The Impact of Individual Auditor Characteristics on the Expanded Audit Report

on the Expanded Audit Report

Rijksuniversiteit Groningen, Msc Accountancy V. A. Porumb

Rijksuniversiteit Groningen, Supervisor

19th of January, 2018 8731 Words

Abstract

Recent regulatory changes to the audit report significantly adjusted its content and scope. Specifically, UK regulators mandate, starting October 2013, that firms would disclose an expanded audit report that includes auditors’ assessment of the main risks of material misstatements. Given the lack of clear guidance on the structure and content of the expanded audit report, this research examines whether auditor personal characteristics influence these dimensions. I draw on a hand-collected sample of 592 observations from all UK premium listed firms from 2013 through 2015. My findings suggest that auditor characteristics (gender, experience, university attended and degree obtained) have a significant influence on the structure and content of the expanded audit report. The results of this study are of interest to researchers, regulators and the auditing profession.

Keywords: Audit quality, Education, Engagement partner, Expanded audit report, Experience, Gender, Personal characteristics, United Kingdom

Data Availability: All data used in this paper is available on request.

I. Introduction

In recent years international auditing regulators have sought to increase transparency around the audit process by decreasing the expectation gap that exists. For instance, one of the more recent regulatory changes in the European Union (EU) is to increase the information content within the audit report by way of the expanded audit report (Smith, 2016). The United Kingdom (UK) was the first European country to mandate the expanded audit report contained in the ISA 700 regulations (FRC, 2013)1. Regulators in the UK introduced the expanded audit report to increase the accountability of auditors, which they proposed would lead to higher audit quality (Peecher et al., 2013). Nick Land, Chairman of the FRC’s Audit and Assurance Council was quoted2 _{(2013): “I am pleased that auditors as well as investors have given their enthusiastic} support for the proposal in the Consultation Paper to supplement the binary pass/fail model of audit report. The provision of a fuller description of the work the auditor has undertaken will give far more insight to investors […]. The improved report will be a better basis for engagement by investors with companies, and we encourage auditors and companies to work together to develop succinct communication to do so.” The expanded audit report includes more

detail about the auditors’ assessment of the firm, as it discloses for example key financial issues and identified risks of material restatements (Bens et al., 2017). Nonetheless, the lack of clear guidance in the principle-based ISA 700 gives auditors the chance to freely shape the new audit disclosure (Fakhfakh, 2016).

The introduction of the expanded audit report in the UK thus provides the opportunity to look more closely to the effects that auditors have on firm reporting (Smith, 2016). Given the fact that ISA 700 (FRC, 2013) is principle based, it is possible for auditors to write in their own style. The additions in the expanded audit report lead to an increased leeway for auditors to freely shape the audit disclosure (Fakhfakh, 2016). Fakhfakh (2016) found that this lack of clear guidance led to wildly different expanded audit reports in the first years after its introduction. This gives rise to research involving the examination of the effects that certain personal characteristics of the signing auditor have on the quality of their audit report. Existing research about the effect of the expanded audit report focusses mainly on investors’ perception on added

1 _{Earlier attempts were made to decrease the expectation gap. For example, the EU issued the 8}th _{Company Law}

directive in 2006 which required audit engagement partners to sign of the audit report.1 _{The United States followed}

a few years later, when the PCOAB proposed that audit firms needed to disclose the names of their engagement partner in the audit report. Attempts were therefore made to match individual auditors to their signed audit report (King, 2012; Carcello, 2013), but results were incomplete as most audit reports at the time were standardized (Weinrich, 2014).

communication value (Bens et al., 2017; Lennox et al., 2015).3 _{Answering the call for research}

by DeFond and Francis (2005) to extend the existing literature from firm level down to individual auditor level, I will assess the impact of individual auditor characteristics on the structure and content of the expanded audit report.

Up until recently it was difficult to investigate this matter given the lack of publicly available data on individual auditors (FRC, 2013). Gul et al. (2013) were the first to investigate the effects of auditor personal characteristics on audit quality, but their research was performed in the institutionally different China. Thus, it is difficult to generalize their findings towards more developed economies (Lennox and Wu, 2017). Bens et al. (2017), Smith (2016) and Lennox et al. (2015) also used the regulatory changes in the United Kingdom regarding the expanded audit report. They suggested that the expanded audit report does indeed increase audit quality when looking at investors’ perception and communication value. Previous papers have nonetheless largely ignored the impact of auditors’ personal characteristics as explanatory factors to the structure and content of the expanded audit report. My paper is coming to address this important gap in the literature.

In order to test my hypotheses, data on signing auditors and the structure and content of their expanded audit reports is hand-collected from all UK premium listed firms for fiscal years ending after October 2013. This led to a database consisting of 604 firm-year observations and 246 individual auditor observations. Four characteristics are used to proxy for an auditors’ style, namely gender, experience, university attended and degree obtained. The structure and content of the expanded audit report is proxied by five variables, specifically number of words in the audit report, materiality, number of key audit matters, number of words key audit matters paragraph and number of words per key audit matter.

Consistent with expectations, I find that individual auditors’ characteristics have a significant impact on the structure and content of the expanded audit reports. Both auditor gender and experience are significant for 4 out of 5 expanded audit report components, university for 1 component and degree for 3 components. I find that female auditors write a more elaborate expanded audit report, while men show a higher risk appetite in their audits by way of higher materiality. The size of the audit report is also influenced by auditor experience, as auditors with less experience write a more elaborate expanded audit report. In contrast, I

3 _{Moreover, Reid et al. (2015) have investigated the effects that the introduction of the expanded audit report has}

on audit quality and costs. In their review, DeFond and Francis (2005) explain the need to extend the existing literature from the firm level down to the individual auditor level.

find that more experienced auditors use a higher materiality than their less experienced counterparts. Materiality is also affected by the university an auditor attended, as auditors from higher regarded universities permit themselves higher materiality levels. Lastly, the type of degree an auditor obtained has an effect on the number of words per key audit matter. This suggests that auditors with a non-financial degree are superior in describing the key audit matters they found.

My research makes several valuable additions to the existing audit literature. First, it is the first research in a Western setting to provide empirical evidence on the relationship between an auditors’ personal characteristics and the structure and content of the expanded audit report. My findings with Western data confirm findings in those papers (Gul et al., 2013; Ye et al., 2014; Liu, 2017) that used Chinese data. Secondly, it adds to the debate between regulators and the auditing profession on the effectiveness of the new regulations that are meant to increase audit quality (Cordoş and Fülöp, 2015; Knechel and Vanstraelen, 2015; Cameran et al., 2017). While the auditing profession is hesitant to implement new regulations that limit their freedom, my results show that these regulation do work to improve communication value. Lastly, my findings are interesting for bodies entrusted with governance as it shows the importance of soft controls (Chtioui and Thiery-Dubuisson, 2011). With the introduction of the expanded audit report came more leeway for auditors to write in their own style. Thus, it is important for governance bodies within audit firms to reflect on the way they discipline for true and just behavior.

The paper continues with a description of the institutional and theoretical background, which leads to the formulation of several hypotheses. Subsequently the research methodology is developed and the results are given. This paper will end with a conclusion in which the results are discussed.

II. Institutional background, theoretical background and hypotheses development

Institutional background

The UK is a strong candidate to test the effect of auditor personal characteristics on the structure and content of the expanded audit report in a Western setting. The UK institutions and auditing profession are rooted in the common law legal tradition (Latridis, 2012). It is one of the largest global financial centers with large stock markets, with multiple cross listing firms from other countries (including firms that are listed in both the UK and US). Additionally, the auditing profession in the UK exists since the beginning of the industrial revolution and is dominated by the international Big 4 audit firms (Peel, 1997). Lastly, the UK has placed strong institutions and regulatory bodies (e.g. FRC) to regulate the quality of financial information.

Theoretical background

An underlying theory is needed to provide guidance for the direction of this research. Most research on accounting topics are related to the agency theory (Smith, 2015), making it a good starting point. Ever since firm ownership and day-to-day operations were separated, an agency problem existed (Jensen, 1976). Agency theory states that there is an information asymmetry between firms’ owners (principals) and the firms’ management (agents). Ever since this information asymmetry was described, research focused on ways to reduce it. DeAngelo (1981) tried to explain differences in audit quality, and thus a reduction in information asymmetry, with audit firm size. He found that larger audit firms have more to lose and thus have put in place stricter quality controls. This line of thought is underscored by Francis (1988). In his paper (Francis, 1988), he finds a positive correlation between agency costs and the need for high quality audit firms. Both these papers rely on the same assumption, namely that the auditing profession plays a crucial role in decreasing information asymmetry. A new chapter started for agency theory with the introduction of the audit committee (Wild, 1996). The audit committee is responsible for appointing an auditor to verify the financial information that a firms’ management discloses (Wu et al, 2016). After the introduction of the expanded audit report, the auditor has to substantiate his/her audit opinion by supplying the underlying reasons (FRC, 2013). Through the risk paragraph of this expanded audit report, the auditor can show his opinion about the principal-agent relation. For both the relation between firms’ management-auditor and auditor-firms’ owners.

Limperg (1985) deepened this relationship between the auditor and the firms’ owners by describing the expectation gap that exists. He observed that the usefulness of the audit report is related to the confidence that users have in the effectiveness of the opinion given in the audit report. Limperg (1985) proposes in his book his theory of inspired confidence, which connects the needs of financial statement users to the abilities and techniques used by auditors to meet these needs. Within this theory of inspired confidence, Limperg (1985) defines the audit role as follows: “The auditor-confidential agent derives his general function in society from the need for expert and independent examination and the need for an expert and independent opinion based on that examination. The function is rooted in the confidence that society places in the effectiveness of the audit and in the opinion of the accountant. This confidence is consequently a condition for the existence of that function; if the confidence is betrayed, the function, too, is destroyed, since it becomes useless”. The core of the theory of inspired confidence is rooted on the fact that the auditor is obliged to work in such a way that he does not betray the expectations of the financial statement users, nor can he be expected to deliver more than can be justified by the work carried out during his audit. The theory expects that auditors seek insight in the expectations of the financial statement users, and models his audit in a way that fulfills these expectations.

So what is needed from a theory of inspired confidence perspective, is that the auditor needs to be an independent expert which can fulfill the expectations by way of his audit report. This need for expertise implies that the work of an auditor cannot be captured solely by rules and regulations, but there is also a need for the auditor to have enough skills. Next to that, an auditor must also be given the space to express his opinion in a way so that he meets the expectations of financial statement users.

Hypotheses development

The personal capabilities needed by an auditor to be considered an expert have been discussed exhaustively in research.4 This paper give a helpful insight into which personal characteristics have an effect on the way an auditor performs the audit. Characteristics that have been used in previous research include age, tenure, busyness, gender, industry expertise, education and big 4 experience. Studying the effects of personal characteristics originally comes from the field of behavioral sciences. Nelson and Tan (2005) translated this to the auditing profession as “Auditors need to perform a variety of tasks to form an overall assurance

or attestation opinion. To do so, various personal attributes of the auditor (e.g., skills and personality) influence the outcome”. Another viewpoint arises from the managerial fixed effect literature, which shows that personal characteristics of CEO’s affects their decisions (Dyreng, 2010). Thus, it can be expected that auditor personal characteristics also influence the decisions they make, in particular towards the structure and content of the expanded audit report.

In this paper, I build on previous research and assess the impact of multiple individual auditor characteristics. Specifically, the gender of the auditor, the amount of experience in the audit profession, the university that the auditor attended and the type of degree the auditor obtained.

The first individual auditor characteristic of interest is gender. Studies related to gender come from the psychology sciences, in which for example Chung and Monroe (2001) found that males are more accurate when solving simple problems and females are more accurate when solving more complex problems. Males and females also differentiate in their moral view, as Borkowski and Ugras (1992) found that men are more utilitarian. Fellner and Maciejovsky (2007) found that females are more risk adverse and conservative than males when making financial decisions. Hardies et al. (2010) show that gender matters in cognitive style, problem solving ability and risk preference. These studies show that males and females are distinctly different in the way they handle situations. Gold et al. (2009) translated this to audit research by finding that female auditors are more influenced by male CFO’s than female CFO’s. Moreover, female auditors are shown to be superior to their male counterparts in detecting errors and fraud in financial statements (Chin and Chi, 2008). Given the findings of the previous literature, I expect that the gender of individual auditors is likely to have a significant impact on the structure and content of the expanded audit report. Thus, the first hypothesis is:

H1: Signing auditors’ gender has a significant effect on the structure and content of the expanded audit report.

The second individual auditor characteristic of interest is experience. The effects of auditor experience are extensively studied in the audit literature. For example, DeAngelo (1981) found that the auditing profession is one in which the learning curve is clearly present. Experience is predominantly tested with two measures, seniority in the audit firm (Abdolmohammadi and Wright, 1987; Lim and Tan, 2010) and years within the audit profession (Simnett, 1996). This

research will use years within the audit profession as the measure for experience. Earlier research suggests that there is a positive relationship between auditor experience and their performance. In their study, Abdolmohammadi and Wright (1987) find that more experienced auditors have more accurate predictions. Experience is also related to higher audit quality, as suggested by Lim and Tan (2010). Thus, it can be argued that auditors with more experience are perceived by the market as of ‘higher quality’. Other measures correlated with experience are knowledge, skill and problem solving ability. The correlation between these and experience have been extensively tested in literature (Libby and Frederick, 1990; Bonner, 1991; Libby and Luft, 1993; Tan and Kao, 1999; Tan et al., 2002). Given the findings of the previous literature, I expect that the experience of individual auditors is likely to have a significant impact on the structure and content of the expanded audit report. Thus, the second hypothesis is:

H2: Signing auditors’ experience has a significant effect on the structure and content of the expanded audit report.

The third and fourth individual auditor characteristics of interest are related to education, namely the university the auditor attended and the type of degree (s)he received. It is long known that education affects an auditors’ performance, as continuous education is mandated in the profession by International Education Standards 7 and 8 (IAESB, 2012, 2014). Knowledge acquisition by the auditor can be facilitated with education (Bonner & Walker, 1994). The effects of education on auditor performance have become an interesting variable in multiple studies lately (Gul et al., 2013; Ye et al., 2014; Liu, 2017). Gul et al. (2013) find that Chinese auditors who attended a Western university are more conservative than those who attended a Chinese university. Ye et al. (2014) findings suggest that education can influence the risk of audit failure. Educational background is also shown to affect the audit fee (Liu, 2017). Given these previous studies, I derive that education has an effect on the way an auditor performs. Thus, the third and fourth hypothesis are:

H3a: Signing auditors’ attended university has a significant effect on the structure and content of the expanded audit report.

H3b: Signing auditors’ type of degree has a significant effect on the structure and content of the expanded audit report.

III. Research methodology

Data collection procedure

This research uses hand-collected data from the annual reports of all UK firms with a premium listing at the London Stock Exchange from 2013 through 2015. This resulted in a database consisting of 604 firm-year observations. The annual reports were hand-collected from company websites and recorded with their company ticker and year. The expanded audit reports were derived from these annual reports and information about the structure and content were recorded. General financial data regarding a firms’ size and complexity were retrieved from DataStream.

The names of the signing auditors were then hand-collected from the accompanying expanded audit report and recorded in the database. This resulted in 592 observations, of which 246 individual auditors. The names of these signing auditors were then searched through the professional network LinkedIn to hand-collect data on their gender, experience as an auditor, the university they attended and the type of degree they achieved. 592 observations of gender where found, 449 observations of experience, 367 observations of university attended and 332 observations about the type of degree.

Data analysis procedure

The data was tested statistically by way of a random effects GLS panel regression. By using panel data, a large number of data points are obtained which increases the degrees of freedom and reduces collinearity between explanatory variables (Hsiao, 2007). This is an essential condition as this research employs multiple dependent and independent variables in its models. The random effects ensure that time constant explanatory variables do not need to be eliminated, making it an appropriate choice when regressing characteristics of a population (Jager, 2008). I estimate the empirical models using the following regression:

Report characteristicst = 0 + 1Auditor characteristicst + 2Sizet-1 + 3Yeart + 4Audit

firmt + 5Industryt + ε (1)

Where Report characteristics represents the following variables, in alternate form: number of words that the expanded audit report exists of, materiality, number of key audit matters, number of words key audit matters paragraph and average number of words per key audit

matter. Auditor characteristics represent the following variables, in alternate form: gender, experience, university and degree.

The R2 _{was expected to be high for accounting research as most control variables are shown}

to have an effect on the audit process5. Thus, the increment in R2 was calculated to uncover the increase in explanatory power of the independent variables of interest (IVI). Following Collins et al. (1997), the incremental R2 and scaled incremental R2 were calculated as:

R2

IVI = R2FULL - R2W/O IVI

%R2IVI = (R2FULL - R2W/O IVI)/ R2W/O IVI

In the first stage, I estimate the regression model including all individual auditor characteristics. The results that follow from this model show the overall effect of the auditor on the structure and content of their expanded audit report which exceed the effects of the control variables. In a second stage, I assess to what degree each individual auditor characteristic influences the expanded audit report structure and content. Each individual auditor characteristic is regressed independently of each other to counteract possible interdependencies. The results that follow from these models show the independent effect of each individual auditor characteristic on the structure and content of the expanded audit report which exceed the effect of the control variables.

The dependent variables are chosen in line with recent research related to the expanded audit report. The number of words that the expanded audit reports exists of is used by Gutierrez et al. (2016) to examine if the introduction of the expanded audit report has an effect on audit fees. Materiality is an important means by which the auditor expresses his confidence on the internal controls (Gutierrez et al., 2016). Thus, the natural logarithm of materiality is also taken as a dependent variable. Another significant part of the expanded audit report is the key audit matters paragraph. Cordoş and Fülöp (2015) used this paragraph as a means to find if annual report users see this addition as an increase in communication value. Based on this, number of key audit matters, total words of key audit matters and average words per key audit matter are used as dependent variables. Table 1 gives the descriptive statistics for the dependent variables.

[Insert table 1 about here]

The independent variables of interest are chosen in line with Gul et al. (2013) and Caglio & Cameran (2017), which are gender, experience, university and degree. The first variable that will be tested is gender. Gender is shown to affect an auditors risk appetite and their problem solving ability (Hardies et al., 2010). Auditors that are male are rated as 0 and auditors that are female are rated as 1. In line with Libby and Frederick (1990), Bonner (1991) and Liu (2017) experience is defined as the amount of years an auditor has worked at an auditing firm. To get more distinct results, auditors are divided into three experience groups. Auditors with experience less than 20 years are rated as 0, auditors with experience from 21 to 30 years are rated as 1 and auditors with experience over 31 years are rated as 2. According to Bertrand & Schoar (2003), an auditors’ education can affect his or her risk preference, substantive knowledge and ethical behavior. Thus, the auditor might be influenced by the university he attended. University quality is derived from the QS World University ranking 2018, see appendix A for an overview. Auditors that attended the top 20 universities are rated as 0, auditors that attended universities placed from 21 to 40 are rated as 1 and auditors that attended universities placed from 41 and up are rated as 2. Industry and financial knowledge could also be influencing factors on the way auditors recognize key audit matters. This is also used by Liu (2016) who based her interpretation of type of degree on the paper by Bonner and Walker (1994). Thus, auditors with a financial degree are rated as 0 and auditors with a non-financial degree are rated as 1, see appendix B for the type of degrees found in my database. Table 2 gives the descriptive statistics for the independent variables.

[Insert table 2 about here]

Following the study by Gul et al. (2013), I need to control for the client and year effects and also the time-varying client characteristics that might affect the outcome of the relationship measured. Year effects were dealt with by creating year dummies. Standard size and complexity measures were used to control for the client effects, namely number of employees, total assets, revenues, profit before taxes, return on assets, current assets to liabilities ratio and leverage. These where based on actual numbers of the firms from 2012 till 2016 and reshaped from wide to long data. They were then all winsorized before calculation their natural logarithm. Industry classifications were added as a measure of complexity to control for possible risk differences (Goodwin & Wu, 2014). Multiple papers have researched the effects of the audit firm on audit quality (DeAngelo and Elizabeth, 1981; Deis and Guiroux, 1992; Choi et al., 2010; Bills et al.,

2016). The general consensus in these papers is that audit firms do impact audit quality through their quality-control measures. In line with Liu (2017), I added audit firm dummies to control for unwanted effects that are the result of audit firm quality-control measures. Table 3 gives the descriptive statistics for the control variables.

IV. Results Auditor characteristics

Table 4 contains the results of the regressions estimating the effects of all auditor characteristics combined on the structure and content of the expanded audit report. All auditor characteristics are included to see their interaction with the structure and content of the expanded audit report and also their interdependence. In all these regressions, client, year, audit firm and industry are included as control variables. This led to an increase in R2 _{of between}

0.0167 (1.76%) for materiality and 0.3134 (49.04%) for number of key audit matters. Significant results were found for all five dependent variables.

For the number of words the expanded audit report is made of, there are three independent variables of interest which are significant. Experience is significant at the 1 percent level, while gender and degree are significant at the 5 percent level. Including all auditor characteristic variables leads to an increase in explanatory power of the model by 16.31%.

As shown in the second column, there are again three independent variables of interest that are significant when regressing them on materiality. Gender and experience are both highly significant at the 0.1 percent level and university is significant at the 1 percent level. As the explanatory power of materiality was already rather high (95.03%), the increase in R2 is only 1.76%.

There is one independent variable of interest significant for number of key audit matters, experience is significant at the 1 percent level. The increase in explanatory power by 49.04% hints to auditor characteristics being an important factor in detecting key audit matters.

For number of words within the key audit matters paragraph, three independent variables of interest are significant. Experience is highly significant at the 0.1 percent level, degree is significant at the 1 percent level and gender is slightly significant at the 5 percent level. The increase in explanatory power of the model (41.49%) is as expected in line with the amount of key audit matters.

The last column containing average number of words per key audit matter, shows significance of two independent variables of interest. Degree is significant at the 1 percent level, while gender is highly significant at the 0.1 percent level. By including all the auditor characteristic variables, the models’ incremental R2 _{is 9.94%.}

While the above model suggests that the auditor characteristics are jointly significant, it is possible that these results are due to only a few highly significant coefficients. Thus, the auditor

characteristics where then dissected into multiple models to provide clarity on which auditor characteristics have an effect on which components of the expanded audit report. This will eliminate the interdependence of the auditor characteristics and show if the individual effects are strong enough on their own to provide significant results. This will strengthen the conclusions drawn based on the results.

[Insert table 4 about here]

Gender

The first auditor characteristic that was tested is gender. Table 5 gives the results of the regressions with only gender as the independent variable of interest. Again, client, year, audit firm and industry effects were controlled. Including only auditor gender as an independent variable gave an increase in R2 of between 0.0004 (0.04%) for materiality and 0.0338 (10.78%) for number of key audit matters. Gender was significant for 4 out of 5 dependent variables in the complete model, taking gender as an individual variable still yields 2 out of 5 significant results. Number of words key audit matters paragraph and average words per key audit matter are no longer significant. Effect of gender on the number of words in the expanded audit report has become more significant, from the 5 percent level in the complete model to highly significant on the 0.1 percent level in gender only model. The opposite happens for materiality, while it was highly significant at the 0.1% level in the complete model it significance slightly decreased to the 1% level in the gender only model. Auditor gender is positively correlated with the number of words the audit report is made up of, indicating that female auditors describe their findings in more words then men. Opposite to that, the significant negative correlation of gender on materiality shows that male auditors compute higher materiality levels when performing an audit. This is in line with previous research indicating that males have a higher risk appetite. Hypothesis 1 can be confirmed, as gender shows correlated variations in the structure and content of the expanded audit report.

[Insert table 5 about here]

Experience

The next auditor characteristic that will be tested is experience. The results of the regressions with only experience as the independent variable are shown in table 6. Again, client, year, audit firm and industry effects were controlled. The increase in incremental R2 ranged

from 0.0002 (0.05%) for average words per key audit matter to 0.0553 (17.93%) for number of words key audit matters paragraph. Experience was significant for 4 out of 5 dependent variables when all auditor characteristics were regressed together. Two of those are still significant when regressing only experience as an explanatory variable. The number of key audit matters and number of words key audit matters paragraph have lost their significance. The effect of experience on number of words in the expanded audit report has become more significant, from the 1 percent level in the complete model towards the 0.1 percent level in the experience only model. Materiality, while still significant, has decreased from highly significant at the 0.1 percent level down to significant at the 5 percent level. These results indicate that experience has a considerable negative correlation to the size of the audit report. Although what is interesting to find is that less experienced auditors write a more elaborate expanded audit report. This is possibly due to the fact that younger auditors have better learning skills and are more able to adapt to changing standards. Contrary to this, the positive correlation of experience on materiality indicates that more experienced auditors use a higher materiality when performing their audit. This can be explained by the fact that more experienced auditors have more knowledge and thus can compensate a higher inherent and control risk with their detection skills6. Given these findings, the second hypothesis can be accepted on the fact that variations in experience correlate with variations in the structure and content of the expanded audit report.

[Insert table 6 about here]

University

The third characteristic of interest is the university the auditor attended. Table 7 shows the results of the regressions that only include university as the independent variable. Again, client, year, audit firm and industry effects were controlled. By including only university as the independent variable, the increase in incremental R2 is between 0.0095 (1.00%) for materiality and 0.1142 (36.44%) for number of key audit matters. University was only significantly correlated with materiality in the complete model, this holds in the university only model. The correlation is significant at the 5 percent level in both the complete model and the university only model. The significantly positively correlation of university on materiality indicates that

the university an auditor attended has an effect on their computation of materiality. This positive correlation implies that auditors from a less prestigious university use a higher materiality when performing their audit and vice versa. One explanation could be that auditors who studied at a university that is regarded lower (higher) quality, feel the need for a lower (higher) materiality threshold to safeguard themselves against missing material misstatements. Hypothesis 3a can also be confirmed, as the university an auditor attended is significantly variation correlated with materiality.

[Insert table 7 about here]

Degree

The last independent variable of interest is type of degree. Table 8 gives the results of the regressions when only including degree as an independent variable. Again, client, year, audit firm and industry effects were controlled. The incremental explanatory power of the degree only model ranged from 0.0107 (1.13%) for materiality to 0.0877 (27.98%) for number of key audit matters. Degree was significant for 3 out of 5 dependent variables in the regressions of complete model. Out of those three, average number of words per key audit matter is still significantly positively correlated at the 1 percent level. The dependent variables total words audit report and number of words key audit matters paragraph lost their significance in the degree only model. This result suggests that auditors that have obtained a degree in a non-financial study are more skilled at describing the key audit matters they found. This could be due to the fact that auditors with a non-financial degree are more literate in actual business circumstances than those that focused on financials during their education. Given these findings, hypothesis 3b can be accepted as variation in the type of degree is significantly correlated with variation in structure and content within the expanded audit report.

[Insert table 8 about here]

Additional tests

It is interesting to see if the audit firm that the signing partner works at has a moderating effect on the freedom the auditor has to write in his own style. Multiple papers have researched the effects of the audit firm on audit quality (DeAngelo and Elizabeth, 1981; Deis and Guiroux, 1992; Choi et al., 2010; Bills et al., 2016). The general consensus here is that there is a difference between BIG 4 audit firms and non-BIG 4 audit firms regarding quality-control

mechanisms. Jeppesen (2007) found that audit firms use standardized work procedures, central control of materialism and socialization to counteract the auditors’ personal characteristics. Thus, it can be expected that individual auditors have less leeway to structure the expanded audit report in their own style. This would suggest that by removing the audit firm dummies, the significant correlation between auditor characteristics and the structure and content expanded audit report would increase. Untabulated results show that the increase in explanatory power of the model ranges from 0.0162 (1.71%) for materiality to 0.13577 (58.57%) for number of key audit matters. When comparing these results to the original complete model, the explanatory power remains largely the same. Only number of words audit report and average number of words per key audit matter show a large negative difference. What is interesting to see is that removing the audit firm control, does not have a significant positive effect on the independent variables of interest. Most significant results remain significant while, to the contrary, two out of four gender correlations lose their significance. This goes against previous findings, which indicated a moderating effect of the audit firm.

Robustness tests

This section discusses the sensitivity and robustness tests that were performed to assure that the results above are not merely the effect of the choices I made. Eight changes were made individually to the existing models and the tests were ran again, the original results did not change significantly.

A potential concern is that the categories used to make the independent variables more measurable are driving the results. While the choices made are in line with previous research and done with the greatest care, it cannot be ruled out that they influence the results. Thus, the independent variables that are categorized were adjusted to detect any changes in the results. The first independent variable is university, based upon the QS World University Rankings 2018. While this is the most used and highest regarded ranking, its’ methods have been questioned7. The second most used ranking, Times Higher Education World University Rankings 2018, was then categorized and compared. This resulted in 9 out of 86 universities and 41 out of 367 observations being in a different category. The tests were rerun with the new university rankings and the results of the university variable were largely unchanged. Untabulated results show that the university variable is still significantly correlated with

materiality. Next the experience variable was examined for possible bias. Experience was also categorized into three categories to enable clearer results. The tests were rerun with the uncategorized experience variable to find out if this would change the results. The untabulated results of the complete model show that they were largely unaffected, only experience to number of words audit report became more significant (from the 1 percent level to the 0.1 percent level). For the experience only model, number of words audit report and materiality lost their significance, while number of key audit matter became significant at the 1 percent level. These robustness tests give additional evidence that the results are not driven by choices made regarding independent variables.

To assure that the results in the main tests are not driven by the grouping variable, the year variable is replaced by both the clients’ industry and audit firm variable. The clients’ industry could be driving the result, as shown for example by Moroney and Simnett (2009) who used the difference in audit risks between industries in their study. Next to that, the audit firm could act as a moderating factor on the way auditors write their expanded audit report (Jeppesen 2007). The original results are largely unchanged when the general industry classification of the audited firms is used as the grouping variable. The significant results for gender and university all stayed significant. For experience, only the number of words audit report is no longer significant. The significance for degree shifted from number of words audit report, number of words key audit matter paragraph and average number of words per key audit matter towards materiality. When the results (untabulated) are clustered by audit firm, a lot of significance is lost. Gender only stays significant at the 5 percent level for average words per key audit matter, while experience remains significant for materiality (5 percent level) and number of key audit matters (0.1 percent level). University remains significant, albeit down from significance from the 1 percent level to the 5 percent level. Lastly, degree is no longer significant for 1 out 3 dependent variables, namely number of words key audit matters paragraph. In sum, it can be concluded that the original test are largely robust to changes in the grouping variable. The results are not driven by a clients’ industry specific risks, while the audit firm only moderately affects the results.

Client size and complexity could influence the results as they impact the risks and materiality judgements an auditor makes (Parmar, 2015) While these effects are controlled for in the original models, it would be wise to distinguish between larger and smaller firms. Thus, the database was divided into the 50 percent smallest and 50 percent largest firms. The untabulated results indicate that the results mostly hold for the largest clients and only partly for the smaller clients. For the largest firms, only the association of gender to number of words

audit report, university to materiality and degree to average words per key audit matter lose their significance. The model with the 50 percent smallest firms is less robust, as 4 out of 12 significant results lost their significance. These results show that the original model is quite robust to client effects with the control variables that were included, albeit more to the larger than the smaller clients.

V. Summary and Conclusions Conclusions

This research examines the impact of individual auditors on their audit report. A regulatory change in the United Kingdom, the introduction of the expanded audit report, gave the opportunity to explore this relationship. I document a clear systematic association between auditor characteristics and the way they write their expanded audit report. Significant variation exists on measures of gender, experience, university attended and degree obtained. Females are shown to write a more elaborate expanded audit report, while men show a higher risk appetite in their audits by way of a higher materiality when performing the audit. The size of the audit report is also influenced by auditor experience, as auditors with less experience write a more elaborate expanded audit report. Contrary to this, I show that more experienced auditors use a higher materiality than their less experienced counterparts. Materiality is also affected by the university an auditor attended, as auditors from higher regarded universities permit themselves higher materiality levels. Lastly, the type of degree an auditor obtained has an effect on the number of words per key audit matter. This suggests that auditors with a non-financial degree are superior in describing the key audit matters they found.

What is interesting to find is that the association found in the characteristic only models are strengthened in the complete model were all auditor characteristics are regressed together. Including all auditor characteristics resulted in existing associations to become more significant, and associations that were not significant to become significant. This hints at a strengthening effect when multiple auditor characteristics coexist. My findings indicate that an auditors’ characteristics enhance each other in ways not seen before.

The results are robust to other measures of independent variables and clustering by other variables. However, the results are less robust to client effects, which show that they are more reliable for bigger than smaller clients.

Limitations

Although the results show that the auditor characteristics used in this research explain differences in the structure and content of the expanded audit report, there is still a portion of this variation that remains unexplained. The auditor characteristics that were used are only a small set of the numerous characteristics that influence the way they compile their expanded audit report. Future research should try to find data related to for example leisure activities, family influences and industry expertise. The collection of characteristics was limited by the

voluntary disclosure of this information by the auditors. Another recommendation for future research would be that they should try to gain more extensive data on auditor characteristics by collecting it straight from the source instead of through social media. Although the interdependencies of individual auditor characteristics is shown in my research, it would be interesting for future research to find out which characteristics strengthen and which weaken each other. Future research could also search for the underlying reasons on why particular characteristics influence the way an auditor writes his expanded audit report. While the moderating effect of audit firms is shown, it would be interesting to find out which controls are used by audit firms to limit the leeway of their signing auditors. Another limitation of this research is that it forced to use data from immediately after the introduction of the expanded audit report. The risk here is that the learning curve resulting from this regulatory change influences the results. Thus, future research could replicate this research over more years to counteract this effect.

References

Abdolmohammadi, M., & Wright, A. (1987). An examination of the effects of experience and task complexity on audit judgments. The Accounting Review 62(1): 1–13.

Bens, D., Chang, W. J. & Huang, S. (2017). The association between the expanded audit report and investor uncertainty. Working paper.

Bertrand, M., & Schoar, A. (2003). Managing with style: The effect of managers on firm policies. The Quarterly Journal of Economics 118 (4): 1169-1208.

Bills, K. L., Cunningham, L. M. & Myers, L. A. (2016). Small audit firm membership in associations, networks and alliances: implications for audit quality and audit fees. The

Accounting Review 91 (3): 767-792.

Bonner, S. (1991). Is experience necessary in cue measurement? The case of auditing tasks.

Contemporary Accounting Research 8(1): 253–269.

Bonner, S. and Walker, P., L. (1994). The Effects of Instruction and Experience on the Acquisition of Auditing Knowledge. Accounting Review 69(1): 157-178.

Borkowski, S.C. and Ugras, Y.J. (1992). The Ethical Attitudes of Students as a Function of Age, Sex and Experience. Journal of Business Ethics 11: 961-979.

Caglio, A. & Cameran, M. (2017). Is it shameful to be an accountant? GenMe perception(s) of accountants’ ethics. Abacus 53 (1).

Chin, C. and Chi, H. (2008) Gender Differences in Audit Quality. American Accounting

Association Annual Meeting.

Choi, J., H., Kim, C. Kim, J., B. & Zang, Y. (2010) Audit office size, audit quality and audit pricing. Auditing: A Journal of Practice & Theory: 29(1): 73-97.

Chtioui, t. & Thiery-Dubuisson, S. (2011) Hard and Soft Controls: Mind the Gap! International

Journal of Business 16(3): 289-302.

Chung, J. & Monroe, G., S. (2001). A research note on the effects of gender and task complexity on the audit judgement. Behavioral Research in Accounting 13: 111-125.

Church, B., Davis, S. & McCraken, S. (2008). The auditor’s reporting model: A literature overview and research synthesis. Accounting Horizons 22 (1): 69-90.

Cordoş, G., S. & Fülöp, M., T. (2015) Understanding audit reporting changes: introduction of Key Audit Matters. Accounting & Management Information Systems 14(1): 128-152. DeAngelo, A. & Elizabeth, L. (1981). Auditor size and audit quality. Journal of Accounting &

DeFond, M., & Francis, J. (2005). Audit research after Sarbanes-Oxley. Auditing: A Journal of

Practice & Theory 24: 5-30.

Deis, D. R. & Guiroux G. A. (1992). Determinants of audit quality in the public sector. The

Accounting Review 67 (3): 462-479.

Dyreng, S., Hanlon M. & Maydew E. (2010). The effects of executives on corporate tax avoidance. The Accounting Review 85 (4): 1163-1189.

Fakhfakh, M. (2016). Empirical Investigation of Readability and Communicative Performance of Auditor's Report Post ISA 700. IUP Journal of Accounting Research & Audit

Practices 15(4): 64-96.

Financial Reporting Council. (2013). International Standard on Auditing (UK and Ireland) 700: The Independent Auditor’s Report on Financial Statements. London, UK.

Francis, J. (2004). What do we know about audit quality? The British Accounting Review 36 (4): 345-368.

Fellner, G. & Maciejovsky, B. (2007). Risk Attitude and Market Behavior: Evidence from Experimental Asset Markets. Journal of Economic Psychology 28: 338-350.

Gold, A., Hunton, J., E. & Gomaa M., I. (2009). The impact of client and auditor gender on auditors’ judgments. Accounting Horizons 23 (1): 1–18.

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.

Hardies, K., Breesch, D. & Branson, J. (2010). Are female auditors still women? Analyzing the sex differences affecting audit quality. Working paper.

Henry, E. (2008). Are investors influenced by how earnings press releases are written? Journal

of Business Communication 45 (4): 363-407.

Hsiao, C. (2007) Panel Data Analysis - Advantages and Challenges. Test 16(1): 1-22.

IAESB (2012). International Education Standard 7: Continuing professional development. New

York, United States of America, IAESB.

IAESB (2014). International Education Standard 8: Professional competence for engagement partners responsible for audits of financial statements. New York, United States of

America, IAESB.

Jager, P., D. (2008). Panel Data Techniques and Accounting Research. Meditari Acccountancy

Research 16(2): 53-68.

Jensen, M., C. & Meckling, W., H. (1976) Theory of the firm: managerial behavior, agency costs and ownership structure. Journal of Financial Economics (3): 305-360.

Jeppesen, L. (2007). Organizational risk in large audit firms. Managerial Auditing Journal 22 (6): 590-603.

King, R., R., Davis, S., M. & Mintchik, N. (2012) Mandatory Disclosure of the Engagement Partner’s Identity: Potential Benefits and Unintended Consequences. Accounting

Horizons 26(3): 533-561.

Latridis, G., E. (2012). Audit quality in common-law and code-law emerging markets: Evidence on earnings conservatism, agency costs and cost of equity. Emerging Markets Review 13: 101-117.

Lennox, C., S., Schmidt, J. J. & Thompson, A. M. (2015). Is the expanded model of audit reporting informative to investors? Evidence from the UK. Working paper.

Lennox, C., S. & Wu, X. (2017). A review of the literature on audit partners. Working paper. Libby, R., & Frederick, D. (1990). Experience and the ability to explain audit findings. Journal

of Accounting Research: 28(2), 348–367.

Libby, R., & Luft, J. (1993). Determinants of judgement performance in accounting setting: Ability, knowledge, motivation, and environment. Accounting, Organizations and

Society: 18(5), 425–450.

Lim, C., & Tan, H. (2010). Does auditor tenure improve audit quality? Moderating effects of industry specialization and fee dependence. Contemporary Accounting Research 27: 923–957.

Limperg Institute (1985). The social responsibility of the auditor. A basic theory on the auditor's function. Amsterdam, The Netherlands, Limperg Institute

Moroney R. & Simnett, R. (2009) Differences in Industry Specialist Knowledge and Business Risk Identification and Evaluation. Behavioral Research in Accounting 21(2): 73-89. Parmar, C., K. (2015) Audit Planning: Materiality Concept. International Journal of Research

in Commerce, Economics and Management 5(2): 41-43.

Peecher, M. E., Solomon I., & Trotman, K. T. (2013). An accountability framework for financial statement auditors and related research questions. Accounting, Organizations

and Society 38 (8): 596-620.

Peel, M., J. (1997). UK Auditor Concentration: A Descriptive Note. Accounting & Business

Research 27(4): 311-322.

Reid, L. C., Carcello, J. V., Li, C. & Neal, T. L. (2015). Impact of auditor and audit committee report changes on audit quality and costs: evidence from the United Kingdom. Working

Simnett, R. (1996). The effect of information selection, information processing and task complexity on predictive accuracy of auditors. Accounting, Organizations and Society: 21(7), 699–719.

Smith, K. W. (2016). Tell me more: a content analysis of expanded auditor reporting in the United Kingdom. Working paper.

Smith, M. (2015) Research Methods in Accounting 3rd Edition. London: Sage Publications Ltd.

Tan, H., & Kao, A. (1999). Accountability effects on auditors’ performance: The influence of knowledge, problem-solving ability, and task complexity. Journal of Accounting

Research: 37(1), 209–223.

Tan, H., N, T., & Mak, B. (2002). The effects of task complexity on auditors’ performance: The impact of accountability and knowledge. Auditing: A journal of Practice & Theory: 21(2), 81–95.

Trotman, K., Wright, A. & Wright, S. (2009). An examination of the effects of auditor rank on renegotiation judgments. Auditing: A Journal of Practice & Theory 28 (1): 191-203. Weinrich, T., R. & Reinstein A. (2014). The PCAOB’s Proposed New Audit Report. The CPA

Journal 84(4): 24-29.

Wild, J., J. (1996). The Audit Committee and Earnings Quality. Journal of Accounting, Auditing

& Finance 11(2): 247-276.

Wu, C., Y., Hsu, H., H. & Haslam, J. (2016). Audit committees, non-audit services, and auditor reporting decisions prior to failure. The British Accounting Review 48: 240-256.

Appendix A QS World University Rankings 2018

UK Rank World Rank University

1 5 University of Cambridge

2 6 University of Oxford

3 7 UCL (University College London)

4 8 Imperial College London

5 23 King's College London

6 23 The University of Edinburgh 7 34 The University of Manchester

8 35 London School of Economics and Political Science (LSE) 9 44 University of Bristol

10 57 The University of Warwick 11 65 University of Glasgow

12 78 Durham University

13 82 The University of Sheffield 14 84 The University of Nottingham 15 84 University of Birmingham 16 88 University of Dublin 17 92 University of St. Andrews 18 101 University of Leeds 19 102 University of Southampton 20 113 Rijksuniversiteit Groningen 21 127 Queen Mary University of London 22 135 Lancaster University

23 135 University of York 24 137 Cardiff University 25 158 The University of Exeter 26 158 University of Aberdeen 27 160 University of Bath 28 161 Newcastle University 29 168 University College Dublin 30 173 University of Liverpool 31 188 University of Reading 32 191 University of Cape Town 33 202 Queen's University Belfast 34 228 University of Sussex 35 234 Loughborough University 36 238 University of Leicester

37 243 National University of Ireland Galway 38 259 Royal Holloway University of London 39 264 University of Surrey

40 267 University of Dundee

41 274 University of East Anglia (UEA) 42 277 University of Strathclyde

43 283 University College Cork 44 296 SOAS University of London 45 308 Birkbeck, University of London 46 312 Heriot-Watt University

47 343 City, University of London 48 346 Brunel University London 49 347 University of Essex

50 361 Oxford Brookes University 51 373 University of Kent

52 373 Aston University

53 391 Dublin City University

54 398 Goldsmiths, University of London 55 431-440 Swansea University

56 441-450 Bangor University 57 451-460 University of Stirling 58 481-490 Aberystwyth University 59 501-550 Kingston University London 60 501-550 University of Limerick 61 551-600 Coventry University 62 551-600 University of Westminster 63 601-650 Keele University 64 601-650 University of Bradford 65 601-650 University of Hull 66 601-650 University of Portsmouth 67 601-650 Ulster University

68 651-700 Dublin Institute of Technology 69 701-750 Bournemouth University

70 701-750 London Metropolitan University 71 701-750 Maynoon University

72 701-750 Middlesex University 73 701-750 University of Brighton 74 701-750 Plymouth University

75 751-800 Northumbria University at Newcastle 76 751-800 University of Hertfordshire

77 751-800 University of Huddersfield 78 751-800 University of Salford

79 801-1000 Edinburgh Napier University 80 801-1000 London South Bank University 81 801-1000 Manchester Metropolitan University 82 801-1000 Nottingham Trent University

83 801-1000 Robert Gordon University 84 801-1000 University of Central Lancashire 85 801-1000 University of East London 86 801-1000 University of Greenwich

Appendix B

Financial degree Non-financial degree

Accountancy

Accountancy and Finance

Accountancy, Accounting and Finance Accounting

Accounting and Business Management Accounting and Computer Science Accounting and Finance

Accounting and Financial Management

Accounting and Marketing Accounting and Mathematics Accounting, Statistics and IT Agriculture Economics Business Administration Business and Accountancy Business and Technology

Business Economics and Accounting Business Studies

Commerce Economics

Economics and Accountancy Economics and Accounting Economics and Politics Economics and Social History Economics History

EMA Accountancy Law and Accountancy

Modern Languages with Business Studies

Philosophy, Politics and Economics Psychology and Economics

Applied Mathematics Astrophysics

Biochemistry Applied Medical Biotechnology

Chemistry

Chemistry with Management Studies Civil Engineering

Classical Studies Classics

Consulting and Coaching for Change Engineering

Engineering and Management Engineering Science

English Language and Literature Geography

History

History and Politics Latin

Law

LLB AKC Law Mathematics

Mathematical Science Mathematics and Statistics Maths and Management sciences Mechanical Engineering

Microbiology Modern Languages Natural Sciences Physics

Physics and Mathematics Pure and Applied Mathematics Statistics

TABLE 1

Descriptive statistics for dependent variables

VARIABLES N mean sd min max

Number of words audit report 602 2,694 882.6 1,208 5,503

Ln of Materiality 489 16.569 0.061 12.766 21.239

Key Audit Matters 490 4.006 1.574 0 10

Number of words KAM 490 1,001 587.3 55 3,135

TABLE 2

Descriptive statistics for independent variables Panel A: descriptive statistics

VARIABLES N mean sd min max

Auditor Gender 591 0.102 0.302 0 1

Auditor Experience 448 26.75 6.156 9 41

Auditor Experience 448 1.234 0.663 0 2

Auditor University 329 0.401 0.701 0 2

Auditor Degree 331 0.441 0.497 0 1

Panel B: frequency table

Freq Freq Freq Freq

Gender (Percent) Experience (Percent) University (Percent) Degree (Percent)

0 531 0 58 0 238 0 185 (89.85) (12.95) (72.34) (55.89) 1 60 1 227 1 50 1 146 (10.15) (50.67) (15.20) (44.11) 2 163 2 41 (36.38) (12.46)

TABLE 3

Descriptive statistics for control variables

VARIABLES N mean sd min max

Number of Employees 463 31,700 67,891 22 514,718

Ln Total Assets 490 22.18 1.830 17.84 27.90

Ln Revenue 486 21.44 1.777 17.06 26.14

Ln Profit Before Taxes 452 19.48 1.400 16.01 23.82

ROA 449 0.101 0.0796 0.00128 0.451

Current Ratio 462 1.557 1.045 0.303 6.540

Leverage 592 77.21 164.0 -660.4 986.5

TABLE 4

Analysis of auditor characteristics effects on expanded audit report #Words Audit Report Materiality # Key Audit Matters #Words KAM Average #Words KAM Gender 357.6* _-0.108*** _{0.211 298.4}* _90.79*** (2.15) (-8.00) (0.54) (2.22) (6.14) Experience 53.85** _0.0912*** _0.344** _92.48*** _16.23 (3.00) (4.33) (2.74) (4.51) (1.26) University 29.16 0.0650** _{0.0124 -3.065 -0.827} (0.36) (2.60) (0.12) (-0.12) (-0.05) Degree 237.5* _{-0.0191 0.315 215.1}** _38.74** (2.25) (-0.97) (1.07) (2.59) (3.26) Employees 0.001 -0.000 0.000 0.001 0.0001 (0.49) (-1.72) (1.67) (0.73) (0.47) Total Assets 215.3** _0.938*** _{-0.198 120.0 46.78}*** (3.12) (4.01) (-0.59) (1.08) (3.92) Revenue 34.05 0.0215 0.272* 34.71 -19.34** (1.01) (0.74) (2.22) (0.71) (-3.12) ProfitBeforeTax -101.3* -0.0247 0.115 -107.3 -31.98* (-2.52) (-0.10) (0.44) (-0.99) (-2.11) ROA 455.6*** 5.237** -5.068* -50.11 305.3* (3.85) (3.06) (-2.11) (-0.09) (1.98) Current ratio -141.2*** 0.0256 -0.394** -93.01*** -3.683 (-4.57) (1.37) (-2.82) (-3.73) (-1.01) Leverage 0.134 -0.000 0.0012*** _{0.132 -0.007} (1.42) (-1.89) (11.40) (0.87) (-0.28)

Year dummies Yes Yes Yes Yes Yes

Auditor dummies Yes Yes Yes Yes Yes

Industry dummies Yes Yes Yes Yes Yes

_cons -52.95 -4.606*** 0.292 -109.0 271.1*** (-0.04) (-6.16) (0.37) (-0.17) (3.56) N R-squared R2 _{w/o IVI} Change in R2 Incremental R2 139 0.5299 0.4556 0.0743 16.31% 139 0.9670 0.9503 0.0167 1.76% 139 0.4671 0.3134 0.1537 49.04% 139 0.4365 0.3085 0.1280 41.49% 139 0.4699 0.4274 0.0425 9.94% t statistics in parentheses * _{p < 0.05,} ** _{p < 0.01,} *** _{p < 0.001}

TABLE 5

Analysis of auditors’ gender effects on expanded audit report #Words Audit Report Materiality # Key Audit Matters #Words KAM Average #Words KAM Gender 164.1*** _-0.109** _{-0.155 38.50 14.78} (4.44) (-3.13) (-0.83) (0.99) (1.71) Employees -0.0000 -0.0000 0.0000 0.0005 0.0000 (-0.02) (-1.33) (1.62) (0.42) (0.09) Total Assets 412.0*** 0.893*** 0.346 258.1*** 37.32 (4.53) (10.65) (1.48) (3.83) (1.45) Revenue 6.394 -0.0168 0.242** 19.05 -16.65 (0.13) (-0.29) (2.95) (1.58) (-1.32) ProfitBeforeTaxes -280.0*** 0.0519 -0.301 -209.0*** -26.34 (-5.06) (0.94) (-1.93) (-4.21) (-1.27) ROA 1610.3*** 5.083*** 0.777 1083.0*** 178.1 (3.97) (15.73) (0.53) (4.83) (0.73) Current Ratio -35.54 -0.00242 -0.114 -20.33 1.541 (-0.80) (-0.29) (-1.06) (-0.86) (1.16) Leverage -0.0879 -0.0000 0.0000 -0.0553 -0.0254** (-1.13) (-0.79) (0.07) (-0.51) (-2.94)

Year dummies Yes Yes Yes Yes Yes

Auditor dummies Yes Yes Yes Yes Yes

Industry dummies Yes Yes Yes Yes Yes

_cons -281.7 -4.202*** _-2.563** _{-770.2 361.1}*** (-0.25) (-21.31) (-2.70) (-1.57) (3.81) N 306 306 306 306 306 R-squared R2 w/o IVI Change in R2 Incremental R2 0.4654 0.4556 0.0098 2.15% 0.9507 0.9503 0.0004 0.04% 0.3472 0.3134 0.0338 10.78% 0.3088 0.3085 0.0003 0.10% 0.4335 0.4274 0.0061 1.43% t statistics in parentheses * _{p < 0.05,} ** _{p < 0.01,} *** _{p < 0.001}

TABLE 6

Analysis of auditors’ experience effects on expanded audit report #Words Audit Report Materiality # Key Audit Matters #Words KAM Average #Words KAM Experience -59.30*** _0.0613* _{0.0624 -34.86 -8.790} (-7.32) (2.34) (1.09) (-1.61) (-1.06) Employees 0.0005 -0.0000 0.0000 0.001 0.0000 (0.30) (-1.74) (1.83) (0.68) (0.32) Total Assets 371.9*** 0.907*** 0.106 245.4** 42.12* (3.34) (7.57) (0.48) (2.72) (2.37) Revenue 34.10 -0.00231 0.205* 11.07 -16.64 (0.70) (-0.03) (2.13) (0.30) (-1.22) ProfitBeforeTaxes -254.9* 0.0290 -0.0216 -184.5 -32.89 (-2.30) (0.39) (-0.13) (-1.73) (-1.69) ROA 1055.4* 5.131*** -1.771 549.7 199.7 (2.35) (9.39) (-1.10) (1.38) (0.84) Current Ratio -61.49*** _{-0.0014 -0.118 -22.87}* _-3.307 (-4.81) (-0.16) (-1.24) (-1.99) (-0.67) Leverage -0.0845 -0.0000 0.0006*** -0.0048 -0.0314 (-0.77) (-1.90) (3.57) (-0.05) (-1.69)

Year dummies Yes Yes Yes Yes Yes

Auditor dummies Yes Yes Yes Yes Yes

Industry dummies Yes Yes Yes Yes Yes

_cons -308.0 -4.430*** _-2.253* _{-635.5 423.7}** (-0.22) (-14.57) (-2.10) (-0.94) (3.13) N 232 232 232 232 232 R-squared R2 w/o IVI Change in R2 Incremental R2 0.5048 0.4556 0.0492 10.80% 0.9509 0.9503 0.0006 0.06% 0.3587 0.3134 0.0453 14.45% 0.3638 0.3085 0.0553 17.93% 0.4276 0.4274 0.0002 0.05% t statistics in parentheses * _{p < 0.05,} ** _{p < 0.01,} *** _{p < 0.001}

TABLE 7

Analysis of auditors’ university attended effects on expanded audit report #Words Audit Report Materiality # Key Audit Matters #Words KAM Average #Words KAM University -46.46 0.0940** _{0.0476 -37.70 -13.38} (-0.74) (2.69) (0.57) (-1.58) (-1.80) Employees 0.000921 -0.0000** _{0.0000 0.0013 0.0001} (0.45) (-2.60) (1.38) (0.83) (0.61) Total Assets 266.8*** 1.004*** 0.170 217.3*** 49.42** (4.39) (5.32) (0.53) (3.90) (2.58) Revenue 67.28** -0.0207 0.332** 46.06* -19.57*** (3.27) (-0.32) (2.87) (2.21) (-5.82) ProfitBeforeTaxes -176.7*** -0.0520 -0.242 -191.8*** -31.94 (-8.50) (-0.39) (-1.05) (-5.00) (-1.42) ROA 1346.9* 5.536*** -1.321 1121.7** 365.9 (2.36) (5.04) (-0.55) (2.77) (1.64) Current Ratio -137.5*** _{0.0240 -0.335}*** _-71.56*** _-0.496 (-3.76) (1.31) (-3.40) (-3.87) (-0.16) Leverage 0.0362 -0.0001 0.00103*** 0.0394 -0.0231 (0.35) (-1.83) (22.51) (0.49) (-1.59)

Year dummies Yes Yes Yes Yes Yes

Auditor dummies Yes Yes Yes Yes Yes

Industry dummies Yes Yes Yes Yes Yes

_cons -304.8 -4.499*** _-1.912** _{-713.2 267.1}*** (-0.23) (-10.92) (-2.92) (-1.66) (3.95) N 159 159 159 159 159 R-squared R2 w/o IVI Change in R2 Incremental R2 0.5056 0.4556 0.0500 10.97% 0.9598 0.9503 0.0095 1.00% 0.4276 0.3134 0.1142 36.44% 0.3874 0.3085 0.0789 25.58% 0.4371 0.4274 0.0097 2,27% t statistics in parentheses * _{p < 0.05,} ** _{p < 0.01,} *** _{p < 0.001}