(À2.64) 0.0001 (0.77) 0.0003 (1.39) À0.0002 (À1.05) À0.0002 (À0.75)
Import 0.0003
a
(2.95) 0.0001 (0.95) À0.0001 (À0.83) 0.0004b (2.46) 0.0002 (1.29) À0.0001 (À0.94)
Industry and year dummies included
N 9856 5106 4750 3880 1943 1937
F value 39.08
a
172.90
a
176.65
a
15.07
a
106.72
a
94.88
a
Adj R
2
0.1131 0.5263 0.5497 0.1127 0.6558 0.6292
Note: jDAj, DAþ, and DAÀ is the absolute, positive, and negative value of abnormal accruals using the cross-sectional modi?ed Jones model proposed by Dechow et al. (1995),
respectively. The other variable de?nitions are same as those in Tables 2 and 6. The t-statistic (in parentheses) is based on the two-way robust standard errors clustered at the
?rm and year levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
13
Schipper (1989) was one of the ?rst to include real earnings management in the
de?nition of earnings management.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 11
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
research (Roychowdhury, 2006; Tseng et al., 2012; Young et al.,
2012). Firstly, the cross-sectional regression of cash ?ows from
operations on sales and change in sales through model (4) is esti-
mated for each year and industry, and the residual term represents
Abn_CFO.
CFO
it
TA
itÀ1
¼ d
0
1
TA
itÀ1
þ d
1
REV
it
TA
itÀ1
þ d
2
DREV
it
TA
itÀ1
þ c
it
(4)
where CFO ¼ cash ?ows from operation; TA ¼ total assets in the
beginning of year; REV ¼ total sales; DREV ¼ the change in sales.
Secondly, the cross-sectional regression of the sumof the cost of
goods sold and change in inventory on sales, change in sales, and
lagged change in sales through model (5) is estimated for each year
and industry, and the residual term represents Abn_Prod.
Prod
it
TA
itÀ1
¼ f
0
1
TA
itÀ1
þ f
1
REV
it
TA
itÀ1
þ f
2
DREV
it
TA
itÀ1
þ f
3
DREV
itÀ1
TA
itÀ1
þ4
it
(5)
where Prod ¼ the sum of the cost of goods sold and change in
inventory.
Thirdly, the cross-sectional regression of the sum of advertising
expenses, R&D expenses, and selling and administrative expenses
on lagged sales through model (6) is estimated for each year and
industry, and the residual term represents Abn_Discexp.
Discexp
it
TA
itÀ1
¼ h
0
1
TA
itÀ1
þ h
1
REV
itÀ1
TA
itÀ1
þ K
it
(6)
where Discexp ¼ the sum of advertising expenses, R&D expenses,
and selling and administrative expenses.
Generally, abnormally low cash ?ows from operation indicate
sales manipulation. Abnormally high production costs suggest a
higher degree of overproduction. Abnormally low discretionary
expenditures imply a reduction of discretionary expenditures
(Young et al., 2012). Thus, this study de?nes a composite real
earnings management index (REM) by the sum of standardized
Abn_CFO, Abn_Prod, and Abn_Dixcexp as follows (Hsu et al., 2013;
Tseng et al., 2012). A higher value of REM suggests a higher degree
of real activities manipulation.
REM
it
¼ ÀstdðAbn_CFOÞ þstdðAbn_ProdÞ ÀstdðAbn_DiscexpÞ
(7)
Table 13 shows the effect of mandatory disclosure requirements
and disclosure types on real earnings management. As shown in
Table 13, the coef?cient of mandatory disclosure requirements (i.e.,
After) is signi?cantly and negatively related to real earnings man-
agement (i.e., REM). Moreover, although ?rms reporting the indi-
vidual amount of auditor fees have lower real earnings
management than those reporting the fee range, there is no sig-
ni?cant difference between two subsamples.
4.3.8. Changes in disclosure types of auditor fees
Wang et al. (2011) indicated that the changes in the disclosure
rankings are negatively associated with changes in discretionary
accruals, especially for the ?rms with ranking upgrading that tend
to engage less in earnings management. This study further analyses
whether the changes in disclosure type of auditor fees are related to
the changes in audit quality by examining the effect of upgrading
disclosure type (i.e., from fee range to individual amount) and
downgrading disclosure type (i.e., from individual amount to fee
range) on changes in accrual-based and real earnings management
after mandatory disclosure requirements. As shown in Table 14, the
disclosure type upgrading of auditor fees (i.e., D_type_up) is
signi?cantly and negatively associated with both changes in
discretionary accruals and real earnings management, implying
that the changes in disclosure types of auditor fees could also affect
audit quality.
4.3.9. Other sensitivity analyses
This study also executes a set of sensitivity analyses for the
robustness of our ?ndings. Firstly, the prior research showed that
better board structure is useful to restrain earnings management
(Chang et al., 2007; Hsu et al., 2013; Lee &Liao, 2004; Lin, Kuo, &Su,
2008; Wang & Chang, 2012; Young et al., 2012). This study includes
board size, board independence, chairman internalization, the
duality of the CEO and chairman, ownership of directors and su-
pervisors, share collateralization as additional control variables.
Secondly, the prior studies indicated how managers manipulate
earnings to meet earnings thresholds. This study employs cash
?ows from operation as a proxy for income before discretionary
accruals, and includes earnings thresholds, such as zero earnings
(avoiding loss) and prior year reported income (avoiding earnings
decline) used by Healy (1985) and Young and Wu (2003), as addi-
tional control variables. On the other hand, this study also employs
earnings levels around zero and earnings changes as the proxy for
earnings thresholds used by Young et al. (2012),
14
and includes
them as additional control variables. Thirdly, this study excludes
?rms with three speci?c conditions. Fourthly, this study excludes
the observations before 2006 to avoid the effect of the ?rst revision
of the disclosure regulation of auditor fees related to removing the
condition for non-audit fees more than 500 thousand Taiwan dol-
lars. Fifthly, this study excludes the observations in 2007 and 2008
Table 13
The regression model of mandatory disclosure requirements of auditor fees, the
disclosure type and control variables on real earnings management.
Variable During 2002e2012 During 2009e2012
Dependent variable: REM
Intercept À1.6459
a
(À13.18) À0.9299
a
(À4.68)
After À0.0991
a
(À8.00)
D_amount À0.0090 (À0.51)
Asset 0.1557
a
(22.58) 0.1070
a
(10.43)
Lev 0.1040
b
(2.41) 0.2085
a
(3.11)
Turnover 0.4574
a
(25.54) 0.3096
a
(11.88)
CFO À6.0121
a
(À69.91) À5.6560
a
(À40.10)
D_loss 0.0783
a
(4.95) 0.0750
a
(3.06)
ROA 0.0002 (0.00) À0.1833 (À0.93)
MB À0.0899
a
(À9.38) À0.1169
a
(À7.44)
Growth À0.5009
a
(À19.51) À0.3926
a
(À9.46)
Age À0.0031
b
(À2.38) 0.0010 (0.45)
D_big4 À0.0053 (À0.25) À0.0574 (À1.58)
D_expert 0.0414
a
(3.08) 0.0314 (1.56)
D_newaudit À0.0258 (À1.14) À0.0049 (À0.10)
Tenure À0.0028
c
(À1.76) À0.0030 (À1.22)
Import 0.0019
b
(2.16) 0.0004 (0.26)
Industry dummies included
N 9856 3880
F value 356.76
a
118.90
a
Adj R
2
0.5436 0.5225
Note: REMrepresents the composite real earnings management index by the sumof
standardized abnormal cash ?ows, abnormal production costs, and abnormal
discretionary expenses, and other variable de?nitions are same as those in Tables 2
and 6. The t-statistic (in parentheses) is based on the two-way robust standard
errors clustered at the ?rm and year levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
14
According to Young et al. (2012), ?rms are identi?ed as benchmark beaters
when they report earnings or a change in earnings between zero and one percent of
lagged total assets.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 12
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
to avoid the effect of the global ?nancial crisis.
15
Finally, this study
examined if the results are sensitive to the survivorship bias by
excluding the delisted ?rms.
16
The untabulated results for these
sensitivity analyses are consistent with those in Tables 5 and 6.
5. Conclusion
The disclosure regulation of auditor fees in Taiwan has been
revised by the regulators three times since 2002. The most
important change in disclosure regulation is that ?rms have to
disclose their auditor fees and have been able to disclose fees in the
form of individual amount or fee range since 2009. This study in-
vestigates whether audit quality, measured in terms of accrual-
based earnings management, will be improved after mandatory
disclosure requirements, and whether there exists differences in
audit quality between ?rms reporting the individual amount of
auditor fees and those reporting the fee range. Using unique Taiwan
data, this study shows that not only accrual-based earnings man-
agement, especially the positive discretionary accruals, but also real
earnings management is lower, after mandatory disclosure
requirements of auditor fees, implying that the enhanced infor-
mation transparency induced by the mandatory disclosure re-
quirements of auditor fees is useful to improve auditor
independence and further induce better audit quality. Furthermore,
?rms that disclose auditor fees in the form of individual amount
have lower (higher) positive (negative) accrual-based earnings
management than those in the form of fee range. Speci?cally, both
accrual-based and real earnings management decrease signi?-
cantly for ?rms with the upgrading disclosure type of auditor fees
from fee range to individual amount. This implies that the more
transparent disclosure type of auditor fees is associated with
enhanced audit quality. These results are robust to various addi-
tional tests, including the control for the yearly effects after
mandatory disclosure requirements, current accruals and prior
discretionary accruals, self-selection of auditors, board character-
istics, and earnings thresholds, and the examination for the
restrained samples, modi?ed Jones model for discretionary accruals
and the truncated regression. Overall, the evidence form this study
is consistent with the notion that the enhanced information
transparency induced by mandatory disclosure requirements
related to auditor fees is associated with enhanced auditor inde-
pendence and thus supports Dye's (1991) theory. Some research
limitations of this study should be noted. Firstly, the empirical re-
sults from listed ?rms in this study could not apply to non-listed
?rms. Secondly, there could be a potential self-selection problem
of disclosure types of auditor fees.
Acknowledgments
The author is grateful for the ?nancial support of Ministry of
Science and Technology of R.O.C under grant no. MOST 104-2410-
H-018-006, and wishes to thank the two anonymous referees for
their valuable comments.
References
Anthony, J. H., & Ramesh, K. (1992). Association between accounting performance
measures and stock prices: a test of the life cycle hypothesis. Journal of Ac-
counting and Economics, 15(2e3), 203e227.
Balsam, S., Krishnan, J., & Yang, J. S. (2003). Auditor industry specialization and
earnings quality. Auditing: A Journal of Practice & Theory, 22(2), 71e97.
Barton, J., & Simko, P. J. (2002). The balance sheet as an earnings management
constraint. The Accounting Review, 77(Suppl. 1), 1e27.
Becker, C. L., DeFond, M. L., Jiambalvo, J., & Subramanyam, K. R. (1998). The effect of
audit quality on earning management. Contemporary Accounting Research, 15(1),
1e24.
Bloom?eld, R. J. (2002). The incomplete revelation hypothesis and ?nancial
reporting. Accounting Horizons, 16(3), 233e243.
Boone, J. P., Khurana, I. K., & Raman, K. K. (2010). Do the Big 4 and the second-tier
?rms provide audits of similar quality? Journal of Accounting and Public Policy,
29(4), 330e352.
Cameron, A. C., & Miller, D. L. (2011). Robust inference with clustered data. In
A. Ullah, & D. E. A. Giles (Eds.), Handbook of empirical economics and ?nance (pp.
1e28). Boca Raton, FL: CRC Press.
Carcello, J. V., & Nagy, A. L. (2004). Audit ?rm tenure and fraudulent ?nancial
reporting. Auditing: A Journal of Practice & Theory, 23(2), 55e69.
Carey, P., & Simnett, R. (2006). Audit partner tenure and audit quality. The Ac-
counting Review, 81(3), 653e676.
Chaney, P. K., Jeter, D. C., & Shivakumar, L. (2004). Self-selection of auditors and
audit pricing in private ?rms. The Accounting Review, 79(1), 51e72.
Chang, W. J., Chen, J. P., & Su, J. W. (2010). The effect of market perception man-
agement on the SFAS no. 35: the analysis of early adoption. Journal of Man-
agement & Systems, 17(2), 311e338.
Chang, R. D., & Fang, C. J. (2006). Effects of the information disclosure and evalu-
ation system on earnings management. Journal of Accounting Review, 42, 1e22.
Chang, R. D., Fang, C. J., & Tseng, Y. C. (2007). Central agency problem and earnings
management: an investigation of board structure and external monitoring
mechanism. Journal of Management, 24(1), 17e39.
Chang, R. D., Shen, W. H., & Fang, C. J. (2009). The effects of the Procomp scandal on
going concern audit opinions: an investigation of the moderating effects. NTU
Management Review, 19(2), 75e108.
Chang, C. C., & Tsao, M. C. (2005). The determinants of audit fees for Taiwanese
listed ?rms. Journal of Contemporary Accounting, 6(2), 125e152.
Table 14
The regression model of changes in disclosure types of auditor fees and control
variables on changes in earnings management during 2009e2012.
Variable Dependent variable
DDA DREM
Intercept 0.0284 (1.03) À0.7102
a
(À3.23)
D_type_up À0.0101
c
(À1.74) À0.0965
a
(À2.59)
D_type_down À0.0014 (À0.24) À0.0345 (À0.96)
Asset 0.0005 (0.26) 0.0621
a
(5.07)
Lev À0.0341
b
(À2.54) À0.2526
a
(À3.00)
Turnover À0.0002 (À0.05) 0.0134 (0.40)
CFO À0.8237
a
(À27.96) À4.6625
a
(À22.69)
D_loss 0.0415
a
(8.27) À0.0735
b
(À2.40)
ROA 0.7495
a
(19.20) 1.3909
a
(5.25)
MB À0.0020 (À0.75) 0.0553
a
(2.63)
Growth À0.0304
a
(À2.99) À0.5927
a
(À9.56)
Age À0.0003 (À0.71) À0.0079
a
(À3.02)
D_big4 À0.0077 (À1.26) À0.0103 (À0.26)
D_expert 0.0045 (1.23) 0.0325 (1.27)
D_newaudit 0.0058 (0.54) 0.0992 (1.33)
Tenure À0.0001 (À0.34) À0.0029 (À1.03)
Import À0.0006
b
(À2.00) À0.0020 (À0.89)
Industry and year dummies included
N 2876 2876
F value 38.51
a
30.67
a
Adj R
2
0.3135 0.2654
Note: DDA is the change in performance-adjusted abnormal accruals proposed by
Kothari et al. (2005), while DREM is the change in the composite real earnings
management index by the sum of standardized abnormal cash ?ows, abnormal
production costs, and abnormal discretionary expenses. D_type_up is a dummy
variable for upgrading disclosure type that equals one if the ?rms disclose auditor
fees in the formof individual amount in the current year and in the formof fee range
in the preceding year. D_type_down is a dummy variable for downgrading disclo-
sure type that equals one if the ?rms disclose auditor fees in the form of fee range in
the current year and in the form of individual amount in the preceding year. The
other variable de?nitions are same as those in Table 2. The t-statistic (in paren-
theses) is based on the two-way robust standard errors clustered at the ?rm and
year levels.
a,b,c
denote signi?cance at the 0.01, 0.05, and 0.10 levels, respectively.
15
Francis, Hasan, and Wu (2013) demonstrated that accounting conservatism is
contributive to ?rm stock performance during the global ?nancial crisis. Because
there is no general agreement on the beginning and ending of the global ?nancial
crisis, the periods of 2008e2009 or 2007e2009 are also excluded, respectively, to
avoid the potential effect of the global ?nancial crisis or related accounting
conservatism. The untabulated results are still consistent.
16
The delisted ?rms are included in the sample to reduce the potential survi-
vorship bias (Chen & Chien, 2011; Elbadry, Gounopoulos, & Skinner, 2015).
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 13
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
Chen, T. C., & Chien, C. C. (2011). Size effect in January and cultural in?uences in an
emerging stock market: the perspective of behavioral ?nance. Paci?c-Basin
Finance Journal, 19(2), 208e229.
Chen, C. Y., Lin, C. J., & Lin, Y. C. (2008). Audit partner tenure, audit ?rm tenure, and
discretionary accruals: does long auditor tenure impair earnings quality?
Contemporary Accounting Research, 25(2), 415e445.
Chen, K. Y., Liu, J. L., & Lin, K. L. (2003). Industry specialist, client satisfaction and
audit fees. Journal of Accounting Review, 37, 31e52.
Chen, K. Y., & Wu, S. Y. (2004). Industry specialists, audit fees and auditor size:
evidence from Taiwan. Taiwan Accounting Review, 5(1), 41e69.
Chi, W., Huang, H., Liao, Y., & Xie, H. (2009). Mandatory audit partner rotation, audit
quality and market perception: evidence from Taiwan. Contemporary Account-
ing Research, 26(2), 359e391.
Chi, W., Lisic, L. L., & Pevzner, M. (2011). Is enhanced audit quality associated with
greater real earnings management? Accounting Horizons, 25(2), 315e335.
Choi, J. H., Kim, J. B., Liu, X., & Simunic, D. A. (2008). Audit pricing, legal liability
regimes, and big 4 premiums: theory and cross-country evidence. Contempo-
rary Accounting Research, 25(1), 55e99.
Choi, J. H., Kim, J. B., & Zang, Y. (2010). Do abnormally high audit fees impair audit
quality? Auditing: A Journal of Practice & Theory, 29(2), 115e140.
Clinton, S. B., White, J. T., & Woidtke, T. (2014). Differences in the information
environment prior to seasoned equity offerings under relaxed disclosure
regulation. Journal of Accounting and Economics, 58(1), 59e78.
Cohen, D. A., Dey, A., & Lys, T. Z. (2008). Real and accrual-based earnings man-
agement in the pre- and post-Sarbanes-Oxley periods. The Accounting Review,
83(3), 757e787.
Cohen, D. A., & Zarowin, P. (2010). Accrual-based and real earnings management
activities around seasoned equity offerings. Journal of Accounting and Eco-
nomics, 50(1), 2e19.
DeAngelo, L. E. (1981a). Auditor independence, low balling, and disclosure regula-
tion. Journal of Accounting and Economics, 3(2), 113e127.
DeAngelo, L. E. (1981b). Auditor size and auditor quality. Journal of Accounting and
Economics, 3(3), 183e199.
DeAngelo, H., DeAngelo, L., & Skinner, D. J. (1994). Accounting choice in troubled
companies. Journal of Accounting and Economics, 7(1e2), 113e143.
Dechow, P. M., Sloan, R. G., & Sweeney, A. P. (1995). Detecting earning management.
The Accounting Review, 70(2), 193e225.
Dechow, P. M., Sloan, R. G., & Sweeney, A. P. (1996). Causes and consequences of
earnings manipulation: an analysis of ?rms subject to enforcement actions by
the SEC. Contemporary Accounting Research, 13(1), 1e36.
DeFond, M. L., Raghunandan, K., & Subramanyam, K. R. (2002). Do non-audit ser-
vices affect auditor independence? Evidence from going-concern audit opin-
ions. Journal of Accounting Research, 40(4), 1247e1274.
DeFond, M. L., & Subramanyam, K. R. (1998). Auditor changes and discretionary
accruals. Journal of Accounting and Economics, 25(1), 35e67.
Dye, R. A. (1991). Informationally motivated auditor replacement. Journal of Ac-
counting and Economics, 14(4), 347e374.
Elbadry, A., Gounopoulos, D., & Skinner, F. (2015). Governance quality and infor-
mation asymmetry. Financial Markets, Institutions & Instruments, 24(2e3),
127e157.
Fan, H. S., Chen, C. L., & Wu, C. Y. (2007). The impacts of client importance and
auditor industrial specialization on earnings quality: evidence from audit
groups. Business Review, 12(1), 75e108.
Francis, B., Hasan, I., & Wu, Q. (2013). The bene?ts of conservative accounting to
shareholders: evidence from the ?nancial crisis. Accounting Horizons, 27(2),
319e346.
Francis, J. R., & Ke, B. (2006). Disclosure of fees paid to auditors and the market
valuation of earnings surprises. Review of Accounting Studies, 11(4), 495e523.
Francis, J. R., Maydew, E. L., & Sparks, H. C. (1999). The role of big 6 auditors in the
credible reporting of accruals. Auditing: A Journal of Practice & Theory, 18(2),
17e34.
Francis, J. R., & Wang, D. (2005). Impact of the SEC's public fee disclosure require-
ment on subsequent period fees and implications for market ef?ciency. Audit-
ing: A Journal of Practice & Theory, 24(Suppl. 1), 145e160.
Frankel, R. M., Johnson, M. F., & Nelson, K. K. (2002). The relation between auditors'
fees for nonaudit services and earnings management. The Accounting Review,
77(Suppl. 1), 71e105.
Ghosh, A., & Moon, D. (2005). Auditor tenure and perceptions of audit quality. The
Accounting Review, 80(2), 585e612.
Gow, I. D., Ormazabal, G., & Taylor, D. J. (2010). Correcting for cross-sectional and
time-series dependence in accounting research. The Accounting Review, 85(2),
483e512.
Guan, Y. D., Chien, S. Y., & Hsu, Y. P. (2008). Accounting frauds and ?nancial
reporting conservatism-observations from Enron/Procomp. Journal of Contem-
porary Accounting, 9(1), 69e104.
Gujarati, D. (1995). Basic econometrics (3rd ed.). New York, NY: McGraw-Hill.
Gupta, M., Pevzner, M., & Seethamraju, C. (2010). The implications of absorption
costing accounting and performance decisions for future ?rm performance and
valuation. Contemporary Accounting Research, 27(3), 889e922.
Healy, P. M. (1985). The effect of bonus schemes on accounting decision. Journal of
Accounting and Economics, 7(1e3), 85e107.
Heckman, J. J. (1979). Sample selection bias as a speci?cation error. Econometrica,
47(1), 153e161.
Heninger, W. G. (2001). The association between auditor litigation and abnormal
accruals. The Accounting Review, 76(1), 111e126.
Hirshleifer, D., & Teoh, S. H. (2003). Limited attention, information disclosure, and
?nancial reporting. Journal of Accounting and Economics, 36(1e3), 337e386.
Hirst, D. E., & Hopkins, P. E. (1998). Comprehensive income reporting and analysts'
valuation judgments. Journal of Accounting Research, 36(Suppl.), 47e75.
Hsu, Y. S., Chen, H. C., & Chen, W. C. (2013). The impact of board characteristics on
the choice of earnings management. Journal of Accounting Review, 57, 39e72.
Huang, S. Y., Lin, C. C., & Tien, K. M. (2012). The relationship among the CPA ?rms
change, audit fee discount and audit quality. Chaoyang Business and Manage-
ment Review, 11(2), 43e76.
Jiang, J. X., & Yang, Y. W. (2005). The effects of audit ?rms' industry specialization
and auditor tenure on earnings quality. Journal of Contemporary Accounting,
6(1), 23e60.
Jones, S., & Sharma, R. (2001). The impact of free cash ?ow, ?nancial leverage and
accounting regulation on earnings management in Australia's old and new
economics. Managerial Finance, 27(12), 18e39.
Kim, J. B., Chung, R., & Firth, M. (2003). Auditor conservatism, asymmetric moni-
toring, and earnings management. Contemporary Accounting Research, 20(2),
323e359.
Kinney, W. R., & Martin, R. D. (1994). Does auditing reduce bias in ?nancial
reporting? A review of audit-related adjustment studies. Auditing: A Journal of
Practice & Theory, 13(1), 149e156.
Kothari, S. P., Leone, A. J., & Wasley, C. E. (2005). Performance matched discretionary
accrual measures. Journal of Accounting and Economics, 39(1), 163e197.
Krishnan, G. V. (2003). Does big 6 auditor industry expertise constrain earnings
management? Accounting Horizons, 17(Suppl.), 1e16.
Lai, K. W. (2009). Audit opinion and disclosure of audit fees. Journal of Accounting,
Auditing & Finance, 24(1), 91e114.
Lee, J. Z., & Chen, J. F. (2004). The effect of audit client's importance on magnitude of
earnings management: from the perspective of audit groups within the big ?ve.
Journal of Accounting Review, 38, 59e80.
Lee, J. Z., Hsu, S. W., & Chen, J. F. (2003). The relationship between non-audit ser-
vices and abnormal accruals. Journal of Accounting Review, 37, 1e30.
Lee, J. Z., & Liao, Y. H. (2004). Board of director characteristics and earnings man-
agement: evidence from the effect of family-controlling on Taiwan corpora-
tions. Taiwan Accounting Review, 5(1), 1e39.
Lee, J. Z., Liao, H. M., & Huang, Y. C. (2010). Knowledge spillover effect of audit and
non-audit services. Web Journal of Chinese Management Review, 13(2), 1e23.
Lee, J. Z., & Lin, H. F. (2005). The relationship between auditor tenure and abnormal
accruals. Management Review, 24(4), 103e126.
Lee, J. Z., & Lin, H. F. (2013). Is the audit quality of big4 better than non-big4? Ev-
idence from ability of CPA ?rm to restrict earnings management: the impor-
tance of controlling self-selection bias. Taiwan Accounting Review, 9(1), 77e110.
Liao, H. M., & Hung, S. W. (2010). Nonaudit service, audit tenure and audit quality.
Journal of Contemporary Accounting, 11(2), 151e178.
Liao, Y. H., Sang, T. S., & Lin, H. F. (2013). The association between substantial non-
audit services and audit quality: evidence from special audit of internal control
system. Taiwan Accounting Review, 9(1), 111e158.
Liao, H. M., Wang, C. C., & Chi, W. C. (2012). What does the complete disclosure of
audit fee information tell us in Taiwan? Taiwan Accounting Review, 8(1), 49e88.
Lin, H. M., Kuo, B. L., & Su, N. H. (2008). The in?uence of internationalization, board
characteristics, and sophisticated institutional monitoring on earnings man-
agement: evidence from information and electronics industries in Taiwan. NTU
Management Review, 19(1), 157e186.
Maines, L. A., & McDaniel, L. S. (2000). Effects of comprehensive-income charac-
teristics on nonprofessional investors' judgments: the role of ?nancial-
statement presentation format. The Accounting Review, 75(2), 179e207.
Mitra, S., Deis, D. R., & Hossain, M. (2009). The association between audit fees and
reported earnings quality in pre- and post-Sarbanes-Oxley regimes. Review of
Accounting and Finance, 8(3), 232e252.
Myers, J. N., Myers, L. A., & 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, 78(3), 779e799.
Nichols, D. R., & Price, K. H. (1976). The auditore?rm con?ict: an analysis using
concepts of exchange theory. The Accounting Review, 51(2), 335e346.
Peng, C. W., & Young, C. S. (2013). The relationship between avoidance of reporting
annual losses and excess production: the role of controlling shareholders.
Journal of Management & Systems, 20(4), 727e753.
Press, E. G., & Weintrop, J. B. (1990). Accounting-based constraints in public and
private debt agreements: their association with leverage and impact on ac-
counting choice. Journal of Accounting and Economics, 12(1e3), 65e95.
Raghunandan, K., & Rama, D. V. (1995). Audit report for companies in ?nancial
distress: before and after SAS no. 59. Auditing: A Journal of Practice & Theory,
14(1), 50e63.
Roychowdhury, S. (2006). Earnings management through real activities manipu-
lation. Journal of Accounting and Economics, 42(3), 335e370.
Schipper, K. (1989). Commentary on earnings management. Accounting Horizons,
3(4), 91e102.
Shiue, M. J., Chang, Y. S., & Kao, C. T. (2008). Self-selection of auditors and audit fees.
Journal of Contemporary Accounting, 9(2), 167e200.
Singhvi, S. S., & Desai, H. B. (1971). An empirical analysis of the quality of corporate
?nancial disclosure. The Accounting review, 46(1), 129e138.
Su, Y. H. (2000). Auditor fees and auditor size: a study of audit market in Taiwan.
Taiwan Accounting Review, 1, 59e78.
Su, Y. H. (2005). Auditor tenure and audit quality. Journal of Contemporary Ac-
counting, 6(2), 187e220.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 14
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002
Taylor, G. K., & Xu, R. Z. (2010). Consequences of real earnings management on
subsequent operating performance. Research in Accounting Regulation, 22(2),
128e132.
Taylor, M. H., & Simon, D. T. (1999). Determinants of audit fees: the importance of
litigation, disclosure, and regulatory burdens in audit engagements in 20
countries. The International Journal of Accounting, 34(3), 375e388.
Thompson, S. B. (2011). Simple formulas for standard errors that cluster by both
?rm and time. Journal of Financial Economics, 99(1), 1e10.
Trompeter, G. (1994). The effect of partner compensation schemes and generally
accepted accounting principles on audit partner judgment. Auditing: A Journal of
Practice & Theory, 13(2), 56e68.
Tseng, C. H., Chang, Y. S., & Shiue, M. J. (2012). The association between the number
of individual auditors' clients and earnings quality. Taiwan Accounting Review,
8(2), 217e256.
Wang, Y. C., & Chang, C. P. (2012). An empirical study of core agency problem on
earnings management. Review of Accounting and Auditing Studies, 2(2),
69e113.
Wang, C. C., Chang, Y. S., & Lin, K. H. (2012). Audit quality and information asym-
metry. Taiwan Accounting Review, 8(1), 89e135.
Wang, L. F., Shiue, M. J., & Tseng, C. H. (2011). The relation between the changes of
information transparency ranking and earnings management. Review of Ac-
counting and Auditing Studies, 1(1), 101e131.
Watkins, A. L., Hillison, W., & Morecroft, S. E. (2004). Audit quality: a synthesis of
theory and empirical evidence. Journal of Accounting Literature, 23, 153e193.
Watts, R. L., & Zimmerman, J. L. (1978). Towards a positive theory of the determi-
nation of accounting standards. The Accounting Review, 53(1), 112e134.
Wu, T. Z., Chiou, J. R., & Cheng, L. (2011). The in?uence of corporate governance on a
CPA's going-concern opinion for ?nancially distressed ?rms: an empirical study
for ?rms listed in the Taiwan stock exchange. NTU Management Review, 21(2),
187e218.
Young, C. S., Tsai, L. C., Chen, C. H., & Liao, S. T. (2012). Board characteristics and real
earnings management. NTU Management Review, 23(1), 363e400.
Young, C. S., & Wu, S. J. (2003). CEO pay-performance sensitivity, performance
threshold and discretionary accounting choices: an empirical study. Journal of
Accounting Review, 36, 55e87.
Zang, A. Y. (2012). Evidence on the trade-off between real activities manipulation
and accrual based earnings management. The Accounting Review, 87(2),
675e703.
C.-S. Chen / Asia Paci?c Management Review xxx (2016) 1e15 15
Please cite this article in press as: Chen, C.-S., The effect of mandatory disclosure requirements and disclosure types of auditor fees on earnings
management: Evidence from Taiwan, Asia Paci?c Management Review (2016),http://dx.doi.org/10.1016/j.apmrv.2015.12.002

doc_973063880.pdf