The Impact of Gender Diversity in Audit Committees on Earnings Management and Key Audit Matters

(1)

The Impact of Gender Diversity in Audit

Committees on Earnings Management and Key

Audit Matters

A quantitative study among UK and Dutch stock listed companies

ABSTRACT

This study examines whether gender diversity in the audit committee has a significant relationship with both earnings management and the number of reported key audit matters in the auditor’s report. Using the adjusted Jones Model, this study was unable to identify a significant association between the presence of female directors and the extent of earnings management for a sample of 758 firm-year observations in the UK and the Netherlands, during the period 2013-2016. Furthermore, the results show no significant association between the audit committee’s gender diversity characteristics and the number of reported key audit matters in the auditor’s report. However, the study did find several relationships among the control variables that lead to interesting directions for possible future research.

KEYWORDS: Gender Diversity, Audit Committee, Earnings Management, Key Audit Matters

R.F. Visser

S3268284

r.f.visser@student.rug.nl University of Groningen Faculty of Economics and Business

January 2018

Assessor: prof. dr. D.A. de Waard Co-Assessor: dr. R.C. Trapp

(2)

1

1 INTRODUCTION

Financial scandals, both in the past and more recent history, have caused public concerns about aggressive earnings management practices, especially by managers of public interest entities. These concerns have resulted in a call for more effective audit committees from regulators, shareholders and other related organizations. For example, the Sarbanes-Oxley Act (SOX 301) states that the audit committee is directly responsible for the appointment, compensation, and oversight of the auditor. This monitoring function of the audit committee is further emphasized after many investors lost their confidence in managers and external audits following the financial debacles of high-profile companies such as Enron, WorldCom, Ahold, and various other frauds related to financial reporting.

The purpose of this study is twofold. Firstly, this paper aims to examine whether gender diversity in the audit committee has a significant impact on the amount of earnings management. Secondly, this paper addresses the research question if gender diverse audit committees have an impact on the number of reported key audit matters in the auditor’s report.

A large amount of academics have researched the phenomenon earnings management. Next to that, various studies have been conducted on the relationship between the audit committee and earnings management. However, most of these studies focused on other aspects than gender diversity. For example, Klein (2002) found a non-linear negative relationship between audit committee independence and earnings management. The study of Bédard, Chtourou, and Courteau (2004) examined the relationship between a firm’s audit committee characteristics and managers’ earnings management. They found that the presence of a financial expert, a committee composed solely of independent directors, and a clear mandate to oversee the financial reporting processes, and the audit are negatively related with the likelihood of aggressive earnings management practices.

Similarly, the relation between earnings management and earnings management and gender diversity has been studied increasingly lately. However, these studies focused on the gender diversity in the board of directors, and not the audit committee specifically. For example, researchers have found that firms with a higher number of (independent) female directors are adopting restrained earnings management practices (Kyaw et al., 2015; Arun et al., 2015). Gull et al. (2017) found that when female directors are appointed to the audit committee they promote effective monitoring of earnings management.

Despite the conclusive evidence that gender diversity in the board has a significant impact on earnings management, mixed results have been presented by researchers who studied earnings management in relation to gender diversity among audit committees. Thiruvadi & Huang (2011) [hereafter: TH (2011)] examined 320 firms for the fiscal year 2003, and found consistent evidence to show that the presence of a female director on the audit committee constrains earnings management by increasing negative (income-decreasing) discretionary accruals. Contradictory evidence is presented by Sun, Liu, and Lan (2011). They find no gender effect with respect to independent audit committees’ effectiveness in constraining earnings management.

In the second part of this study the extended auditor’s report is addressed. The specific subject of interest is the key audit matters section of this report. Auditor standards state that the auditor is required to:

“Communicate key audit matters that in the auditor’s professional judgment were of most significance in the audit of the financial statements” (IFAC, 2017, p. 2).

(4)

3

Key audit matters are a relatively new subject for academics, as it is only mandated since 2013. Therefore, little research has been conducted on the factors that determine these key audit matters. As a result, an important question in both the academic world and society arises on what drives auditors to report key audit matters.

To summarize abovementioned, the following research questions (RQ) are formulated: RQ1: Does female presence in the independent audit committee constrain earnings management? RQ2: Does female presence in the independent audit committee affect the number of reported key

audit matters?

1.1 Academic contribution

The first part of this paper is based on the study by TH (2011) and extends the study of TH (2011) and related literature in several ways. Firstly, by focusing on other environmental elements and data sampling. The TH (2011) study used data from the United States (US), whereas this paper will use data gathered from the United Kingdom (UK) and the Netherlands. Secondly, this paper extends the existing literature and provides empirical evidence to fill the current knowledge gap regarding gender diversity and its implications for earnings management. Furthermore, this paper addresses the importance of women on corporate boards, in particular the audit committee. The outcomes could also provide insights for regulators and policy makers in the presence of the contemporary discussion about the appointment of women on important corporate positions.

The second part of this paper contributes to the existing knowledge gap regarding the determinants of reported key audit matters. Since there is little research on this topic in the literature, this paper sheds more light on this question and tries to fill a modest part of the knowledge gap with regard to the determinants of key audit matters. Hence, this paper responses to the call of Köhler et al. (2016) to conduct research to provide any further insights on this subject as they state:

“With regard to the identification of KAM, future research could also elaborate on the impact of the communication process between the auditor and the Audit Committee and on the impact of Audit Committee characteristics” (Köhler et al., 2016, p. 213).

The remainder of this paper is structured as follows. Chapter 2 provides an overview of the relevant background, the theoretical framework and the hypothesis development. This is followed, in chapter 3, by a description of the research methodology and, in chapter 4, by an analysis of the results. Last, chapter 5 reports on the conclusion and discussion of the paper.

(5)

4

2 THEORETICAL FRAMEWORK

In this chapter the underlying theory of this research is discussed. Furthermore, background information is provided and the hypothesis are formulated in order to answer the main research questions of this study.

2.1 Earnings Management, Agency Theory, and the Audit Committee

Managers may engage in earnings management as a result of perhaps the widest known and most studied auditing subject: agency theory (Jensen & Meckling, 1976). Agency theory explains the relationship between the separation of ownership and control in which the manager acts as ‘agent’ and the shareholder as ‘principal’. Agency theory states that the manager (agent) is more likely to pursue their self-interest above the interest of the shareholder (principal). As a result, shareholders introduced (internal) governance mechanisms which align the interest of the agent with the principal’s. In other words, to reduce this conflict, agents have linked the manager’s compensation plan to certain accounting numbers so that their interests are better aligned. However, manipulation of these accounting numbers may arise as managers still have the incentive to pursue their own interest. This is supported by a study of Healy (1985), who found an association between discretionary accruals (attributed to earnings management) and the existence of earnings-based bonus plans, which suggests that management compensation may be an incentive for managers to engage in earnings management. Similarly, Fudenberg and Tirole (1995) argue that managers have incentives to manage earnings for their job security. As a result, managers may use positive discretionary accruals to increase their accounting earnings, and thus, their compensation plan.

When internal controls are insufficient, external control mechanisms may become effective. In this occasion, the audit committee could become useful as a corporate governance mechanism in controlling aggressive earnings management. Klein (2002) states that the main goal of the audit committee is that:

“…the audit committee’s role as arbiter between the two parties is to weigh and broker divergent views of both parties to produce ultimately a balanced, more accurate report. Equivalently, its role is to reduce the magnitude of positive or negative discretionary accruals” (p. 378).

TH (2011) argue that when the audit committee is very effective: (i) the management will be less aggressive; and (ii) the auditor will be more conservative, and thus make use of negative discretionary accruals to decrease the accounting earnings of the client. Furthermore, they argue that conservative earnings reporting is viewed as good financial reporting since it provides higher earnings quality, and also reduces the probability of potential shareholder litigation at the same time.

2.2 Gender theory and the influence of gender diversity

As mentioned in the introduction, researchers have focused on different characteristics of either the board of directors, or the audit committee. As a result of the recent debate about (gender) diversity and women on corporate boards1_{, gender diversity became more prominent in current research.}

1 For a clear discussion about the topic of women on boards and the development of related studies, I refer to the study of Seierstad et al. (2017).

(6)

5

Harrison and Klein (2007) describe three distinctive types of diversity: separation, variety, and disparity. Harrison and Klein (2007) state that: “suggesting that men and women have qualitatively

different caches of knowledge … scholars invoke gender diversity as variety” (p. 1209). Hence, gender

diversity can be defined as the variety inherent in the board or audit committee’s composition. This variety can be measured on a number of dimensions: gender, age, ethnicity, nationality, educational background and industrial expertise, among others. This study focuses on gender, which is arguably the most debated issue in the current era.

Abundant arguments are valid to state that female directors provide benefits and therefore should be included in the audit committee, this is elaborated in the remainder of this paragraph. Therefore, the remainder of the paragraph contains a discussion and overview of research on the following topics: (i) gender diversity on boards, (ii) differences in ethical behavior by males and females, and (iii) gender diversity in the audit committees.

Firstly, board diversity in the context of agency theory is considered in the study of Carter et al. (2003). The results of Carter et al. (2003) suggest that greater diversity may increase the independence as women are more inclined to ask questions that would not be asked by male directors. Furthermore, it is argued that board diversity can enhance problem-solving as the variety of perspectives that emerges from a more diverse board means that more alternatives are evaluated (Campbell & Mínguez-Vera, 2008). By taking a broader view, a better understanding of the complexities that are faced is obtained, and therefore the decision-making process will improve. As shortly mentioned in the introduction, Arun et al. (2015) examined how the presence of female directors on the corporate board influenced earnings management practices. Using 1,217 firm-year observations, they found that firms with a higher number of (independent) female directors are adopting restrained earnings management practices in the UK. Subsequently, they examined whether this relationship exists in different types of firms and found that low-debt firms tend to be more conservative than high-debt firms. The study of Kyaw et al. (2015) examined the effects of board gender diversity on earnings management in European countries. The findings reveal that a gender diverse board mitigates earnings management in countries where gender equality is high. Considerable evidence is provided by Peni and Vähämaa (2010) who suggest that firms with female CFOs are associated with income-decreasing discretionary accruals. This is supported by the study of Francis et al. (2015), who focus on the transition of male to female CFOs and compare the firms’ degree of accounting conservatism between pre- and post-transition periods. Overall, the study provides strong support for the notion that female CFOs are more risk averse than male CFOs. As a result, female CFOs tend to adopt more conservative financial reporting policies.

Secondly, the study of Marta, Singhapakdi and Kraft (2008) argues that “Some ethics studies

have shown women to be more ethically perceptive, and also that they tend to arrive at more ethical judgments” (p. 594). In addition, Chonko and Hunt (1985) found a greater perceptiveness of ethical

problems among female managers. To conclude the ethical characteristic, in a meta-analysis study conducted with data from more than 20,000 respondents across 66 samples, Franke, Crown and Spake (1997) concluded that women do show significantly higher ethical standards than men. However, based on social role theory, they find that these differences tend to decline with work experience.

Lastly, consistent evidence has shown that audit committees with at least one female director were more likely to meet more often than all-male audit committees (Thiruvadi, 2012). DeZoort et al. (2002) state that a greater meeting frequency is associated with a reduced incidence of financial reporting problems and greater external audit quality. Thus, as it is argued that more frequent audit

(7)

6

committee meetings lead to better (i.e. more effective) audit committee outcomes, it could be beneficial to employ female members as they are positively associated with a higher meeting frequency. More recent, Gull et al. (2017) investigate the relationship between female directors and earnings management by considering their specific attributes. They found that business expertise and audit committee membership are key attributes of female directors that promote the effective monitoring of earnings management. An important implication of their findings is that the decision to appoint women on corporate boards should be based on specific criteria, rather than blind implementation of gender quotas.

To summarize the aforementioned, I argue that when female members are present in the audit committee, the audit committee is supposed to be more effective, and therefore less earnings management occurs. To answer the first research question, the following hypothesis is formulated:

Hypothesis 1: There is a significant association between the presence of female members in the

audit committee and the amount of earnings management.

2.3 Key audit matters and the extended auditor’s report

The second research question of this paper addresses the key audit matters section in the extended auditor’s report. The extended auditor’s report has been introduced in 2013 (FRC, 2013), and since then been innovated and developed further by the audit firms (FRC, 2015; FRC, 2016a). The UK and the Netherlands have been pioneers regarding the implementation of this new report. In 2013 and 2014, respectively, the UK and the Netherlands have issued new auditing standards which mandated public interest entities, and other entities that are required, to communicate through the extended auditor’s report (FRC, 2016b; NBA, 2017).

Lennox, Schmidt and Thompson (2016) measured, using short-window market reactions, investors’ responses to the new risk disclosures (i.e. key audit matters). Their results indicate that investors do not find these disclosures incrementally informative. According to Lennox et al. (2016) this can be explained by the lack of information content that is attributable to investors being informed about the risks before the extended auditor’s report was introduced. In other words, the results suggest that investors already knew about the risks before the risks were disclosed by the auditors. Hence, investors were already informed about the vast majority of financial reporting risks before the extended auditor’s report was introduced.

Gutierrez et al. (2016) examined changes in audit fees, audit quality, and investors’ reaction two years pre and two years post the new report requirements in the UK. After examining the pre-post change they find that audit fees increased by approximately four percent. However, they note that a caveat of their findings is that there could be other compliance costs, absorbed by clients or auditors, which are not reflected in the audit fees. On the other hand, they do find that that the length of the auditor’s report, the length of the risks section, and the number of risks are positively associated with audit fees. This suggests that an association between auditor’s risk and effort and the report’s content exists.

In a study targeting the effects of the communication of key audit matters on auditor liability Gimbar, Hansen and Ozlanski (2016) find that precise accounting standards and Critical Audit Matters (CAM is the French version of KAM) result in a lower propensity for negligence verdicts by jurors and that the use of either imprecise standards or CAMs leads to increased auditor liability.

(8)

7

To the best of my knowledge, the paper of Sierra-Garcia et al. (2017) is the first paper that provides empirical evidence of a significant relationship between the number of key audit matters and client specific characteristics. The study focusses on the determinants of the number of key audit matters as reported by the auditor in the extended auditor’s report. The results provide evidence to support that auditors’ and clients’ characteristics are determinants of the number of reported key audit matters.

To summarize aforementioned, and with regard of the gender theory from section 2.2, it is uncertain how, and if, the audit committee’s characteristics influence the number of reported key audit matters in the auditor’s report. Therefore, the following hypothesis is formulated in order to answer the second research question:

Hypothesis 2: There is a significant association between the presence of female members in the

(9)

8

3 METHODOLOGY

In the previous chapter the theoretical framework and hypothesis were described, this chapter will explain the examination and testing of these hypothesis. In section 3.1 the dependent, independent, and control variables, as well as the measurements of the models are discussed. Section 3.2 elaborates on the data collection and sample distributions. This chapter concludes with a detailed explanation of the regression models in section 3.3.

3.1 Variables

Dependent variables

As explained in the previous chapters, this study uses a different dependent variable for each hypothesis. The first dependent variable of this study uses is discretionary current accruals, which are used as a measurement of the amount of earnings management of the firms.

Prior research has shown that managers are more likely to manage earnings through accruals since it is more difficult to be detected by outsiders (Dechow, Sloan, & Sweeney, 1995; Jones, 1991; Kothari, Leone, & Wasley, 2005). In line with Ashbaugh, LaFond, and Mayhew (2003), this study focuses on current accruals, since research suggests that management has the most discretion over current accruals (Becker et al., 1998). The most frequent used method for determining the discretionary current accruals is the modified Jones Model (Dechow et al., 1995). The modified Jones Model is explained in detail in section 3.3.

The dependent variable for the second hypothesis is the number of reported key audit matters. A similar variable is used in the studies of Bédard et al. (2014) and Lenox et al. (2016). The number of key audit matters is measured as the sum of individually counted key audit matters as reported by the auditor, represented in the auditor’s report. The key audit matters variable is stated as #KAM.

Independent variables

The gender diversity of the audit committee is the independent variable of both hypothesis. For that reason, the impact of gender diversity can be measured for both earnings management and the number of key audit matters.

The gender diversity of the audit committee is measured via the gender diversity variables, expressed as GENDER-VARS. In order to be robust, three different proxies are used to measure the gender diversity in audit committees; percentage of female members (FPCT), presence of female members (FEMD), and number of female members (FSIZE). FPCT is measured by dividing the total number of audit committee members by the number of female members. FEMD is a dummy variable in which 1 represents the presence of a female member, and 0 indicates that no female member is present in the audit committee. Finally, FSIZE is measured as the total number of female members in the audit committee and MSIZE is its male equivalent.

Control variables of discretionary accruals

In order to detangle the effect of the gender variables on the discretionary accruals some control variables are used in this study. Following prior research (i.e. TH, 2011; Ashbaugh et al., 2003), this study controls for size via the natural logarithm of the market value per year-end (LogMV). To avoid problems of scale, the variable is transformed using the natural logarithm. The market to book ratio

(10)

9

per year-end (MB) controls for firm complexity, and last year’s accruals (L1ACCRUAL) is included as control variable to capture the reversal effect of prior accruals.

The leverage (LVRG), performance (LOSS) and cash flow from operations (CFO) control for the effects of the firm’s risk and performance on earnings management. LVRG is measured as total assets less book value, divided by total assets. LOSS is a dummy variable in which the value 1 represents a loss (i.e. negative pretax income) in the observed year, and 0 otherwise. CFO is measured via the financial statements of the firms and represents the net cash flow from operating activities scaled by beginning of year total assets.

Furthermore, the model controls for the additional risk of obtaining new capitals (FINCG).

FINCG is a dummy variable and the value 1 is granted if the number of common shares outstanding

or the long-term debt increased by at least 10 percent in the observed year relative to prior year, and 0 otherwise.

Prior research has shown that big-four audit firms have significant influence on the amount of earnings management. For that reason a dummy variable (BIG4) is created and is given the value 1 if one of the big-four audit firms (Deloitte, EY, KPMG, and PwC) is the independent auditor of the firm, and 0 otherwise. Hence, BIG4 controls for the effects of auditor’s quality on discretionary accruals. Control variables of key audit matters

To measure the effects of the gender variables on the number of key audit matters, several control variables have been included in the model. Corresponding to the discretionary accrual model, this study controls for audit firms using the dummy variable BIG4. In contrast to Sierra-Garcia et al. (2017), the control variable BIG4 includes all four big-four auditors instead of the largest three.

Based on Hay et al. (2006), the control variable AUDITFEES is included in the model. This variable controls for the audit fees charged by the auditor to the client. To avoid problems of scale, the variable (AUDITFEES) is transformed using the natural logarithm. The materiality amount, as disclosed by the auditor in the auditor’s report, is controlled for via MATERIALITY (Gutierrez et al., 2016). Again, the natural logarithm is used to avoid problems of scale. The size of the firm is controlled via SIZE, which is measured as the natural logarithm of total assets of the observed year (Becker et al., 1998; Hay et al., 2006).

To control for the complexity of the audit, the current ratio (CURRENTRATIO) of the firm is used. The current ratio is measured as total assets divided by total current liabilities. To capture profitability, return on assets (ROA) is used. ROA is measured as net income before tax divided by total assets.

The variables LEVERAGE and LOSS remain unchanged for the key audit matters model with respect to the discretionary accrual model, and control for the firm’s risk and performance.

3.2 Data collection and sample distribution

Data collection

As mentioned in the introduction, companies from the UK and the Netherlands are subject to this study. The Dutch listed companies are separated in three different indices. The first 25 companies are listed on the Amsterdam Exchange Index (AEX), the second 25 companies are listed on the Amsterdam Midkap Index (AMX), and the last 25 companies are listed on the Amsterdam Small Cap Index

(11)

10

(AScX). Hence, in total 75 Dutch listed companies are included in the sample. The UK listed companies are divided over two indices. The first 100 companies are listed on the Financial Times Stock Exchange 100 Index (FTSE 100), the second part of UK companies are the top 150 listed companies of the FTSE 250. Combining both UK and Dutch companies, this results in a grand total of 325 companies. The reference date for the sample is set at fiscal year 2015 to make the results more generalizable for the present.

The start of data collection for the research project began in 2013. From that year forward, data has been hand collected from the companies’ annual reports. This period includes four fiscal years, and thus data for almost 1,300 firm-year observations is available. The dataset used for this research is combined from two data sources. On the one hand, the company and market specific information has been retrieved from the database of Thomson Reuters Datastream. On the other hand, data regarding the auditor and the composition of the audit committee is retrieved from the abovementioned hand collected database, available via the University of Groningen.

Inevitable to data collection for research purposes, not all data was available for each individual company. Moreover, some changes in the composition of the listed companies have emerged as a result of initial public offerings, bankruptcies, or merger and acquisitions. To make optimum use of the available data, different samples have been created for each hypothesis and therefore a large as possible sample size can be tested.

Sample distribution for hypothesis 1

With regard to the first hypothesis, the total sample contains 1,298 firm-year observations. In order with prior studies, firms in the financial sector are excluded because of their distinct industrial characteristics. These firms are excluded based on their Standard Industrial Classification (SIC) code. All observations in the range of SIC 6000-6999 were excluded. Furthermore, observations were excluded when data of at least one of the following variables was not available: BIG4, L1ACCRUAL,

LogMV, FINCG, MB, LVRG, LOSS, CFO, AC SIZE, MSIZE and FSIZE. After excluding all

observations with missing data, 758 firm-year observations remained to be tested. Sample distribution for hypothesis 2

The same 1,298 firm-year observations were the starting point for the second hypothesis. Again, the financial sector firms were excluded first. Hereafter, observations were deleted when data of at least one of the following variables was not available: #KAM, BIG4, AUDIT FEES, MATERIALITY, SIZE,

LVRG, CURRENT RATIO, ROA, LOSS, AC SIZE, MSIZE and FSIZE. After these observations were

excluded, 629 firm-year observations remained to be tested.

3.3 Regression models

Discretionary accruals

A multiple (logistic) regression model, based on the TH (2011) study, is used to test the first hypothesis. As in TH (2011) and Ashbaugh et al. (2003), this model controls for the association between firm performance and discretionary accruals by calculating ROA in Estimation Discretionary Current Accruals (REDCA). REDCA includes a control variable for firm performance (lagged ROA) in the model used to estimate discretionary accruals.

(12)

11

The calculation of REDCA starts by estimating the following cross-sectional current accrual regression model parameters:

𝐶𝐴 = 𝛾₁(1/𝐿𝐴𝐺1𝐴𝑆𝑆𝐸𝑇) + 𝛾₂ (∆𝑅𝐸𝑉) + 𝛾₃𝐿𝐴𝐺1𝑅𝑂𝐴 + 𝜀 (1) Where current accruals (CA) is net income before extraordinary items plus depreciation and amortization minus operating cash flows scaled by beginning of year total assets. LAG1ASSET is total assets at the beginning of the year (i.e. prior year-end total assets), and ∆REV equals change in net sales from the prior year scaled by beginning total assets.

The parameter estimates from the current accruals equation (1) are those obtained from the original Jones Model. These parameters are then used to calculate the expected current accruals with a performance control (ECAPC), in which ∆REC equals change in net receivables from the prior year scaled by beginning total assets. According to the modified Jones Model the following equation is established:

𝐸𝐶𝐴𝑃𝐶 = 𝑦̂1 (1/𝐿𝐴𝐺1𝐴𝑆𝑆𝐸𝑇) + 𝑦̂2 (∆𝑅𝐸𝑉 − ∆𝑅𝐸𝐶) + 𝑦̂3 𝐿𝐴𝐺1𝑅𝑂𝐴 (2)

REDCA equals CA minus ECAPC. Therefore, REDCA controls for firm-specific performance. REDCA

can be denoted as:

𝑅𝐸𝐷𝐶𝐴 = 𝐶𝐴 – 𝐸𝐶𝐴𝑃𝐶 (3)

Then, the following regression model can be used to test the association between the gender variables and the Discretionary Current Accruals Performance Adjusted (DCA_PA = REDCA):

𝐷𝐶𝐴_𝑃𝐴 = 𝛽₀ + 𝛽₁𝐵𝐼𝐺4 + 𝛽₂𝐿1𝐴𝐶𝐶𝑅𝑈𝐴𝐿 + 𝛽₃𝐿𝑜𝑔𝑀𝑉 + 𝛽₄𝐹𝐼𝑁𝐶𝐺 + 𝛽₅𝑀𝐵 +

𝛽6𝐿𝑉𝑅𝐺 + 𝛽7𝐿𝑂𝑆𝑆 + 𝛽8𝐶𝐹𝑂 + 𝛽9𝐺𝐸𝑁𝐷𝐸𝑅 − 𝑉𝐴𝑅𝑆 + 𝜀 (4)

Number of reported key audit matters

The linear regression model of Sierra-Garcia et al. (2017) is used as starting point for the regression model used to test the second hypothesis. The gender variables (GENDER-VARS) have been included in the equation to test the relationship of the gender diversity in audit committees and the number of key audit matters. This results in the following linear regression model:

#𝐾𝐴𝑀 = 𝛽₀+ 𝛽₁𝐵𝐼𝐺4 + 𝛽₂𝐴𝑈𝐷𝐼𝑇 𝐹𝐸𝐸𝑆 + 𝛽₃𝑀𝐴𝑇𝐸𝑅𝐼𝐴𝐿𝐼𝑇𝑌 + 𝛽₄𝑆𝐼𝑍𝐸 + 𝛽₅𝐿𝑉𝑅𝐺 + 𝛽₆𝐶𝑈𝑅𝑅𝐸𝑁𝑇 𝑅𝐴𝑇𝐼𝑂 + 𝛽₇𝑅𝑂𝐴 + 𝛽₈𝐿𝑂𝑆𝑆 + 𝛽₉𝐺𝐸𝑁𝐷𝐸𝑅 − 𝑉𝐴𝑅𝑆 + 𝜀 (5)

(13)

12

TABLE 1

Definition of variables

Variable Definition Hypothesis 1

DCA_PA Discretionary accruals measured by REDCA model.

BIG4 If the sample firm is audited by big-four firms 1, otherwise 0.

L1ACCRUAL Last year’s total current accruals (net income before extraordinary items plus depreciation and amortization minus operating cash flows) scaled by beginning of year total assets.

LogMV Market value of firms as of year-end, measured via the natural logarithm.

FINCG If the number of common shares outstanding or the long-term debt increased by at least 10 percent 1, otherwise 0.

MB Ratio of market value to book value as of year-end.

LVRG Total assets less book value divided by total assets.

LOSS If a firm had a loss for the observed year 1, otherwise 0.

CFO Cash flow from operations scaled by beginning of year total assets.

Hypothesis 2

#KAM Number of key audit matters in the extended auditor’s report. AUDITFEES Audit fees charged by the auditor to the client, measured via the

natural logarithm.

MATERIALITY Materiality amount disclosed in the audit report, measured via the natural logarithm.

SIZE Total assets in year t, measured via the natural logarithm.

CURRENTRATIO Total current assets divided by total current liabilities.

ROA Return on assets, measured as net income before extraordinary items divided by beginning of year total assets.

Gender variables (GENDER-VARS)

FPCT Percentage of female members to total number of members.

FEMD If audit committee has a female director 1, otherwise 0.

FSIZE Number of female audit committee members.

(14)

13

4 EMPIRICAL RESULTS

In the previous chapter the measurement methods and variables of the regression models were discussed. This chapter provides the results of the performed statistical analysis. At first, in section 4.1, the descriptive statistics and the regression results of hypothesis 1 are discussed. Section 4.2 reports on the descriptive statistics and regression results of hypothesis 2.

4.1 Earnings Management

Descriptive statistics

Table 2 shows the descriptive statistics of the sample that is used to test hypothesis 1. The average of the discretionary accrual measure (DCA_PA) is negative (mean = -0.032) indicating that, in general, managers tend to report negative (income-decreasing) discretionary accruals, and thus are more conservative in reporting. This finding is consistent with the studies of TH (2011) and Sun et al. (2011). The mean of BIG4 is 0.984 which indicates that 98.4 percent of the firms from the sample are audited by a big-four auditor. The mean values for MB, LVRG, and FINCG are, respectively, 3.43, 0.63, and 0.40. Negative pretax income is reported for little over 13 percent of the sample (LOSS = 0.132). These findings are consistent with prior research.

The average audit committee consists of 3.84 members (ACSIZE), of which 2.75 are male (MSIZE) and 1.08 female (FSIZE). On average 26.7 percent of the audit committee members is female (FPCT). The mean value of presence of female director (FEMD) is 0.745, which indicates that on 74.5 percent of the committees at least one female director is present. These values differ from prior literature as they are significantly larger in this study.

TABLE 2

Descriptive statistics of hypothesis 1

Variable N Mean SD Min Max

DCA_PA 758 -0.032 0.064 -0.51 0.31 BIG4 758 0.984 0.125 0 1 L1ACCRUAL 758 0.000 0.071 -0.49 0.40 LogMV 758 21.961 1.365 15.23 25.44 FINCG 758 0.404 0.491 0 1 MB 758 3.431 4.610 -16.5 26.53 LVRG 758 0.629 0.233 0.07 2.38 LOSS 758 0.132 0.339 0 1 CFO 758 0.103 0.087 -0.57 0.63

(15)

14

TABLE 2 (continued)

ACSIZE 758 3.840 1.090 2 9

MSIZE 758 2.757 0.894 0 6

FSIZE 758 1.083 0.846 0 4

FEMD 758 0.745 0.436 0 1

FPCT 758 0.267 0.192 0 1

Table 3 provides the Pearson correlation coefficients between the continuous independent variables to test for multicollinearity. I do not find high correlations between the variables, therefore multicollinearity seem not to be an issue. To be robust, a VIF-test (variance inflation factor) is conducted additionally. A VIF-value higher than 10 may indicate multicollinearity. The results of the VIF-test are presented in table 4. All scores are below 10 and therefore multicollinearity is not an issue.

TABLE 3 Pearson correlations (N = 758) Variable A B C D E A DCA_PA 1.000 B L1ACCRUAL 0.135 1.000 C LogMV -0.027 0.060 1.000 D MB -0.054 0.001 0.132 1.000 E CFO -0.721 -0.068 0.160 0.196 1.000 TABLE 4 VIF-values (N = 758)

Variable VIF Variable VIF

CFO 1.15 _MB 1.06 LOSS 1.15 _FPCT 1.04 LogMV 1.08 _BIG4 1.03 L1ACCRUAL 1.07 _FINCG 1.01 LVRG 1.06 Mean VIF 1.07

The data of both models is tested via a Breusch-Pagan / Cook-Weisberg test for heteroscedasticity. Since the p-values of variables for both models are below the 1 percent significance level (p < 0.01), the null hypothesis – that data is homoscedastic – is rejected. Hence, the alternative means that the data is heteroscedastic. In order to deal with heteroscedastic data, robust standard errors are used in the regression of both models.

(16)

15 Regression results

Table 5 shows the regression results of the relationship between the gender diversity in the audit committee and the amount of discretionary accruals in order to test hypothesis 1.

The results show that the coefficient of percentage of female members in the audit committee (FEMD) is positive, but the t-statistic is not significant (p > 0.10). These results remain stable when other proxies for gender diversity (FEMD and FSIZE) are tested (p > 0.10).

Furthermore, the results show that the variables L1ACCRUAL and FINCG are positive and also not significant (p > 0.10), this is however not similar to prior research. Dummy variable BIG4 is negative but not significant (p > 0.10). The market value (LogMV) and Market-to-book ratio (MB) are positive and significant (p < 0.05). The variables LOSS and CFO are negative and highly significant at the 1 percent level (p < 0.01).

The coefficients for all gender variables are insignificant at the conventional levels. Hence, there is no evidence to support the hypothesis (H1) that female directors in the audit committee are associated with a decrease in earnings management.

TABLE 5

Regression results on Earnings Management

Model 1 Model 2 Model 3

Variable Coefficient Coefficient Coefficient

BIG4 -0.028 -0.027 -0.027 L1ACCRUAL 0.030 0.031 0.031 LogMV 0.002** 0.003** 0.003*** FINCG 0.000 0.000 0.000 MB 0.001** 0.001** 0.001** LVRG -0.025** -0.025** -0.025** LOSS -0.033*** -0.033*** -0.033*** CFO -0.585*** -0.584*** -0.584*** FPCT 0.006 FEMD -0.001 MSIZE -0.002 FSIZE 0.000 Intercept 0.016 0.014 0.014 N 758 758 758 R2 0.58 0.58 0.58

(17)

16

4.2 Key Audit Matters

Descriptive statistics

Table 6 shows the descriptive statistics of the sample that serves to test hypothesis 2. The average number of reported key audit matters (KAM) is 4.2 (mean = 4.169). Table 7 shows the frequency distribution of the number of reported key audit matters. The most frequent observed number is 4, with 158 observations, which represents 25 percent of the sample. The lowest number of reported key audit matters is 1 and a maximum of 9 is observed.

In this sample 98.6 percent of the firms are audited by a big-four auditor (BIG4 = 0.986). The mean leverage ratio (LVRG) is 0.606 and the mean current ratio (CURRENTRATIO) is 1.527. The average return on assets (ROA) is 9.0 percent. Almost 11 percent of the sample reported a negative pretax income (LOSS) during the years 2013-2016.

Logically, the audit committee gender variables are in line with the descriptive statistics of the first hypothesis. The average audit committee consists of 3.97 members (ACSIZE), of which 2.79 are male (MSIZE) and 1.17 female (FSIZE). On average 28.5 percent of the audit committee members are female (FPCT). The mean value of number of female directors (FEMD) is 0.790, which indicates that for this sample on 79.0 percent of the committees at least one female director is present.

TABLE 6

Descriptive statistics for hypothesis 2

#KAM 629 4.169 1.497 1 9 BIG4 629 0.986 0.119 0 1 AUDITFEES 629 14.194 1.294 11.16 17.75 MATERIALITY 629 16.723 1.276 13.82 21.24 SIZE 629 22.060 1.505 17.31 26.65 LVRG 629 0.606 0.233 0.06 2.20 CURRENTRATIO 629 1.527 1.170 0.28 12.31 ROA 629 0.090 0.238 -0.49 2.97 LOSS 629 0.110 0.313 0 1 ACSIZE 629 3.968 1.017 2 9 MSIZE 629 2.795 0.875 1 6 FSIZE 629 1.173 0.841 0 4 FEMD 629 0.790 0.408 0 1 FPCT 629 0.285 0.187 0 0.8

(18)

17

TABLE 7

Frequency table #KAM

#KAM Frequency Percentage

1 12 1.9 2 66 10.5 3 143 22.7 4 158 25.1 5 143 22.7 6 64 10.2 7 29 4.6 8 11 1.8 9 3 0.5 Total 629 100.0

Table 8 provides the Pearson correlation coefficients between the continuous independent variables to test for multicollinearity. I find the highest correlations between MATERIALITY and AUDITFEES (0.755), SIZE and AUDITFEES (0.779), and SIZE and MATERIALITY (0.849). All coefficients are significant at the 1 percent level (p < 0.01). To mitigate the concern for multicollinearity, I drop the variables and run the regression tests again. The results do not substantially change and therefore multicollinearity seem not to be an issue. On the additional VIF-test on hypothesis 2, the result of table 9 show that multicollinearity is again not an issue in this research.

TABLE 8 Pearson correlations (N = 629) Variable A B C D E F G A #KAM 1.000 B AUDITFEES 0.457 1.000 C MATERIALITY 0.331 0.755 1.000 D SIZE 0.434 0.779 0.849 1.000 E LVRG 0.222 0.245 0.140 0.160 1.000 F CURRENTRATIO -0.148 -0.235 -0.127 -0.082 -0.456 1.000 G ROA -0.241 -0.248 -0.107 -0.358 0.045 -0.001 1.000 TABLE 9 VIF-values (N = 629)

Variable VIF Variable VIF

SIZE 5.61 _LVRG 1.34 MATERIALITY 4.73 _LOSS 1.08 AUDITFEES 3.04 _BIG4 1.06 ROA 1.43 _FPCT 1.03 CURRENTRATIO 1.39 Mean VIF 2.3

(19)

18 Regression results

Table 10 presents the results of hypothesis 2 on the relationship between the gender diversity variables and the number of reported key audit matters.

The results show that the coefficient of percentage of female members in the audit committee (FPCT) is negative, but the t-statistic is not significant (p > 0.10). These results hold when other proxies for gender diversity (FEMD and FSIZE) are tested (p > 0.10). Furthermore, audit fees (AUDITFEES) are positive and highly significant (p < .01), indicating that audit fees are higher for firms with more reported key audit matters. The materiality used by the auditor (MATERIALITY) is negative and significant (p < .05), indicating that lower materiality levels are used by the auditor for firms with more key audit matters. The control variables size (SIZE) and leverage (LVRG) are positive and highly significant (p < .01). The current ratio (CURRENTRATIO) is negative and not significant (p > .10). Return on assets (ROA) is negative and significant (p < .05). The dummy variable for pretax income (LOSS) is positive and significant (p < .05), indicating that firms who report a loss also tend to report significantly more key audit matters.

Since I find insignificant coefficients for all gender diversity proxies, I find no evidence to conclude that there is a significant association between the presence of female directors in the audit committee and the number of reported key audit matters in the auditor’s report.

TABLE 10

Regression results on Key Audit Matters

Model 1 Model 2 Model 3

Variable Coefficient Coefficient Coefficient

BIG4 -1.291*** -1.317*** -1.302*** AUDITFEES 0.337*** 0.346*** 0.329*** MATERIALITY -0.223** -0.219** -0.236*** SIZE 0.311*** 0.304*** 0.322*** LVRG 0.767*** 0.741*** 0.771*** CURRENTRATIO -0.056 -0.064 -0.054 ROA -0.425*** -0.401*** -0.412** LOSS 0.352** 0.3423** 0.360** FPCT -0.044 FEMD -0.195 MSIZE 0.042 FSIZE 0.037 Intercept -2.841*** -2.672*** -2.921*** N 629 629 629 R2 0.27 0.28 0.27

(20)

19

5 CONCLUSION AND DISCUSSION

In the previous chapter the results have been presented. Based on these results, section 5.1 of this chapter will answer and discuss the research questions which were formulated in the introduction. Section 5.2 concludes with the limitations of this study and provides possible directions for future research.

5.1 Conclusion and discussion

Conclusion

The purpose of this study was twofold, as the impact of gender diversity on two different aspects was researched. Firstly, the relationship between gender diversity in the audit committee and the amount of earnings management of the firms was examined. Secondly, a relationship was sought between gender diversity and reported key audit matters in the auditor’s report.

The first part of this paper tried to answer the following research question: Does female presence

in the independent audit committee constrain earnings management? Based on prior research (i.e. TH,

2011) I did expect a significant negative relationship between the gender variables and the discretionary accruals. To answer this research question, I formulated the following hypothesis: There

is a significant association between the presence of female members in the audit committee and the amount of earnings management. In order to test the hypothesis, the modified Jones Model was used

to measure the discretionary current accruals of 758 firm-year observations in the UK and the Netherlands during the period from 2013 until 2016. Contrary to what I expected for this research, this study finds no relationship between gender diversity in the audit committee and the amount of earnings management of the firms. However, these findings are similar to Sun et al. (2011) who also found no gender effects with respect to independent audit committees’ effectiveness in constraining earning management.

The second part of this paper sought for determinants of the number of reported key audit matters, as reported by the independent auditor in the extended auditor’s report. In order to do so, the second research question was stated: Does female presence in the independent audit committee affect

the number of reported key audit matters? As there is little known on this particular subject, it was

unclear whether there is a relation, and if so, in what direction. However, based on gender theory it could be expected that female presence reduces the number of key audit matters as women are more risk averse and ethical than men (Francis et al., 2015; Marta et al., 2008; Franke et al., 1997). Therefore the following hypothesis was developed: There is a significant association between the presence of

female members in the audit committee and the number of reported key audit matters. To test this

relationship, 629 firm-year observations in the UK and the Netherlands were examined via a linear regression model for the period from 2013 until 2016. The results indicate that the number of key audit matters are not related to the audit committee’s gender variables. This implies that the auditor does not consider the gender of the audit committee members in determining the number of key audit matters. Discussion

While the results of this study suggest that there are no significant differences in beliefs between female and male audit committee members, there may be several possible causes for the observed null result with regard to the first hypothesis. For example, it is possible that female directors are more ethical

(21)

20

and intent to constrain earnings management, but that they are not able to do so. For instance, it could occur that they are unable to influence the other members of the audit committee. Furthermore, Chong (2006) argues that earnings management is a logical result of flexibility in financial reporting decisions and is not always considered bad if management uses earnings management to generate a stable financial performance. Finally, the number of female members in audit committees has increased extensively in the last couple of years. While TH (2011) found female members on only 20 percent of the audit committees, I find significantly higher percentages of female presence ranging between 75 and 79 percent of the sample. Consequently, this may have biased the results as other factors from the non-female firms could affect the results. Unfortunately, it was not possible to control for these effects in this study.

However, the results of this study provide several interesting insights. Firstly, one might wonder why firms, on average, tend to report negative discretionary accruals. While this indicates that firms are more conservative and suppressing their profits, it is unclear as to why they do this. A possible explanation could be that when females are more often appointed in the board of directors, the board adopts a more conservative reporting strategy as females are more risk averse then men.

Secondly, although I find no significant relationship between gender diversity in audit committees and the number of reported key audit matters, some control variables show noteworthy results. Particularly the audit fees and materiality variables are interesting. I found a highly significant negative relationship between materiality and the number of reported key audit matters. This indicates that a lower materiality level results in a higher number of reported key audit matters. Furthermore, I found a highly significant positive relationship between audit fees and the number of reported key audit matters. In other words, when higher audit fees are charged, the number of reported key audit matters increases. This suggests that a firm is required to pay a higher audit fee, to obtain more key audit matters. However, this relation remains still unclear as the direction of this relationship has to be further investigated. For example, it is also arguable that an audit with a higher number of key audit matters results in a higher audit fee.

5.2 Limitations and future research

Limitations

The results of this study should, however, be interpreted with caution because of its own limitations. Although I tried to control for as many external factors as possible, it is still possible that I have omitted other control variables that have affected the results. Moreover, I used a discretionary accruals model that has been used in previous studies, and such models are usually restricted to measurement errors. Therefore the results of the study depend on the capability of both the model and the researcher to accurately estimate the discretionary current accruals. Hence, there are several possible causes that may have generated these results.

Despite of the mentioned limitations, this study extends the literature regarding earnings management and the literature regarding the determinants of the number of reported key audit matters by considering the gender of audit committee members.

(22)

21 Future research

The discussion of this paper highlights several interesting directions for future research. For instance, it could be interesting to find out more about the underlying reason why firms tend to use negative (income-decreasing) discretionary accruals. Did the firms recover from the financial crisis and are they performing on the same high levels as before, but still remain cautious? Or could it be the influence from the more risks averse women on the boards?

Secondly, the relationships between the number of reported key audit matters, materiality, and audit fees is intriguing as there is still little research available on this subject. It would be interesting to conduct interviews or a case study within audit firms to investigate the auditor’s reporting process and to find out what really drives the auditors to report key audit matters. This could be developed in a way that not only the number of reported key audit matters is investigated, but also the determinants of specific key audit matters, such as significant fraud risks, are discovered.

(23)

22

6 REFERENCES

Arun, T. G., Almahrog, Y. E., & Aribi, Z. A. (2015). Female directors and earnings management: Evidence from UK companies. International Review of Financial Analysis, 39, 137-146. Ashbaugh, H., LaFond, R., & Mayhew, B. W. (2003). Do nonaudit services compromise auditor

independence? Further evidence. The Accounting Review, 78(3), 611-639.

Becker, C. L., DeFond, M. L., Jiambalvo, J., & Subramanyam, K. R. (1998). The effect of audit quality on earnings management. Contemporary accounting research, 15(1), 1-24.

Bédard, J., Chtourou, S. M., & Courteau, L. (2004). The effect of audit committee expertise, independence, and activity on aggressive earnings management. Auditing: A Journal of

Practice & Theory, 23(2), 13-35.

Bédard, J., Gonthier-Besacier, N., and Schatt, A. (2014). Costs and benefits of reporting Key Audit Matters in the audit report: The French experience. In International Symposium on Audit Research. Available at: https://documents.bsb-education.com/pdf/cig2014/ ACTESDUCOLLOQUE/BEDARD_GONTHIER_BESACIER_SCH ATT.pdf, retrieved: December, 2017.

Campbell, K., & Mínguez-Vera, A. (2008). Gender diversity in the boardroom and firm financial performance. Journal of business ethics, 83(3), 435-451.

Carter, D. A., Simkins, B. J., & Simpson, W. G. (2003). Corporate governance, board diversity, and firm value. Financial review, 38(1), 33-53.

Chong, G. (2006). Is income smoothing ethical? Journal of Corporate Accounting & Finance, 18(1), 41-44.

Chonko, L. B., & Hunt, S. D. (1985). Ethics and marketing management: An empirical examination. Journal of Business Research, 13(4), 339-359.

Dechow, P. M., Sloan, R. G., & Sweeney, A. P. (1995). Detecting earnings management.

Accounting review, 193-225.

DeZoort, F. T., Hermanson, D. R., Archambeault, D. S., & Reed, S. A. (2002). Audit committee effectiveness: A synthesis of the empirical audit committee literature. Journal of accounting

literature, 21.

Financial Reporting Council (FRC). (2013). ‘Revision to ISA (UK and Ireland) 700: Requiring the auditor’s report to address risks of material misstatement, materiality and a summary of the audit scope’, Consultation Paper, February. London.

Financial Reporting Council (FRC). (2015). ‘Extended auditor’s reports: a review of experience in the first year’, Report, March. London.

Financial Reporting Council (FRC). (2016a). ‘Extended auditor’s reports: a further review of experience’, Report, January. London.

Financial Reporting Council (FRC). (2016b). International Standard on Auditing (UK) 700 (Revised June 2016), Report, June. London.

Francis, B., Hasan, I., Park, J. C., & Wu, Q. (2015). Gender differences in financial reporting decision making: Evidence from accounting conservatism. Contemporary Accounting

Research, 32(3), 1285-1318.

Franke, G., Crown, D., & Spake, D. (1997). Gender differences in ethical perceptions of business practices: A social role theory perspective. Journal of Applied Psychology, 82(6), 920-934.

(24)

23

Fudenberg, D., & Tirole, J. (1995). A theory of income and dividend smoothing based on incumbency rents. Journal of Political economy, 103(1), 75-93.

Gimbar, C., Hansen, B., & Ozlanski, M. E. (2016). The effects of critical audit matter paragraphs and accounting standard precision on auditor liability. The Accounting Review, 91(6), 1629-1646.

Gull, A. A., et al. (2017). Beyond gender diversity: How specific attributes of female directors affect earnings management, The British Accounting Review (2017). Available at: http://dx.doi.org/10.1016/j.bar.2017.09.001. Retrieved: December, 2017.

Gutierrez, E., M. Minutti-Meza, K.W. Tatum, and M. Vulcheva. 2016. Consequences of changing the auditor’s report: Evidence from the UK. Working paper, Universidad de Chile, University of Miami, and Florida International University.

Harrison, D. A., & Klein, K. J. (2007). What's the difference? Diversity constructs as separation, variety, or disparity in organizations. Academy of management review, 32(4), 1199-1228. Hay, D. C., Knechel, W. R., & Wong, N. (2006). Audit fees: A Meta‐analysis of the effect of supply

and demand attributes. Contemporary accounting research, 23(1), 141-191.

Healy, P. M. (1985). The effect of bonus schemes on accounting decisions. Journal of accounting and

economics, 7(1-3), 85-107.

International Federation of Accountants (IFAC). (2017). ISA 701 (new). Communicating Key Audit Matters in the Independent Auditor’s Report. Retrieved from: https://www.ifac.org/system/files/publications/files/ISA-701_2.pdf. Retrieved: September, 2017.

Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of financial economics, 3(4), 305-360.

Jones, J. J. (1991). Earnings management during import relief investigations. Journal of accounting

research, 193-228.

Klein, A. (2002). Audit committee, board of director characteristics, and earnings management.

Journal of accounting and economics, 33(3), 375-400.

Kothari, S. P., Leone, A. J., & Wasley, C. E. (2005). Performance matched discretionary accrual measures. Journal of accounting and economics, 39(1), 163-197.

Köhler, A. G., Quick, R., & Willekens, M. (2016). The New European Audit Regulation Arena: Discussion of New Rules and Ideas for Future Research. International Journal of Auditing,

20(3), 211-214.

Kyaw, K., Olugbode, M., & Petracci, B. (2015). Does gender diverse board mean less earnings management? Finance Research Letters, 14, 135-141.

Lennox, C. S., Schmidt, J. J., & Thompson, A. (2016). Is the expanded model of audit reporting informative to investors? Evidence from the UK. Working Paper, The University of Texas at Austin.

Marta, J., Singhapakdi, A., & Kraft, K. (2008). Personal characteristics underlying ethical decisions in marketing situations: A survey of small business managers. Journal of Small

Business Management, 46(4), 589-606.

(Koninklijke) Nederlandse Beroepsorganisatie van Accountants (NBA). (2017). Nadere voorschriften controle- en overige standaarden. Available at: https://www.nba.nl/tools/hra-2017/?folder=3567. Retrieved: October, 2017.

(25)

24

Peni, E., & Vähämaa, S. (2010). Female executives and earnings management. Managerial

Finance, 36(7), 629-645.

Seierstad, C., Warner-Søderholm, G., Torchia, M., & Huse, M. (2017). Increasing the number of women on boards: The role of actors and processes. Journal of Business Ethics, 141(2), 289-315.

Sierra-Garcia, L., Gambetta, N., García-Benau, M., Orta-Pérez, M. (2017). Understanding the Determinants of Key Audit Matters: The Case of the United Kingdom. Working paper, Universidad Pablo de Olavide.

Sun, J., Liu, G., & Lan, G. (2011). Does female directorship on independent audit committees constrain earnings management? Journal of Business Ethics, 99(3), 369-382.

Thiruvadi, S. (2012). Gender differences and audit committee diligence. Gender in

Management: An International Journal, 27(6), 366-379.

Thiruvadi, S., & Huang, H. W. (2011). Audit committee gender differences and earnings management. Gender in Management: An International Journal, 26(7), 483-498.

The Impact of Gender Diversity in Audit Committees on Earnings Management and Key Audit Matters