Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Article information:
To cite this document:
C.A. JubbK.A. HoughtonS. Butterworth, (1996),"Audit fee determinants: the plural nature of risk", Managerial Auditing
Journal, Vol. 11 Iss 3 pp. 25 - 40
Permanent link to this document:
http://dx.doi.org/10.1108/02686909610115222
Downloaded on: 18 June 2015, At: 23:24 (PT)
References: this document contains references to 74 other documents.
To copy this document: permissions@emeraldinsight.com
The fulltext of this document has been downloaded 1547 times since 2006*
Nathalie Gonthier-Besacier, Alain Schatt, (2007),"Determinants of audit fees for French quoted firms", Managerial Auditing
Journal, Vol. 22 Iss 2 pp. 139-160 http://dx.doi.org/10.1108/02686900710718654
Iain Gerrard, Keith Houghton, David Woodliff, (1994),"Audit Fees:: The Effects of Auditee, Auditor and Industry Differences",
Managerial Auditing Journal, Vol. 9 Iss 7 pp. 3-11 http://dx.doi.org/10.1108/02686909410067534
Rani Hoitash, Ariel Markelevich, Charles A. Barragato, (2007),"Auditor fees and audit quality", Managerial Auditing Journal,
Vol. 22 Iss 8 pp. 761-786 http://dx.doi.org/10.1108/02686900710819634
Access to this document was granted through an Emerald subscription provided by emerald-srm:198285 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please
visit www.emeraldinsight.com/authors for more information.
Addresses concerns regarding perceptions and measurement of risk and the resultant
confusion relating to understanding of the market for
audit services. Examines the
theoretical justification for a
plural approach to dealing
with risk in audit fee
models. Reviews the relevant
literature and undertakes a
factor analysis of 229 Western Australian firms in search
of evidence of plurality.
Argues for recognition of the
idea that risk in the audit
context is composed of two
separate but related concepts: audit risk and business
risk.
Introduction
Understanding the operation of the audit
market is of importance to both the
consumers and providers of audit services,
and those attempting to regulate that market.
Properly specified models of the determinants of audit fees, such as those introduced
by Simunic[1] and others, are one of the most
powerful means by which this market can be
analysed.
To date, most of these models have posited
theoretically and supported empirically the
view that there is a positive relationship
between some concept of risk and audit
fees[1-15]: that is to say, as audit risk increases
so too does the fee charged by the auditor for
audit services to the auditee. However, the
audit risk concept has been ill-defined in the
relevant literature. As a result, major differences have existed between researchers as to
perceptions of risk and the way in which it
has been measured. Such confusion makes it
difficult to compare the various audit fee
models across studies and introduces a
degree of confusion into our understanding of
the market for audit services.
This paper aims to address these concerns.
Specifically, it argues that the audit fee literature should recognize the theoretical auditing literature which notes that risk, in the
audit context (not necessarily identical to the
notion of risk often used in the finance literature) is composed of two distinct but related
concepts; audit risk and business risk.
This paper examines the theoretical justification for a plural approach to dealing with
risk in audit fee models. It reviews the
literature to determine the theoretical justification for, and operationalization of, the most
frequently utilized risk measures. Factor
analysis of selected audit fee determinants is
then undertaken on a sample of 229 listed
companies in search of evidence of this plurality. Comparison is made between the labels
(audit and business) applied by
researchers to these various measures and
the factor analysed loadings. Finally, an audit
fee model adopting variables based on the
results of factor analysis is tested empirically.
The next section reviews the relevant literature and develops the research question to
Literature review
There is now a substantial and growing literature concerned with the multivariate modelling of audit fees. Many of these studies contain common variables, including measures
for size, complexity and risk. Several constructs and measures have been used to operationalize each of these variables, however,
for reasons already noted (and developed
below) the present study focuses on the risk
variable alone.
Definitions
It is commonly accepted within the relevant
literature that, in any audit engagement,
there are two forms of risk which are relevant
to the auditor (see for example [16-18]). One of
these generally is labelled business risk.
This is the risk related to the business of
managing the audit firm and is defined as
the probability that an auditor will suffer loss
or injury in his professional practice[19,
p. 60]. Such loss or injury may arise through
law suits against the auditor, sanctions
imposed by external regulators, diminution
of the auditors professional reputation, possible loss of clients, time and costs incurred
in defending the auditors position and nonrealization of audit fees[19,20].
The second form of risk is labelled audit
risk. Audit risk is the likelihood that an
auditor will render an inappropriate opinion
on an auditees financial statements[21]. It is
generally accepted within both the professional and scholarly literature to be a function of a joint probability, multiplicative[22]
model involving a further set of components
of risk known generally as the audit risk
model.
Despite acceptance of this model as a planning tool within the professional literature
([23-25] in Australia, [26-28] in the USA, [29] in
[ 25 ]
[ 26 ]
**
NS
NS
***
**
***
***
+
NS
NS
NS
**
***
NS
**
***
***
**
+
NS
NS
NS
NS
***
**
NS
***
NS
NS
NS
ROI
or ROE
NS
NS
NS
NS
Quick
or current
ratio
**
***
***
NS
NS
**
NS
Debt
ratio
NS
NS
**
NS
NS
Loss in
last three
years
NS
NS
**
**
Variability
Current
in
Unsystematic
cost
profitability
risk
accounting
***
Ownership
***
Ins. co
rating
Notes: Not included; *** Significant at the p = 0.01 level; ** Significant at the p = 0.05 level; * Significant at the p = 0.10 level; NS Included but not significant; + Included but regression results
not reported
Simunic[1]
Francis[2]
Firth[3]
Simon[4]
Palmrose[5]
Francis and
Stokes[6]
Francis and
Simon[7]
Simon and
Francis[8]
Chung and
Lindsay[9]
Ettredge and
Greenberg[52]
Low et al.[11]
Turpen[10]
Maher et al[55]
Gist[12]
Chan et al[13]
Pearson and
Trompeter[15]
Anderson and
Zeghal[14]
Butterworth and
Houghton[56]
Study
Audit
opinion
Inventories
receivables
total assets
Table I
The operationalization of risk in audit fee studies
[ 27 ]
[ 28 ]
Audit qualification
At a theoretical level it is unclear which of
the two aspects of risk the audit qualification is likely to capture. On one hand, it can
be argued that a qualification increases auditor business risk. This follows for two reasons. First, it may indicate the existence of
financial or other uncertainty surrounding
the auditee[57]. This may lead the audit firm
to suffer damage as a result of being involved
in that audit[1]. Second, it may indicate the
necessity for additional audit work. Palmrose
argues that qualifications require the
accumulation of a greater amount of evidence
to achieve the auditors desired level of assurance. In addition, circumstances that give
rise to modifications may signal the need to
increase the desired level of assurance
because of the increased risk of adverse
actions against the auditor [5, p. 100].
On the other hand, it can be argued that a
qualification helps to protect the auditor from
a charge of negligence and so reduces auditor
business risk. Nothing seems to raise the
ire of shareholders and regulators as much as
an unqualified audit opinion followed soon
thereafter by auditee collapse. It can be
argued that a qualification takes the surprise out of such a phenomenon and reduces
the chance that an auditor will be involved in
subsequent litigation.
Another view sees qualification not as a
business risk measure at all but rather as
an audit risk measure. This is because of
its link with the definition of audit risk
used in the professional and scholarly literature in terms of the risk of giving an inappropriate opinion (see for example [23,33]).
Thus, the exact nature of the qualification
measure when used as a proxy for risk is
unclear, although it is testable empirically.
Proportion of inventory and receivables to
total assets. Simunic was the first to recognize inventory and receivables as risky
balance sheet assets, in that they require
specific auditing procedures and even then
may not be well audited since their valuation
requires a forecast of future events [1, p. 173].
Since these items generally represent a material portion of the balance sheet they usually
warrant substantial audit effort and the risk
of a material misstatement being missed by
the auditor is higher than for other accounts.
It is argued here that the proportion of inventory and receivables to total assets measurement is a proxy for audit risk .
The use of this variable in audit fee studies
creates problems in that it could be a proxy
for (and highly correlated with) auditee complexity. This is so because when the auditee
becomes more complex and its operations
more dispersed, it may be harder for the audi-
tees central management to maintain adequate control over inventory and debtors.
Therefore, as complexity increases, so too
could the value of this ratio. Firth[3] used
such a ratio to control for auditee complexity
in his study and Simon and Francis[8] found
a significant positive correlation between this
variable and their complexity variable. On
the other hand, although not directly relevant
since inventory plus receivables over total
assets was not included, Taylor and Baker[58]
in a factor analysis of a variety of variables
affecting audit fees (but not explicitly including risk) found that the number of subsidiaries as a measure of complexity did not
load on the same factor as either current
assets or total assets. (Wallace[18] also factor
analyses a number of variables she argues to
be related to risk. However, none of the
variables included here form part of that
analysis.)
Thus the ratio of an auditees receivables
and/or inventories to the auditees total
assets has received overlapping use in the
literature and its suitability as a measure of
risk (especially in studies which concurrently measure complexity) can be
questioned. Once again, the question as to
which construct this measure relates (business risk, audit risk or complexity) is
empirically testable[59,60].
Leverage, return on investment, loss history
and liquidity. It is argued that these measures
generally are associated with the potential
for, or actual level of, auditee financial distress. As such they are likely to be proxies for
auditee business risk and hence (indirectly)
auditor business risk[1, p. 173]. This
approach is justified on the basis that an
auditors practice, generally, is vulnerable to
damage through involvement with a particular auditee if the stakeholders in that auditee
have suffered economic loss[61-63]. Such
economic loss is more likely to arise when the
auditee is in financial distress or experiences
failure. In addition, the tendency for
distressed companies to experience more
financial statement errors[64], window
dress financial reports[65] and be associated
with greater non-routine management
turnover[66] can be argued to lead to a
greater likelihood of audit failure and hence
involvement in audit-related litigation[67].
The link between auditee financial distress
and audit fees is strengthened by findings of
an association between financial distress and
auditor switching[68,69] since audit fees have
been found to be a factor influencing auditor
switching in several studies (e.g. [69,70]).
Despite the link between business risk
and auditee distress it is observed that, to
date, only univariate measures of financial
[ 29 ]
[ 30 ]
Data
The data used in this study come from the
1988 Annual Reports of 433 Western
Australian-headquartered companies maintained by the (then) Department of Corporate
Affairs in Perth, Western Australia (now the
Western Australian Division of the
Australian Securities Commission). These
data represent the (then) complete population
of publicly listed companies within the jurisdiction of that body. From this population the
data selection process yielded 229 companies
as shown in Table II.
Companies for which complete data were
unavailable were omitted. Financial Services
sector companies were also omitted. This
industry sector previously has been signalled
as characteristically different from other
industry sectors due to the nature of its asset
base[75]. Some previous audit fee studies
similarly omit this sector from the sample
examined (e.g. [1,3,56]). Palmrose[5] reported
that this sector gave rise to a fee model with
coefficients different from the remainder of
her sample.
Mining companies constituted the largest
industry sector within the sample (61 per
cent), with other companies distributed
across remaining industry groupings in proportions varying from 1-12 per cent. Table III
shows the distribution of companies in the
sample across industry groups while Table IV
gives descriptive details for the mining and
non-mining categories and the total group of
companies.
Mining companies, on average, demonstrate approximately half the average total
Table II
Data relating to publicly-listed companies in
jurisdiction of Department of Corporate Affairs,
Perth (1988)
433
349
Financial
characteristics
Inventory ($m)
305
293
268
229
Table III
Summary of sample coverage by industry
Total
sample
R&D high-tech
R&D others
Retail
Manufacturing
Consultancy
Leisure and
tourism
Construction
Mining
(extractive)
Transport
Health
8
7
28
26
8
(4)
(3)
(12)
(11)
(4)
5
1
(2)
(0)
140
2
4
(61)
(1)
(2)
Total
229
1
2
3
4
Total Mining Nonmean mean mining Diff.a
n = 229 n = 140 n = 89 2 vs 3
5.5
1.0
12.6
ns
Receivables ($m)
8.3
5.9
12.1
ns
102.4
50.3
182.7
ns
Inventory+receivables/ 0.11
0.06
0.18
ns
Stock exchange
industry
classification
Table IV
Descriptive statistics company attributes
Total liabilities
61.0
20.4
123.6
ns
Leverage (TL/TA)
37%
17%
68%
0.000
33.7
126.5
ns
2.0
0.9
3.7
ns
Adult opinion
Total
(%)
Percentage
qualified (1988)
31.9
27.9
38.2 0.109
Percentage
qualified (1987)
28.4
21.4
39.3 0.005
Percentage
qualification (1988)
3.9
1.4
7.7
Major qualification
(1987)
7.4
1.4
16.5 0.000
majorc
0.038
Notes:
a Two-tailed t-test
b 2 test
c Major here refers to adverse, except for and subject to
opinions. In the analysis, all qualifications were used
Methodology
The audit fee model takes the form of a multiple linear regression function as originally
developed by Simunic[1] and used extensively
in the literature (e.g. [4,7,8,56]):
FEEi = b0 + b1 AUDi + b2NASi + b3SIZEi
+ b4COMPi + b5RISKi +
where:
FEEi = Audit fee charged to auditee i
(log10)
AUDi = Size of auditee is auditor (Big 8
or non-Big 8)
NASi = Amount of non-audit services
provided to auditee i by its
auditor (log10)
SIZEi = Size of auditee i (log10)
COMPi = Complexity of auditee i (log)
RISKi = Risk to the auditor of auditee i.
(This is the variable of interest
and may be split into business
and audit risk, depending on
factor analysis results.)
[ 31 ]
[ 32 ]
Other variables
The remaining variables were measured as
follows (with transformations as indicated) to
better meet the assumptions underlying
regression analysis:
Audit fee. The audit fee charged to auditees
is measured by the log10 of the dollar value
of audit fees paid by the firm to its auditor.
In Australia this dollar value must be disclosed by law. (See Schedule 5 to Australian
Corporations Regulations, cl. 27[79]). The
regulations also require that firms disclose
the dollar amount of other services (NAS)
purchased from the auditor. Since fees are
likely to vary at a decreasing rate with
increases in client size a log10 transformation is used.
Auditor size. As with other studies in this
field, auditor size is measured using a
dummy variable coded 1 if the auditor is a
member of the first tier of auditors (the Big
8, or, more recently the Big 6); otherwise
coded 0.
Provision of non-audit services. The log10 of
the dollar value of non-audit services purchased by the auditee from its auditor. As
with the audit fee, this amount must by law
Table V
Correlations between dependent and independent variables
Audit
fee
(log10)
Audit fee
(log10)
Size
(log10TA)
Complexity
(log10sub+1)
Big 8/non
Big 8
Non audit
fees (log10)
Distress
score
Opinion
(1987)
Opinion
(1988)
Inventory+
receivables/
total assets
Size
(log10)
Complexity Big 8/
(log10
non
sub + 1) Big 8
Non
audit fee Distress
(log10)
score
Opinion
(1987)
Inventory
Opinion receivables/
(1988) total assets
0.72**
0.69**
0.53**
0.35**
0.29**
0.26**
0.62**
0.55**
0.44**
0.31**
0.04
0.15*
0.13
0.01
0.01
0.02
0.16*
0.07
0.11
0.01
0.18**
0.07
0.23** 0.09
0.13
0.01
0.29**
0.37
0.05
0.31**
0.067
0.21**
0.01
0.44**
0.10
0.01
Notes:
* < 0.05
** < 0.01
be disclosed by the auditee. Simunic[1],
Simon[4] and Butterworth and
Houghton[56] have previously found this
variable to be significant. It captures the
possibility that the greater the amount of
non-audit services provided by the auditor,
the greater will be the potential spill-over
effects or the greater will be the potential
for audit fee/non-audit services fee tradeoffs, and the greater will be the effect on the
audit price.
Client size. Consistent with other researchers, (see for example [7]) the size of an
auditee is measured using the log10 of the
book value of the auditees total assets.
Client complexity. Most researchers in this
field have measured the complexity of an
auditee by using a continuous measure of
its organizational dispersion (the legal
form of the entity may be a relevant factor
here) such as the number of consolidated
subsidiaries in the auditees group of companies. The rationale behind this type of
measure is that the greater the dispersion
that is shown by a firms structure, the
more intricate and sophisticated become
the operational forms adopted by the
firm[58]. Consistent with this, the present
study uses the number of companies in the
auditees group (log 10 holding company
plus subsidiaries) as the measure for
complexity.
Results
The following section consists of two parts.
First, the results of an analysis to assess the
extent of the independence of the two risk
concepts are discussed. Second is an analysis
of multiple regression results arising from
inclusion in the model of both risk concepts. The following then discusses implications arising from this study and makes suggestions for future research.
[ 33 ]
Table VI
Factor analysis (varimax rotation) of variables: auditor and auditee characteristics, audit qualification, distress and inventory + receivables/total assets
Independent
variable
Size
(log10TA)
Non-audit
fees (log10)
Complexity
(log10sub+1)
Big 8/non Big 8
(dummy)
Audit opinion
(1987)
Audit opinion
(1988)
Distress score
Inventory+
receivables/
total assets
Panel A
concurrent
opinion
Panel B
lagged
opinion
Panel C
concurrent
opinion
Factors
Factors
Factors
0.81
0.80
0.79
0.78
0.78
0.79
0.65
0.68
0.66
0.66
0.66
0.65
0.56
0.78
0.81
0.50
0.71
0.81
0.65
0.87
0.92
0.82
0.82
Notes: Variables measures: Size = log10 book value of auditees total assets. Non-audit service fees (NAS) = log10 of
dollar value of NAS provided. Complexity = log10 of subsidaries + 1. Auditor size = dummy coded 1 if auditor is in the Big
8. Audit opinion 1987 (1988) = all audit qualifications received in 1987 (1988). Distress score = Z-type score.
Inventory + receivables/total assets = book value of auditees inventory and net receivables/book value of total assets
receivables over total assets represented a
second factor. The concurrent audit qualification, not able to be predicted as to risk
category, loaded with the distress score. Thus
it seems that the presence of an audit qualification represents a business risk to the
auditor rather than a protection against such
risk. Interestingly, a lagged audit qualification loaded almost equally to both the audit
risk and business risk categories. It would
seem logical that the lagged opinion, when
qualified, influences fees ex ante through both
anticipation of additional audit effort and as
a litigation litmus. These findings imply that
the theoretical equivocation as to the exact
nature of the risk which qualification
measures is, at least partly, justified.
All other variables (Big 8/non Big 8, NAS,
size and complexity) loaded on a further
factor. This is to be expected since they all, to
some extent, proxy for size and are significantly correlated (Table V). It is interesting to
note that the complexity measure loaded
separately from the risky balance sheet
measure, despite the correlation between
these two measures in this ( r = 0.31) and
previous studies (e.g. [8]).
The result confirms the confusion in the
literature. If risk is discussed in audit fee
studies, generally it has been labelled audit
[ 34 ]
6.728
1.994
4.326
5.638
6.044
Complexity
(subsidaries + 1)
(log10)
Non-audit fees
(log10)
Audit opinion
(1988)
Audit opinion
(1987)
Distress score
Inventory +
receivables/
total assets
Constant
0.000
0.000
0.000
0.047
0.000
0.000
Significance
level
0.000
0.082
0.285
0.000
0.066
0.000
0.000
Significance
level
73.61 73.03
91.84 124.48
0.000 0.000
4.856
1.747
1.071
4.534
1.845
8.546
8.214
0.000
0.000
0.077
0.163
0.002
0.374
0.000
0.000
Significance
level
70.73 74.12
92.83 94.26
0.000 0.000
6.443
5.746
1.778
1.401
3.838
2.095
6.884
9.776
C1
Audit and
business risk
(concurrent)
6.044
5.638
4.326
1.994
6.728
9.397
0.000
0.000
0.000
0.047
0.000
0.000
Significance
level
A2
Audit risk
only (opinion
lagged)
5.341
1.660
2.870
4.539
2.038
8.674
8.629
0.000
0.098
0.004
0.000
0.043
0.000
0.000
Significance
level
B2
Business risk
only (opinion
lagged)
6.744
5.480
1.822
2.517
3.945
2.203
7.060
10.036
0.000
0.000
0.070
0.013
0.000
0.029
0.000
0.000
Significance
level
C2
Audit and
business risk
(lagged)
Notes: Variable measures: Audit fees = log10 of published audit fees. Size = log10 book value of auditees total assets. NAS = log10 of dollar value of NAS provided. Complexity = log10 of subsideries +
1. Auditor size = dummy coded 1 if auditor is in the Big 8. Audit opinion = dummy coded 1 if auditee receives an audit qualification in 1987(1988) distress score + z typescore. Audit risk = inventory
+ receivables/total assets.
73.03 70.60
124.48 88.83
0.000 0.000
9.397
Size
(log10 TA)
R2 (%) (Adj.)
F
Significance
Independent
variable
B1
Business risk
only (opinion
concurrent)
A1
Audit risk
only (opinion
concurrent)
Table VII
Regression of size, complexity, auditor size, non-audit fees and risk on audit fees
[ 35 ]
[ 36 ]
[ 37 ]
[ 38 ]
51
52
53
54
55
56
57
58
59
60
61
62
[ 39 ]
63
65
66
67
68
69
70
71
72
73
74
[ 40 ]
75
76
77
78
79
80
81
82
83
Further reading
Grobstein, M. and Craig, P.W., A risk analysis
approach to auditing, Auditing: A Journal of
Practice and Theory, Spring 1984, pp. 1-16.