Do women in board represent better accounting quality?

M.Sc. in Accountancy & Control, Variant Accountancy

Faculty of Economics and Business, University of Amsterdam

Master Thesis

Do women in board represent better accounting

quality?

Student Name : Shiery Melinda Handoko (10395555)

Date : August 31, 2014

1st _{Supervisor : dr. R.S. Boomsma} 2nd _{Supervisor : dr. ir. S.P. van Triest}

ABSTRACT

This thesis studies the effect of gender in the board at various positions as CFO, CEO, Audit Committee, and Compensation Committee on earnings management during pre and post global financial crisis in 2008. The presence of women in the board is negatively associated with earnings management proxied by discretionary accruals during the two periods except for their position as Audit Committee which is, on the contrary, positively associated with earnings management. The presence of women as CFO is the most effective in reducing earnings management relative to the other positions as CEO, Compensation Committee and, of course, the Audit Committee. The roles of women as Audit Committee and Compensation Committee, and as CFO to be more negatively associated with earnings management after the crisis period compared with that in the pre-crisis period are, however, not supported by the t-test; while the earnings management is even less if women present as both CFO and CEO.

Keywords: Corporate Governance, Board of Directors, Accounting Quality, Earnings Management, Women in Board

LIST OF CONTENTS

2.2. Interaction between Board of Directors and 8

Earnings Management

2.3. Women in Board 9

3. Hypothesis Development 12

4. Research Methodology 14

4.1. Measures of Earnings Management 14

4.2. Sample Selection 15

5. Results 16

5.1. Jones Model Coefficients 16

5.2. Role of Female CFO on Earnings Management during 17 Pre and Post Crisis Periods

5.3. Interaction of Women as CFO and CEO on Earnings 20 Management during Pre Crisis Period

5.4. Interaction of Women as CFO and CEO on Earning 22 Management during Post Crisis Period

5.5. Role of Women in Audit Committee 25

5.6. Role of Women in Compensation Committee 27

6. Discussion, Conclusions, and Limitations 30

6.1. Discussion 30

Role of Women in Different Possitions as CFO, CEO, 30 Audit Committee and Compensation Committee

The Difference of Female Role in the Board between 32 Pre and Post Crisis Periods

6.2. Conclusions 33

6.3. Limitations 34

LIST OF TABLES

Table 1a. Analysis results using modified Jones Model (Pre-Crisis) 17

Table 1b. Analysis results using modified Jones Model (Post Crisis) 17

Table 2a. Descriptive statistics of discretionary accruals of all male CFOs 18 and that with female CFOs during pre and post-crisis periods

Table 2b. Result of the t-test on discretionary accruals of all male CFOs (CFO=0) 18 compared to that with the presence of female CFOs (CFO=1) during

pre and post-crisis periods

Table 3a. Descriptive statistics of discretionary accruals on interaction of 20 male/female CFOs and male/female CEOs during pre-crisis period

Table 3b. Result of the t-test on discretionary accruals on the interaction of 21 male/female CFOs and that of male/female CEOs during

pre-crisis period

Table 4a. Descriptive statistics of discretionary accruals on interaction of 23 male/female CFOs and male/female CEOs during post-crisis period

Table 4b. Result of the t-test on discretionary accruals on interaction of 23

male/female CFOs compared to that of male/female CEOs during post-crisis period

Table 5a. Descriptive statistics of discretionary accruals of all men (AC=0) 26 and with women present (AC=1) in Audit Committee during

pre and post crisis period

Table 5b. Result of the t-test on discretionary accruals on of all men (AC=0) 26

compared to that with women present (AC=1) in Audit Committee during pre and post crisis periods

Table 6a. Descriptive statistics of discretionary accruals of all men (CC=0) 28 and with women present (CC=1) in Compensation Committee

during pre and post Crisis Period

Table 6b. Result of the t-test on discretionary accruals of all men (CC=0) 29

compared to that with women present (CC=1) in Compensation Committee during pre and post crisis periods

LIST OF FIGURES

Figure 1. Comparison of minimum and maximum discretionary accruals of all 19 male CFOs against that with the presence of female CFOs during

pre and post crisis periods.

Figure 2. Interaction between male/female CFOs and male/female CEOs on 22 discretionary accruals (data on minimum values) for its actual DA

and absolute DA during pre-crisis period.

Figure 3. Interaction between male/female CFOs and male/female CEOs on 22 discretionary accruals (data on maximum values) during

pre-crisis period.

Figure 4. Interaction between male/female CFOs and male/female CEOs on 24 discretionary accruals (data on minimum values) for its actual DA

and absolute DA during post-crisis period.

Figure 5. Interaction between Male/Female CFOs and Male/Female CEOs on 24 discretionary accruals (data on maximum values) for its actual DA

and absolute DA during Post-Crisis period.

Figure 6. Comparison of minimum and maximum discretionary accruals of 27 all men against women present in the Audit Committee during

pre and post crisis periods.

Figure 7. Comparison of minimum and maximum discretionary accruals of 29 all men against that women present in the Compensation Committee

1. Introduction

The event of accounting scandals such Enron, WorldCom and Ahold following the ongoing financial crisis in 2008 has exposed the quality of financial statements. The fact that it is a tool in communicating firm performance to the shareholders, the role of financial reporting is at stake. Society blames the lack of its quality on involved parties in the financial reporting process such regulators, auditors, and governance. This thesis focuses on corporate governance particularly board gender composition.

Separation of ownership and control results in agency conflict between managers and shareholders (Jensen and Meckling, 1976). Agency conflict arises because managers have better information regarding firms and they are less likely to report information to shareholders that cause harm to their personal interest such as report their poor performance (Armstrong et al. 2010). As a consequence, managers are more inclined to produce false accounting, aggressive earnings management and other reporting failure resulting loss of transparency, integrity, and accountability thus low financial reporting quality. In order to mitigate the agency conflict, board of directors should ensure the actions of the management are aligned with the objectives of the firm and ensure the quality of financial reporting (Cohen et al. 2004). However, some claim that board of directors are merely cosmetic and not effective on firm value (Ahern & Dittmar, 2011). Moreover, financial crisis which commenced around 2008 has implied that board members in financial sectors fail to do their role. Various board characteristics such as outside directors, board size, financial expertise and composition of gender; and their effect on financial reporting quality have been empirically examined. I am interested to investigate gender composition because boards around the world are now under increasing pressure to choose female directors (Adams and Ferreira 2008). In addition, a new law was introduced in Norway to require 40 percent directors must be women and firms have to comply the law by January 2008 otherwise the firms would be forced to dissolve (Ahern and Dittmar 2011). In this relation, effective leadership style is an issue for the success of the company.

At present, effective leadership styles are perceived to have similar pattern with feminine characteristics to be more collaborative, emotional, and considerate including participative, transformational, and democratic (Adams & Hambright 2004; Emmerik et al. 2010) compared to masculine characteristics having behaviors of authority and discipline (Burke

et al. 2006). Nonetheless, board diversity especially composition of gender is currently a debated topic. Although female characteristics are perceived as valuable for firms’ long term success and women have now great opportunities to hold seats on the board, men are still reluctant to accept women despite the facts that many female directors have a significant impact on board inputs and company outcomes (Adams and Ferreira 2008). Considering this phenomenon, I am interested to search gender issue in the board assuming the presence of women in the board will make the leadership style more effective which coupled with commitments of the female leaders and will lead to positive outcomes in corporate monitoring and oversight (Ittonen et al. 2009). Consequently, financial reporting quality will be improved by the presence of women in board.

Prior literature study the association between particular board position in relation to gender and earnings management. Peni and Vähämaa (2010) report that female CFOs correspond to income decreasing accruals. Thiruvadi and Wang, 2011 document income decreasing accruals are increasing when female is present in Audit Committees. These two findings support the argument that female directors are risk averse and more likely to be conservative (Francis et al. 2013). While previous researches study mainly on one particular board position, this thesis examine the presence of women in four positions (as CFO, CEO, Audit Committee and Compensation Committee) on earnings management during pre and post global financial crisis. The main question of this research which I used as the title of this thesis is “Do women in board represent better accounting quality?”. To answer this main question, five hypotheses are proposed which will be explained in Chapter 3 based on the Literature Review in Chapter 2.

Chapter 4 describes research methodology which includes statistical analysis [the t-test, descriptive statistics] to test the hypotheses and the regression analysis to estimate earnings management based on modified Jones Model (Jones 1991; Dechow et al. 1995). Chapter 5 presents the results of this research in answering the main research question whether women in board represent better accounting quality based on the five hypotheses. Final chapter (Chapter 6) presents discussion, conclusions and limitations of this research thesis.

2. Literature Review

2.1. Earnings Management

Earnings management is an interesting subject to discuss by which a question rises whether the observed earnings management is a dysfunctional behavior (Bushman and Smith 2001). Scott (2009) broadly state that earnings management are: "the choice by a manager of accounting policies, or actions affecting earnings, so as to achieve some specific reported objective". In addition, Healy and Wahlen (1999) define "earnings management occurs when managers use judgment in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers". Bushman and Smith (2001) argue that any incentives for earnings management could be mitigated by offering contracts of flat wage, instead of nonlinear bonus schemes existing at the heart of the earnings management results. Also, they observed that earnings management would arise endogenously in an equilibrium where the contractor rationally anticipates possibilities of earnings management and reflects them in contract design.

Earnings management’s analysis frequently focuses on management’s use of discretionary accruals (Dechow et al. 1995). In measuring discretionary accruals, it is often started by measuring total accruals which is followed by using a model to separate non-discretionary and discretionary components. Discretionary accruals as the proxy of earnings management are then calculated as the difference of total accruals and non-discretionary accruals.

2.2. Interaction between Board of Directors and Earnings Management

Board of directors is one of most vital parts in corporate governance structure since the board has the vital decision authority including hiring or firing executives, dividend or option policies, and executive compensation. The main duty is to monitor and to supervise the management’s decision in operating the company on behalf of shareholders. The role of board of directors is crucial in monitoring management and in setting up mechanisms that align the executives’ objectives with the shareholders’ interests (Armstrong et al.

2010). On the other hand, the function of board of directors is being questioned and considered ineffective while many suggestions have been introduced to improve the firms’ governance to rebuild trust (Van den Berghe and Levrau 2004).

There has been a big debate about earnings management and governance, and researchers have isolated a number of pure market forces that discipline managerial behavior (Bushman and Smith 2001). Nevertheless, despite these market forces, a residual demand remains for additional governance mechanisms that can be set up to the particular interests of individual firms. Then, this demand becomes an interest of research area such as directors, compensation contracts, concentrated ownership, debt contracts, and the role of securities law in disciplining managers to act in the interests of capital suppliers (Bushman and Smith 2001).

Armstrong et al. (2010) argue that corporate boards generally consist of both outside and inside directors. In addition, outside directors are considered to be more independent and have superior objectivity in monitoring management’s behavior. Adams and Ferreira (2008) states that “Directors are classified as independent if they have no business relation with the firm, are not related or interlocked with management, and are not current or former employees. Directors who are not independent are either classified as inside directors, who are current employees, or affiliated directors, who have significant business or family relations with the firm”. Outside directors on the audit committee like directors in general are also likely to bring greater independence in monitoring financial reporting activities of the management; their commitment may be partially associated with financial incentives being a board member and perhaps more importantly from their desire to secure significant personal reputation (Armstrong et al. 2010). As such, it is believed that independent directors should be negatively associated with earnings management.

2.3. Women in Board

High level positions in many organizations are generally dominated by male leaders. It is now, however, more and more women are becoming top executives but women are perceived to be outsiders by male leaders, and it is often difficult for male leaders accepting female leaders (e.g. Thompson 2000).

In many cases, roles of leadership are considered being masculine-oriented having behaviors of discipline and authority while women who have feminine characteristics are perceived to be more emotional and collaborative in nature (Adams & Hambright 2004). Men are perceived being stronger, dominant, and motivated to manage their environment and, thus, they tend to have “initiating structure” leadership behavior while women leadership behavior is perceived being “consideration” (Emmerik et al. 2010). In this relation there are two settings of behavior related to leadership styles consisting of (1) considerate-people oriented, and (2) initiating structure-task oriented (e.g. Bass 1991, Hofstede 1998, and Burke et al. 2006). Consideration behavior is related with job satisfaction and organizational commitment of the employee while initiating structure aims for performance and profit. According to Burke et al. (2006) initiating structure focuses on two task objectives of (1) directive leadership and (2) autocracy leadership. On the other hand, consideration behavior is characterized by maintaining closed relationship, empowerment and group cohesion that creates subordinates being self management. In addition, motivation being created generates continued support to subordinates particularly during difficult conditions (Burke et al. 2006). Moreover, Dunn (2012) argues that inclusion of women in the board of directors comprising all men positively affect the performance of organization.

Two famous contrasting leadership styles are transactional and transformational. Transactional leadership style is considered to maintain status-quo which the leaders provide rewards and impose punishments to subordinates in exchange of their efforts and commitments. In contrast, transformational leadership style is characterized by the leaders who encourage subordinate to get involved in the decision making process as well as encourage the subordinates to their full potential including to develop their capacity to become leaders by themselves (Tichy 1977; Cole et al. 2009). In practice, however, leadership style could be a combination of few styles since a particular style might only be effective in a certain condition while ineffective in different situation.

Nowadays, effective leadership styles are believed to those having similar pattern with feminine characteristics which includes transformational, participative and democratic. Due to the changing situation, a paradigm shift has occurred with the change of time course in which women are no longer to be inferior to men. According to Stufft & Coyne (2009), many women have emerged as powerful leaders in business and agencies which

disqualifies the long perception of women to have only feminine characteristics being “nurturing, compassionate, emotional, expressive, communal, passive, uncertain, subjective, and supportive” in contrast to men of being “intelligent, powerful, competent, objective, independent, methodical, and driven”.

Having the trend of leadership perception in which the preferred leadership style corresponds to feminine characteristics, women are supposed to have greater chance than men in the higher positions. On the contrary, it is difficult for male leaders to accept the actual gender (i.e. women) being leaders (Thompson 2000). Choices to nominate female directors by firms could be influenced by the firm characteristics (Adams and Ferreira 2008). Moreover, a relatively more feminine and less initiating structure leadership behavior rather than masculine and initiating structure behavior is only practiced when women is greater in number within the group of higher positions (Emeric et al. 2001). This condition seems to change since boards around the world are under increasing pressure to choose female directors (Adams and Ferreira 2008). In 2003, female directors of Norwegian’s firms were only nine percent and a new law was introduced to require 40 percent directors be women and firms not comply by January 2008 would be forced to dissolve (Ahern and Dittmar 2011). In my research, however, I do not take gender ratio or the number of women in the management executives into account in studying the effect of women in the board on earnings management.

3. Hypothesis Development

Gender issue deals with stereotype characteristics of female and male in nature by which leadership has been usually perceived to be masculine characteristics being authority, discipline and initiating structure. On the other hand, feminine characteristics is associated with being more considerate, emotional and collaborative. Gender equity is currently considered as an important factor to strengthen organizational culture that affects changes in organizational culture from more masculine to be more oriented on employment equity of gender as well as from initiating structure-task behavior to consideration behavior.

Considering the trend of leadership style and effectiveness being associated with feminine behavior, I assumed that the actual gender (male or female executives as CFO and CEO) will affect the leadership effectiveness as well as their commitment in managing organization and, hence, the level of earnings management as measured (proxied) by discretionary accruals. Thus, the following hypotheses are proposed:

H1: Female CFO is negatively associated with earnings management H2: Female CEO is negatively associated with earnings management

I consider that Audit Committee (AC) constraint earnings management since AC has to oversee the external auditors' work and to make sure the quality of financial reporting (Thiruvadi and Wang, 2011). In addition, I predict Compensation Committee (CC) also reduces earnings management behavior. Compensation Committee is responsible for designing and monitoring executives pay. Executives pay (i.e cash and/or stock based pay) is perceived to be a mechanism to align the interest between managers and shareholders in a way that motivate managers to work harder to earn high bonus. However, managers are more inclined to manipulate earnings report instead of working hard. The role of CC comes into play when they set stronger pay performance sensitivity to CEO compensation, increases the incentives of Audit Committee to carefully monitoring financial reporting (Laux and Laux, 2009). I also assume that the leadership style of AC and CC will also be influenced by the presence of women in AC and CC with the proposed hypotheses are as follows:

H3: Female Audit Committee is negatively associated with earnings management

I also assumed that female CFO, who has primary responsible of firm's financial reporting system, will also have more commitments to the organization after the post crisis period which causes less earnings management for which the following hypothesis is proposed:

H5: Female CFOs have stronger negative association with earnings management in the post-crisis period than in the pre-crisis period.

In this research, the gender ratio or number of women in the board are not taken into account other than the presence of women in the management executives (CFO, CEO), Audit Committee (AC), and Compensation Committee (CC).

4. Research Methodology

The main question of this research is also used as the title of this thesis, which is “Do women in board represent better accounting quality?”. To answer this main research question, five hypotheses were developed which have been explained in Chapter 3 (Hypothesis Development). To test the hypotheses, I used a simple analysis using descriptive statistics as well as the t-test using the following formula:

where,

t is the calculated t-test value

x1 and x2 are variables being compared (i.e. absolute values of discretionary accruals)

s1 and s2 are variances of variables x1 and x2, respectively

n1 and n2 are number of samples of variables x1 and x2, respectively

Significant or very significant differences between variables x1 and x2 exist if t >t0.05 or t >

t0.01, respectively. The values of t0.05 and t0.01 are obtained from the statistical table (e.g.

Berman 2007).

Financial reporting quality is measured by earnings management before and after global financial crisis in 2008 which is approached by discretionary accruals as the proxy of the earnings management. Earnings management is the dependent variable of independent variables being the presence of women in board (as CFO, CEO, Audit Committee, and Compensation Committee) during pre and post financial crisis. The earnings management itself was then calculated from discretionary accruals as the proxy based on regression analysis following modified Jones Model (Jones 1991; Dechow et al. 1995).

4.1. Measures of Earnings Management

Dechow et al. (1995) explains that modified Jones Model is the most powerful model to detect earnings management. Therefore, I use modified Jones Model in determining Discretionary Accruals (DA). Firstly, we need to calculate Total Accruals (TA) which is computed as follows:

TA = ΔCA – ΔCash – ΔCL + ΔSTD - Depr where,

TA : total accruals

ΔCA : change in total current assets

ΔCash : change in cash and cash equivalents ΔCL : change in total current liabilities ΔSTD : change in short-term debt

Depr : depreciation and amortization expense for firm i in year t

Secondly, I estimated Non Discretionary Accruals (NDA) using the following regression model:

Estimates of the firm specific parameters α1, α2, α3 are generated using the following model in the estimation period and obtaining the coefficients through a linear regression.

where,

ΔREVt : revenues in year t less revenues in year t-1 ΔRECt : receivables in year t less receivables in year t-1

PPEt : gross property plan and equipment in year t scaled by total assets At-1 : total assets in year t-1

The level of DA for a particular firm is calculated as the difference between the firm’s TA and its Non-DA.

4.2. Sample Selection

The samples consist of S&P 500 firms from 2006 to 2011 since I want to examine the effect of female representation on board to earnings management three years before global financial crisis and three years after the crisis. The governance data was obtained from

Risk Metrics and merged with Compustat to retrieve financial reporting information, yielded 1137 firm years.

5. Results

5.1. Jones Model Coefficients

Table 1a and b show the regression results of modified Jones Model for pre crisis period (2006-2008) and post crisis period (2009-2011), respectively. As it has been previously mentioned in Research Methodology section, the coefficient will be used to calculate Non Discretionary Accruals (NDA) using the following regression model:

After knowing the value of non-discretionary accruals, we compute discretionary accruals as a measure of earnings management:

Table 1a. Analysis results using modified Jones Model (Pre-Crisis)

Variable Coefficient Std. Error t-Statistic Prob.

α -0.038969 0.005970 -6.527619 0.000000

At-1 26.695060 19.350767 1.379535 0.168259

ΔREVt 0.010340 0.002393 4.321880 0.000018

PPE -0.023472 0.006182 -3.796961 0.000162

R-squared 0.061191 Mean dependent var -0.039084

Adjusted R-squared 0.056360 S.D. dependent var 0.063712

S.E. of regression 0.061891 Akaike info criterion -2.720092

Sum squared resid 2.233181 Schwarz criterion -2.690279

Log likelihood 802.346995 Hannan-Quinn criter. -2.708475

F-statistic 12.666430 Durbin-Watson stat 2.174097

Prob(F-statistic) 0.000000

Table 1b. Analysis results using modified Jones Model (Post Crisis)

Variable Coefficient Std. Error t-Statistic Prob.

α -0.016404 0.005456 -3.006566 0.002764

At-1 30.295720 18.163935 1.667905 0.095908

ΔREVt -0.001894 0.002755 -0.687429 0.492104

PPE -0.035465 0.005038 -7.039507 0.000000

R-squared 0.093783 Mean dependent var -0.038828

Adjusted R-squared 0.088804 S.D. dependent var 0.053350

S.E. of regression 0.050927 Akaike info criterion -3.109620

Sum squared resid 1.416056 Schwarz criterion -3.078276

Log likelihood 859.145633 Hannan-Quinn criter. -3.097371

F-statistic 18.834974 Durbin-Watson stat 2.050806

Prob(F-statistic) 0.000000

5.2. Role of Female CFO on Earnings Management during Pre and

Post Crisis Periods

The role of female CFO on earnings management was examined during pre and post crisis periods by comparing discretionary accruals (DA) calculated from all male CFOs (CFO=0) against those with the presence of female CFOs (CFO=1) during the two periods. The descriptive statistics of the discretionary accruals by each category are presented in Table

of samples of each category in Table 2a, I conducted the t-test as described in the Research Methodology section and the results are presented in Table 2b. Since the t-test is mainly based on the averaged data, I conducted further analysis by using the minimum and maximum data in Table 2a by considering that those values to some extent may represent negative and positive earnings management by using maximum and minimum data, respectively. The results of this analysis are shown in Figure 1.

Table 2a. Descriptive statistics of discretionary accruals of all male CFOs and that with

the presence of female CFOs during pre and post-crisis periods

Descriptive Statistics Pre-Crisis Post-Crisis

Minimum -0.359 -0.207

Maximum 0.339 0.444

Male CFO (CFO=0) Average -0.039 -0.016

Standard Deviation 0.062 0.051

No. of samples (n) 566 543

Minimum -0.200 -0.073

Maximum 0.046 -0.005

Female CFO (CFO=1) Average -0.049 -0.040

Standard Deviation 0.055 0.030

No. of samples (n) 21 7

Table 2b. Result of the t-test on discretionary accruals of all male CFOs (CFO=0)

compared to that with the presence of female CFOs (CFO=1) during pre and post-crisis periods.

[note: the same letters following the average values (in bold fonts) indicate non significant difference, ns; while * and ** are significantly different at 0.05 and 0.01 levels, respectively, and df is degree of freedom]

Pre-Crisis Post-Crisis _d f t t0.05 t0.01 CFO=0 0.039 a 0.016 b 1107 6.758** 1.962 2.581 CFO=1 _0.049 a _0.040 a _{26 0.5451}ns 2.056 2.779 d_f 585 548 t _0.814ns 2.078* t0.05 1.972 1.973 t0.01 2.602 2.604

The absolute value of DA decreases (both for minimum and maximum DA) with the presence of female CFOs during both pre and post crisis periods (Figure 1). This feature qualitatively supports hypothesis H1: Female CFO is negatively associated with earnings management. In addition to this finding, considering DA as the proxy of earnings management, is that negative earnings management (from minimum DA, Figure 1) during pre-crisis period tends to be less in value during post-crisis period; while positive earnings management (from maximum DA, Figure 1) during pre-crisis period tends to be higher for all male CFOs but to be lower with the presence of female CFOs during post-crisis period. This feature shows that the presence of women as CFO has a role in reducing earnings management from pre crisis to post crisis period which is qualitatively consistent with the hypothesis H5: Female CFOs have stronger negative association with earnings management in the post crisis period than that in the pre crisis period.

After all, considering the t-test results as shown in Table 2b in which the values of absolute DA are not significantly different (P>0.05) between all male CFOs (CFO=0) and that with the presence of female CFOs (CFO=1) both during pre and post crisis periods, thus, the hypothesis H1: Female CFO is negatively associated with earnings management, is rejected.

Figure 1. Comparison of minimum and maximum discretionary accruals of all male

CFOs against that with the presence of female CFOs during pre and post crisis periods.

5.3. Interaction of Women as CFO and CEO on Discretionary Accruals

during Pre- Crisis Period

The analysis results of interaction between male/female CFO and CEO on discretionary accruals (DA) during pre-crisis period are presented in Table 3a which shows the descriptive statistics. From the data in Table 3a, the t-test was conducted based on the available data of average, standard deviation and number of samples. The results are presented in Table 3b showing the interaction between the role of CFO and CEO on DA with the presence of women in those positions. First, the highest value of DA is in all male CFO and CEO (i.e. CFO=0 and CEO=0) with the value of DA = 0.309 compared to less than DA = 0.05 when women are present as CFO or CEO. The interesting feature is that female CFO significantly reduces DA when no women is present as CEO, and female CEO significantly decreases DA when no women is present as CFO. Overall, the t-test results presented in Table 3b support hypothesis H2: Female CEO is negatively associated with earnings management, particularly when no women as CFO.

Table 3a. Descriptive statistics of discretionary accruals on interaction of male/female

CFOs and male/female CEOs during pre-crisis period Descriptive Statistics Male CEO

(CEO=0)

Female CEO (CEO=1)

Minimum -0.359 -0.304

Maximum 0.339 0.206

Male CFO (CFO=0) Average -0.309 -0.035

Standard Deviation 0.061 0.071

No. of samples (n) 499 67

Minimum -0.200 0.037

Maximum 0.046 0.037

Female CFO (CFO=1) Average -0.049 0.037

Standard Deviation 0.056 -

Table 3b. Result of the t-test on discretionary accruals on the interaction of male/female CFOs and that of male/female CEOs during pre-crisis period

[note: the same letters following the average values (in bold fonts) indicate non significant difference, ns; while * and ** are significantly different at 0.05 and 0.01 levels, respectively, and d_f is degree of freedom]

CEO=0 CEO=1 _d f t t0.05 t0.01 CFO=0 _0.309 c _0.035 a 564 30.526** 1.973 2.603 CFO=1 _0.049 a _0.037 a 19 _0.958ns 2.051 2.769 d_f 517 66 t 20.287** _0.234ns t0.05 1.974 2.012 t0.01 2.605 2.686

The analysis results on minimum data provide negative values which suggest the negative earnings management. In terms of absolute value of the minimum DA, it represents the magnitude of the DA regardless whether it is negative or positive earnings management. By using the absolute value, it shows that without any female CFO and CEO, thus, the DA is the highest compared to that if female CFOs or CEOs are present. Figure 2 indicates that female CFO decreases the absolute DA significantly while additional presence of female CEO further decreases absolute DA but in a smaller magnitude. In this case, the role of female is very significant as CFO and even more important than that of being CEO in reducing absolute DA.

In the case of maximum data, all values are positive and hence they equal to absolute values which represent positive earnings management (Figure 3). Consistent with the negative earnings management, the presence of female CFO and CEO decreases DA but the magnitude as CFO is also more significant than that of CEO.

The two findings on negative and positive earnings management support hypothesis H2: Female CEO is negatively associated with earnings management which is consistent with the results of t-test.

Figure 2. Interaction between male/female CFOs and male/female CEOs on discretionary accruals (data on minimum values) for its actual DA (left) and absolute DA (right) during Pre-Crisis period.

Figure 3. Interaction between male/female CFOs and male/female CEOs on discretionary

accruals (data on maximum values) during pre-crisis period.

5.4. Interaction of Women as CFO and CEO on Earnings Management

during Post-Crisis Period

Descriptive statistics of discretionary accruals as the proxy of earnings management for the interaction between female CFO and CEO during post-crisis period are presented in Table 4a. The results of t-test are presented in Table 4b which shows that no significant difference on DA between all male CEO and that with the presence of female CEO. Also, there is no significant difference between all male CFO and that with the presence of female CFO except that when there is female CEO but the value of DA is even smaller if all CFO are males. Therefore, the t-test suggest to reject H1 (Female CFO is negatively associated with earnings management) but different analysis shows differently.

Table 4a. Descriptive statistics of discretionary accruals on interaction of male/female

CFOs and male/female CEOs during post-crisis period Descriptive Statistics Male CEO

(CEO=0)

Female CEO (CEO=1)

Minimum -0.207 -0.120

Maximum 0.444 0.255

Male CFO (CFO=0) Average -0.016 -0.020

Standard Deviation 0.050 0.062

No. of samples (n) 492 51

Minimum -0.073 - 0.038

Maximum -0.005 - 0.038

Female CFO (CFO=1) Average -0.041 - 0.038

Standard Deviation 0.033 -

No. of samples (n) 6 1

Table 4b. Result of the t-test on discretionary accruals on interaction of male/female

CFOs compared to that of male/female CEOs during post-crisis period

[note: the same letters following the average values (in bold fonts) indicate non significant difference, ns; while * and ** are significantly different at 0.05 and 0.01 levels, respectively, and d_f is degree of freedom]

CEO=0 CEO=1 _d f t t0.05 t0.01 CFO=0 _0.016 a,b _0.020 b 541 _0.446 ns 1.973 2.604 CFO=1 _0.041 a _0.038 a 5 _0.223ns 2.571 4.032 d_f 496 50 t _1.830ns 2.073* t0.05 1.974 2.025 t0.01 2.606 2.715

I also transformed the data into absolute values to easily measure the magnitude of DA for further analysis, both for minimum and maximum data. Although the t-test does not support the hypothesis, however, consistent with the previous findings the magnitude of DA (using minimum data) decreases with the presence of female CFO and it decreases more if female CEO is present (Figure 4). The role of women as CFO is also more significant than that of women being CEO. The magnitude of decreasing DA is even pronounced if maximum data is used (Figure 5).

Figure 4. Interaction between male/female CFOs and male/female CEOs on discretionary

accruals (data on minimum values) for its actual DA (left) and absolute DA (right) during post-crisis period.

Figure 5. Interaction between male/female CFOs and male/female CEOs on discretionary

accruals (data on maximum values) for its actual DA (left) and absolute DA (right) during post-crisis period.

5.5. Role of Women in Audit Committee

The role of women in Audit Committee (AC) was analyzed by comparing the descriptive statistics of Audit Committee which comprises all men (AC=0) and that with the presence of women (AC=1) during both pre and post-crisis periods. Similarly with the previous test, I also conducted t-test based on the absolute discretionary accruals data presented in Table 5a to search the difference of gender’s role (the role of female members in the AC) and that of pre-crisis against post-crisis. I assumed that the presence of women in the Audit Committee will also have negative influence on the earnings management as stated in the hypothesis H3: Female Audit Committee is negatively associated with earnings management.

Based on the descriptive statistics as presented in Table 5a and t-test being conducted, there is very significant difference (P<0.01) between pre and post crisis periods both for all male AC and that with the presence of women in the AC. In addition, significant difference in DA between all male AC and that with the presence of women in AC only occurs during pre-crisis period but the magnitude is smaller for all male AC. Therefore, on the contrary to the hypothesis (H3), the presence of female Audit Committee is not negatively associated with earnings management. The earnings managements even tends to be higher during pre crisis periods with the presence of women in the Audit Committee as measured by discretionary accruals which is presented in Figure 6. Therefore, the hypothesis H3 : Female Audit Committee is negatively associated with earnings management is rejected. Interesting feature is that positive earnings management increases while negative earnings management decreases in absolute values from pre-crisis period to post-crisis period both for all males as well as that with the presence of female in the Audit Committee (Figure 6).

Table 5a. Descriptive statistics of discretionary accruals of all men (AC=0) and with

women present (AC=1) in Audit Committee during pre and post crisis periods

Gender Descriptive Statistics

Pre-crisis Discretionary Accrual Post-crisis Discretionary Accrual Minimum -0.323 -0.207 Maximum 0.192 0.255

All Male Audit Committee (AC=0) Average -0.032 -0.018 Standard Deviation 0.054 0.051 No. of samples (n) 239 181 Minimum -0.359 -0.186 Maximum 0.339 0.444 Female present in Audit Committee (AC=1) Average -0.044 -0.016 Standard Deviation 0.066 0.051 No. of samples (n) 348 369

Table 5b. Result of the t-test on discretionary accruals on of all men (AC=0) compared to

that with women present (AC=1) in Audit Committee during pre and post crisis periods.

[note: the same letters following the average values (in bold fonts) indicate non significant difference, ns; while * and ** are significantly different at 0.05 and 0.01 levels, respectively, and df is degree of freedom]

Pre-Crisis Post-Crisis _d f t t0.05 t0.01 AC=0 _0.032 b _0.018 c 418 2.716** 1.976 2.610 AC=1 _0.044 a _0.016 c 715 6.330** 1.969 2.595 d_f 585 548 t 2.414* _0.432 ns t0.05 1.972 1.973 t0.01 2.602 2.604

Figure 6. Comparison of minimum and maximum discretionary accruals of all men

against women present in the Audit Committee during pre and post crisis periods.

5.6. Role of Women in Compensation Committee

Table 6a presents the descriptive statistics of discretionary accruals, as the proxy of earnings management, of all men and that with women present in the Compensation Committee. Similar to the analysis of the role of women in the Audit Committee, I also conducted the t-test based on the absolute value of discretionary accruals data during pre and post crisis periods which results are presented in Table 6b. The role of women in Compensation Committee (CC) was also analyzed by comparing the descriptive statistics of discretionary accruals which comprises all men (CC=0) and that with the presence of women (CC=1) in the Compensation Committee using minimum and maximum data. Similar to the Audit Committee, positive earnings management increases while negative earnings management decreases in absolute values from pre-crisis period to post-crisis period both for all men and that with the presence of women in the Compensation Committee. On the other hand, while the presence of women in the Audit Committee is not negatively associated with discretionary accruals, the presence of women in the

during pre-crisis and post-crisis in qualitative manner (Figure 7). However, the t-test does not show significant difference on the role of women on the Compensation Committee except that very significant difference between pre and post crisis periods for both all men and that with the presence of women in the Compensation Committee (Table 6b). Therefore, the hypothesis H4: Female Compensation Committee is negatively associated with earnings management is rejected but different feature indicates differently as shown in Figure 7.

Table 6a. Descriptive statistics of discretionary accruals of all men (CC=0) and with

women present (CC=1) in Compensation Committee during pre and post crisis periods.

Gender Descriptive Statistics

Pre-crisis Discretionary Accrual Post-crisis Discretionary Accrual Minimum -0.359 -0.207 Maximum 0.339 0.444 All Male Compensation Committee (CC=0) Average -0.040 -0.018 Standard Deviation 0.063 0.058 No. of samples (n) 296 241 Minimum -0.304 -0.151 Maximum 0.224 0.255 Female present in Compensation Committee (CC=1) Average -0.038 -0.015 Standard Deviation 0.059 0.045 No. of samples (n) 291 309

Table 6b. Result of the t-test on discretionary accruals of all men (CC=0) compared to

that with women present (CC=1) in Compensation Committee during pre and post crisis periods

[note: the same letters following the average values (in bold fonts) indicate non significant difference, ns; while * and ** are significantly different at 0.05 and 0.01 levels, respectively, and df is degree of freedom]

Pre-Crisis Post-Crisis _d f t t0.05 t0.01 CC=0 _0.032 a _0.018 c 535 4.205** 1.973 2.604 CC=1 _0.044 a _0.016 c 598 5.345** 1.972 2.601 d_f 585 548 t _0.397 ns _0.662ns t0.05 1.972 1.973 t0.01 2.602 2.604

Figure 7. Comparison of minimum and maximum discretionary accruals of all men

against that women present in the Compensation Committee during pre and post crisis periods.

6. Discussion, Conclusions and Limitations

6.1. Discussion

Role of Women in Different Positions as CFO, CEO, Audit Committee and Compensation Committee

The summary results of hypothesis tests using the t-test are presented in Table 7. The only t-test result which supports the hypothesis is on Table 3b for H2 : “Female CEO is negatively associated with earnings management”. In addition, using data of minimum and maximum values on Table 2a, both negative and positive earnings managements viz-a-viz absolute discretionary accruals as the proxy of earnings management are the highest if there is no female CFO both during pre and post crisis periods (Figure 1). The quality of audit statement is assumed to be better if there is independent directors (including CFO or CEO) which is, in this relation, to reduce the absolute discretionary accruals (DA) as the proxy of earnings management. Female directors are commonly seen to be out of the group of male directors (Thompson 2000), hence, their presence could act as the outside directors who are perceived to be more independent than the inside directors. As such, we may consider that the role of female directors both as CFO and CEO will improve audit quality. On the contrary, in this research thesis, based on the t-test the presence of women as CFO does not reduce earnings management as indicated by Table 2b by which the hypothesis H1: Female CFO is negatively associated with earning management is rejected. On the other hand, smaller earnings management represented by DA is observed when female CFO or CEO are present and the smallest earnings management exists when women are present as both CFO and CEO when minimum and maximum data instead of averaged values are compared (Figures 4 and 5).

This findings might be related to the trend of leadership style and effectiveness that are believed to be associated with feminine behavior. The characteristic of women is that they tend to be more cautious compared to men, which results in less risk in managing the fund. In this relation, the results show that variation of data represented by standard deviations of discretionary accruals for all male CFO and CEO are always higher than that if female directors are present (Tables 2a and 3a). This indicates that female directors are acting more securely than that of male directors. It might also suggest that the cautions and

commitments as well as consideration to people of the feminine behavior are generally perceived better than initiating structure characteristic of masculine behavior. In this thesis, however, the magnitude of earnings management as expressed by DA are only in agreement with H2 according to the t-test which is based on the averaged value instead of minimum and maximum data being employed and analyzed here. This feature suggests that varying results may be associated with the different approaches in the analysis as well as the assumptions for the underlying processes.

Although feminine behavior is perceived to be better than masculine behavior in managing the organization, the t-test does not support hypothesis H3 in which “Female Audit Committee is negatively associated with earnings management” (Table 5b, Figure 6). Further t-test (Table 6b) also rejects H4 that the presence of women in the Compensation Committee is negatively associated with earnings management although Figure 7 which is based on minimum and maximum data shows the contrary.

The overall picture shows that effectiveness of female roles in the management is not across the different positions. It suggests that the presence of women in the board is negatively associated with earnings management provided they are in the right positions. Based on the t-test as well as by the qualitative analysis based on minimum and maximum data of DA, the most effective position for women in the board is as CFO, CEO followed by Compensation Committee. The effectiveness might be related to the authority in directly manage as well as monitor the financial position. In this research, I found that the role of women as CFO on earnings management represented by discretionary accruals data is more profound than that of female in other positions. This might suggest that female director who has direct control on finance (as CFO and CEO) is more effective than female director who has indirect control on finance (as CC and AC). Indeed, the smallest earnings management represented by absolute discretionary accruals is found if female is present as both CFO and CEO which might suggest that female directors can act as more independent directors who can both manage and monitor the firm’s finance as well as audit report. As a result, this may give rise to a better audit quality.

Table 7. Summary of hypothesis test results (note: numbers/letters in the bracket

indicate the supporting Tables of the t-test results.

Hypothesis Accepted Rejected

H1: Female CFO is negatively associated with earnings

management Yes (3b) Yes (2b)

H2: Female CEO is negatively associated with earnings

management Yes (3b)

H3: Female Audit Committee is negatively associated with

earnings management Yes (5b)

H4: Female Compensation Committee is negatively associated

with earnings management Yes (6b)

H5 : Female CFOs has stronger negative association with earnings management in the post-crisis period than in the pre-crisis period.

Yes (4b)

The Difference of Female Role in the Board between Pre and Post Crisis Periods

The roles of women in board have been shown to be negatively associated with earnings management, both during pre-crisis and post-crisis periods for their position as CFO, CEO and Compensation Committee (all figures except Figure 6), while as AC it is positively associated with earnings management (Figure 6). In particular to female CFO/CEO, women are most effectively to decrease earnings management compared to the other positions in the board that might be associated with their feminine behavior associated with job satisfaction and organizational commitment of employee. Furthermore, commitments by female leaders have generated continued support to subordinates particularly during difficult times (Burke et al. 2006). As many women have emerged as powerful leaders in business which disqualifies the long perception of women being less powerful than men (Stufft & Coyne 2009), their role as CFO/CEO associated with earnings management are determining. This might explain even smaller earnings management during post-crisis period as observed in this thesis research (see Figure 1). In addition, this might be related to the statement of Armstrong et al. (2010) of “CFO is a key individual with substantial decision rights over financial reporting, and thus it seems reasonable that a high quality

CFO with appropriate incentives could have a positive influence on the quality of financial reporting”. In addition, the less amount of earnings management with the presence of women in the board might also be related to the fact that women are more risk averse than men to which female directors prefer to choose less-equity based incentives in order to decrease risk of aggressive reporting, and female CFOs are likely to take less risk, dividend payout, discretionary accruals and earning volatility (e.g. Eckel and Grossman 2008; Gneezy and Rustichini 2003).

Although outside directors on the Audit Committee may bring more independence in monitoring financial reporting activities of the managements (Armstrong et al. 2010), the presence of women in Audit Committee (AC) is insignificant and even positively associated with positive earnings management during post-crisis period (Figure 6). However, their presence in the Compensation Committee is similar to that as CFO but at a lesser extent (Figures 1 and 7). This feature might be related to insignificant women role on the Audit Committee, or it is caused by the insignificant role of the Audit Committee itself. This is based on the consideration that all women present in other positions tend to be negatively associated with earnings management.

6.2. Conclusions

The background of this research thesis was on the role of women in the board on the accounting quality during pre and post global financial crisis. This research examines the effect of the presence of women in the positions of CFO, CEO, Audit Committee and Compensation Committee on the earnings management. Earnings management was measured by discretionary accruals as its proxy by using S&P 500 firms from 2006 to 2011.

The presence of women on board is negatively associated with earnings management depending on their positions. Women in CFO, CEO and Compensation Committee tend to be negatively associated with earnings management while their presence in Audit Committee is, on the contrary, positively associated with earnings management. The effectiveness of women’s role in the earnings management and hence the audit quality is not across the board and most effective position of which women significantly reduce

earnings management in the post-crisis period than that in the pre-crisis period. The effect on the earnings management is even reduced when women are present both as CFO and CEO; and possibly with the presence of women in Compensation Committee. Although the presence of women in Compensation Committee is negatively associated with earnings management but its interactions with female positions in CFO and CEO were not examined.

6.3. Limitations

This research thesis employed data of S&P 500 firms from 2006 to 2011 while governance data was obtained from Risk Metrics merged with Compustat to retrieve financial reporting information which provided 1137 firm years. Since the number of women in the board is relatively small as also observed by Adams and Ferreira (2008), eventually the number of data for the combination of women present in different positions (i.e. both as CFO and CEO) is very limited and I found only one combination for both CFO and CEO during pre and post crisis periods, respectively. This feature might limit the sensitivity and accuracy of the analysis. Further research using a relatively big number of data with a more evenly distributed women present in various positions in the board might be necessary.

The positively associated relationship between women in Audit Committee and earnings management raises the questions whether it is related to insignificant women role on the Audit Committee, or insignificant role of the Audit Committee itself on the earnings management considering all in other positions in which women present tend to be negatively associated with earnings management although the t-test does not support it. Further study might be required to answer this issue. Another question that needs further research is interaction between the presence of women in Compensation Committee and their presence as both CFO and CEO. Indeed, for this research the distribution of data of women across the board with various types of position will be crucial for ensuring sensitivity and accuracy of the analysis.

Another limitation is that this research only employed S&P 500 firms from 2006 to 2011. Out of five hypothesis being proposed, only one hypothesis which is supported by the t-test although the qualitative analysis based on extreme values (maximum and minimum data)

shows the opposite. The results might be different if data from different years or different companies are used. Therefore, further research might be needed to validate these findings of this research thesis using different set of data with different years and companies.

References

Adams, R.B. and D. Ferreira. 2008. Women in The Boardroom and Their Impact on Governance and Performance. Journal of Financial Economics. pp: 1-45.

Adams, K. L. & Hambright, W. G. 2004. Encouraged or Discouraged? Women Teacher Leaders Becoming Principals. Clearing House, 77 (5), 209-211.

Ahern, K.R. and A. K. Dittmar, 2011. The Changing of the Boards: The Impact on Firm Valuation of Mandated Female Board Representation. The Quarterly Journal of Economics. pp: 1-15.

Armstrong, C.S., W.R. Guay, and J.P.Weber, 2010. The role of Information and Financial Reporting in Corporate Governance and Debt Contracting. Journal of Accounting and Economics 50, 179–234

Berman, E.M. 2007. Essential Statistics for Public Managers and Policy Analysts. 2nd ed. CQ Press. Washington, DC. pp: 179-197.

Bass, B. M. 1991. Bass & Stogdill’s Handbook of Leadership (3rd ed.). New York, NY: Free Press.

Burke, C.S., Stagl, K.C., Klein, C., Goodwin, G.F., Salas, E., & Halpin, S.M. 2006. What Type of Leadership Behaviors Are Functional in Teams? A meta-analysis. The Leadership Quarterly, 17, 288–307.

Bushman, R. M. and A. J. Smith. 2001. Financial Accounting Information and Corporate Governance. Journal of Accounting and Economics 32, 237–333.

Cohen, J., Krishnamoorthy, G. & Wright, A. 2004. The corporate governance mosaic and financial reporting quality. Journal of Accounting Literature, 87-152

Cole, M.S., Bruch, H. & Shamir, B. 2009. Social Distance as a Moderator of The Effects of Transformational Leadership: Both Neutralizer and Enhancer. Human Relations, 62 (11):1697–1733.

Dechow, P.M., R.G. Sloan and A.P. Sweeney. 1995. Detecting Earnings Management. The Accounting Review 70 (2): 193-225.

Dunn, P. 2012. Breaking The Boardroom Gender Barrier: The Human Capital of Female Corporate Directors. Journal of Management & Governance, 16 (4): 557 – 570. Eckel, C.C. and P.J. Grossman. 2008. Forecasting Risk Attitudes; an Experimental Study

Using Actual and Forecast Gamble Choices. Journal of Economic Behavior & Organization, 68 (1): 1- 17

Emmerik, H., Wendt, H. & Euwema, M.C. 2010. Gender Ratio, Societal Culture, and Male and Female Leadership. Journal of Occupational and Organizational Psychology, 83, 895–914.

Gneezy, U., M. Niederle and A. Rustichini. 2003. Performance In Competitive Environments: Gender Differences. Quarterly Journal of Economics, 118 (3): 1049–1074.

Healy, P., Wahlen, J. (1999), “A review of the earnings management literature and its implications for standard setting.” Accounting Horizons, 13(4): 365-383.

Hofstede, G. 1998. Identifying Organizational Subcultures: An Empirical Approach. Journal of Management Studies, 35 (1), 1-12.

Ittonen, K., Miettinen, J. and Vahamaa, S. (2009), “Does female representation in audit committees affect audit fees? Quarterly Journal of Finance and Accounting, 49, (3/4), 113-139.

Jensen, M.C. and W.H. Meckling. 1976. Theory of The Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economic. 3 (4): 305:360.

Jones, J. 1991. Earnings Management during Import Relief Investigations. Journal of Accounting Research 29: 193-228.

Laux, C. and V. Laux. 2009. Board committees, CEO compensation, and earnings management. The Accounting Review 84 (3): 869-891

Peni, E. and S. Vähämaa. 2010. Female executives and earnings management. Managerial finance, 36 (7); 629 – 645.

Scott, W. (2009), “Financial Accounting Theory.” 5th Edition.

Stufft, D., & Coyne, K. 2009. Educational Leadership: Where are the women? Advancing Women in Leadership Journal, 29 (11), 1-7.

Thiruvadi. S., & Huang. H. 2011. Audit Committee Gender Differences and Earnings Management. Gender in Management: An International Journal, 26 (7), 483-498. Thompson, M. D. 2000. Gender, Leadership Orientation, And Effectiveness: Testing The

Theoretical Models of Bolman & Deal and Quinn. Sex Roles, 42 (11/12), 969-992. Tichy, N. M. 1997. The Leadership Engine: How Winning Companies Build Leaders at

Every Level. New York: Harper Business.

Van den Berghe, L.A.A. and A. Levrau. 2004. Evaluating Boards of Directors: What Constitutes a Good Corporate Board? Corporate Governance. 12 (4): 461-478.