INTRODUCTION
The Sarbanes-Oxley Act (SOX) enacted into law in 2002 is considered
one of the most significant pieces of legislation since the Securities
Acts of 1933 and 1934. An important change to the existing regime came
with Sections 302 and 404 which require management to provide an
assessment of the design and effectiveness of firms' internal
controls, as well as auditors to provide an opinion on management's
assessment of controls. In addition, Auditing Standard No. 2 now
requires that auditors provide a separate opinion on firms'
internal controls based on an independent evaluation. As a result of
these mandates and the corresponding increase in scrutiny of internal
controls, a number of companies received adverse opinions on their
internal controls, which likely impacted to some extent the relationship
with their auditors. Despite internal control problems, a client and an
auditor may decide to continue their engagement and work through the
problems together. Alternatively, a client may change audit firms
because of irreparable damage to the relationship due to the conflict,
or in order to seek another firm that may help the client earn an
unqualified opinion. This study contributes to the growing literature on
internal controls and the impact of SOX by investigating the factors
that affect the decision to switch auditors, either by dismissal or
resignation, following the issuance of an adverse opinion on internal
controls over financial reporting (ICOFR) by auditors.
Prior research shows that auditor switches are related to the
issuance of qualified audit opinions and going-concern reports (Chow
& Rice, 1982; Mutchler, 1984). In addition, auditor turnover is more
likely given internal control deficiencies disclosure pursuant to
Section 302 and 404 of SOX (Ashbaugh-Skaife, Collins & Kinney Jr.,
2007; Ettredge, Heintz, Li & Scholz, 2007). We extend this line of
research by focusing on a sample of firms that received an adverse
opinion on internal controls, and examine the specific factors that
affect auditor turnover, including the severity and nature of the
internal control deficiencies, the amount of auditor-related fees, the
length of the auditor-client relationship, and the type of audit firm
that expressed the negative opinion.
Deficiencies in internal controls have been associated with poor
accrual quality (Ashbaugh-Skaife, Collins, Kinney Jr. & LaFond,
2008), poor board and audit committee quality (Krishnan, 2005; Zhang,
Zhou & Zhou, 2007; Hoitash, Hoitash & Bedard, 2009), firm risk
(Ashbaugh-Skaife, Collins, Kinney Jr. & LaFond, 2009), and the cost
of equity capital (Ogneva, Subramanyam & Raghunandan, 2007;
Ashbaugh-Skaife, Collins, Kinney Jr. & LaFond, 2009). In addition,
firms with more severe internal control weaknesses tend to be smaller,
financially weaker, have more complex operations and fewer resources (Ge
& McVay, 2005; Ashbaugh-Skaife, Collins, & Kinney Jr., 2007;
Doyle, Ge & McVay, 2007). These firms also experience a higher drop
in share price when control problems are disclosed (Hammersley, Myers
& Shakespeare, 2008). These prior studies suggest that firms with
disclosed deficiencies in internal controls are significantly
disadvantaged relative to other firms in their access to audit services
because they pose risks that auditors may be unwilling to take.
Consistently, Raghunandan and Rama (2006) and Hogan and Wilkins (2008)
find that firms with internal control deficiencies pay higher audit
fees. Such firms have a strong incentive to change auditors, and auditor
switches have been shown to be associated with a decrease, or less of an
increase in audit fees (Simon & Francis, 1988; Ettredge &
Greenberg, 1990). However, a majority of firms with internal control
deficiencies remain with their incumbent auditors (Hall & Bennett,
2010), posing the question: what are the factors that prompt firms with
internal control weaknesses to switch auditors? In this paper, we
further examine what motivates a firm to dismiss their auditor or the
auditor to resign from an engagement. We also explore the factors that
affect firms' decisions to switch from one Big Four auditor to
another, versus switching from a Big Four to a non-Big Four auditor.
We find that the number of material weaknesses in internal controls
disclosed in an ICOFR examination, which are the most severe internal
control deficiencies, increases the likelihood of an auditor switch.
When the effect of the type of control weakness is examined, we find
that only entity-level weaknesses, perceived to be more severe than
account-specific deficiencies, affect auditor switching. Our results
show that the amount of auditor-related fees, the length of the
client-auditor relationship and the presence of a Big Four auditor also
affect the probability of an auditor change. When auditor dismissals and
resignations are examined separately, some interesting results emerge.
First, only resignations are affected by the severity of the internal
control weaknesses, implying that auditors shift away from potentially
risky firms, while clients do not seem to dismiss auditors solely in
light of severe problems. Second, high audit fees are an important
factor in dismissals, which is not surprising considering the weak
incentive for auditors to resign when a client is paying high fees.
Interestingly, we find that the number of disclosed material weaknesses
affects the likelihood of a switch from a Big Four to a non-Big Four
auditor, but not the change from one Big Four to another Big Four
auditor. This suggests that firms with more severe problems turn to
smaller auditors potentially looking for less conservative treatment.
Moreover, we find that firms are likely to switch from a Big Four to a
non-Big Four auditor when the auditor-related fees are high, perhaps to
decrease their future audit cost.
The results of this study should be of interest to audit firm
managers, audit committee and board members, investors, regulators and
other stakeholders. Following SOX, a large number of firms switched
auditors, and researchers and professionals alike have been trying to
explain this trend (Turner, Williams & Weirich, 2005). The new
internal controls requirements enacted in Sections 302 and 404 spurred a
great deal of research and debate about the effects of the disclosure of
significant control deficiencies. This study contributes to the
literature on this topic because it is the first of its type to focus on
a unique set of firms with internal control deficiencies and the factors
that affect the decision of these firms to switch auditors.
The remainder of the paper proceeds as follows. Section II
describes the contextual and institutional background and develops the
hypotheses to be tested. Section III describes the sample
characteristics. Section IV presents descriptive statistics and results
of univariate statistical analysis. Section V presents the models
utilized and provides results of multivariate analyses. Section VI and
VII describe our findings with respect to dismissals versus
resignations, and the effects of the presence of a Big Four audit firm
on the analyses. The final section offers our conclusions based on this
study and considerations for future research.
BACKGROUND AND HYPOTHESIS DEVELOPMENT
Internal control over financial reporting is defined by the Public
Company Accounting Oversight Board (PCAOB) as a process purported
"to provide reasonable assurance regarding the reliability of
financial reporting and the preparation of financial statements for
external purposes in accordance with generally accepted accounting
principles" (AS No. 2, PCAOB, 2004, para. 7). Prior to the
enactment of SOX in 2002, the reporting requirements for internal
control deficiencies were limited. The only statutory regulation over
all SEC registrants that required companies to maintain a system of
internal controls to ensure financial reporting according to GAAP and to
safeguard corporate assets was the Foreign Corrupt Practices Act (FCPA)
of 1977. However, this Act did not require managers to evaluate and
report on the effectiveness of existing internal controls. The only
required public disclosure of internal control problems before SOX was
in the firm's 8-K when disclosing a change in auditors (SEC, 1988).
Similarly, auditors were not obliged to publicly disclose any problems
with their clients' internal controls, although auditing standards
required that if the auditor discovered any deficiencies, referred to as
"reportable conditions", such problems should be communicated
to the audit committee or someone else with similar authority (Krishnan,
2005). These communications were disclosed publicly only if the firm
changed auditors within two years of the discovery of control problems.
Therefore, SOX substantially changed public disclosure requirements
of internal control deficiencies for both the client and the auditor.
First, Section 302 of SOX requires that a firm's CEO and CFO
evaluate and provide in periodic filings "their conclusions about
the effectiveness of their internal controls based on their
evaluation" (302 (a) (D)). In addition, the officers should
disclose to the auditor and the audit committee "all significant
deficiencies in the design or operation of internal controls" (302
(a) (5) (A)). Section 404 goes one step further and requires a formal
internal control report from auditors, which should "contain an
assessment ... of the effectiveness of the internal control structure
and procedures of the issuer for financial reporting" (404 (a)
(2)). In addition, "each registered public accounting firm that
prepares or issues the audit report for the issuer shall attest to, and
report on, the assessment made by the management of the issuer"
(404 (b)). Auditing Standard (AS) No. 2, which was later superseded by
AS No. 5, adds an additional requirement for a separate opinion on the
issuer's internal controls based on the auditor's independent
review.
Three types of internal control weaknesses are defined by the PCAOB
in AS No. 2 based on the likelihood that a misstatement of a particular
magnitude will not be prevented or detected: control deficiency,
significant deficiency, and material weakness. Each of these weaknesses
has a different level of severity with control deficiencies being the
lowest and material weaknesses being the most severe. Only material
weaknesses are required to be disclosed under Sections 302 and 404, and
they are automatically accompanied by an adverse opinion on internal
controls by the auditor. Therefore, firms with such deficiencies are the
focus of our study.
A number of researchers have examined the effects of audit,
financial, and litigation risk on audit engagements. For example,
Johnstone and Bedard (2004) analyze a single large audit firm and find
that this auditor substitutes high-risk clients for low-risk clients,
and that audit risk factors, including factors related to internal
controls, are more important in the auditor's client portfolio
management decisions than financial risk factors, such as factors
related to the client's overall economic condition. In addition,
Shu (2000) and Krishnan and Krishnan (1997) find that litigation risk is
an important factor affecting auditor resignations, and Jones and
Raghunandan (1998) find that in a period of increasing litigation risk,
larger audit firms are less likely to audit high-risk firms. However,
Landsman, Nelson and Rountree (2009) find evidence that auditor switches
in the post-Enron era are less likely to be associated with higher
client risk and are more likely to be due to misalignment between the
auditor and the client firm. Therefore, it is important to examine
separately the effect of disclosed material weaknesses in internal
controls on auditor dismissal or resignations.
Adverse audit opinions on internal controls are likely to create
conflict between auditors and clients, and the resolution of the
conflict may depend on the severity of the internal control issues. On
one hand, the client firm may try to work with the auditor, relying on
the audit team's expertise to resolve the control problems. On the
other hand, the client may disagree with the auditor, and may blame the
auditor for being too conservative. Consistent with this idea, Krishnan
(1994) finds evidence that switching companies receive more conservative
treatment from their successor auditor than non-switching firms, and
that the switch rate is higher when the predecessor's audit opinion
is based on more conservative standards. This suggests that if a firm
believes it is treated unfairly, it will choose to change auditors and
look for less conservative treatment elsewhere. However, in the presence
of a great number of material weaknesses, the auditor may perceive the
client to be too risky and choose to resign from the engagement. Hence,
our first hypothesis tests whether the number of reported material
weaknesses affects the probability of auditor turnover (all hypotheses
are stated in the alternative form):
H1 Firms are more likely to switch auditors after receiving an
adverse ICOFR opinion if they report a high number of material
weaknesses in internal controls.
Internal control weaknesses may be classified as account-specific
and entity-level deficiencies. Account-specific weaknesses are
considered less severe because they are related to specific financial
statements accounts, such as inventory, receivables and intangibles. In
contrast, entity-level deficiencies affect broader areas, such as
revenue recognition and segregation of duties and indicate an
organization-wide weak control environment. We expect both types of
weaknesses to increase the probability of auditor turnover, but the
degree of their incremental effect is an empirical question, tested with
the following hypotheses:
H1a Firms are more likely to switch auditors after receiving an
adverse ICOFR opinion if they report a high number of account-specific
weaknesses in internal controls.
H1b Firms are more likely to switch auditors after receiving an
adverse ICOFR opinion if they report a high number of entity-level
weaknesses in internal controls.
In the presence of an adverse ICOFR opinion, the conflict between
the client and the auditor may be exacerbated by the amount of fees
charged by the auditor. Therefore, fees may be an additional factor that
affects the likelihood of auditor realignment. Prior research by Hogan
and Wilkins (2008) shows that the presence of internal control
deficiencies is associated with higher audit fees prior to the
disclosure of the deficiencies, which they interpret as evidence of
audit firms exerting higher effort when auditing a client with weak
internal controls.
Raghunandan and Rama (2006) show that audit fees have increased
post-SOX, and they find some evidence that firms with material
weaknesses pay higher audit fees than firms without internal control
problems. Ettredge, Li and Scholz (2007) use early post-SOX data and
find that firms paying higher audit fees are more likely to dismiss
their auditor. In addition, Simon and Francis (1988) and Ettredge and
Greenberg (1990), among others, show that audit fees decrease following
an auditor switch. Hence, we surmise that firms may be willing to change
auditors following an adverse ICOFR opinion to decrease their
auditor-related costs. This leads to our second main hypothesis:
H2 Firms are more likely to switch auditors after receiving an
adverse ICOFR opinion if their auditor-related fees are higher.
While the strength and nature of the relationship between a client
and an auditor may not be observed or measured, it is likely that longer
lasting engagements are associated with the auditor having better
knowledge of the client's operations and organizational
environment. Prior research suggests that a company's financial
statements are affected by negotiations between auditors and their
clients, which are affected by the auditor-client history (Gibbins,
Salterio & Webb, 2001; Hatfield, Agoglia & Sanchez, 2008).
Moreover, accounting disclosures may be viewed as being the product of
the joint efforts of the auditor and the client (Antle & Nalebuff,
1991).
Management theorists and sociologists have studied professional
service provider/client relationships in an effort to understand the
dynamics affecting behaviors exhibited between and within the exchange
agents. The ties between audit firms and clients go beyond the
contractual relationship, because professional accountants possess a
specific body of knowledge outside the technical capacity of the client,
whose output is intangible but quite valuable. These ties are strong
because one or both parties make investments in human and social capital
to enhance the longevity of the relationship (Levinthal & Fichman,
1988). Embeddedness theory holds that the various economic actors or
exchange agents are embedded in social affiliations that create economic
value. This value is enhanced as the parties build trust and share
private information. Embeddedness theory acknowledges that people often
guide their choices based on past interactions and continue to deal with
those they trust, while economic theory holds that behavior is affected
primarily by the forces of the market (Granovetter, 1985; Uzzi &
Lancaster, 2004). This implies that the nature of the auditor-client
relationship is likely to affect the decision to change auditors even in
the presence of an adverse ICOFR opinion (Hall & Bennett, 2010). The
longer this relationship lasts, the stronger the bond between the client
and the auditor and the more likely that the firms stays with the
incumbent auditor. Therefore, the length of the client-auditor
relationship is likely to have a negative effect on the probability of
an auditor change, even when the auditor gives an adverse audit opinion.
Hence, we test the following hypothesis:
H3 Firms are less likely to switch auditors after receiving an
adverse ICOFR opinion if they have a longer relationship with their
auditor.
Shu (2000) finds that firms are more likely to employ a small
(versus Big Four) auditor following an auditor resignation if the
resignation is due to increased litigation risk. This suggests that
higher risk clients, including firms with internal control weaknesses,
may be more likely to shift away from Big Four auditors. In addition,
Ettredge, Heintz, Li and Scholz (2007) find that smaller firms tend to
dismiss their Big Four auditors and subsequently, hire a smaller
auditor. However, it is not clear ex ante how the presence of a Big Four
auditor affects switching within the sample of firms that received an
adverse internal controls opinion. Therefore, we test the following
hypothesis in our sample:
H4 Firms are more likely to switch auditors after receiving an
adverse ICOFR opinion if their auditor is one of the Big Four.
SAMPLE
The sample for this study was collected from Audit Analytics and
includes firms that received an adverse internal control opinion in one
reporting year, followed by an unqualified opinion in the next year. We
restrict the sample in this way to avoid firms with continuing internal
controls issues. The sample period includes adverse ICOFR opinions from
year 2004, when Section 404 reporting requirements first went into
effect for accelerated filers, to year 2007. This procedure yields a
total of 765 valid firms, of which 649 remained with their auditor and
116 switched auditors in the year after the adverse opinion. The two
categories of firms are labeled "loyal" and
"switcher". Tables 1 through 3 provide relevant sample
characteristics for each of these groups.
Panel A of Table 1 lists the number of firms that received adverse
ICOFR opinions by fiscal year. The table shows that year 2007 is
underrepresented in the sample, relative to the other years, 5.5 percent
vs. 28.9 percent for both 2004 and 2006 and 36.7 percent for 2005, which
is due to the limited availability of data for 2008 at the time the
sample is collected. We include these firms in the sample since the set
of years examined has no direct bearing on the hypotheses being tested.
When the sets of fiscal years are examined with respect to the behavior
of clients (loyal versus switcher), the data show that the majority of
firms remain with their auditors; 649 remain loyal, while 116 switch. In
addition, there seems to be a slight increase in loyalty from 2004 to
2007. Namely, the percent of firms experiencing first an adverse and
then an unqualified opinion on ICOFR that remained with their existing
auditors increased from 80.9 percent in 2004 (179 of 221) to over 88
percent in 2006 (196 of 221) and 2007 (37 of 42).
Panels B and C of Table 1 provide information about size, measured
by average market capitalization and average total assets between the
year of the adverse ICOFR opinion and the year of the clean opinion.
While all size brackets are well-represented in the sample, both panels
suggest that loyal companies tend to be larger than switchers, 42.1
percent and 46.3 percent of loyal firms are in the largest bracket (more
than $750 million of average market capitalization and average total
assets, respectively), while only 21.1 percent and 27.9 percent of
switcher firms are in this bracket. At the same time 46.8 percent and
45.0 percent of switching firms tend to be smallest in terms of market
capitalization and average total assets, respectively, compared with
only 27.8 percent and 24.7 percent of loyal firms. Overall, switching
firms in the sample are smaller than loyal firms.
Table 2 presents the client company sample by industry under the
SIC code system. Client companies in manufacturing represent the largest
portion (37.1 percent) of the sample, followed by Services (19.6
percent) and Finance, Insurance and Real Estate (17.6 percent). Overall,
there does not seem to be a concentration of loyal or switcher firms
within industry.
Table 3 presents the sample distribution based on auditor in the
year of the adverse ICOFR opinion (Panel A) and new auditor in the year
following the adverse opinion for switcher firms (Panel B). The table
shows that 592 (77.4 percent) of sample firms employ a Big Four auditor
in the year of the adverse opinion, although a lesser percentage of
these firms change their auditor in the following year (16.0 percent
switch vs. 84.0 percent remain loyal). However, the switch rate of firms
with non-Big Four auditors is even lower at 12.1 percent. Interestingly,
within the switcher sub-sample, 61.0 percent of firms that were
originally with a Big Four auditor switch to a non- Big Four auditor. In
addition, from the firms that change auditors but remain with a Big
Four, the majority (83.8 percent) dismiss the auditor, while only 16.2
percent have their auditor resign. In the total sample of switchers only
21 (18.1 percent) firms were with a non-Big Four auditor in the year of
the adverse ICOFR opinion and most (61.9 percent) switched to another
non-Big Four auditor.
DESCRIPTIVE STATISTICS AND UNIVARIATE RESULTS
Descriptive statistics for the variables of interest are presented
in Panel A of Table 4. In addition, the panel presents the results of
univariate tests that assess the comparisons between loyal and switcher
firms. Variable definitions are found in Appendix A. The results show
that switcher firms have, on average, a higher number of material
weaknesses, a mean of 2.19 for switchers and 1.67 for loyals, and the
difference is statistically significant, providing initial support for
H1. Moreover, these firms have higher numbers of both account-specific
and entity-level weaknesses, as H1a and H1b predict, and the difference
in the number of entity-level weaknesses is highly statistically
significant (t-statistic of 2.88), although the t-test for a difference
in means shows only marginal significance for account-based weaknesses
(t-statistic of 1.90). Switcher firms also pay higher fees to their
former auditor as a percentage of total assets and the differences are
statistically significant, providing preliminary support for H2. Loyal
firms have longer relationships with their auditors, a mean of almost 65
months by the end of the adverse ICOFR opinion year, versus a mean of 51
months for switcher firms. The difference is significant at the 0.01
level, lending support to H3. While more switchers employ Big Four
auditors (81.9 percent versus 76.6 percent for loyal firms), the
difference is not statistically significant. Finally, as suspected from
the sample descriptive statistics, loyal firms are larger than
switchers. Therefore, with the exception of the variable Big Four
Auditor, the descriptive statistics in Table 4, Panel A, support our
predictions about the factors that affect auditor switching behavior.
Panel B of Table 4 provides Pearson correlation coefficients among
the variables of interest. There are several interesting insights that
stand out from this panel. First, Material Weak is significantly
negatively associated with Tenure, which suggests that firms with longer
relationships with their auditors tend to have a lower number of
material weaknesses. This is interesting because it may suggest that
auditors with a stronger relationship with their clients may apply less
conservative standards in their evaluation of internal controls.
Alternatively, auditors may tend to stay longer with clients that are
less risky, or those with stronger internal controls. Firms with Big
Four auditors have a higher number of account-specific weaknesses,
perceived to be of lower severity, while company size is not correlated
with the number of account-specific weaknesses. In contrast, the
correlation between a Big Four auditor and entity-level weaknesses is
only marginally significant at the 0.10 percent level, while client
company size is significantly correlated with the number of entity-level
weaknesses. This suggests that firms employing Big Four auditors are
more likely to have account-specific weaknesses, and larger clients are
more likely to have a higher number of entity-level weaknesses but not
necessarily a higher number of account-specific weaknesses. Overall, the
univariate results are consistent with our hypotheses outlined in
Section II. Next, we test our predictions using multivariate analysis.
MULTIVARIATE ANALYSIS
We model a firm's decision to change its audit firm using
logistic regression including the following constructs. Additional
factors that may be associated with auditor changes were considered,
including growth, leverage, distress, management change and unfavorable
audit opinions. To control for the potential confounding effect of these
variables, we repeated the analyses including these constructs (none of
these variables are significant in any of the model specifications and
the results presented in this paper are unaffected). We construct the
following regression model to test our hypotheses:
Prob(Switch) = f([[beta].sub.0] + [[beta].sub.1] MaterialWeak +
[[beta].sub.2] PctTotalfees + [[beta].sub.3] Tenure + [[beta].sub.4]
BigFourAuditor + [[beta].sub.5] AveMcap + [[summation].sup.K.sub.k=1]
[[gamma].sub.k]Industry), (1)
where Switch is an indicator variable that is equal to one if the
firm changes its auditor, and zero otherwise; MaterialWeak is the number
of material internal control weaknesses; PctTotalfees is total fees paid
to the auditor, including audit and non-audit fees, divided by total
assets; Tenure is the log of the length of the relation in months
between the firm and the auditor until the end of the year of the
adverse ICOFR opinion; BigFourAuditor is equal to one if the audit firm
is one of the Big Four, zero otherwise; and AveMCap is the log of the
average market capitalization in the year of the adverse opinion and the
following year. In addition, nine industry indicator variables are
included in the model. All variables are summarized in Appendix A.
Equation (1) includes the variable MaterialWeak for the number of
material internal control weaknesses, both account-specific and
entity-level. To gain insight into which internal control weakness has
an effect on the decision to change the audit firm, we run a model with
two separate variables for the number of account-specific and
entity-level weaknesses:
Prob(Switch) = f([[beta].sub.0] + [[delta].sub.1] AccruleReasons +
[[delta].sub.2] EntityReasons + [[delta].sub.3] PctTotalfees +
[[delta].sub.4] Tenure + [[delta].sub.5] BigFourAuditor +
[[delta].sub.6] AveMcap + [[summation].sup.K.sub.k=1]
[[gamma].sub.k]Industry), (2)
where AccruleReasons is the number of account-specific weaknesses
and EntityReasons is the number of entity-level weaknesses as specified
by Audit Analytics. In its detailed list of internal controls
deficiencies reasons, Audit Analytics includes issues that are
classified as material weaknesses as well as those that are less severe.
Hence, the account-specific and entity-level weaknesses may add up to a
number that is higher than the number of material weakness. All other
variables are as defined above and in Appendix A.
We present the results of our main analysis in Table 5. The first
column presents the results from Equation (1) and the second column
displays the results from Equation (2). Consistent with H1, the
coefficient on MaterialWeak is positive and statistically significant at
the 0.05 level, suggesting that firms that have a higher number of
material weaknesses are more likely to change auditors after an adverse
ICOFR opinion. When examining account-specific and entity-level
weaknesses separately, only the coefficient on EntityReasons is
significant, but only marginally, with a p-value of 0.06. Therefore,
there is no support for H1a that firms with a higher number of
account-specific weaknesses are more likely to switch auditors and only
limited support for H1b that the number of entity-level weaknesses
increases the probability that the firm changes its auditor.
The second hypothesis of this study posits that higher
auditor-related fees increase the probability that a firm switches
auditors after an adverse ICOFR opinion. Table 5 provides support for
this hypothesis. The importance of this variable is examined by
measuring the relative magnitude of fees as a percentage of total
assets. Specifically, the coefficient on PctTotalfees is positive and
statistically significant at the 0.05 level, suggesting that higher fees
increase the probability of a change in audit firm subsequent to an
unfavorable ICOFR opinion, in support of H2.
Further, H3 predicts that the longer the relation between a firm
and its auditor, the less likely that the company will switch auditors
when an adverse opinion on internal controls is issued. The coefficient
on Tenure is negative and highly statistically significant, suggesting
that this is indeed the case. Finally, H4 suggests that firms that
employ a Big Four auditor and receive an adverse ICOFR opinion are more
likely to change their auditor. The coefficient on BigFourAuditor is
positive and highly significant, providing support to H4.
AUDITOR DISMISSALS AND RESIGNATIONS
Next, we examine sub-samples of firms that dismiss their auditors,
and firms whose auditors resign, since Krishnan and Krishnan (1997)
suggest that factors affecting resignations and dismissals may differ.
In the first two columns of Table 6, the test sample includes firms that
dismissed their auditor after receiving adverse ICOFR opinion (Dismiss
sample) and the control sample includes the loyal firms. In the last two
columns, the firms that had their auditor resign are examined (Resign
sample). Hence, the dependent variable in Table 6 is equal to one if the
firm dismissed its auditor (Columns 1 and 2) or the auditor resigned
(Columns 3 and 4), and zero if the firm remained loyal. Several
interesting results emerge from the analysis of dismissals versus
resignations. First, the number of reported material weaknesses affects
only the probability that the auditor resigns and not the probability
that the audit firm is dismissed. The coefficient on MaterialWeak is
positive and highly significant at the 0.001 level for the Resign
sample, while it is negative and insignificant for the Dismiss sample.
Moreover, when the weaknesses are separated into account-specific and
entity-level, the coefficient on EntityReasons is positive and highly
significant (p-value 0.006) suggesting that the auditor is more likely
to resign when the firm has more entity-level internal control issues,
while the number and nature of material weaknesses do not appear to
affect the decision of the firm to dismiss their auditor. This also
suggests that the results in Table 5 for EntityReasons are driven by the
Resign sample. Overall, the results suggest that firms do not seem to
blame the auditor for disclosing their internal control issues and
therefore, do not appear to subsequently dismiss the auditor, although
auditors seem to prefer to avoid firms with multiple material weaknesses
by resigning, due likely to the significant legal risks associated with
such clients.
Second, fees do not have an effect on the probability that the
auditor resigns, although higher fees increase the probability that the
auditor is dismissed. The latter result is consistent with the evidence
found by Ettredge, Li and Scholz (2007), who examine a wider population
of firms in the early post-SOX era. The coefficient on PctTotalfees is
insignificant for the Resign sample and is positive and statistically
significant at the 0.05 level for the Dismiss sample. This implies that
higher relative fees increase the probability that the auditor is
dismissed after an adverse ICOFR opinion is expressed. Perhaps an
adverse internal controls opinion creates an irreparable conflict when
the auditor-related fees are high, which may lead to engagement
termination. This is an interesting result worthy of further
investigation.
Third, the results in Table 6 also show that after the effect of
the severity and nature of internal controls issues is taken into
consideration, Big Four auditors are more likely to be dismissed after
unfavorable ICOFR opinions, while Big Four auditors are not more likely
to resign, all else equal. This suggests that in the presence of a weak
internal controls environment, firms are more likely to dismiss a Big
Four auditor. Big Four auditors do not seem to resign from a client
following the disclosure of internal controls weaknesses, however, this
may be due to the fact that the sample is restricted to firms receiving
adverse internal controls opinions, which may limit the power of the
test in this sample. Another reason for the lack of significance of
BigFourAuditor in the Resign sample may be that Big Four auditors have a
large number of clients and they may be able to afford the opportunity
to minimize the additional risk associated with clients experiencing
internal control problems.
FIRMS WITH BIG FOUR AUDITORS
Next, we analyze the subset of firms that had a Big Four auditor in
the year of the adverse opinion and then switched either to another Big
Four or to a non-Big Four auditor. Again, the control sample includes
firms that did not switch auditors. Two interesting results are shown in
Table 7. First, the number of material weaknesses affects only the
probability that the client switches from a Big Four to a non-Big Four
and not to another Big Four firm. The coefficient on MaterialWeak is
positive and highly statistically significant with a p-value of 0.002.
This implies that firms with a higher number of material internal
control weaknesses are more likely to turn to a smaller auditor
subsequent to an adverse ICOFR opinion, possibly looking for more
lenient treatment in future years, since Big Four auditors are
considered to be more conservative. This result is consistent with Shu
(2000) who find that high-litigation risk firms are more likely to be
dropped by a large audit firm and to subsequently engage a smaller audit
firm, and Raghunandan and Rama (1999) who show that a large audit firm
is less likely to become a successor auditor when the predecessor has
resigned.
Second, the coefficient estimates for PctTotalfees provide
contrasting results for the two sub-samples. On the one hand, the higher
the fees, the higher the probability that the client goes from a Big
Four to a non-Big Four audit firm, which suggests that the client may be
trying to lower its audit cost subsequent to receiving an adverse ICOFR
opinion. This result is consistent with Ettredge, Heintz, Li and Scholz
(2007) who find that smaller firms that pay higher audit fees tend to
dismiss their Big Four auditor and subsequently hire a non-Big Four
auditor. On the other hand, the coefficient on this variable in the
sub-sample of firms that switch from one Big Four auditor to another is
negative and marginally significant (p-value 0.054), implying that the
higher the fees, the lower the probability that the client switches from
one Big Four to another Big Four firm. This suggests that when internal
controls have been assessed as weak and the audit fees are high, the
client is better off not changing its auditor. This is an interesting
result that should be examined more extensively in future research.
However, it should be noted that the sample size of switcher firms is
quite small and the results should be interpreted with caution.
CONCLUSION
Recent changes in the regulatory environment have significantly
expanded the disclosure requirements pertaining to firms' internal
controls. On one hand, Sections 302 and 404 of SOX require that managers
provide an assessment of the design and effectiveness of their
firms' internal controls and auditors express an opinion on this
assessment. On the other hand, Auditing Standard No. 2 requires that
auditors provide a separate opinion on internal controls based on their
independent examination. Disclosing internal control problems is not
looked upon favorably in the market (Hammersley, Myers &
Shakespeare, 2008), and hence, adverse ICOFR opinions are likely to
impact auditor-client relationships negatively, which may result in
engagement termination. In this paper, we study the factors that affect
auditor switches following adverse ICOFR opinions. This study is unique
in that it focuses on firms that received adverse opinions in one year,
followed by an unqualified opinion in the next, isolating the sample
from firms with longstanding or endemic internal control weaknesses.
We find that the probability of firms switching auditors in the
year following an adverse ICOFR opinion increases with the number of
material weaknesses, which are the most severe problems in internal
controls. An examination of the type of internal control issues firms
face shows that only entity-level and not account-specific weaknesses
increase this probability. This suggests that firms with more severe
internal controls issues are more likely to change auditors in an effort
to achieve an unqualified opinion. In addition, the amount of
auditor-related fees and the presence of a Big Four audit firms also
increase the probability of a switch, while the length of the
auditor-client relationship decreases the probability of a switch.
Several interesting results emerge when dismissals and resignation
are examined separately. First, severe internal control issues affect
only the probability of an auditor resigning, implying that auditors try
to stay away from risky firms. Second, the magnitude of auditor-related
fees and the presence of a Big Four auditor only increase the
probability of an auditor dismissal. This suggests that when faced with
an adverse internal controls opinion, firms tend to dismiss their
auditor when they are paying high fees and their incumbent auditor is
one of the Big Four. Dismissal may be due to a firm's desire to
decrease their audit costs and/or look for less conservative treatment
from a smaller audit firm. Auditor tenure decreases the probability of
both auditor dismissal and resignation, although its effect on
dismissals is much stronger. This implies that clients and auditors have
an investment in their relationship that is strengthened over time and
is less likely to be terminated as a result of an adverse opinion.
An examination of the switches from a Big Four to another Big Four
or to a non-Big Four audit firm reveals that the number of material
weaknesses and the amount of auditor-related fees increase the
probability of a switch to a smaller auditor. This is consistent with
the idea that firms tend to switch from large audit firms to smaller
auditors, either to get less conservative treatment or to decrease their
audit cost. In addition, firms tend to stay loyal, rather than switch to
another Big Four firm when they have a longer tenure with their current
auditor.
Overall, the results suggest that the number and severity of the
internal control deficiencies, the amount of auditor-related fees,
auditor tenure, and the presence of a Big Four auditor affect the
probability of an auditor switch in the year following an adverse ICOFR
opinion. However, these factors affect auditor dismissals and
resignations differently suggesting that it is important to consider the
underlying reason for the switch. Switching from one Big Four audit firm
to another or to a non-Big Four auditor is also affected by the severity
of internal control weaknesses, audit costs, and the length of the
client/auditor relationship.
The results of this study are relevant and useful to a variety of
audiences. First, the results provide evidence of the significant
effects of the Sarbanes-Oxley legislation, especially the effects of
Sections 302 and 404 on client-auditor relationships. The sample is
current and unique, focusing the results on the critical point in the
client-auditor relationship where a decision to switch or not is likely
to occur. Therefore, these findings should be useful to audit firm
managers, audit committee and board members, investors, regulators and
other stakeholders interested in the impact of SOX on firm behavior. In
addition, the study provides insight into auditor switching behavior,
prompted either by the client or the audit firm. It also draws attention
to the prevalence of loyal or non-switching behavior which can be best
explained by embeddedness theory. This is important because while
clients may engage in opinion-shopping or audit cost-minimizing
behavior, and audit firms may take on risk-reducing actions, some
relationships are maintained and sustain throughout adverse conditions.
REFERENCES
Antle, R., & B. Nalebuff (1991). Conservatism and
auditor-client negotiations. Journal of Accounting Research,
29(Supplement), 31-54.
Ashbaugh-Skaife, H., Collins, D., Kinney Jr., W. R., & R.
LaFond (2009). The effect of SOX internal control deficiencies on firm
risk and cost of equity. Journal of Accounting Research, 47(1), 1-43.
Ashbaugh-Skaife, H., Collins, D., Kinney Jr., W. R., & R.
LaFond (2008). The effect of SOX internal control deficiencies and their
remediation on accrual quality. The Accounting Review, 83(1), 217-250.
Ashbaugh-Skaife, H., Collins, D., & W. R. Kinney Jr. (2007).
The discovery and reporting of internal control deficiencies prior to
SOX-mandated audits. Journal of Accounting and Economics, 44(1-2),
166-192.
Chow, C., & S. Rice (1982). Qualified audit opinions and
auditor switching. The Accounting Review, 57(2), 326-335.
Doyle, J., Ge, W., & S. McVay (2007). Determinants of
weaknesses in internal control over financial reporting. Journal of
Accounting and Economics, 44(1-2), 193-223.
Ettredge, M., & R. Greenberg (1990). Determinants of fee
cutting on initial audit engagements. Journal of Accounting Research,
28(1), 198-210.
Ettredge, M., Heintz, J., Li, C., & S. Scholz (2007). Auditor
realignments accompanying implementation of SOX 404 reporting
requirements. Working paper, University of Kansas.
Ettredge, M., Li, C., & S. Scholz (2007). Audit fees and
auditor dismissals in the Sarbanes-Oxley era. Accounting Horizons,
21(4), 371-386.
Ge, W., & S. McVay (2005). The disclosure of material
weaknesses in internal control after the Sarbanes-Oxley Act. Accounting
Horizons, 19(3), 137-158.
Gibbins, M., Salterio, S., & A. Webb (2001). Evidence about
auditor-client management negotiation concerning client's financial
reporting. Journal of Accounting Research, 39(3), 535-563.
Granovetter, M. (1985). Economic action and social structure: The
problem of embeddedness. American Journal of Sociology, 91, 481-510.
Hall, L., & V. Bennett (2010). Client/auditor loyalty,
embeddedness, and adverse internal control opinions. The BRC Academy
Journal of Business, 1(1), 1-24.
Hammersley, J., Myers, L., & C. Shakespeare (2008). Market
reactions to the disclosure of internal control weaknesses and to the
characteristics of those weaknesses under Section 302 of the Sarbanes
Oxley Act of 2002. Review of Accounting Studies, 13(1), 141-165.
Hatfield, R., Agoglia, C., & M. Sanchez (2008). Client
characteristics and the negotiation tactics of auditors: Implications
for financial reporting. Journal of Accounting Research, 46(5),
1183-1207.
Hogan, C., & M. Wilkins (2008). Evidence on the audit risk
model: Do auditors increase audit fess in the presence of internal
control deficiencies? Contemporary Accounting Research, 25(1), 219-242.
Hoitash, U., Hoitash, R., & J. Bedard (2009). Corporate
governance and internal control over financial reporting: A comparison
of regulatory regimes. The Accounting Review, 84(3), 839-867.
Johnstone, K., & J. Bedard (2004). Audit firm portfolio
management decisions. Journal of Accounting Research, 42(4), 659-690.
Jones, F., & K. Raghunandan (1998). Client risk and recent
changes in the market for audit services. Journal of Accounting and
Public Policy, 17(2), 169-181.
Krishnan, J. (1994). Auditor switching and conservatism. The
Accounting Review, 69(1), 200-215.
Krishnan, J. (2005). Audit committee quality and internal control:
An empirical analysis. The Accounting Review, 80(2), 649-675.
Krishnan, J., & J. Krishnan (1997). Litigation risk and auditor
resignation. The Accounting Review, 72(4), 539-560.
Landsman, W., Nelson, K. K., & B. Rountree (2009). Auditor
switches in the pre- and post- Enron eras: Risk or realignment? The
Accounting Review, 84(2), 531-558.
Levinthal, D.A., & M. Fichman (1988). Dynamics of
interorganizational attachments: Auditor-client relationships.
Administrative Science Quarterly, 33(3), 345-369.
Mutchler, J. F. (1984). Auditors' perceptions of the
going-concern opinion decision. Auditing: A Journal of Practice &
Theory, 3, 17-30.
Ogneva, M., Subramanyam, K. R., & K. Raghunandan (2007).
Internal control weakness and cost of equity: Evidence from SOX Section
404 disclosures. The Accounting Review, 82(5), 1255-1298.
Public Company Accounting Oversight Board (PCAOB) (2004). An audit
of internal control over financial reporting performed in conjunction
with an audit of Financial Statements Auditing Standard No. 2.
Washington, D.C.: PCAOB.
Raghunandan, K., & D. Rama (1999). Auditor resignations and the
market for audit services. Auditing: A Journal of Practice & Theory,
18(1), 124-134.
Raghunandan, K., & D. Rama (2006). SOX Section 404 material
weakness disclosures and audit fees. Auditing: A Journal of Practice
& Theory, 25(1), 99-114.
Securities and Exchange Commission (SEC) (1988). Disclosure
amendments to Regulation S-K, Form 8-K and Schedule 14A regarding
changes in accountants and potential opinion shopping situations
Financial Reporting Release No. 31 (April). SEC Docket 1140-1147.
Washington, D.C.: SEC.
Shu, S. (2000). Auditor resignations: Clientele effects and legal
liability. Journal of Accounting & Economics, 29(2), 173-205.
Simon, D. T., & J. R. Francis (1988). The effects of auditor
change on audit fees: Tests of price cutting and price recovery. The
Accounting Review, 63(2), 255-269.
Turner, L., Williams, J., & T. Weirich (2005). An inside look
at auditor changes. The CPA Journal, 75(11), 12-21.
Uzzi, B. & R. Lancaster (2004). Embeddedness and the price
formation in the corporate law market. American Sociological Review,
69(3), 319-344.
Zhang, Y., Zhou, J., & N. Zhou (2007). Audit committee quality,
auditor independence, and internal control weaknesses. Journal of
Accounting and Public Policy, 26(3), 300-327.
Maya Thevenot, SUNY Fredonia
Linda Hall, SUNY Fredonia
Appendix A: Variable Definitions
Variable Definition
MaterialWeak Number of material weaknesses identified in
assessment of internal controls as reported by Audit
Analytics.
AccruleReasons Number of accounting rule (GAAP/FASB) application
failures identified in assessment of internal
controls as reported by Audit Analytics.
EntityReasons Number of entity-level weaknesses identified in
assessment of internal controls as reported by Audit
Analytics.
PctTotalfees Sum of total audit and non-audit fees divided by
total assets in the year of the adverse ICOFR
opinion.
Tenure Log of the number of months the firm maintained the
same auditor until the end of the year of the adverse
ICOFR opinion.
BigFourAuditor Coded one if the firm engaged one of the Big Four
audit firms in the year of the adverse ICOFR opinion.
AveMcap Log of the average market capitalization from the
year of the adverse ICOFR opinion to the year of the
clean opinion.
AveTotalAssets Log of the average total assets from the year of the
adverse ICOFR opinion to the year of the clean
opinion.Table 1: Sample Characteristics
Panel A: Loyal Switcher Total
Opinion years
n Percentage n Percentage n Percentage
2004 179 27.6% 42 36.2% 221 28.9%
2005 237 36.5% 44 37.9% 281 36.7%
2006 196 30.2% 25 21.6% 221 28.9%
2007 37 5.7% 5 4.3% 42 5.5%
Total 649 100.0% 116 100.0% 765 100.0%
Panel B: Loyal Switcher Total
Average Market
Capitalization n Percentage n Percentage n Percentage
Less than $250
million 168 27.8% 51 46.8% 219 30.7%
$250-$500
million 111 18.3% 26 23.9% 137 19.2%
$500-$750
million 71 11.7% 9 8.3% 80 11.2%
More than $750
million 255 42.1% 23 21.1% 278 38.9%
Total 605 100.0% 109 100.0% 714 100.0%
Panel C: Loyal Switcher Total
Average Total
Assets n Percentage n Percentage n Percentage
Less than $250
million 155 24.7% 50 45.0% 205 27.7%
$250-$500
million 107 17.0% 18 16.2% 125 16.9%
$500-$750
million 75 11.9% 12 10.8% 87 11.8%
More than $750
million 291 46.3% 31 27.9% 322 43.6%
Total 628 100.0% 111 100.0% 739 100.0%
The sample was collected from Audit Analytics and includes firms
that received an adverse opinion on internal controls over
financial reporting (ICOFR) in one reporting year, followed by an
unqualified opinion in the next year. The sample period includes
adverse ICOFR opinions from year 2004 to the beginning of year
2007. Year 2007 is underrepresented in the sample due to the lack
of available data for year 2008 at the time the study was
conducted. Firms with missing data were excluded. Loyal and
Switcher designate firms that continue to employ their incumbent
auditor and firms that change auditors in the year following the
adverse ICOFR opinion, respectively.
Table 2: Industry Distribution
SIC Code by Loyal Switcher Total
Division
n Percentage n Percentage n Percentage
Agric.,
Forestry, 0 0.0% 2 1.7% 2 0.3%
Fishing 01-09
Mining 10-14 29 4.5% 3 2.6% 32 4.2%
Construction
15-17 7 1.1% 2 1.7% 9 1.2%
Manufacturing
20-39 239 36.8% 45 38.8% 284 37.1%
Trans, Comm,
Electric, Gas 59 9.1% 9 7.8% 68 8.9%
40-49
Wholesale Trade
50-51 14 2.2% 3 2.6% 17 2.2%
Retail Trade
52-59 62 9.5% 6 5.2% 68 8.9%
Finance, Ins.,
Real Estate 115 17.7% 20 17.2% 135 17.6%
60-67
Services 70-89 124 19.1% 26 22.4% 150 19.6%
Total 649 100.0% 116 100.0% 765 100.0%
Industry classification is based on the SIC code system.
Table 3: Auditor Statistics
Panel A: Auditor distribution in the year of the adverse opinion
Auditor Loyal Switcher Total
n Percentage n Percentage n Percentage
Big Four 497 76.6% 95 81.9% 592 77.4%
Non-Big Four 152 23.4% 21 18.1% 173 22.6%
Total 649 100.0% 116 100.0% 765 100.0%
Panel B: Auditor Characteristics of switching firms
Year-to-Year Dismiss Resign Switcher
Auditor
n Percentage n Percentage n Percentage
Big Four to
Big Four 31 37.8% 6 17.7% 37 31.9%
Big Four to
Non-Big Four 40 48.8% 18 52.9% 58 50.0%
Non-Big to
Non-Big Four 5 6.1% 8 23.5% 13 11.2%
Non-Big to
Big Four 6 7.3% 2 5.9% 8 6.9%
Total 82 100.0% 34 100.0% 116 100.0%
Panel A shows the sample distribution based on the type of audit firm
that expressed an adverse ICOFR opinion. Big Four includes the four
largest audit firms.
Panel B shows the sample distribution of firms that switched auditors
in the year after receiving an adverse ICOFR opinion. Dismiss and
Resign designate firms that dismissed their auditors and firms whose
auditors resigned, respectively.
Table 4: Descriptive Statistics Panel A: Summary Statistics
Loyal Switcher
Variable Mean Median SD Mean
Material Weak 1.670 1.000 1.414 2.190
AccruleReasons 2.020 2.000 1.503 2.353
EntityReasons 3.125 3.000 1.384 3.603
PctTotalfees 0.004 0.003 0.005 0.007
Tenure (in months) 64.978 74.000 27.616 50.862
Tenure (log) 4.040 4.304 0.583 3.754
Big Four Auditor 0.766 1.000 0.424 0.819
AveMcap
(in billions) 2.661 0.581 16.465 0.920
AveTotalAssets
(in billions) 7.906 0.671 54.732 1.744
AveMcap (log) 20.276 20.181 1.400 19.678
AveTotalAssets
(log) 20.448 20.325 1.749 19.760
Switcher Mean Difference
Variable Median SD t-stat
Material Weak 1.000 1.851 -2.88 ***
AccruleReasons 2.000 1.775 -1.90 *
EntityReasons 3.000 1.693 -2.88 ***
PctTotalfees 0.004 0.010 -2.70 ***
Tenure (in months) 56.000 23.463 5.80 ***
Tenure (log) 4.025 0.722 4.04 ***
Big Four Auditor 1.000 0.387 -1.34
AveMcap
(in billions) 0.274 2.063 2.49 **
AveTotalAssets
(in billions) 0.291 4.652 2.77 ***
AveMcap (log) 19.428 1.175 4.74 ***
AveTotalAssets
(log) 19.488 1.577 4.17 ***
***, **, * Significant beyond the 1, 5, and 10 percent levels,
respectively of a two-tailed test.
Variable definitions are summarized in Appendix A.
Panel B: Pearson correlations
Material AccruleReas EntityReas PctTotalfees
Weak
Switch 0.124 0.077 0.119 0.161
(0.001) (0.033) (0.001) (<.001)
Material 0.541 0.640 0.132
Weak (<.001) (<.001) (<.001)
AccruleReas 0.468 0.056
(<.001) (0.123)
EntityReas 0.057
(0.121)
PctTotalfees
Tenure
BigFourAud
Tenure BigFourAud AveMCap
Switch -0.167 0.046 -0.155
(<.001) (0.208) (<.001)
Material -0.102 -0.041 -0.064
Weak (0.005) (0.263) (0.087)
AccruleReas -0.048 0.087 0.020
(0.184) (0.016) (0.597)
EntityReas -0.034 0.067 0.083
(0.343) (0.062) (0.026)
PctTotalfees -0.098 -0.112 -0.361
(0.007) (0.002) (<.001)
Tenure 0.475 0.207
(<.001) (<.001)
BigFourAud 0.363
(<.001)
Correlation coefficients that are significant beyond the 5 percent
level are presented in bold. Variable definitions are summarized in
Appendix A.
Table 5: Logistic Regression Results
Dependent Variable = 1 if Switch, 0 if Loyal
Independent Variable Logit p-value Logit p-value
Estimate Estimate
Intercept 7.425 <.001 7.599 <.001
MaterialWeak 0.135 0.034
AccruleReasons 0.005 0.949
EntityReasons 0.152 0.068
PctTotalfees 40.296 0.049 41.878 0.048
Tenure (log) -1.013 <.001 -1.011 <.001
BigFourAuditor 1.588 <.001 1.544 <.001
AveMcap (log) -0.353 <.001 -0.373 <.001
Industry indicator variables Included Included
Number of switch firms 109 109
Number of total observations 713 713
Likelihood ratio 75.1597 76.0047
<.0001 <.0001
Pseudo R-square 10.00% 10.11%
Parameter estimates that are significant beyond the 5 percent level
are presented in bold.
Variable definitions are summarized in Appendix A.
Table 6: Logistic Regression Results by Dismiss and Resign
Dependent Variable = 1 if Dismiss or Resign, respectively, 0 if Loyal
Dismiss Sample
Independent Variable Logit p-value Logit p-value
Estimate Estimate
Intercept 6.768 0.005 6.763 0.005
MaterialWeak -0.018 0.855
AccruleReasons -0.033 0.746
EntityReasons 0.035 0.748
PctTotalfees 44.868 0.048 44.592 0.049
Tenure (log) -1.093 <.001 -1.099 <.001
BigFourAuditor 1.981 <.001 2.000 <.001
AveMcap (log) -0.329 0.004 -0.331 0.004
Industry indicator
variables Included Included
Number of Dismiss
or Resign firms 75 75
Number of total
observations 679 679
Likelihood ratio 56.3271 56.4324
<.0001 <.0001
Pseudo R-square 7.96% 7.98%
Resign Sample
Independent Variable Logit p-value Logit p-value
Estimate Estimate
Intercept 6.614 0.063 6.586 0.062
MaterialWeak 0.349 <.001
AccruleReasons 0.057 0.627
EntityReasons 0.334 0.006
PctTotalfees 12.240 0.775 22.001 0.593
Tenure (log) -0.721 0.036 -0.672 0.052
BigFourAuditor 0.886 0.099 0.654 0.220
AveMcap (log) -0.413 0.012 -0.441 0.008
Industry indicator
variables Included Included
Number of Dismiss
or Resign firms 34 34
Number of total
observations 638 638
Likelihood ratio 44.8277 43.1699
<.0001 <.0001
Pseudo R-square 6.79% 6.54%
Parameter estimates that are significant beyond the 5 percent level
are presented in bold. The first four columns present result for the
sample of firms that dismissed their auditor and the sample of firms
that did not change auditors serve as a control group. The last four
columns present the results for the sample of firms whose auditors
resigned and the same control group.
Variable definitions are summarized in Appendix A.
Table 7: Logistic Regression Results by Switch
Dependent Variable = 1 if Switch to the respective auditor, 0 if Loyal
Switch from Big 4 Switch from Big 4
to Big 4 to non-Big 4
Independent Variable Logit p-value Logit p-value
Estimate Estimate
Intercept -1.647 0.580 7.097 0.018
MaterialWeak 0.055 0.648 0.243 0.002
PctTotalfees 145.100 0.054 63.947 0.014
Tenure (log) -0.642 0.019 -0.193 0.435
AveMcap (log) 0.095 0.491 -0.483 0.001
Industry indicator variables Included Included
Number of switch firms 35 53
Number of total observations 639 657
Likelihood ratio 19.7103 54.2924
0.0728 <.0001
Pseudo R-square 3.04% 7.93%
Parameter estimates that are significant beyond the 5 percent level
are presented in bold. The first column presents the results for the
sample of firms that switched from one Big Four to another Big Four
audit firm and the sample of firms that did not change auditors serve
as a control group. The last column presents the results for the
sample of firms that switched from a Big Four to a non-Big Four audit
firm and the sample control group.
Variable definitions are summarized in Appendix A.