Earnings quality and perceived auditor independence: irrelevance of nonaudit services time period.
Liu, Carol ; Parkash, Mohinder ; Singhal, Rajeev 等
Abstract
Substantial growth in consulting services provided by auditors in
1990's has been highlighted as a threat to auditor independence,
which has been argued to be the cause of changes in investors'
perception of auditor independence. Research based on audit and nonaudit
fees disclosed in 2000's provide consistent results. This study
proposes that the perceived auditor independence decreased with increase
in auditor provided nonaudit services even before recent controversial
environmental changes. Based on nonaudit fees from 1978-80, the evidence
indicates that smaller earnings response coefficients are associated
with higher auditor provided nonaudit services. These results suggest
impairment of auditor independence by the provisions of nonaudit
services irrespective of the time period.
Keywords: Auditor Independence, Nonaudit Services, Consulting
Services, Earnings Response Coefficients, ERC
I. INTRODUCTION
Since the Securities Acts of 1933 and 1934, the financial
statements of public corporations are required to be audited by
independent public accountants. Auditor independence has formally been
recognized as an essential key feature for effective auditing and
efficient capital markets (Watts and Zimmerman (1983)). The relationship
between auditor independence and auditor provided nonaudit (1) services
continue to be a controversial issue (2). The SEC strongly believes that
the incumbent auditor providing nonaudit services creates financial
dependencies and the nature of nonaudit services may put them in
managerial roles thus eroding the earnings quality (SEC (2000a)). But
the accounting profession has taken the position that providing nonaudit
services to audit clients can increase the audit quality because of
knowledge-spillover without compromising independence as auditors have
economic incentives to protect their professional capital and integrity
(Antle, Griffen, Teece, and Williamson (1997)). SEC Chairman Levitt has
commented in this context (Levitt and Dwyer (2002, p. 129)) that:
"No auditor would ever admit that he allowed bad numbers
because he wanted to bring more consulting business ... (but) the
coincidence of accounting misdeeds, company restatements, and billions
in shareholders losses were too striking to dismiss as
happenstance."
Further, SEC has taken the position that "an auditor's
independence is impaired either when the accountant is not independent
in fact, or when in light of all relevant facts and circumstances, a
reasonable investor would conclude that the auditor would not be capable
of acting without bias" (SEC (2000b)). The sentiment is echoed by
Turner (2000), a former chief accountant when he says that "auditor
independence is really about only one thing--investors' confidence
in numbers."
Levitt (2000) emphasized the substantial growth in consulting
services provided by auditors in 1990's as a threat to auditor
independence at least in appearance if not in fact. The SEC believed
that earnings and financial reporting quality eroded in 1990's,
which led to mandatory disclosure of audit as well as nonaudit fees paid
to auditors in proxy filings after February 5, 2001. In addition, the
SEC banned auditors from providing certain types of nonaudit services.
Recent scandals like failure of Enron also raised concerns about
nonaudit services which further led to enactment of Sarbanes-Oxley Act
of 2002. (3) Based on the audit and nonaudit fees disclosed in
2000's, Francis and Ke (2006) and Ghosh, Kallapur and Moon (2008)
report a negative relation between ERCs and the extent of nonaudit fees
paid to auditors indicating compromised auditor independence. The
contention of both the SEC's position and the analysis by Francis
and Ke (2006) and Ghosh et al. (2008), has been the deterioration of
financial reporting environment with the recent significant increases in
nonaudit fees and the change in investors' perception of auditor
independence. The current study addresses the issue if recent financial
reporting environment changes and increases in nonaudit fees are the
reasons for investors' perception of compromised auditor
independence. The current study utilizes nonaudit fees data from 1978-80
to show that the nonaudit fees relative to audit fees was perceived to
indicate a compromise of auditor independence much earlier than the
recent boost in audit provided nonaudit fees. The current study is the
first research that utilizes data from a non-controversial time period
and addresses directly the effect of investors' perception of
auditor independence while providing nonaudit services on valuation of
earnings. (4)
The firm's earnings receive credibility from auditor's
reputation that varies with auditor quality. Auditor size and brand have
been used to represent auditor quality. Teoh and Wong (1993) report
larger earnings response coefficients (ERCs) for Big 8 clients than for
non-Big 8 clients to conclude that investors assign higher audit and
earnings quality of Big 8 clients as compared to non-Big 8 clients.
Prior research also suggests that external parties and regulators
associate joint audit and nonaudit purchases with beliefs about impaired
auditor independence and audit quality (Parkash and Venable (1993)).
Since reduction in perceived independence reduces audit credibility and
quality, the current study empirically examines the cross-sectional
differences in ERCs according to auditor provided nonaudit services
after controlling for auditor brand name. The empirical evidence
presented in the study indicates that smaller ERCs are associated with
higher auditor provided nonaudit services suggesting perceived
impairment of auditor independence which is consistent with results
reported by Francis and Ke (2006) using more recent data from very high
sensitive time period.
It is also posited that auditor's industry specialization in
providing total or recurring nonaudit services contributes positively to
the auditor and earnings quality. The study provides empirical evidence
indicating that higher ERCs are directly related to the use of industry
specialist provider of total or recurring nonaudit services. The results
suggest that the decrease in ERCs due to nonaudit services is mitigated
when industry specialist provider of nonaudit services is used. Based on
the assertion that recurring and nonrecurring nonaudit services have
different economic bonding with differing implications for auditor
independence (Beck, Frecka, and Solomon (1988a)), the current study
provides empirical results indicating that negative effects of recurring
nonaudit services on earnings response coefficients are larger than
negative effect of nonrecurring nonaudit services on earnings response
coefficients.
The empirical analysis in this study utilizes a unique data set of
nonaudit fees manually collected from proxies when ASR No 250 required
disclosure of specific audit services for tax services; pension and
personnel services; system services; merger and acquisition services;
and other services in addition to total nonaudit fees as percentage of
total audit fees. These detailed disclosures were made only under ASR
No. 250 requirements. The detailed disclosure allows the division of
nonaudit services into recurring and nonrecurring nonaudit services,
which is essential for assessment of auditor independence while
analyzing the impact of auditor provided nonaudit services. (5) The
results of the current study provide a valid basis for comparative
analysis when data from other regimes are used where accounting
reporting environment is more controversial.
The current study extends the results of Teoh and Wong (1993) who
reported differences in credibility of audited earnings reports because
of auditor quality as represented by Big 8 and non-Big 8 distinction.
The results from this study provide evidence of differential market
valuation of audited earnings because of differences in nonaudit
services procured from auditors after controlling for the auditor brand
name.
The study contributes to the existing literature on auditor
independence because of nonaudit services by analyzing the
investors' perception whereas other studies have examined the
effect of nonaudit services on the auditor independence in fact with
mixed results. This paper also makes a contribution to the growing
literature on the explanation of cross-sectional differences in the
relation of stock market reactions with earnings announcements.
The remaining paper is organized in four sections. The second
section reviews the related literature and motivates the hypotheses. The
sample and measures of variables used in the study are described in the
third section. The regression results to test the hypotheses are
presented in the following section, with conclusions appearing in the
last section.
II. LITERATURE REVIEW AND HYPOTHESES
Audit and Nonaudit Services Combination
In agency context, stockholders and creditors rely on management to
safeguard their assets, but they cannot directly observe if managers act
in accordance with these expectations. Hence, accounting variables and
other financial disclosures become proxies for direct observations with
an external third party attesting to the accuracy of the disclosures
(Fama and Jensen 1983a, 1983b). In order for the attestation to provide
value, the auditor must be independent from management. In the mind of
outside observers, actions that imply a conflict of interest impair the
independence of the auditor. One such action is the provision of
nonaudit services (POB (1986)). Increases in economic bonding associated
with the provision of such services give auditors added incentive to
maintain ongoing relationships (DeAngelo (1981); Simunic (1984); Beck et
al. (1988a)). To the extent that nonaudit fees are generated from audit
clients, auditors face increased economic incentives to bear higher risk
levels in audit engagements. DeAngelo (1981) modeled the audit pricing
structure and found that future quasi-rents from audit engagements
create an economic bond that provides incentives for the auditor and
client to maintain their relationship resulting in possible compromises.
Subsequently, Simunic (1984) analyzed the joint demand for nonaudit and
audit services and further showed economic bonding to the client when
attendant costs savings were retained. Beck et al. (1988a) showed
increases in bonding dependent upon the magnitude of nonaudit startup
and switching costs, audit cost savings due to knowledge spillovers, and
the recurring or nonrecurring nature of the nonaudit engagement. If
disclosure of nonaudit services is made, the assessment of independence
made by external financial statement users can be affected resulting in
negative economic consequences. But the predicted effects of audit and
nonaudit combination on auditor independence are ambiguous with mixed
results of the research on the issue (Gunther and Moore (2002)).
SEC and Nonaudit Services
In 1978, SEC registered firms were required to disclose the amount
of auditor provided nonaudit services as a percentage of audit fees (ASR
No. 250) but as of 1982 this requirement was rescinded (ASR No. 304).
The commission implemented the disclosure requirement because it felt
that the reports of external accountants substantially increased the
reliability of filed financial statements. In order to assess the
independence of the accountants, SEC wanted to consider all relevant
evidence bearing on the audit relationship. In its rescission of ASR No.
250, SEC cited the responses of comments to the proposal to rescind ASR
No 250 to conclude that there was no evidence of investor interest in
disclosure of auditor provided nonaudit services. But recently, SEC has
expressed concerns that providing nonaudit services to clients impairs
auditor independence (see Levitt (1999, 2000)). Similar concerns are
echoed in the Independent Standards Board's Conceptual Framework
(ISB (2000)). The SEC has adopted a new rule requiring firms to disclose
audit and nonaudit fees paid to auditors. SEC argues that this
disclosure will provide investors with the information to assess if
nonaudit fees relative to audit fees raise a question about
auditor's independence (see SEC (2000a)).
Research Supporting Auditor Independence with Nonaudit Services
In 1979, the Public Oversight Board (POB) issued a report
concluding that available empirical evidence did not reveal any actual
instances where the furnishing of nonaudit services impaired
independence. In a study sponsored by the American Institute of
Certified Public Accountants (AICPA), Antle et al. (1997) report that
there is no evidence that the supply of nonaudit services threatened
auditor independence. They find that the provision of nonaudit services
by auditors is not a significant factor in auditors' losses in
litigation or in pricing their liability insurance. They further report
that they do not find any empirical evidence of investors being
concerned that providing nonaudit services impairs auditors'
independence. Palmrose (1999) analyzes the provision of nonaudit
services as basis of complaints in litigation against auditors and
without any significant results indicating investors are not concerned
about auditors providing nonaudit services. Arrunada (1999) argues that
if an auditor has a large number of clients the provision of nonaudit
services can actually increase independence. Ashbaugh et al. (2003)
analyzed going concern opinions to find no systematic evidence to
support auditors violating their independence as a result of clients
purchasing higher nonaudit services. Chung and Kallapur (2003) suggest
that client importance determines the incentives for auditor to
compromise its independence. Based on their empirical results they do
not find auditor independence impairment as a function of different
client fees ratios.
In the experimental setting, Dopuch, King and Joyce (1991) find
that there is no indication that market participants perceived
auditors' reports to be less credible because the same auditors
provide nonaudit services to the clients. McKinley, Pany, and Reckers
(1985) examine whether providing nonaudit services for audit clients
affects financial statements users' perceptions of auditor
independence. They observe that financial statements users believe that
Big 8 auditors were more independent but providing nonaudit services did
not significantly affect loan decisions. Pany and Reckers (1988)
presented identical loan and investment packages to loan officers and
financial analysts with nonaudit variable systematically manipulated
across individuals. The results of this study indicate that auditor
provided nonaudit services has little effect on perceptions of auditor
independence.
In summary, the conclusion of above studies is that there is no
effect of the purchase of nonaudit services from auditor on auditor
independence.
Research Indicating Lack of Auditor Independence with Nonaudit
Services
Parkash and Venable (1993) point out (p. 113) "External
parties and regulators have associated joint (audit and) nonaudit
purchases with beliefs about impaired auditor independence."
Frankel et al. (2002) report a negative significant market reaction to
proxy statements filed by firms reporting higher than expected nonaudit
fees. They also present empirical evidence indicating direct relation
between the extent of nonaudit fees and earnings management concluding a
lack of auditor independence because of nonaudit services. Pany and
Reckers (1983) and Shockley (1981) report that their subjects perceive
auditor independence as being affected negatively when nonaudit services
are provided by the auditors. Swanger and Chewning (2001) report a
decreased perception of auditor independence when the auditor also
performs internal audit function. Beck et al. (1988b) investigate the
effects of providing nonaudit services on auditor tenure. The results
indicate that auditor tenure is longer and less variable when high
levels of nonaudit services are purchased from the incumbent auditor.
Parkash and Venable (1993) present results to indicate that auditees
have incentives to limit nonaudit purchases because of the agency costs
imposed when an audit's monitoring value is diminished.
Specifically, firms with high agency costs demand high auditor
independence and restrict the purchase of nonaudit services.
Collectively, these studies provide evidence that the purchase of
nonaudit services from auditors does impact the auditor independence.
Overall the evidence is mixed on the issue of whether nonaudit
services purchased from auditor adversely affects auditor independence.
Stock Returns-Earnings Relationship and Auditor Quality
In the accounting literature, a considerable amount of evidence has
accumulated on the relation between stock returns and unexpected
earnings to assess the information content of a firm's reported
earnings (see Foster (1986)). The recent research in this area has
focused on the explanation of cross-sectional differences in ERCs
(Kormendi and Lipe (1986); Easton and Zmijewski (1989); Lipe (1986);
Hoskin, Hughes, and Ricks (1986)). The researchers have started
examining the effects of the quality of earnings numbers on the relation
of the stock price reaction to earnings surprises. The degree to which
announced earnings results in stock price movements as measured by ERC
could indicate the quality of earnings and financial statements. Imhoff
and Lobo (1992) observed that companies with low analysts' earnings
forecasts consensus tend to have lower ECR to conclude that the
dispersion in analysts' forecasts proxy the noise in accounting
numbers. Researchers have argued that higher audit quality can reduce
the perceived uncertainty and noise in reported earnings resulting in
higher ERC.
Holthausen and Verrecchia (1988) present a model, which predicts
that the magnitude of stock price response increases with the precision
of the information. Teoh and Wong (1993) have simplified Holthausen and
Verrecchia model to predict that the stock price response increases with
the quality in the earnings signal when holding constant differences in
prior uncertainty of investors about underlying value of the firm. They
tested this prediction by using auditor brand name as a proxy for
earnings quality. They find firms audited by Big 8 firms have higher
ERCs than firms audited by non-Big 8 firms. Moreland (1995) reports
client ERC declined after their (Big 8) auditors were subject to SEC
sanctions because of enforcement actions. Hackenbrack and Hogan (2002)
argue that disclosure of reasons for auditor changes signal information
about earnings precision leading to changes in ERCs following auditor
changes. Specifically they observe ERCs tend to increase when firms
change auditors to increase service and monitoring. They also find that
ERCs decline where firms change auditors for other reasons, such as
auditor resignation implying conflict in auditor-client relationship.
The reduction of ERCs suggests that audit conflicts decrease earnings
quality because of diminished audit quality. Based on their multivariate
analysis, Myers, Myers, and Omer (2003) suggest higher earnings quality
with longer auditor tenure.
Collectively, the literature discussed above suggests that the
perceived independence decreases with increases in nonaudit services
from auditor. If investors believe that higher auditor provided nonaudit
services decrease auditor independence reducing audit credibility as
well as earnings quality, higher nonaudit services result in smaller
ERC.
Therefore, it is hypothesized that ERC decreases with higher
auditor provided nonaudit services after controlling for auditor brand
name. Francis and Ke (2006) evaluated if audit and nonaudit fees
disclosure required by the SEC in proxy filings after February 5, 2001
affected investors' valuation of earnings surprise and observed
consistent results. Ghosh et al. (2008) also used 2000-02 audit and
nonaudit data to report a negative association of ERCs with the ratio of
nonaudit and total audit and nonaudit fees and with client importance.
Industry Specialist for Providing Nonaudit Services
Another stream of literature argues that in addition to brand name,
an industry specialist offers a higher level of assurance and audit
quality than does a non-specialist (e.g., see Craswell, Francis, and
Taylor (1995)). Owhoso, Messier and Lynch (2002) show that industry
experienced auditors are better able to detect errors within their
industry specialization than outside their specialization. Researchers
have observed that auditor's industry specialization contributes
positively to the credibility provided by the auditor (Craswell et al.
(1995)). Parkash and Venable (1993) indicate that higher nonaudit
services are purchased from the auditor who is an industry specialist in
providing nonaudit services while integrating concerns because of higher
agency costs. They described an industry specialist to be the auditor
who is the largest overall provider of the nonaudit services in the
industry. Krishnan (2003) reports his finding, which "is consistent
with the notion that specialist auditors mitigate accruals-based
earnings management more than nonspecialist auditors and, therefore,
influence the quality of earnings." Hence it is hypothesized that
higher ERCs are directly related to the use of industry specialist
provider of nonaudit services.
Recurring and Nonrecurring Nonaudit Services
Beck et al. (1988a) present an analytical model to show that
recurring and nonrecurring nonaudit services have different economic
bonding leading to differing implications for auditor independence.
Recurring nonaudit services provide continuous stream of fees whereas
nonrecurring nonaudit services are generally single engagement or
irregular engagements resulting in one time or sporadic fees. Because
recurring nonaudit services provide continuous rents to the auditor, it
is expected that recurring services imply greater auditor dependencies
than with nonrecurring services. Therefore, it is hypothesized that
negative effect on ERC is higher with recurring nonaudit services than
with nonrecurring nonaudit services.
III. DATA AND MEASURES
Sample
In proxy statements filed between September 30, 1978 and January
28, 1982, SEC registrants were required to disclose the amount of
nonaudit services as a percentage of audit fees. Under ASR No.250, in
addition to total nonaudit services (NONADT) firms also disclosed the
nature of specific audit services that were at least three percent of
audit fees. Specific audit services disclosed were for tax services;
pension and personnel services; system services; merger and acquisition
services; and other services. These detailed disclosures were made only
under ASR No. 250 requirements. In conformance with prior research (Beck
et al. (1988a); Parkash and Venable (1993)) recurring nonaudit services
(RECUR) are measured by combining the totals for nonaudit services of
tax services; pension and personnel services; and system services.
Nonrecurring services (NONRECUR) combine the total for nonaudit services
of merger and acquisition services; and other services not classified as
RECUR. The division of audit services as recurring and non-recurring
nonaudit services is possible only with data disclosed during ASR No.
250 regime. The total population is composed of those firms filing proxy
statements within the above time period. The study uses data for the
three fiscal years (1978-80) during which ASR No.250 was in effect. The
sample is limited to Fortune 500 companies. These companies are chosen
because they were more likely to employ a large brand name auditor, thus
signaling their desire for a high quality audit. The sample is limited
to firms with Big 8 auditors to control the demand for quality audits as
represented by auditor brand name.
The nonaudit data were collected manually from proxy statements
from a university library. Firm specific financial data were collected
from COMPUSTAT and returns data were obtained from CRISP. Financial
analysts' forecast information was collected from the Institutional
Brokers Estimate System (IBES). This gave a sample of 777 observations.
Observations from different years are reported in Table 1. Table 1
includes the distribution of total, recurring and nonrecurring nonaudit
services. The means of total and recurring nonaudit services
consistently decreased over three years which is consistent with an
observation made by Beck et al. (1988a). Mean nonrecurring nonaudit
services decreased from 1978 to 1979 but increased in 1980. It is
interesting to observe that highest nonaudit services purchased during
this period was 268% of audit fees. The highest recurring and
nonrecurring nonaudit fees were 263% and 173% of audit fees,
respectively. The industry specialist classification is based on
industry participation within the available sample. INDSP includes the
auditor providing the largest mean nonaudit services within each
industry. Also included are the second and possibly third largest
providers when distinctive differences were observed between the second
and third or between the third and remaining providers. INDSP takes the
value one when the auditor is an industry specialist, zero otherwise.
Two separate industry specialists' classifications are defined in
terms of total and recurring nonaudit services. Table 2 reports the
distribution of total and recurring nonaudit services by auditor and
industry membership based on two-digit SIC codes.
Seven industry classifications as defined in Parkash and Venable
(1993) are used in this study. (6) Although there is no difference of
nonaudit services between industries, there is a distinctive pattern
within each industry as reported in Table 2, Panel A. The industry
specialist rules identified two industry specialists within six of the
industries and three within one remaining industry. [INDSP.sub.a] takes
the value one when the auditor is an industry specialist for providing
total nonaudit services, zero otherwise. In Panel B of Table 2, again
there is no difference of recurring nonaudit services between industries
but pattern similar to that of total nonaudit service reported in Panel
A are observed. The industry specialists in four industries remain
unchanged. In IND4 there are three specialists and IND6 there is only
one auditor classified as specialist. [INDSP.sub.r] takes the value one
when the auditor is an industry specialist for providing recurring
nonaudit services, zero otherwise.
To empirically test the stated hypothesis the following regression
is estimated.
CAR = a + [b.sub.1] UE + [b.sub.2] UE*NONADT + [b.sub.3]
UE*NONADT*[INDSP.sub.a] + [b.sub.4] UPDATE + [b.sub.5] LMV + [b.sub.6]
UE*BETA + [b.sub.7] UE*GROWTH + e (1)
To test the hypothesis that ERC decreases with the extent of
nonaudit services procured from auditor the coefficient [b.sub.2] has to
be negative. To test the hypothesized effect of using industry
specialist for providing nonaudit services the coefficient [b.sub.3] has
to be positive. To control the auditor name brand the sample is limited
to Fortune 500 firms with Big 8 clients. As suggested in the extant
literature, control variables UPDATE, LMV, interaction of UE and BETA,
and interaction of UE and GROWTH are included. The variables used in the
regression (1) are discussed below.
Measures
Cumulative abnormal returns, CAR, are risk-adjusted returns over a
holding period of three days around earnings announcement day
'a'. The cumulative abnormal returns over the event period
(three days) is computed as:
CAR = [a+1.summation over (d=a-1) [U.sub.d]], and, [U.sub.d] =
[r.sub.d] -[alpha]' -[beta] [r.sub.md], where [U.sub.d] is abnormal
returns on day 'd'; day a is earnings announcement day;
[r.sub.d] is daily returns on individual firm's common stock on day
d; [r.sub.md] is daily return on the market on day d; and [alpha],'
[beta]' are ordinary least square regression parameters estimated
using returns of 150 days including the period from 105 to 31 days
before earnings announcement day and from 31 days to 105 days after
earnings announcement day.
Unexpected earnings (UE) are defined as:
UE = ([EPS.sub.a] - [AF.sub.t]) / [P.sub.t], where [EPS.sub.a]:
reported earnings per share for the current year announced on day a,
[AF.sub.t]: analysts' most recent forecasts (at time t) before
earnings announcement, and [P.sub.t]: share price at time t.
The impact of earnings announcements (the event) is measured from
one day before through one day after the earnings announcement day
'a'. But the unexpected earnings are based on the most recent
forecasts that are made on day 't'. Therefore, these forecasts
are t-(a-2) days old for the event period. This introduces a problem due
to measurement error in the independent variables (see Brown, Hagerman,
Griffen, and Zmijewski (1987)). To mitigate this problem, Easton and
Zmijewski (1989) propose to use the stock returns from the day after the
forecast through the day before the event period, as an additional
explanatory variable. Analogous to their procedure, an additional
variable UPDATE is included which is constructed as:
UPDATE = [a-2.summation over (d=t+1)] [U.sub.d],
where [U.sub.d] is the excess returns using market model depicted
above.
The market value of firms's equity has been suggested to
represent pre-disclosure information (Atiase (1985); Collins, Kothari,
and Rayburn (1987); and Kross and Schroder (1989)), which if not
controlled can also introduce measurement errors. Therefore, the log
transformation of market value of firm's equity at time t, LMV, is
used as control for pre-disclosure information. LMV is calculated as:
LMV = LN(#SHARES * [P.sub.t]), where #SHARES is number of shares
outstanding at time t, [P.sub.t] is the share price at time t, and LN is
natural log transformation.
INDSP takes the value one when the auditor is an industry
specialist, zero otherwise.
Positive coefficient of UE, and negative coefficients of UPDATE and
LMV are expected.
IV. RESULTS
The cross-section estimation results of regression (1) are
presented in Tables 3 and 4. White (1980) test rejected the null
hypothesis that the regression residuals are homoskedastic. Adjusted
t-statistics based on the heteroskedasticity consistent covariance
matrix (White (1980)) were also calculated. Hypothesis testing also
assumes a normal distribution of residuals, which are found not to be
normally distributed. Therefore, t-statistics based on the Jackknife
distribution free nonparametric estimate of standard error are also
calculated (Miller (1974a, 1974b)). Unadjusted and adjusted t-values are
generally higher than Jackknife t-values, therefore, only Jackknife
t-statistics are reported for regression analysis in Tables 3 and 4. As
positive coefficients of UE, UE*NONADT*INDSP, UE*GROWTH and negative
coefficients of UE*NONADT, UPDATE, LMV, UE*BETA are predicted,
one-tailed t-statistics are appropriate to test the significance of
these coefficients.
To test the first two hypotheses, the total nonaudit services,
NONADT and industry specialist to provide total nonaudit services
[INDSP.sub.a] are included in regression analysis and results are
reported in Table 3. To support the first hypothesis that ERC decreases
with higher auditor provided nonaudit services the coefficient [b.sub.2]
has to negative. To support the second hypothesis that higher ERCs are
directly related to the use of industry specialist provider of nonaudit
services [b.sub.3] has to be positive. The results reported in Table 3
support both the hypotheses. The coefficient [b.sub.2] is negative and
significant at better than 0.01 level (one-tailed test) indicating ERC
decreases with increase in nonaudit services. The coefficient [b.sub.3]
is positive and significant at better than 0.01 level (one-tailed test)
indicating ERCs are directly related to the use of industry specialist
provider of nonaudit services. The results suggest that smaller ERC
indicate investors become more uncertain of earnings quality because of
nonaudit service decreasing auditor independence. But the decrease in
ERC due to nonaudit services is mitigated when industry specialist
provider of nonaudit services is used.
To test the differing effects of recurring and nonrecurring
nonaudit services, regression (1) is restated to replace NONADT by RECUR
and NONRECUR. It is reasonable to assume that firms hire specialists to
purchase recurring nonaudit services if distinction is made between
recurring and nonrecurring nonaudit services. Therefore, [INDSP.sub.r]
is used for industry specialist. The revised regression is stated as:
CAR = a + [b.sub.1] UE + [b.sub.21] UE*RECUR + [b.sub.22]
UE*NONRECUR + [b.sub.3] UE*RECUR*[INDSP.sub.r] + [b.sub.4] UPDATE +
[b.sub.5] LMV + [b.sub.6] UE*BETA + [b.sub.7] UE*GROWTH + e (2)
If providing any kind of nonaudit services implies decreased
independence then both coefficients [b.sub.21] and [b.sub.22] should be
negative which would be consistent with first hypothesis. Consistent
with second hypothesis that higher ERCs are directly related to the use
of industry specialist provider of recurring nonaudit services [b.sub.3]
has to be positive. To support third hypothesis that negative effect on
ERC is higher with recurring nonaudit services than with nonrecurring
nonaudit services magnitude of [b.sub.21] should be higher than
[b.sub.22]. To test the third hypothesis, the Jackknife t-value for
([b.sub.21]-[b.sub.22]) based on the standard error of ([b.sub.21] -
[b.sub.22]) computed from its distribution obtained from regression
estimation repetitions with three observations excluded in each
repetition. The results are reported in Table 4. Both coefficients
[b.sub.21] and [b.sub.22] are negative but only [b.sub.21] is
statistically significant at the 0.01 level (one-tailed test). Similar
to the results reported in Table 3, the coefficient [b.sub.3] is
positive. The difference between [b.sub.21] and [b.sub.22] is negative
and significant at the 0.01 level (one-tailed test). indicating that
negative effect of recurring nonaudit services on ERC is larger than
negative "effect of nonrecurring nonaudit services on ERC.
Together, the results suggest that (i) both recurring and nonrecurring
nonaudit services indicate reduced auditor independence, (ii) recurring
nonaudit services indicate higher independence compromise than
nonrecurring nonaudit services, and (iii) the use of industry specialist
provider of recurring nonaudit services mitigates the effect
independence due to auditor provided recurring nonaudit services.
V. CONCLUSIONS
Teoh and Wong (1993) reported cross-sectional differences in
earnings response coefficients due to differential audit quality
represented by Big 8 and non-Big 8 distinction, a common measure of
auditor brand name. The current study proposes that the perceived
auditor independence decreases with increases in nonaudit services from
auditor. If investors believe that higher auditor provided nonaudit
services decrease auditor independence, reducing audit and earnings
quality, it is hypothesized that higher nonaudit services result in
smaller earnings response coefficients, which is supported by the
empirical evidence presented in the study. This is consistent with
SEC's concern of impairment of auditor independence when auditor
provides nonaudit services to the audit client. The results of this
study also compliments the results reported by Francis and Ke (2006) to
suggest that the relationship holds irrespective of the sensitivity of
the time periods.
It is also posited that auditor's industry specialization in
providing total or recurring nonaudit services contributes positively to
the auditor and earnings quality. The study provides empirical evidence
indicating that higher earnings response coefficients are directly
related to the use of industry specialist provider of total or recurring
nonaudit services. The results suggest that smaller earnings response
coefficients indicate investors become more uncertain of earnings
quality because of nonaudit service decreasing auditor independence but
the decrease in earnings response coefficients due to nonaudit services
is mitigated when industry specialist provider of nonaudit services is
used. Based on the assertion that recurring and nonrecurring nonaudit
services have different economic bonding with differing implications for
auditor independence, the current study provides empirical results
indicating that negative effects of recurring nonaudit services on
earnings response coefficients are larger than negative effect of
nonrecurring nonaudit services on earnings response coefficients.
Even with no factual auditor impairment, any perception of reduced
auditor independence creates concerns for the accounting profession and
regulators. The auditor provided nonaudit services appears to be such a
situation and the current study suggests that the restriction on
nonaudit services imposed by laws like the Sarbanes-Oxly Act (Public Law
107-204, July 30, 2002) can benefit the market's perception of
auditor independence. In the spirit of self-regulation, it may be in the
accounting profession's interest to suggest self-imposed additional
restrictions to maintain auditor independence.
References
Antle, R., P.A. Griffen, D. J. Teece, and O. E. Williamson (1997),
An Economic Analysis of Auditor Independence for Multi-client,
Multi-service Public Accounting firm, Report Prepared for the AICPA by
The Law & Economics Consulting Group, Inc.
Arrunada, B. (1999), The Economics of Audit Quality: Private
Incentives and the Regulation of Audit and Non-Audit Services, Boston:
Kluwer Academic Publishers.
Ashbaugh, H., R. LaFond, and B. Mayhew (2003), "Do Nonaudit
Services Compromise Auditor Independence?" The Accounting Review,
Vol. 78, 611-639.
Atiase, R. (1985), "Predisclosure Information, Firm
Capitalization, and Security Price Behavior Around Earnings
Announcements," Journal of Accounting Research, Vol. 23, 21-36.
Beck, P. J., T. J. Frecka, and I. Solomon (1988a), "A Model of
the Market for M.AS and Audit Services: Knowledge Spillovers and
Auditor-Auditee Bonding," Journal of Accounting Literature, Spring,
50-64.
Beck, P. J., T. J. Frecka, and I. Solomon (1988b), "An
Empirical Analysis of the Relationship between MAS Involvement and
Auditor Tenure: Implication for Auditor Independence," Journal of
Accounting Literature, Spring, 65-84.
Berton, L. (1989), "SEC is Studying Independence of Firms
doing Consulting and Accounting Work," The Wall Street Journal,
November 10, A12.
Brown, L., R. Hagerman, P. Griffen, and M. Zmijewski (1987),
"An Evaluation of Alternative Proxies for the Market's
Assessment of Unexpected Earnings," Journal of Accounting and
Economics, Vol. 9(2), 159-193.
Chung, H. and S. Kallapur (2003), "Client Importance Nonaudit
Services, and Abnormal Accruals," The Accounting Review, Vol. 78,
931-955.
Collins, Daniel W., S.P. Kothari, and J. Rayburn (1987), "Firm
Size and the Information Content of Prices With Respect to
Earnings," Journal of Accounting and Economics, Vol. 9, 111-138.
Commission on Auditors' Responsibilities (2003), Report,
Conclusions, and Recommendations, Craswell, A., J. Francis, and S.
Taylor (1995), "Auditor Brand Name Reputations and Industry
Specializations," Journal of Accounting and Economies, Vol. 20,
297-322.
DeAngelo, L. (1981), "Auditor Independence, Low-Bailing, and
Disclosure Regulation," Journal of Accounting and Economics, Vol.
3, 113-127.
DeFond, M., K. Raghunandan, and K. R. Subramanyam (2002), "Do
Non-Audit Service Fees Impair Auditor Independence? Evidence from Going
Concern Audit Opinions," Journal of Accounting Research, Vol. 40,
1247-1274.
Dopuch, N., R. King, and B. Joyce (1991), "The Impact of MAS
on Auditors' Independence: An Experimental Markets Study,"
Journal of Accounting Research, Vol. 29, 60-106.
Easton, P. and M. Zmijewski (1989), "Cross-sectional Variation
in the Stock Market Response to Accounting Earnings Announcements",
Journal of Accounting and Economics Vol. 11, 117-141.
Fama, E. F., and M. C. Jensen (1983a), "Separation of
Ownership and Control," Journal of Law and Economics, Vol. 26,
301-325.
Fama, E. F., and M. C. Jensen (1983b), "Agency Problems and
Residual Claims," Journal of Law and Economics, Vol. 26, 327-350.
Foster, G. (1986), Financial Statement Analysis, Prentice Hall, NJ.
Frankel, R., M. Johnson, and K. Nelson (2002), "The Relation
between Auditors' Fees for Nonaudit Services and Earnings
Management," The Accounting Review, Vol. 77 (Supplement), 71-105.
Francis, J., and B. Ke (2006), "Disclosure of fees paid to
auditors and the market valuation of earnings surprises," Review of
Accounting Studies, Vol. 11(4), 495-523.
Ghosh, A., S. Kallapur, and D. Moon (2008), "Audit and
Non-audit Fees and Capital Market Perceptions of Auditor
Independence," Journal of Accounting and Public Policy, Vol. 28,
369-385.
Gunther, J., and R. Moore (2002), "Auditing the Auditors:
Oversight or Overkill?" Economic & Financial Policy Review,
Vol. 1 (5), 1-19.
Hackenbrack, K., and C. Hogan (2002), "Market Response to
Earnings Surprises Conditional on Reasons for an Auditor Change,"
Contemporary Accounting Research, Vol. 19(2), 195-224.
Holthausen, R., and R. Verrecchia (1988), "The Effects of
Sequential Information Releases on the Variance of Price Changes in an
Intertemporal Multi-asset Market," Journal of Accounting Research,
Vol. 26 (Spring), 82-106.
Hoskin, R., J. Hughes and w. Ricks (1986), "Evidence on the
Incremental Information Content of Additional Firm Disclosures Made
Concurrently with Earning," Journal of Accounting Research, Vol. 24
(Supplement), 1-32.
Imhoff, E., and G. Lobo (1992), "The Impact of Ex ante
Earnings Uncertainty on Earnings Response Coefficient," The
Accounting Review, Vol. 67 (April), 427-39.
Independence Standards Board (ISB) (2002), "A Conceptual
Framework for Auditor Independence," DM 00-1, New York, NY.
Kormendi, R. and R. Lipe (1987), "Earnings Innovations,
Earnings Persistence, and Stock Returns", Journal of Business, Vol.
60 (3), 323-345.
Krishnan, G. (2003), "Does Big 6 Auditor Industry Expertise
Constrain Earnings Management?" Accounting Horizons 17
(Supplement), Vol. 17, 1-16.
Kross, W. and D. Schroeder (1989), "Firm Prominence and The
Differential Information Content of Quarterly Earnings
Announcements", Journal of Businesss Finance and Accounting, Vol.
16, 55-74.
Levitt, A. (1999), Quality Information: The Lifeblood of our
Markets, remarks by Arthur Levitt at the Economic Club of New York, NY.
Levitt, A. (2000), Renewing the Covenant with Investors, remarks by
Arthur Levitt at the New York University Center for Law and Business,
New York, NY.
Levitt, A. and P. Dwyer (2002), Take on the Street: What Wall
Street and Corporate America Don't Want You to Know, Pantheon
Books, New York City.
Lipe, R. (1986), "The Information Contained in the Components
of Earnings," Journal of Accounting Research, Vol. 24,
(Supplement), 37-64.
MacDonald, E. (2001), "Was that Auditor's Opinion Really
Independent?" Forbes, March 19, 109-111.
McKinley, S., K. Panny, and P. Reckers (1985), "An Examination
of the Influence of CPA Firm Type, Size, and MAS Provision on Loan
officer Decisions and Perceptions," Journal of Accounting Research,
Vol. 23(2), 887-896.
Moreland, K. (1995), "Criticisms of Auditors and the
Association Between Earnings and Returns of Client Firms,"
Auditing: A Journal of Practice & Theory, Vol. 14 (Spring), 94-104.
Miller, R. (1974a), "The Jackknife--A review,"
Biometrika, Vol. 61(April), 1-16.
Miller, R. (1974b), "An Unbalanced Jackknife," Annals of
Statistics, Vol. 2 (September), 880-891.
Myers, J., L. Myers, and T. Omer (2003), "Exploring the Term
of the Auditor-Client Relationship and the Quality of Earnings: A Case
for Mandatory Auditor Rotation?" The Accounting Review, Vol. 78
(July), 779-799.
Owhoso, V., W. Messier, Jr., and J. Lynch (2002), "Error
Detection by Industry-Specialized Teams during Sequential Audit
Review", Journal of Accounting Research, Vol. 40 (June), 883-900.
Palmrose, Z. V. (1999), "Empirical Research on Auditor
Litigation: Consideration and Data," Studies in Accounting
Research, Vol. 33, American Accounting Association: Sarasota, FL.
Pony, K. and P. Reckers (1983), "Auditors Independence and
Non-audit Services: Director Views and Their Policy Implications,"
Journal of Accounting and Public Policy, Vol. 2 (Spring), 43-62.
Pany, K. and P. Reckers (1988), "Auditor Performance of MAS: A
Study of Its Effects on Decisions and Perceptions," Accounting
Horizons, Vol. 2, 31-38.
Parkash, M., and C. Venable (1993), "Auditee Incentives for
Auditor Independence: The Case of Nonaudit Services", The
Accounting Review, Vol. 68 (January), 113-133.
Public Oversight Board (POB) (1979), Scope of Services by CPA
Firms, AICPA: New York, NY.
Public Oversight Board (POB) (1986), Public Perceptions of
Management Advisory Services Performed by CPA Firms for Audit Clients,
AICPA: New York, NY.
Public Oversight Board (POB) (2000), Panel on Audit Effectiveness:
Report and Recommendation, AICPA: Stamford, CT.
Securities and Exchange Commission (1978), ASR No.250, U.S.
Government Printing Office.
Securities and Exchange Commission (1981), ASR No.296, U.S.
Government Printing Office.
Securities and Exchange Commission (1982), ASR No.304, U.S.
Government Printing Office.
Securities and Exchange Commission (2000a), Final Rule: Revision of
the Commission's Auditor Independence Requirements, Washington D.C.
Securities and Exchange Commission (2000b), Hearing on Auditor
Independence, Washington D.C.
Shockley, R. (1981), "Perceptions of Auditors'
Independence: An Empirical Analysis," The Accounting Review, Vol.
56 (October), 785-800.
Simunic, D. A. (1984), "Auditing, Consulting, and Auditor
Independence," Journal of Accounting Research, Vol. 18 (Autumn),
679-702.
Swanger, S. and E. Chewning Jr. (2001), "The Effects of
Internal Audit Outsourcing on Financial Analysts' Perceptions of
External Auditor Independence," Auditing: A Journal of Practice
& Theory, Vol. 20 (September), 115-129.
Teoh, S., and T. Wong (1993), "Perceived Auditor Quality and
the Earnings Response Coefficient," The Accounting Review, Vol. 68
(April), 346-66.
Turner, L. (2000), "Independence Matters," Speech at the
28th annual National Conference on Current SEC Developments.
Watts, R., and J. Zimmerman (1983), "Agency Problems,
Auditing, and the Theory of Firm: Some Evidence," Journal of Law
and Economics, Vol. 26 (October), 613-33.
Weil, J., and J. Tannebaum (2001), "Big Companies Pay Audit
Firms more for Additional Services." Wall Street Journal, April 10,
C1.
White, H. (1980), "A Heteroskedasticity-Consistent Covariance
Matrix Estimator and a Direct Test for Heteroskedasticity,"
Econometrica, Vol. 48 (May), 817-838.
CAROL LIU
Louisiana State University, Baton Rouge, LA 70803
MOHINDER PARKASH AND RAJEEV SINGHAL
Oakland University, Rochester, MI 48309
Notes
(1.) The terms nonaudit services and consulting services provided
by auditors have been used interchangeably in accounting discipline. The
accounting profession and regulators refer nonaudit services as all
services provided by auditors that do not relate to audit work. In this
sense nonaudit services are deemed to be more exhaustive than consulting
services.
(2.) The issue has been included on the agenda of various
regulatory and investigative agencies: the Commission on Auditors'
Responsibilities (1978), the Securities and Exchange Commission (SEC;
see Accounting Series Releases (ASRs) Nos. 250, 296 and 304; and Berton
(1989)); the Public Oversight Board (1979, 1986, 2000); and various
Congressional hearings held in the 1970's and 1980's.
(3.) The press also pointed out that companies were paying large
amounts of consulting fees to their auditors (Weil and Tannenbaum
(2001)) and resulting in audit failure (MacDonald (2001)).
(4.) The extant research has tested auditor independence because of
nonaudit services indirectly by analyzing outcome of audit reports or
characteristics of audited financial statements. For example, a recent
study by Frankel, Johnson, and Nelson (2002) reports more earnings
management with higher levels of nonaudit services implying auditors
have compromised to give clients higher discretion to manage earnings.
Ashbaugh, LaFond, and Mayhew (2003), Chung and Kallapur (2003) and
DeFond, Raghunandan, and Subramanyam (2002) find no evidence of auditor
compromise to challenge claims made by Frankel et al. (2002). The
results reported utilizing indirect approach are mixed.
(5.) Recently, SEC has required listed companies to disclose
nonaudit services in proxies filed after February 5, 2001. The
disclosure requirements are limited to reporting only two categories of
nonaudit services: (i) for financial information system design and
implementation and (it) all other nonaudit services. The disclosed
information does not allow the division of nonaudit services into
recurring and nonrecurring nonaudit services.
(6.) Seven industry classifications are defined as: IND1--Metal,
Ores, Mining Works & Products; IND2--Consumer Goods; IND3--Paper and
Publishing; IND4--Petroleum and Chemicals; IND5-Electric Machinery;
IND6--Motor Vehicles & Other Machinery; and IND7--Other
Nonregulated.
Table 1
1978-80 Yearly Distribution of Total, Recurring, and Nonrecurring
Nonaudit Fees as a Percentage of Audit Fees
Year N Total Nonaudit Recurring Nonaudit
Mean Range Mean Range
Minimum Maximum Minimum Maximum
1978 254 25.5% 0.0% 195.0% 19.6% 0.0% 195.0%
1979 257 23.6% 0.0% 268.0% 18.2% 0.0% 263.0%
1980 266 22.1% 0.0% 207.0% 16.4% 0.0% 134.0%
1978-80 777 23.7% 0.0% 268.0% 18.0% 0.0% 263.0%
Year Nonrecurring Nonaudit
Mean Range
Minimum Maximum
1978 5.9% 0.0% 176.0%
1979 5.4% 0.0% 66.0%
1980 5.7% 0.0% 73.0%
1978-80 5.7% 0.0% 176.0%
# Total Nonaudit includes nonaudit services of tax services; pension
and personnel services; system services, merger & acquisition services;
and other services.
Recurring Nonaudit includes nonaudit services of tax services; pension
and personnel services; and system services.
Nonrecurring Nonaudit includes nonaudit services of merger and
acquisition services; and other services not classified as Recurring
Nonaudit.
Table 2
Panel A: 1978-80 Distribution of Total Nonaudit Fees as a Percentage
of Audit Fees by Industry and Auditor
Industries
INDI # IND2 IND3 IND4 IND5
Arthur Andersen 66.2%@ 47.8%@ 39.5%@ 37.4%@ 28.4%
Arthur Young 14.5% 12.0% 23.4%@ 26.8%@ 33.7%@
Coopers & Lybrand 13.0% 12.8% 16.5% 22.2% 26.7%
Deloitte Haskin & Sells 18.5% 35.8%@ 14.7% 18.7% 24.8%
Ernst & Whinney 15.4% 17.5% 19.9% 14.2% 13.2%
Peat Marwick 11.0% 10.0% 1.3% 23.2% 23.9%
Price Waterhouse 23.2%@ 18.0% 12.1% 23.4% 33.8%@
Touche Ross 15.0% 21.3% 16.7% 0.0% 59.2%@
Industry Mean 23.1% 23.2% 21.6% 24.6% 27.2%
Industries
Auditor
IND6 IND7 Mean
Arthur Andersen 39.8%@ 72.2%@ 43.6%
Arthur Young 33.3%@ 11.3% 22.6%
Coopers & Lybrand 14.4% 11.3% 17.3%
Deloitte Haskin & Sells 14.3% 31.7%@ 19.5%
Ernst & Whinney 20.5% 3.0% 16.6%
Peat Marwick 17.9% 19.3% 18.7%
Price Waterhouse 17.2% 17.9% 20.3%
Touche Ross 15.5% 18.3% 23.5%
Industry Mean 22.0% 25.1% 23.7%
Panel B:
1978-80 Distribution of Recurring Nonaudit Fees as a Percentage of
Audit Fees by Industry and Auditor
Industries
IND1 IND2 IND3 IND4 IND5
Arthur Andersen 65.9%@ 42.5%@ 31.7%@ 29.1%@ 22.0%
Arthur Young 9.5% 6.1% 18.4%@ 18.2%@ 28.4%@
Coopers & Lybrand 7.7% 10.6% 10.0% 15.1% 24.7%
Deloitte Haskin & Sells 15.2% 23.1%@ 10.7% 10.0% 14.8%
Ernst & Whinney 8.8% 11.5% 14.0% 9.7% 7.8%
Peat Marwick 10.3% 6.0% 0.7% 18.2%@ 19.9%
Price Waterhouse 19.3%@ 11.6% 9.9% 14.7% 30.0%@
Touche Ross 14.8% 11.8% 15.7% 0.0% 15.2%
Industry Mean 17.9% 17.5% 16.8% 17.4% 20.3%
Industries
Auditor
IND6 IND7 Mean
Arthur Andersen 31.6%@ 70.1%@ 36.9%
Arthur Young 12.2% 5.8% 15.3%
Coopers & Lybrand 10.4% 7.2% 13.0%
Deloitte Haskin & Sells 11.1% 31.7%@ 13.5%
Ernst & Whinney 17.9% 2.3% 11.8%
Peat Marwick 14.2% 13.8% 14.5%
Price Waterhouse 11.8% 14.0% 15.2%
Touche Ross 12.8% 8.5% 13.0%
Industry Mean 16.7% 21.0% 17.9%
# Seven industry classifications are defined as: IND1 - Metal, Ores,
Mining Works & Products; IND2 - Consumer Goods; IND3 - Paper and
Publishing; IND4 - Petroleum and Chemicals; IND5 - Electric Machinery;
IND6 - Motor Vehicles & Other Machinery; and IND7 - Other Nonregulated.
Bold numbers indicate industry specialist classification
Note: Bold numbers indicated with @ industry specialist classification.
Table 3
Regression Results of Excess Returns on Unexpected Earnings, Other
Control Varibles and Total Nonaudit Services
CAR# = a + [b.sub.1] UE + [b.sub.2] 2 UE*NONADT + [b.sub.3]
UE*NONADT*[INDSP.sub.a] + [b.sub.4] UPDATE + [b.sub.5] LMV
+ [b.sub.6] UE*BETA + [b.sub.7] UE*GROWTH + e.
Variable Coefficient Coeffi- Jackknife t-value Significance
Coefficient/Predicted cients
Sign
Intercept 0.0033 0.60 0.75
a/?
UE 0.4872 4.61 0.01
[b.sub.1]/+
UE*NONADT -1.0408 -3.52 0.01
[b.sub.2]/-
UE*NONADT*[INDSP.sub.a] 0.7766 3.05 0.01
[b.sub.3]/+
UPDATE -0.0991 -7.53 0.01
[b.sub.4]/-
LMV -0.0008 -1.67 0.10
[b.sub.5]/-
UE*BETA -0.3238 -3.60 0.01
[b.sub.6]/-
UE*GROWTH 0.0058 1.47 0.15
[b.sub.7]/+
N 777
Adjusted R-square 0.084
# CAR is excess returns over a holding period of three days around
earnings announcement day UE is unexpected earnings defined as the
difference of reported earnings per share and analysts' most recent
forecasts deflated by price per share
NONADT is the ratio of total nonaudit fees to total audit fees.
[INDSP.sub.a] takes the value 1 when the auditor is an industry
specialist for total nonaudit services, 0 otherwise
UPDATE is excess returns from forecast date to event period
LMV is natural log of total market value of outstanding stock
BETA is market beta
GROWTH is ratio of market value of stock and book value of equity
Table 4
Regression Results of Excess Returns on Unexpected Earnings,
Other Control Varibles with Recurring and Nonrecurring Nonaudit
Services
CAR = a + [b.sub.1] UE + [b.sub.21] UE*RECUR + [b.sub.22]
UE*NONRECUR + [b.sub.3] UE*RECUR*[INDSP.sub.r] + [b.sub.4]
UPDATE + [b.sub.5] LMV + [b.sub.6] UE*BETA + [b.sub.7] UE*GROWTH + e.
Jackknife
Variable Coefficient t-value
Coefficient/Predicted Sign
Intercept (aP) 0.0034 0.60
UE ([b.sub.1]/+) 0.4989 4.61
UE*RECUR ([b.sub.21]/-) -1.1253 -3.99
UE*NONRECUR ([b.sub.22]/-) -0.7070 -1.24
UE*RECUR*[INDSP.sub.r] ([b.sub.3]/+) 0.7576 2.94
UPDATE ([b.sub.4]/+) -0.0993 -7.53
LMV ([b.sub.5]/-) -0.0008 -1.65
UE*BETA ([b.sub.6]/-) -0.3471 -3.60
UE*GROWTH ([b.sub.7]/+) 0.0078 1.49
N 777
Adjusted R-square 0.089
Jackknife t([b.sub.21] - [b.sub.22]) -3.43
Significance 0.01
Variable Significance
Coefficient/Predicted Sign
Intercept (aP) 0.75
UE ([b.sub.1]/+) 0.01
UE*RECUR ([b.sub.21]/-) 0.01
UE*NONRECUR ([b.sub.22]/-) 0.20
UE*RECUR*[INDSP.sub.r] ([b.sub.3]/+) 0.01
UPDATE ([b.sub.4]/+) 0.01
LMV ([b.sub.5]/-) 0.10
UE*BETA ([b.sub.6]/-) 0.01
UE*GROWTH ([b.sub.7]/+) 0.15
N
Adjusted R-square
Jackknife t([b.sub.21] - [b.sub.22])
Significance
# CAR is cummulative abnormal returns over a 3-days period around
earnings announcement day
UE is unexpected earnings defined as the difference of reported
earnings per share and analysts' most recent forecasts deflated by
price per share
NONADT is the ratio of total nonaudit fees to total audit fees.
[INDSP.sub.r] takes the value 1 when the auditor is an industry
specialist for recurring nonaudit services, 0 otherwise
UPDATE is cummulative abnormal returns from forecast date to event
period
LMV is natural log of total market value of outstanding stock
BETA is market beta
GROWTH is ratio of market value of stock and book value of equity
