University of Groningen
Do the different characteristics of the expanded auditor’s
report have an effect on audit quality in the UK?
1
MSc Thesis – Accountancy & Controlling track
Name: L.M. Koekebakker
Student number: s2499479 Supervisor: dr. C.A. Huijgen
University: University of Groningen Country: The Netherlands
Date: June, 2019
2
ABSTRACT
In 2013, the UK implemented a new ISA 700, which is referred to as the expanded auditor’s
report. This new regulation requires auditors to report about the risks of material misstatements,
materiality, and audit scope. This study examines the effects of the different report characteristics
on audit quality by using a hand-collected data set from UK premium listed companies between
the years 2013-2017. Audit quality is measured by the absolute value of discretionary accruals.
The final sample contains 1.443 observations. This study provides empirical evidence for the fact
that the higher number of critical audit matters (CAMs), average number of words per CAM and
number of words of the audit report that are disclosed by the auditor are related to higher absolute
values of discretionary accruals. However, it can be concluded that when the risks are properly
identified and described in the audit report, and fall within the limits of materiality, the audit quality
does not necessarily have to be compromised and even may indicate a higher audit quality.
Key words: expanded auditor’s report, report characteristics, audit quality, United Kingdom JEL classification: M42, M48
3
INTRODUCTION
In the financial crisis there has been a lot of public criticism on the role of the auditor. The
auditor gives an opinion on whether the financial statements of a company present fairly, in all
material respects, the financial position of the company in accordance with the applicable
accounting standards. These audit reports are communicated to several parties, including
investors, which helps them to make important decisions. Until recently, these auditor’s reports
did not contain any information about the audit or the company other than the auditor’s opinion.
This traditional audit report consist of a “pass or fail” statement regarding the financial statements’ compliance with the applicable financial reporting framework. According to Gutierrez et al. (2018)
“The standard unqualified auditor’s report does not communicate potentially useful information about the audit, such as complex issues that require significant judgement and the ways in which
the audit addresses them”. Auditors gain private information about the company that can be useful to investors when they would disclose this information (PCAOB, 2013). Auditors are in the unique
position to bring risks that were not identified earlier to the attention of investors because they
have a lot of private information about the client’s financial statements. The auditor can also help
the investor to prioritize the most significant risks, given that the annual reports are rather
comprehensive these days. Because the auditor is independent form the management many
investors wanted auditors to provide more information in the report (PCAOB, 2013).
Investors claimed that the traditional audit reports were uninformative, because nearly all
public companies got unqualified opinions (Lennox, 2005; Gray et al., 2011; Church et al., 2008).
4
reporting and audit risk (PCAOB, 2017). In response to this request, standard setters and regulators
all over the world have implemented significant changes to the audit reporting standards to
improve the information value of the audit report. The IAASB, European Commission, Financial
Reporting Council (FRC) and PCAOB have implemented or are in the process of implementing
an expanded model for the reporting (Lennox et al., 2018). Because of this, the report has evolved
from the two-paragraph audit report to a multi-paragraph report and the historical binary
“pass/fail” model is no longer used. The primary objective of the expanded auditor’s report according to the FRC (2012c) was “to improve the information value of an audit for users of
financial statements by promoting greater transparency about the judgements made by
management and auditors in the process of preparing and auditing financial statements”.
The exact terminology of the new report differs by standard setters and jurisdiction. In this
paper, the introduction of the new ISA 700 is referred to as the expanded auditor’s report. This expanded auditor’s report requires auditors to disclose the assessed risks of material misstatement (RMM) that have the greatest effect on the overall audit strategy, efforts of the audit engagement
teams and the allocation of resources during the audit (FRC, 2013). Auditors also have to report
about the materiality and the scope of the audit. The United Kingdom (including Ireland) was the
first country that implemented the new requirements of the international Standard on Auditing
(ISA) 700 in June 2013. The changed requirements are mandatory only for premium listed
companies on the London Stock Exchange (LSE) for fiscal years ending on, or after September
5
The implementation of the new regulations led to a lot of research about the possible impact
of the expanded auditor’s report on different areas. Although the FRC did not expect that the ISA 700 would have other effects than increasing the transparency of the audit report, several (working)
papers have found some other effects of the expanded auditor’s report that will be discussed in the
next section. Interestingly, little research has been done about the effect of the expanded auditor’s
report on the audit itself. The research that is currently being done about this topic focusses
primarily on the existence and not the content of the expanded auditor’s report in the pre- versus
post-ISA 700 era. Therefore, this study is focusing on the structure and content of the expanded
auditor’s report, which include (1) the number of critical audit matters (CAMs), (2) average number of words per CAM, (3) number of words in the audit report, and (4) materiality level.
Therefore, the main research question is: Do the different characteristics of the expanded auditor’s
report have an effect on audit quality in the UK?
This study makes several contributions. First of all, it provides insight that is useful to
standard-setters and regulators by examining the effect of the regulatory change regarding the
format and the content of the audit report. Because the changes in the ISA 700 about the expanded
auditor’s report are quite recently adopted by companies in the UK, there is no published research yet about the effects of the characteristics of the expanded auditor’s report on audit quality. The
results of the papers that find evidence that the new regulations have other effects besides the
increase of transparency can be used by the FRC and other parties to see if regulatory effort is
paying off and to what extent. Secondly, this study also helps in the effort to resolve conflicting
findings in recent working papers, which will be discussed in the next section, examining the
6
papers are partly researching the effect of the expanded auditor’s report on audit quality and are contradicting. This paper will solely focus on the direct relationship between the different
characteristics of the new requirements of ISA 700 and the audit quality. Thirdly, this study uses
a new and more recent data set which is hand-collected from all UK premium listed firms for fiscal
years ending after October 2013. Taking this together it can be expected that this paper will
contribute to the existing papers.
The next section of the paper provides an overview of the prior literature and develops the
hypotheses. Section 3 discusses the research methodology. Section 4 contains the results of this
study and the additional/robustness tests that are performed. Section 5 will provide the discussion
7
THEORETICAL FRAMEWORK
As stated before, the auditor gives an opinion on whether the financial statements of a
company present fairly, in all material respects, the financial position of the company in
accordance with the applicable accounting standards. Information about the financial statements
is important to several parties. Debt financing, for example, introduces a conflict between debt,
equity and management and leads to agency costs (Christensen et al., 2016; Jensen and Meckling,
1976; Myers, 1977; Smith and Warner, 1979). First, according to the agency theory, agency cost
arise because of the information asymmetry between the firms’ owners / investors and firms’
management. The information asymmetry is due to the gap between information that is publicly
available to the financial statement user and information that is not publicly available. Second, as
the debt level increases, the shareholders’ and manager’s incentives to take on excessive risk will also become higher (Christensen et al., 2016). The auditing profession and the role of accounting
information plays a crucial role in decreasing these conflicts.
Academic research on the communication value of the audit report has noted this
information gap between the firms’ owners / investors and firms’ management, but also an
expectation gap (Smith, 2017) between these parties and the audit profession itself. According to
the IAASB (2011) the expectations gap is the gap between what the auditor’s responsibility is and
what financial statement users believe the auditor’s responsibility should be (‘difference between what users expect from the auditor and the financial statement audit, and the reality of what an
audit is’). Investors stated that they prefer more information in the audit report to assist with understanding their underlying investments because of the lack in this communication value of the
8
current “pass/fail” audit reports (PCAOB, 2017). In response to this request, standard setters and regulators all over the world have implemented significant changes to the audit reporting standards
to improve the information value of the audit report. The primary objective of the regulatory
change (ISA 700) was therefore “to enhance the relevance and value of the audit for users and the
public by stimulating greater transparency about the judgments made by management and auditors
in the course of preparing and auditing information” (Gutierrez et al., 2018; FRC 2012c). According to the FRC, the information about the risks, scope and materiality provided in the
expanded auditor’s report will increase the transparency about the information that is not publicly available. At the same time, this expanded auditor’s report can improve the financial statement
users’ understanding of the audit process by the tailored approach for auditors to include important information (Smith, 2017). The discussion of the risks of material misstatements, materiality, and
audit scope can give users insight into the decision process of the auditor during the financial
statement audit. The new requirements thus imply that disclosure about these additional matters
will enhance the communication value and decision-usefulness of the audit report. But because
this expanded auditor’s report gives a better understanding of the reasons behind the auditor’s opinion and improves the users' understanding of the audit process, this disclosure can also
eventually reduce and narrowing the audit expectations gap (De Muylder et al., 2012).
Several studies are conducted about the effect of this regulatory change. Some papers are
examining the effect on information and communication value (Lennox et al., 2018; Smith, 2017).
Other papers are more interested in other topics, such as decision usefulness (Gutierrez et al.,
2018), debt contracting terms (Porumb et al, 2018), audit fees (Gutierrez et al., 2018; Reid et al.,
9
market measures (equity and debt market) to compare the audit reports in the pre- versus post-ISA
700 era. Another interesting effect of the new regulation concerning the audit report, which was
not in scope or limited in scope in other studies, is the effect on the audit itself and the possible
improvement of the audit quality.
EXPANDED AUDITOR’S REPORT AND AUDIT QUALITY
Audit quality can be defined as “greater assurance that the financial statements faithfully
reflect the firm’s underlying economics, conditioned on its financial reporting system and innate characteristics” (DeFond and Zhang, 2014). Auditors with high quality do not only examine at whether the accounting choices of the client are in technical compliance with regulation, they also
look at how faithfully the financial statements reflect the underlying economics of the firm. But
most studies define audit quality as some variation of “the market assessed joint probability that a given auditor will both detect a breach in the client’s accounting system, and report the breach” (DeAngelo, 1981).
The few (working) papers that examine the effect of the new requirements according to
ISA 700 and audit quality provide some substantiation for why it can be expected that the expanded
auditor’s report can influence audit quality, but their results are contradicting. The assumption that the new regulation influence the audit quality sounds plausible, since DeFond and Zhang (2014)
find evidence that litigation, reputation and legislation have an effect on audit quality. Moreover
the audit quality framework is seeing regulators as key interactors within the Financial Reporting
Supply Chain that influences audit quality. But why can regulation (and specifically this new
10
intervention can change auditor’s incentives to provide higher audit quality, because regulation is there to create a minimum floor for the audit quality of public companies. Therefore, when
regulation is adjusted, the minimum floor for the audit quality will also change. According to
Gutierrez et al. (2018) this change in regulation can lead to more attention and additional oversight
of the FRC. Moreover, as discussed before, due to the fact that the new regulation requires auditors
to disclose more information, the investors and regulators can potentially better understand the
work that is done by the auditors. This leads to more scrutiny of the auditor’s work, especially
when it comes to risk disclosures (Gutierrez et al., 2018; Christensen et al., 2014). It means there
would be improved financial statement user’s ability to assess the audit quality. Due to the fact that auditors report more in the expanded auditor’s report and can attract additional oversight from
the investors and regulators, the legal liability that auditors perceive can increase. In combination
with that auditors may feel more accountable for their work, it is possible that they will perform
additional procedures and increase their effort and skepticism (Reid et al., 2018; Bédard et al.,
2018), which are drivers of audit quality (IAASB 2012). Due to the greater accountability and
legal liability, it is possible that auditors gather more and better evidence regarding the related
items (Bédard et al., 2018). According to Reid et al. (2018), another possible explanation could be
the “threat of disclosure”. Because of the regulatory change, the negotiation dynamics between the auditor and the management could change. Wells Fargo (2016) explains it by saying that due to
the fact the new expanded auditor’s report has more to say than pass or fail, auditors gain a stronger hand. “Simply the ability to say something there is an additional tool for the auditor and will lead to the fact that they will win more arguments with the management” (Wells Fargo, 2016). This
“threat of disclosure” will have a possible effect on the financial reporting quality, which is closely linked to audit quality (DeFond and Zhang, 2014). Smith (2017) does indeed find evidence from
11
practitioner feedback that the implementation of ISA 700 leads to more readable and useful audit
reports. These more readable and useful audit reports are associated with higher audit quality. This
means that the expanded auditor’s report is a positive driver of audit quality (Smith, 2017). Doxey (2014) and Reid et al. (2018) also find evidence that investors perceive higher financial reporting
quality and lower misstatement probability after the new requirements, which is closely linked to
audit quality (DeFond and Zhang, 2014).
On the other hand, it has been argued in the literature that the regulatory change does not
have an effect on audit quality. The accountability and legal liability effects may not be large
enough to encourage auditors to gather more and better audit evidence and increase auditor
skepticism (Bédard et al, 2018). The only new requirement is that the auditor has to report on what
they have done during the year, to make the auditor’s work more transparent. It may not affect the
underlying work undertaken by auditors or their behavior (Reid et al., 2018; Gutierrez et al., 2018).
According to Reid et al. (2018) they simply have to report what they have always done. The
expanded auditor’s report can also become a boilerplate (standard text) which does not result in increased effort or attention (Bédard et al., 2018). Gutierrez et al. (2018) even states that the
expanded auditor’s report can have a negative effect, because companies may use benchmarking to influence the materiality, that the “expectation gap” can increase, and there could be more costs that the audit firm might not charge to the client. These factors will eventually impact the audit
quality negatively (Gutierrez et al., 2018).
All the above mentioned studies primarily focus on the implementation and existence of
12
a cross-section analysis using the content, but do not find that the variation in the expanded reports’
has a noticeable effect on the investors’ reaction. This paper therefore examines if the different characteristics of the expanded auditor’s report, such as (1) the number of critical audit matters (CAMs), (2) average number of words per CAM, (3) number of words in the audit report, and (4)
materiality level can find a different outcome on the audit quality, because it can be reasonably be
expected that the length of the audit report and per CAM and the number of CAMs that are
mentioned in the report are related to the work undertaken by the auditor as discussed before. Also
the materiality is an important means by which the auditor expresses his confidence on the internal
controls, which is also related to the work of the auditor. Because of the inconsistency in prior
literature about the direction of the effects, this is left open. This leads to the following hypotheses:
H1: The number of critical audit matters (CAMs) stated in the expanded auditor’s report has an
effect on audit quality.
H2: The average number of words per critical audit matter (CAM) in the expanded auditor’s report
has an effect on audit quality.
H3: The number of words in the expanded auditor’s report has an effect on audit quality.
13
METHODOLOGY
To test the hypotheses discussed in the previous section, the audit quality is the dependent
variable and the audit report characteristics are the independent variables. There are different ways
of measuring audit quality, but a common method is by calculating the absolute value of
discretionary accruals as proxy for audit quality. This measure is extensively used in prior research
(Carcello and Li, 2013; Lawrence et al., 2011; Lamoreaux, 2016; Boone et al., 2017; Fung et al.,
2017). This is also in line with literature which are used in this paper (Gutierrez et al., 2018; Bédard
et al., 2018; Reid et al., 2018). A higher absolute value of the discretionary accrual is viewed as a
lower audit quality. This study uses the absolute value due to the fact there is no further information
about the incentives for the company to manage the earnings either upwards or downwards. The
discretionary accruals are calculated using the following modified Jones model below. The
definitions of the symbols are provided in appendix A.
(1) TACCRt = ∆CAt - ∆Casht - ∆CLt + ∆ST Debtt – Depr. Exp.t
(2) TACCRt (At−1) = α1 1 (At−1) + α2 (∆REVt−∆RECt) (At−1) + α3 PPEt (At−1) (3) NDACCRt (At−1) = ᾱ1 1 (At−1) + ᾱ2 (∆REVt−∆RECt) (At−1) + ᾱ3 PPEt (At−1) (4) DACCRt = TACCRt - NDACCRt
All variables are scaled by total assets in the previous period. To calculate the ᾱ1, ᾱ2 and ᾱ3
an OLS regression is performed for each industry shown in Table 1. All the accounting data for
the premium listed firms at the London Stock Exchange from 2013 through 2017 are downloaded
from the Thomas Reuter Datastream database. Because some data was not available, a number of
14
observations of 392 companies. To collect the independent variables (the content of the expanded
auditor’s report) the audit report is downloaded from Orbis. The (1) number of critical audit matters (CAMs), (2) average number of words per CAM, (3) number of words in the audit report, and (4)
group materiality level (£) are extracted by hand from these reports. After excluding missing
observations, due to PDFs that were protected or reports that were not available because of
bankruptcy, the final sample is reduced to 1,443.
The outcome of the measured relationships can also contain other effects because of other
year and company characteristics. Therefore this paper further needs to control for these effects
and time-varying company characteristics. The time-varying company characteristics that are
controlled for are Size, Big4, Leverage, Book-to-Market (BTM) ratio, Loss and ROA. Size controls
for differences in terms of firm size. Larger clients are more likely to have higher earnings quality,
so it can be expected that the client Size, measured as log of total assets, will be negatively
correlated with accruals (Becker et al., 1998; Corbella et al., 2015; Behn et al., 2008). Big4 is a
dummy variable and controls for differences in terms of the audit firm who performed the audit.
Prior research has shown that the size of the audit firms influence the audit quality (DeAngelo,
1981; Francis and Yu, 2009). Audit quality is higher on average in larger Big4 offices, because
larger offices are more likely to issue going-concern audit reports and can act more independent
from their clients. Therefore it can be expected that Big4 is negatively correlated with accruals.
The controls Leverage and Loss are a proxy for the likelihood of financial distress. Prior research
has shown that financial distress is an incentive to use accruals to manage earnings in order to
avoid violating debt covenants or other market reactions (Corbella et al, 2015; DeFond &
15 Leverage and Loss are positively correlated with accruals. Book-to-Market (BTM) and ROA control
for growth opportunities and loss to tease out the effect of increased audit risk due to loss-making
companies. Matsumoto (2002) and Hribar & Nichols (2007) found that riskier firms and growth
firms may have greater incentives to manage earnings in order to meet market expectations. So a
higher market-to-book and ROA leads to higher accruals. In this study the BTM ratio is taken
instead of market-to-book because of the better normal distribution. This means that it can be
expected that BTM is negatively correlated with accruals. For more information about all the
definitions of the variables and how they are measured see Table 2. For the summary of the
descriptive statistics of all variables of the final sample see Table 3. All continuous variables are
winsorized at a 1% level to mitigate the effect of extreme outliers in the data.
To test the hypotheses stated in section 2, the following empirical model is used to perform
an OLS regressions:
(5) ABS_DACCRt = α0 + α1Report Characteristics + α2Sizet + α3Big4 + α4Leveraget
+ α5BTMt + α6Losst + α7ROAt + IndustryDUM + YearDUM + εt
Where Report Characteristics represent the following variables: (1) number of critical
audit matters (CAMs), (2) average number of words per CAM (Words_CAMs), (3) number of
words in the audit report (Words_AR), and (4) group materiality level (Materiality). The model
includes year and industry fixed effects. Industry 8 is left out because this industry does not contain
16
RESULTS
In Table 4 the Pairwise correlations for all variables are presented. The two highest
numbers in this matrix are for LOSS and ROA with -0.6390 (p<0.01), meaning that when the
company reports a loss the ROA will decrease significantly, and Words_CAMs and Words_AR
with 0.5939 (p<0.01), which suggests that when the average words per CAM increases the total
amount of words in the audit report will increase simultaneously. This makes sense due to the fact
that the critical audit matter section is part of the audit report. Both correlations are below 0.8 and
the pairwise correlations of the other variables are below 0.5. This suggest that multicollinearity
is not a serious issue in this study.
Table 5 contains an overview of the results of the regressions for estimating the effects of
report characteristics on the absolute value of discretionary accruals (as proxy for audit quality).
As shown in the table, most coefficients are quite low. This makes sense due to the fact that the
dependent variable itself is a small number which can only takes values between 0 and 1 while the
independent variables for example takes values up to 6.425 (Words_AR). An increase in the
independent variables will only cause a small effect on the dependent variable. In column (1) only
the control variables are taken into account when determining the effect on audit quality. The
coefficient of firm size is negative and significant (-0.004, p<0.01), suggesting that the larger the
firm the lower the absolute discretionary accruals which indicates a higher audit quality. This is in
line with the expectation discussed in the previous section that larger firms are more likely to have
higher earnings quality (Becker et al., 1998; Corbella et al., 2015; Behn et al., 2008). The variables
leverage and loss are positive and significant (0.013, p<0.5 and 0.018, p<0.01), suggesting that
17
also in line with the expectation discussed in the previous section that financial distress is an
incentive to use accruals to manage earnings in order to avoid violating debt covenants or other
market reactions (Corbella et al, 2015; DeFond & Jiambalvo, 1994; Jaggi & Lee, 2002; Behn et
al., 2008). All the other control variables do have the expected sign as discussed in the
methodology section, but are not statistically supported.
In column (2) up to and including column (5) the report characteristics number of critical
audit matters (CAMs), the average number of words per CAM (Words_CAMs), the number of
words in the audit report (Words_AR) and group materiality level (Materiality) are tested
separately on the effect on absolute discretionary accruals. All of them are controlled for Size,
Big4, Leverage, BTM, Loss, ROA and include year and industry fixed effects. Column (2) shows
that the number of critical audit matters (CAMs) that are included in the audit report have a positive
and significant effect on the absolute discretionary accruals (0.003, p<0.01). This is in line with
the expectation that the number of critical audit matters (CAMs) stated in the expanded auditor’s report has an effect on the absolute value of discretionary accruals (H1). However, at first sight
this is not in accordance with the explanation that when the number of CAMs increases, the auditor
puts more effort into the critical audit matter section and that the more underlying work undertaken
by the auditor will increase the audit quality. It seems to have the opposite effect and support the
explanation of the negative effects on audit quality of Gutierrez et al. (2018), discussed in the
theory section. This also applies to the average number of words per CAM in column (3) and the
total amount of words in the audit report in column (4). Although the coefficient in column (3) of
3.23e-05 implies a very small effect, this effect is positive and significant (p<0.01), suggesting that
an increase in the average number of words per CAM leads to an increase in absolute discretionary
18
(CAM) in the expanded auditor’s report has an effect on the absolute discretionary accruals (H2). The results in column (4) show that the coefficient of the total number of words in the audit report
is even smaller, but significant (5.42e-06, p<0.01). This is also in line with H3 that expects that
the number of words in the expanded auditor’s report has an effect on the absolute discretionary
accruals: when the audit report contains more words, the absolute discretionary accruals increase.
Although the direction of these three effects do not seem to correspond to the detailed possible
explanation about the increased effort and skepticism of the auditor, it does make sense when
looking deeper into an explanation. It is actually quite understandable that the reported risks in the
critical audit matter section and total words in the auditor’s report are positively related with the
absolute value of discretionary accruals. If the company practices more earnings management, the
auditor will have more findings and report these findings. However, by using absolute
discretionary accruals as proxy for audit quality we say that when the discretionary accruals are
high, the audit quality is low as the auditors must restrain this way of earnings management. We
hereby forget, however, that if this earnings management is within the limits of materiality, the
quality of the audit is not necessarily compromised. If all the findings are properly mentioned and
described in detail in the auditor’s report it is therefore too easy to conclude that the audit quality
is actually worse. If the auditors have detected unusual accruals, it can also be concluded that, if it
falls within the approved and reported limits of materiality, the auditor has done more work which
does not have to be adjusted by the client automatically. As a result, the value of discretionary
accruals remains equal, but this does not necessarily mean a lower audit quality. Therefore, this
paper will accept the hypotheses partly. The report characteristics affect the absolute value of
discretionary accruals but this does not automatically means that the audit quality decreases if the
19
Column (5) shows the result for the last report characteristic, materiality. The coefficient
is positive (4.11e-05), but not significant (p>0.1). In other words, when the materiality level (scaled
by total assets) increases, the absolute value of discretionary accruals increases as well, which
implies that the audit quality decreases. An explanation for this direction could be that the higher
the materiality the higher level of earnings management is tolerated. This means that the work
undertaken by the auditor is less when materiality is higher. But because the result is not
significant, H4 (which state that the materiality level given in the expanded auditor’s report has an
effect on the audit quality) cannot be supported. Materiality does not have a significant effect on
the audit quality, meaning that the expectation that the level of materiality influence the underlying
work of the auditor is not statistically supported.
In column (6) all variables are taken together when measuring the effect on absolute
discretionary accruals. The total number of CAMs (CAMs) and the average number of CAMs
(Words_CAMs) stated in the audit report are still positive and significant (0.004, p<0.01 and
3.74e-05, p<0.01). The total number of words in the audit report (Words_AR) stays positive (9.24e-07)
but is not significant anymore (p>0.1). This can be due to the fact that the Words_CAMs and
Words_AR are correlated with each other as discussed in the beginning of this section. The fact
that the total number of words in the audit report is not significant anymore if controlled for, among
other things, the average number of words per CAM and the fact that the coefficient is also
weakened could be indicating that a more lengthy audit report does not necessarily contains more
information. The critical audit matter section contains the real important information. More words
in the audit report does not add additional valuable information when controlling for the other
20
ADDITIONAL/ ROBUSTNESS TESTS
Some additional and robustness tests will be performed to see if the results change or hold
in different circumstances or when using different testing methods. Tests that will be performed
are related to the direction of the accruals, unexpected part of the report characteristics, another
proxy for audit quality (abnormal working capital accruals), control for audit partner
characteristics and control for year effects. All tests will be explained separately below.
DIRECTION ACCRUALS
To see if the results of the effects of the different report characteristics shown in Table 5
differs between the direction of the discretionary accruals, the absolute value of discretionary
accruals are divided into negative and positive discretionary accruals. As shown in Table 6 the
final sample of 1.443 consists of 1.046 negative earnings management accruals and 397 positive
earnings management accruals. In an ideal situation this should be fifty-fifty. Negative earnings
management is also referred to as applying conservative accounting policies, which seems less
harmful than positive earnings management. Positive earnings management is referred to as
aggressive accounting policies due to the fact it uses accounting policies to present the firms
performance in a more favorable light than the underlying reality. Aggressive earnings
management attract much more attention in, for example, the media than negative earnings
management. These outcomes of the separate regressions in Table 6 show that the effect of the
report characteristics is only significant for the negative part of the accruals. This additional test
reinforces the possible explanation given previously. Here was stated that the audit quality does
not necessarily have to be lower if there is earnings management. If the auditor has established and
21
expanded auditor’s report and if this earnings management falls within the limits of materiality, this earnings management does not necessarily be adjusted by the company. This result shows that
when it comes to conservative accounting the auditor apparently permits this earnings management
but does describe the risk(s) in the critical audit matter section in the expanded auditor’s report. If
it concerns aggressive accounting, this no longer applies. Apparently the auditor did not permit the
earnings management and the company had to adjust it.
UNEXPECTED REPORT CHARACTERISTICS
It is possible that the effect of the separate report characteristics (1) number of critical audit
matters (CAMs), (2) average number of words per CAM (Words_CAMs), (3) number of words in
the audit report (Words_AR), and (4) group materiality level (Materiality) on the absolute value of
discretionary accruals is in the unexpected part of these report characteristics. The unexpected part
of the report that was not expected on the basis of, for example, the size of the company and the
industry in which it operates can cause managers to use earnings management. Therefore this
additional test is performed to examine the effects of the unexpected report characteristics,
whereby the expected report characteristics are calculated as follows:
(6) Ex_Report Characteristicst = α0 + α1Sizet + α2Big4 + α3Leveraget + α4BTMt +
α5Losst + α6ROAt + IndustryFE + YearFE + εt
Where Ex_Report Characteristicst represent the following variables: (1) expected number
of critical audit matters (Ex_CAMs), (2) the expected average number of words per CAM
(Ex_Words_CAMs), (3) the expected number of words in the audit report (Ex_Words_AR), and (4)
22
results will be highlighted, because many of them are rather straight forward. The number of
critical audit matters, average number of words per critical audit matter and total number of words
in the audit report for example will increase as the company size increases. The average number
of words per critical audit matter and the total number of words in the audit report will increase
during the years, because auditors will gain more experience with the company and the
accompanying expanded auditor’s report. Also every other industry contains less words in the
audit report compared to the Oil & Gas industry, because the Oil & Gas industry is a complex
industry which needs more explanation, especially the estimation of proven oil reserves. Finally,
it can be seen that when a company has higher leverage or reported a loss, the number of CAMs
described in the audit report will increase. This makes sense due to the fact this companies have
more risks than companies who have lower leverage or reported a profit.
After calculating the expected report characteristics, the unexpected report characteristics
are calculated as follows:
(7) Unex_Report Characteristicst = Report Characteristicst – Ex_Report Characteristicst
Where Unex_Report Characteristicst represent the following variables: (1) unexpected
number of critical audit matters (Unex_CAMs), (2) the unexpected average number of words per
CAM (Unex_Words_CAMs), (3) the unexpected number of words in the audit report
(Unex_Words_AR), and (4) the unexpected group materiality level (Unex_Materiality). The results
of the unexpected report characteristics on the absolute discretionary accruals are shown in Table
8. All the results confirm the effects documented in the main tests, showing that the significant
23 ABNORMAL WORKING CAPITAL ACCRUALS
As discussed in the methodology section, a common method for measuring audit quality is
by calculating the absolute value of discretionary accruals. But because the outcome is industry
sensitive and the negative and positive accruals are not evenly distributed as shown in Table 6, this
paper conducts a robustness test by using another proxy for abnormal accruals to see if the results
hold. Another proxy for abnormal accruals that can be used is the absolute value of abnormal
working capital accruals (ABS_AWCACCR). According to DeFond and Park (2001) this “measures
the difference between realized working capital and a proxy for the market’s expectations of the level of working capital needed to support current revenue levels”. It can be expected that this
difference is the portion of working capital accruals is unlikely to be sustained and, therefore, will
be reversed in future earnings. It is a relatively straight forward model in which no industry
estimation has to be made and in which the depreciation component is omitted. The noncash
working capital and abnormal working capital accruals will be calculated using the following
models below. The definitions of the symbols are provided in appendix B.
(8) WCt = CAt – Casht – CLt + ST Debtt
(9) AWCACCRt (At−1) =
WCt – [(WCt−1 / REVt−1) × REVt] (At−1)
The final sample size after the calculation of the absolute value of the abnormal working
capital accruals is 1.433 firm-year observations. The ABS_AWCACCR has a minimum of 0.000
and maximum of 0.216 with a mean of 0.0337, meaning that this proxy has almost the same
24
The results of the report characteristics on the absolute value of abnormal working capital
accruals are provided in Table 9. As shown in this table, all coefficients of the report characteristics
stay positive. The total number of CAMs that are described in the expanded auditor’s report and total amount of words in the expanded auditor’s report are significantly supported when tested separately. The average number of words per CAM is not significant anymore when tested
separately, but is significant in the last column (5) when all report characteristics taken together.
What is striking here is the fact that materiality is positive and significant when tested separately
and together with the other report characteristics. This additional test confirms that materiality has
an effect on audit quality. Taken together, this robustness test confirms and reinforces the main
results in this paper.
AUDIT PARTNER CHARACTERISTICS
Because this paper examines the structure and content of the expanded auditor’s report, an
interesting study to take into account is the working paper of te Lintelo et al. (2018) because it
provides evidence for the fact that individual characteristics of the signing partner of the expanded
auditor’s report have a significant impact on the structure and content of the expanded auditor’s report used in this paper to determine the effect on audit quality. According to te Lintelo et al.
(2018) “Auditor experience, gender, university and degree type all have a significant impact on
the number of CAMs disclosed in the report. We find that female auditors write a more elaborate
expanded audit report and that the size of the audit report is also influenced by the attended
university”. They also find evidence that the university and type of degree of the auditor does influence the number of words per CAM and that more experienced auditors impose higher levels
25
background and experience, can also influence the audit quality. It can therefore be expected that
these characteristics of the signing partner of the expanded auditor’s report influence the outcomes
of this study. Therefore, an additional test is conducted with the audit partner characteristics as
control variables.
These control variables of the characteristics of the signing partner include 391 auditor’s
names that were hand collected from the accompanying expanded audit report. After collecting
the names, these names of each signing partner are searched on the professional network site
‘LinkedIn’. The data on their experience, gender, the university they attended and the type of degree they achieved were hand-collected from this network site. The definition of these control
variables and how they are measured are provided in Table 2. It was not possible to find or identify
all of the signing partners and/or characteristics. The final sample for this additional test which
includes the audit partner characteristics contains 664 observations. For the summary of the
characteristics of all variables see Table 3. To perform an OLS regression, the following model is
used:
(10) ABS_DACCRt = α0 + α1Report Characteristics + α2Audit Partner Characteristics
+ α3Sizet + α4Big4 + α5Leveraget + α6MTBt + α7Losst + α8ROAt + εt
Where Report Characteristics represent the same variables as in model 6 and following te Lintelo
et al. (2018) the Audit Report Characteristics represent the variables: (1) auditor experience
(Experience), (2) gender of the signing auditor (Gender), (3) graduated university (University),
and (4) type of degree obtained (Degree). The model includes year and industry fixed effects. The
26
shows the same results as the last column (6) of the main results in Table 5 (which shows the
results of all report characteristics together on the absolute value of discretionary accruals with a
sample of 1.443 firm-year observations). The second column (2) shows the results of the regression
when using the same 664 firm-year observations for which the audit partner characteristics are
available, but do not include the audit partner characteristics yet. The results hardly change
regarding the most important relations. As shown, the total number of CAMs and average words
per CAM in the expanded auditor’s report stay positive and significant. The materiality is still positive but not significant. Only the total words in the audit report changes direction, but is still
not significantly supported. The last column (3) provides the results when controlled for the audit
partner characteristics. Experience, Gender and University are negative but not significant. Degree
does have a significant effect and is positive (0.0106, p<0.01), meaning that when an audit partner
has a non-financial degree the audit quality decreases. When reviewing the directions of the
coefficients and p-values of the independent variables, there are no differences compared to
column (2). The effect of the number of CAMs and average words of CAMs is slightly reduced
by the characteristics of the signing audit partner, but are still significant (p<0.1, p<0.05). The total
number of words in the audit report and the materiality are both insignificant again. Therefore, the
expectation that the audit partner characteristics can influence the outcomes of this study can be
rejected.
YEAR EFFECTS
It can be argued that the effect of the content of the different characteristics of the expanded
auditor’s report will decrease over time due to the fact that the information value of the audit report will decline (Lennox et al., 2018; Smith, 2017). In the first year of the adoption of the new ISA
27
700, all information in the critical audit matter section has to be written down for the first time. In
the following years it can be the case that the auditor is only adjusting the text where necessary.
The additional test for unexpected report characteristics show that auditors will probably just add
some text every year instead of rewriting the whole text to save some time. As discussed in the
theory section, the expanded auditor’s report can become a boilerplate which does not result in increased effort or attention (Bédard et al., 2018). Therefore, it could be expected that the audit
quality will be the highest in the first year(s). The results of this additional test are shown in Table
11. As can be seen, the effect of the number of CAMs, average number of CAMs, total number of
words in the audit report and materiality level on the absolute value of discretionary accruals does
not decrease during the years. This means that it is possible that investors perceive a lower audit
quality during the years, because the audit report does not contain new information. However, this
seems to have no increased effect on the earnings management applied by the company’s managers
28
DISCUSSION AND LIMITATIONS
In this section it is important to pay attention to the fact that this paper initially aimed to
measure the audit quality by using the absolute value of discretionary accruals. This method has
been widely used and accepted by previous literature. It therefore can be concluded that when the
management of a company manage earnings the audit quality would be poor, since the auditor
should prevent that the company’s financials are not reflecting reality. Hereby it is not taken into account that when this earnings management is within the limits of materiality, the quality of the
audit is not necessarily compromised. If all the findings are properly mentioned and described in
detail in the auditor’s report it is therefore too easy to conclude that the audit quality is actually worse. If the auditors have detected unusual accruals, it can also be concluded that, if it falls within
the approved and reported limits of materiality, the auditor has done more work which does not
have to be adjusted by the client. As a result, the value of discretionary accruals remains equal, but
this does not necessarily mean a lower audit quality. Future research can prove to find additional
evidence for this explanation by for example measuring the effect on audit quality in a qualitative
way. It is most likely that auditors themselves notice that they are doing more underlying work
since the new legislation according ISA 700 forces them to report more in the audit report. Another
point of attention is the fairly low R-squares in all the performed regressions. The R-squared is the
proportion of the variance in the dependent variable that is predictable from the independent
variables. The low p-values in combination with low R-squares mean that the report characteristics
still provide information about audit quality, but the data points are far away from the regression
line and highly variable. Further research can search for other explanations to mitigate and lower
the noise in the data. Finally, future research may answer the question whether the results found
29
SUMMARY AND CONCLUSIONS
This study examines the effects of different report characteristics of the expanded auditor’s
report after the implementation of the new ISA 700 on audit quality in the UK. Whereby the
absolute value of discretionary accruals is used as proxy for audit quality. This study provides
empirical evidence for the positive relationship between the total number of CAMs disclosed in
the expanded auditor’s report and absolute value of discretionary accruals. Results also show that the average number of words per CAM disclosed in the expanded auditor’s and the total number
of words in the auditor’s report are positively related to the absolute value of the discretionary accruals. These results are in line with the expectation that certain report characteristics are related
to the absolute value of the discretionary accruals. Although at first glance this higher value of
discretionary accruals, better known as earnings management, can conclude that this means that
the audit quality is poorer, this does not have to be the case when looking deeper into an
explanation. In principle the idea behind the use of this proxy is that the auditor must prevent
management from giving an unfaithful representation of the company finances. However, the
found results make sense because when earnings management is higher, the auditor reported more
risks than when earnings management is low. However, if this earnings management falls within
the limits of materiality, this does not necessarily have to be adjusted by the company which
indicate a low audit quality. On the contrary, it is a good thing that the auditor has noticed this
earnings management and reported this in the audit report. It is therefore too easy to conclude that
the quality of the audit is compromised. An additional test that splits the accruals into aggressive
earnings management and conservative earnings management reinforces this explanation by
showing that the effect is only statistically supported when it comes to conservative earnings
30
management, since aggressive earnings management attract much more attention in the media.
The results do not hold when it comes to aggressive earnings management, meaning that the
auditor did not allow this and the company had to adjust their financials. It is therefore still possible
that the auditors performed additional procedures and increase their effort and skepticism (Reid et
al., 2018; Bédard et al., 2018), which are drivers of audit quality (IAASB 2012).
When using the absolute value of abnormal working capital accruals as another proxy for
audit quality the relation for materiality is statistically supported, meaning that when the
materiality is higher, there is more room for earnings management. Because the auditor does not
have to perform additional procedures on the findings below materiality it can be concluded that
the audit quality is in fact lower. Other additional analysis show that the effects of the different
report characteristics on the absolute value of discretionary accruals is in both the expected and
unexpected part of the total amount of CAMs, average number of words per CAM and total amount
of words in the audit report estimated per industry, size, etc. It can also be concluded that the
results do not change by controlling for year effects and audit partner characteristics.
Taken together it can be concluded that this study finds evidence for the fact that different
characteristics of the expanded auditor’s report have an effect on audit quality in the UK. It shows
that the new regulations regarding ISA 700, which requires auditors to report about the risks of
material misstatements, materiality, and audit scope, have other effects besides improving the
31
REFERENCES
Becker, C. M., DeFond, M., Jiambalvo, J., & Subramanyam, K. R. (1998). The effect of audit
quality on earnings management. Contemporary Accounting Research, 15 (1), 1-24.
Bédard, J., Gonthier-Besacier, N., & Schatt, A. (2018). Consequences of expanded audit reports:
Evidence from the justifications of assessments in France. Working paper: available at:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3175497
Behn, B. K., Choi, J. H., & Kang, T. (2008). Audit quality and properties of analyst earnings
forecasts. The Accounting Review, 83 (2), 327-349.
Boone, J. P., Khurana, I. K., & Raman, K. K. (2017). Spatial competition in local audit markets
and the fall-out on Deloitte from the 2007 PCAOB censure. Auditing: A Journal of Practice and
Theory, 36 (2), 1–19.
Carcello, J. V., & Li, C. (2013). Costs and benefits of requiring an engagement partner signature:
Recent experience in the United Kingdom. The Accounting Review, 88 (5), 1511–1546.
Christensen, B. E., Glover, S. M., & Wolfe, C. J. (2014). Do critical audit matter paragraphs in the
audit report change nonprofessional investors' decision to invest? Auditing: A Journal of Practice
32
Christensen, H. B., Nikolaev, V. V., & Wittenberg-Moerman, R. (2016). Accounting information
in financial contracting: The incomplete contract theory perspective. Journal of Accounting
Research, 54 (2), 397-435.
Church, B., Davis, S., & McCracken, S. (2008). The auditor’s reporting model: A literature
overview and research synthesis. Accounting Horizons, 22 (1), 69–90.
Corbella, S., Florio, C., Gotti, G., & Mastrolia, S. A. (2015). Audit firm rotation, audit fees and
audit quality: The experience of Italian public companies. Journal of International Accounting,
Auditing and Taxation, 25, 46-66.
DeAngelo, A., & Elizabeth, L. (1981). Auditor size and audit quality. Journal of Accounting &
Economics, 3 (3), 183-199.
DeFond, M., & Jiambalvo, J. (1994). Debt covenant violation and manipulation of accruals.
Journal of Accounting and Economics, 17, 145-176.
DeFond, M., & Park, C. W. (2001). The reversal of abnormal accruals and the market valuation of
earnings surprises. The Accounting Review, 76 (3), 375-404.
DeFond, M., & Zhang, J. (2014). A review of archival auditing research. Journal of Accounting
33
De Muylder, J., Hardies, K., & Breesch, D. (2012). The expectations gap: Two remedies
investigated. Working paper: available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=
2190404.
Doxey, M. (2014). The effects of auditor disclosures regarding management estimates on financial
statement users’ perceptions and investments. Working paper, University of Alabama.
Financial Reporting Council (FRC). (2012c). Consultation paper. Proposed revisions to
International Standards on Auditing (UK and Ireland) to give effect to the FRC effective company
stewardship proposals. London, England: The Financial Reporting Council Limited.
Financial Reporting Council (FRC). (2013). Standard. The independent auditor’s report on financial statements. International Standard on Auditing (UK and Ireland) 700 (Revised June
2013). London, England: The Financial Reporting Council Limited.
Francis, J. R., & Yu, M. D. (2009). Big 4 office size and audit quality. The Accounting Review, 84
(5), 1521-1552.
Fung, S. Y. K., Raman, K. K., & Zhu, X. K. (2017). Does the PCAOB international inspection
program improve audit quality for non-US-listed foreign clients? Journal of Accounting and
34
Gray, G., Turner, J., Coram, P., & Mock, T. (2011). Perceptions and misperceptions regarding the
unqualified auditor’s report by financial statement preparers, users, and auditors. Accounting
Horizons, 25 (4), 659-684.
Gul, F. A., Wu, D., & Yang, Z. (2013). Do individual auditors affect audit quality? Evidence from
archival data. The Accounting Review, 88 (6), 1993–2023.
Gutierrez, E. F., Minutti-Meza, M., Tatum, K. W., & Vulcheva, M. (2018). Consequences of
adopting an expanded auditor’s report in the United Kingdom. Review of Accounting Studies (in press).
Hribar, P., & Nichols, D. (2007). The use of unsigned earnings quality measures in tests of earnings
management. Journal of Accounting Research, 44, 1017-1053.
International Auditing And Assurance Standards Board (IAASB). (2011). Enhancing the Value of
Auditor Reporting : Exploring Options for Change. New York, NY.
International Auditing and Assurance Standards Board (IAASB). (2012). Invitation to Comment:
Improving the Auditor’s Report. New York, NY: International Federation of Accountants.
Jaggi, B., & Lee, P. (2002). Earnings management response to debt covenant violations and debt
35
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs,
and ownership structure. Journal of Financial Economics, 3, 305–60.
Lamoreaux, P. T. (2016). Does PCAOB inspection access improve audit quality? An examination
of foreign firms listed in the United States. Journal of Accounting and Economics, 61 (2), 313–
337.
Lawrence, M., Minutti-Meza, M., & Zhang, P. (2011). Can Big 4 versus Non-Big 4 differences in
audit-quality proxies be attributed to client characteristics? The Accounting Review, 86, 259–286.
Lennox, C. S. (2005). Audit quality and executive officers’ affiliations with CPA firms. Journal
of Accounting and Economics, 39, 201-231.
Lennox, C. S., Schmidt, J. J., & Thompson, A. M. (2018). Is the expanded model of audit reporting
informative to investors? Evidence from the UK. Working paper. Available at:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2619785.
Matsumoto, D. A. (2002). Management's incentives to avoid negative earnings surprises. The
Accounting Review, 11, 483-514.
Myers, S. C. (1977). Determinants of Corporate Borrowing. Journal of Financial Economics, 5,
36
Porumb, V. A., Karaibrahimoglu, Y. Z., Lobo, G.J., Hooghiemstra, R. & De Waard, D. (2018). Is
more always better? Disclosures in the expanded audit report and their impact on loan contracting.
Working paper. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=
3216492.
Public Company Accounting Oversight Board. (2013). Proposed auditing standards – the auditor’s
report on an audit of financial statements when the auditor expresses an unqualified opinion; the
auditor’s responsibilities regarding other information in certain documents containing audited financial statements and the related auditor’s report; and related amendments to PCAOB standards. PCAOB Release No. 2013-005.
Public Company Accounting Oversight Board. (2017). The auditor’s report on an audit of financial
statements when the auditor expresses an unqualified opinion; and related amendments to PCAOB
standards. PCAOB Release No. 2017-001.
Reid, L. C., Carcello, J. V., Li, C., & Neal, T. L. (2018). Impact of auditor and audit committee
report changes on audit quality and costs: evidence from the United Kingdom. Working paper.
Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2647507&_ga=
1.120943243.20942536 12.1441123208.
Smith, JR. C. W., & Warner, J. B. (1979). On financial contracting: An analysis of bond
37
Smith, K. W. (2017). Tell me more: a content analysis of expanded auditor reporting in the United
Kingdom. Working paper. Accessible at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=
2821399.
Te Lintelo, D., Porumb, V. A., Huijgen, C., & Karaibrahimoglu, Y. (2018). The personal touch:
Audit partner characteristics and the expanded audit report. Working paper.
Wells Fargo. (2016). Comment letter on PCAOB Release No. 2016-003; Rulemaking Docket
38
TABLE 1
TABLE 1: INDUSTRIES
Groups Industry Obs.
0 - 999 Oil & Gas 96
1000 - 1999 Basic Materials 141 2000 - 2999 Industrials 640 3000 - 3999 Consumer Goods 235 4000 - 4999 Health Care 96 5000 - 5999 Consumer Services 523 6000 - 6999 Telecommunications 30 7000 - 7999 Utilities 39 9000 - 9999 Technology 102
39
TABLE 2
TABLE 2: VERIABLE DEFINITION
Variable Definition
ABS_DACCR The absolute value of the discretionary accruals as proxy for audit quality (data from
Datastream);
CAMs The total number of CAMs disclosed in the audit report (hand-collected);
Words_CAMs The average number of words in words per CAM disclosed in the CAMs paragraph of the audit
report (hand-collected);
Words_AR The total number of words in the audit report (hand-collected);
Materiality The group materiality amount (£) disclosed in the audit report scaled by the total assets
(hand-collected);
Size The natural logarithm of total assets (data from Datastream);
Big4 A dummy variable that takes the value of 1 if the company is audited by one of the big-4 audit
firms, 0 otherwise (hand-collected);
Leverage The total liabilities divided by the total assets (data from Datastream);
BTM The ratio of the company’s book-to-market value (data from Datastream);
Loss A dummy variable that takes the value of 1 if the company reported a negative earnings, 0
otherwise (data from Datastream);
ROA The return on assets of a company (data from Datastream).
Additional analysis
Experience The amount of years an auditor has worked at an auditing firm, in this paper defined as the
partner’s experience (hand-collected);
Gender A dummy variable that takes the value of 1 if the audit partner is male and 0 if the audit partner
is female (hand-collected);
University
A variable representing audit partner’s education quality. University quality is derived from the QS World University ranking 2018, see appendix C for an overview. Auditors that attended the top 20 universities are coded as 0, auditors that attended universities placed from 21 to 40 are coded as 1 and auditors that attended universities placed from 41 and up are coded as 2 (hand– collected);
Degree
A dummy variable that takes the value of 1 if the audit partner has a non-financial degree and 0 if the audit partner has a financial degree. See appendix D for an overview of the studies that are defined as non-financial and financial (hand-collected);
40
TABLE 3
TABLE 3: DESCRIPTIVE STATISTICS
Variable Obs. Mean Median Std. Dev. Min. Max.
ABS_DACCR 1,443 0.049 0.038 0.047 0.000 0.238 CAMs 1,443 3.699 4.000 1.407 1.000 8.000 Words_CAMs 1,443 320.993 311.458 136.842 45.833 725.000 Words_AR 1,443 3190.170 3153.500 1148.115 1383.000 6425.000 Materiality 1,443 5.766 4.851 4.161 0.678 23.706 Size 1,443 13.820 13.585 1.705 10.198 18.461 Big4 1,443 0.945 1.000 0.228 0.000 1.000 Leverage 1,443 0.569 0.530 0.221 0.090 1.342 BTM 1,443 0.531 0.420 0.513 -0.268 2.959 Loss 1,443 0.173 0.000 0.379 0.000 1.000 ROA 1,443 0.075 0.077 0.105 -0.367 0.377 Additional analysis Experience 664 26.258 26.000 5.707 15.000 38.000 Gender 664 0.887 1.000 0.317 0.000 1.000 University 664 0.383 0.000 0.651 0.000 2.000 Degree 664 0.459 0.000 0.499 0.000 1.000
41
