I found that earnings quality was lower in the period after the introduction of the board gender quota in Norway

MANDATED BOARD GENDER QUOTA AND EARNINGS QUALITY: EVIDENCE FROM NORWAY

Master thesis, Msc A&C, specialization Accountancy University of Groningen, Faculty of Economics & Business

August 25, 2018 LARS BROUWER Student number: S2508435

Sleedoornweg 181 9674 JN Winschoten tel.: +31 06-42141590 e-mail: larsbrouwer@live.nl

Supervisor/university Dr. Shibashish Mukherjee

Faculty of Business & Economics, Rijksuniversiteit Groningen

Second assessor/university Dr. Nazim Hussain

Faculty of Business & Economics, Rijksuniversiteit Groningen

Acknowledgement: I thank Shibashish Mukherjee greatly for his continuous support, advice and helpful comments when writing this thesis.

1 ABSTRACT

This thesis examines the relationship between earnings quality and the Norwegian board gender quota as board gender diversity is among the most debated issues in both the academic and corporate world. I found that earnings quality was lower in the period after the

introduction of the board gender quota in Norway. I, however, found no evidence this was caused by the quota. This thesis contributes to the debate on board gender quota as it can be used as a starting point for future research.

2 INTRODUCTION

The representation of women on corporate boards receives widespread attention, from both the media, governing organizations such as the EU and the academic community. Norway could be seen as the pioneer of board gender equality, as it passed a bill in December 2003 requiring all public-limited firms to have at least 40% women on their boards of directors by July 2005. At the time only 9% of directors were female. At first, the law was based on voluntary compliance, however as this failed, the law became compulsory in January 2006.

Firms would be allowed a 2 year transition period after which every firm that didn’t comply with the 40% quota would be dissolved. In April 2008 all firms would comply with the law.

However, women are still underrepresented in boardrooms with global estimates ranging from 12 to 15%. (Catalyst, 2017) Although progress regarding women in top positions at EU companies is being made, this is rather slow and most progress is concentrated in a few EU member states such as Norway.¹ This is why the European Commission is expected to advocate for a quota that requires 40% of a company’s non-executive board members to be female and companies that fail to meet this quota will have to prioritize female candidates (The guardian, 2017). These developments suggest that female board representation is perceived to be valuable as it has led to governing organizations intervening in the board composition process.

The question, however, remains whether such a quota is actually useful as the academic community is still divided on the effects of board gender diversity on firm performance. A recent study by Adams and Ferreira (2009), however, did provide some evidence that can be used to justify the current appeal for more gender diversity in the boardroom. They found that female directors provide better oversight over managers. Srinidhi, Gul and Tsui (2011) built on the research done by Adams and Fereira to find the first piece of empirical evidence that female representation on boards results in a tangible benefit for the firm. They found that firms with higher levels of monitoring, as a result of more female representation in the boardroom, have improved quality of earnings.

In this thesis, however, I will explicitly examine the effect of a board gender quota such as that of Norway. I choose to focus on a mandated quota because governing organizations such as the EU seem to find a hard quota the preferred method to stimulate firms to diversify their boards. The research done by Srinidhi, Gul and Tsui, although providing us with evidence in

1 http://ec.europa.eu/newsroom/document.cfm?doc_id=46280

3

favor of more board gender diversity, does not necessarily imply a hard gender quota is the solution. This is because their research was based on a voluntary setting, i.e. the firms were free to choose between male and female directors. Ahern and Dittmar (2012) found evidence from Norway that mandatory female representation may differ from voluntary representation, because companies were force to add lesser experience women to their boards. I will build on the research done by Srinidhi, Gul and Tsui, by examining the effect of a mandated board gender quota on the quality of earnings.

4 HYPOTHESIS DEVELOPMENT

The discussion for more gender diversity in the board room can be substantiated by using financial as well as ethical considerations (Campbell & Mínguez-Vera, 2007). Bernardi, Bosco and Columb (2009) found an association between the number of women on the board of directors of Fortune 500 companies and their presence on the Forbes’ World’s Most Ethical Companies list. They suggest that, because of the nature of the ethical analyses used to compose the list, companies with a higher percentage of women are perceived to be more ethical.

A lot of research has been done on the financial consequences of increasing the percentage of women on the board of directors. Srinidhi, Gul and Tsui (2007) found a positive relationship between gender diversity of boards and the informativeness of the respective company’s stock price. Their analyses suggest that diverse boards are more likely to disclose information voluntary in large firms and provide an incentive to investors to privately collect information in small firms. This relationship was especially strong for firms with weak corporate

governance. This suggests that gender diverse boards can be used to reduce the negative impact of poor governance. Studies on the relationship between board gender diversity and firm performance show mixed results. A meta-analysis on the relationship between board gender diversity and firm performance supports this. In their analysis they found a small positive relationship between board gender diversity and the accounting result. They found no evidence for a relationship between board gender diversity and market performance. They did however find a positive relationship between board gender diversity and board activities.

They found that boards with higher female representation perform more monitoring and are more involved with board strategy (Byron & Post 2015). From these studies we can conclude that research on the impact of gender diversity in the board of directors is of interest to not only society, but also to various stakeholders of a company.

A study by Adams and Ferreira (2009) contributed to this research. They found evidence that female directors provide better oversight over managers because they allocate more efforts to monitoring, and that women are more likely to participate in committees that are charged with reporting quality. They also found that female directors are less likely to have attendance problems and improve the attendance behavior of their male colleagues. Finally they found that diverse boards are more likely to hold CEOs accountable for poor stock performance. A study by McInerney-Lacombe, Billimoria, and Salipante (2008) found that gender-diverse

5

boards are more likely to discuss tougher issues. A study by Brabeck (1984) supports this idea. She found that women are much more like to ‘blow the whistle’. In the boardroom this may translate to women being more likely to report behavior associated with earnings management.

These findings are of particular interest from an agency theory perspective. It may suggest that a gender diverse board helps to negate the negative effects explained by agency theory.

By providing better oversight over managers, holding CEOs accountable for poor stock performance and discussing tougher issues, gender diverse boards might reduce information asymmetry and disclose poor alignment between principals and agents

This may suggest there exists a relationship between the representation of women on the board and the quality of reporting due to the better monitoring and better communication, which is of particular interest because high quality of earnings can be seen as a result of good corporate governance. Companies with weak corporate governance are more likely to

manage earnings and companies with strong governance have higher quality earnings. It is thus important for companies to display high quality of earnings, as this is representation of good corporate governance.

Srinidhi, Gul and Tsui studied the effect of gender diversity on reporting quality in their 2011 paper and found a positive relationship between the participation of women on boards and the quality of earnings. They suggest that the greater oversight provided by gender diverse boards is the reason for this increase in earnings quality. These findings are of particular interest to companies that want to improve the monitoring function of their board as well as increase earnings quality to satisfy investors. Increasing the gender diversity of the board is a way to help realize those goals.

The study done by Srinidhi, Gul and Tsui, however, was based on the assumption that companies were free to choose their board members. They were correct in assuming this because their sample did not include any firm year observations that were subject to a mandatory gender quota. This assumption, however, cannot always be made as research suggests mandatory female representation differs from voluntary female representation. A mandatory quota may result in a situation where companies have to keep quota compliance in mind when deciding between candidates. This can result in a firm choosing for a female, instead of a male addition to the board even though the female is less experienced. Ahern and Dittmar (2012) found evidence from Norway that mandatory female representation may differ

6

from voluntary representation. They found that because firms are not necessarily free to choose their board members, they were forced to choose women that had less experience than their male counterparts. Another study by Farrel and Hersch (2003) supports this idea. They looked into the additions to corporate boards and found that gender effects these additions, mainly because of internal or external pressure to increase gender diversity. Their results show that when a woman leaves the board, the probability of adding a new woman to the board increases. They also found that, the higher the percentage of women already on the board, the smaller the probability of a new board member being female. Finally, they found that as companies met their minimal gender diversity expectations, they no longer actively sought new female directors. This implies that there is a preference towards women until a certain diversity target is met. These results support the findings of the study by Ahern and Dittmar (2012) that when firms are confronted with gender diversity expectations or quota, gender becomes an important factor when adding a new member to the board.

The study by Ahern and Dittmar (2012) has some interesting implications. They found that, because of the small pool of available female board candidates, some board characteristics changed. The new female directors were younger on average and less experienced. As a result firms grew in size, made more acquisitions and accounting results were lower. These results may be due to less effective monitoring by the board. Another study by Bohren and Staubo (2012) studied Norwegian firms after the mandated 40% gender quota. They found that half the firms would exit to an organizational form not exposed to the law. They argue that such a mandated law may produce firms with inefficient boards.

This may suggest that the results from the previous research on the relationship between board gender diversity and the quality of earnings do not apply to a setting where gender diversity is mandated, because of the negative effects such a mandated quota has on the monitoring capabilities of the board. This research will try to close this gap in existing

literature by looking at the effect of board gender diversity on earnings quality in a mandatory setting. It will do so by testing the following hypothesis:

H1: A mandated board gender quota reduces earnings quality of Norwegian firms.

7 RESEARCH DESIGN

Measures

My choice for an earnings quality measurement is based on the fact that I am interested in the monitoring capabilities of the board. Literature on earnings quality identifies 2 ways of measuring the quality of the earnings. The first being a market based measure, which looks at the association between the quality of earnings and stock-based measures such as price. The second measure is about the ability of current earnings to predict future earnings and cash flows.

I choose the second type of measure because monitoring of the board stimulates managers to put more effort in estimating earnings as greater caution is required. The monitoring also discourages managers from overstating earnings when they try to reach benchmarks, or understating them when they already have reached such a benchmark by managing earnings.

(Srinidhi et al., 2011) Earnings management will thus serve as a proxy for the earnings quality.

Earnings management is often measured by the amount of discretionary accruals a manager uses. Discretionary accruals are accruals subject to management’s choices. I will use a model that estimates the nondiscretionary component of the reported income. (Dechow, Sloan &

Sweeney, 1995) By then subtracting the nondiscretionary component from the total accruals, I can calculate the discretionary component.

The first step in this process is to calculate the total accruals by solving the following equation:

𝑇𝐴𝐶𝐶_𝑡= ∆𝐶𝐴_𝑡− ∆𝐶𝑎𝑠ℎ − ∆𝐶𝐿_𝑡+ ∆𝐷𝐶𝐿_𝑡− 𝐷𝐸𝑃_𝑡 (1)

Where,

TACCt = total accruals in year t,

ΔCAt = change in current assets in year t, ΔCasht = change in cash in year t,

ΔCLt = change in current liabilities in year t,

ΔDCLt = change in short term debt included in current CL in year t,

8

DEPt = depreciation and amortization expenses in year t.

The next step is to split the total accruals in a discretionary and a nondiscretionary part, as I am only interest in the discretionary accruals. To do this I will need to estimate certain firm specific parameters which can be done by using an ‘estimation period’, a period during which no discretionary is expected. I choose to use the modified Jones model for this estimation introduced by Dechow, Sloan and Sweeney in 1995. This model eliminates a major limitation of the original Jones model, which is the assumption that revenues are nondiscretionary.

(Jones, 1991) The modified Jones model corrects for this limitation by assuming that all changes in credit sales during the event period (the period during which earnings management is expected to occur) results from earnings management. They made this assumption based on the premise that it is easier for managers to exercise discretion over credit sales compared to cash sales. In the modified Jones model, nondiscretionary accruals are estimated as:

𝑁𝐷𝐴_𝑡 = 𝛼_𝑡(_𝐴¹

𝑡−1) + 𝛼₂(𝛥𝑅𝐸𝑉_𝑡− 𝛥𝑅𝐸𝐶_𝑡) + 𝛼₃(𝑃𝑃𝐸_𝑡) (2) Where,

ΔNDAt = nondiscretionary accruals in year t,

ΔREVt = revenues in year t less revenues in year t-1 scaled at total assets at t-1,

ΔRECt = net receivable in year t less net receivables in year t-1 scaled at total assets at t- 1,

PPEt = gross property plant and equipment in year X scaled by total assets at t-1, At-1 = total assets at t-1 and

α1, α2 and α3 = firm specific parameters.

To estimate the firm specific parameters I will use the method used by the original Jones model, as there should be no difference between the two models since during the estimation period no earnings management is expected to occur. These estimates are generated using the following model:

𝑇𝐴_𝑡 = 𝑎₁(_𝐴¹

𝑡−1) + 𝑎₂(𝛥𝑅𝐸𝑉_𝑡) + 𝑎₃(𝑃𝑃𝐸_𝑡) + 𝜐_𝑡 (3) Where,

TAt = total accruals scaled by lagged total assets,

9 a1, a2 and a3 = OLS estimate of α1, α2 & α3 and υt = the residual.

By using an OLS regression to calculate a1, a2 and a3 I can then use those as estimates for α1, α2 and α3 in the modified jones model and calculate the nondiscretionary accruals. The discretionary accruals can then be calculated by solving the following equation:

𝐷𝐴_𝑡 = 𝑇𝐴𝐶𝐶_𝑡− 𝑁𝐷𝐴_𝑡 (4)

Where,

DAt = discretionary accruals in year t.

To measure the effect of the board gender quota on the quality of earnings I will use the difference in differences method popularized by Card & Krueger in 1994. The difference in differences method mimics an experimental research design using observational data by measuring the effect of a treatment on a treatment group and comparing it to a control group.

The current situation in Norway is perfect for such a method. The introduction of the mandated quota in 2008 allows for a clear distinction between a ‘before’ and ‘after’ period.

As such Norway will serve as my treatment group, and Sweden as the control group. Sweden has no mandated board gender quota, which is the most important determinant for choosing Sweden as the control group. My choice for Sweden as the control group is further based on the available data and the similarities between it and Norway. Sweden and Norway have a relatively similar population and GDP per capita, they are neighboring countries and are part both of Scandinavia. Another reason as to why Sweden will serve as a good control group is the fact that both Norway and Sweden score similarly on the Hofstede² country comparison index.

Table 1 summarizes the main variables of interest. All data comes from Boardex and

Compustat, except for one variable. I hand collected director level data. I did this by looking up publically available data. I used the following data sources: Linkedin, Bloomberg, annual reports, Wikipedia, firm websites and news articles. I collaborated with four other graduate

2https://www.hofstede-insights.com/product/compare-countries/

10

students to collect data on: profession, educational background, financial background, nationality and family status. I used this data to create a firm level variable that controls for the financial expertise of the board.

Note that data on earnings quality is available for roughly half of the total number of observations for both countries. The average value of the earnings quality proxy shows an interesting difference between the countries. On average, earnings quality in Norway is lower than in Sweden as the mean of discretionary accruals is higher. The mean amount of financial expertise on the board is also higher in Norway than it is in Sweden. Furthermore, this data indicates that Board characteristics are very similar between the two countries.

TABLE 1 Summary of variables

NORWAY SWEDEN

VARIABLE OBSERVATIONS MEAN ST. DEV. OBSERVATIONS MEAN ST. DEV.

EARNINGS QUALITY

566 0.108 0.238 974 0.062 0.084

ROA 1042 0.025 0.151 1677 0.036 0.160

TOBINS Q 1000 0.752 0.984 1681 1.032 1.334

BOARD TENURE

1091 0.673 0.572 1763 0.758 0.465

BOARD

INDEPENDENCE

1091 0.368 0.344 1763 0.506 0.307

BOARD SIZE 1091 1.931 0.332 1763 2.086 0.338

LOSS DUMMY 1034 0.264 0.441 1688 0.198 0.399

TOTAL ASSTES 1042 6.207 2.070 1709 6.344 2.203

GDP PER CAPITA

1091 11.382 0.031 1763 10.838 0.062

BOARD FINANCIAL EXPERTISE

1091 0.262 0.272 1763 0.175 0.182

Sample

My sample of the treatment group consists of 1091 firm observations. The Swedish control group consists of 1780 observations. Table 2 provides the sample details.

TABLE 2 Sample details

YEAR NORWAY SWEDEN TOTAL

2000 43 76 139

2001 54 89 143

11

2002 61 96 157

2003 66 108 174

2004 69 112 181

2005 71 114 185

2006 71 115 186

2007 72 117 189

2008 72 117 189

2009 68 110 178

2010 74 117 191

2011 76 116 192

2012 72 115 187

2013 70 114 184

2014 78 121 199

2015 74 126 200

TOTAL 1091 1763 2854

12 RESULTS

To measure the effect of the mandated board gender quota on the quality of earnings I use the following model:

EM = β0 + β1TREATMENT + β2POST + β3(TREATMENT * POST) + β4ROA + β5TOBINSQ + β6TENURE + β7INDEPENDENCE + β8SIZE + β9LOSS + β10ASSETS + β11GDP +

β12BRD_FIN_EXP + ε

Where,

EM = earnings management estimated using the modified Jones model, TREATMENT = treatment group dummy which equals 1 for Norwegian observations and 0 otherwise,

POST = treatment period dummy which equals 1 for all observations in year>2007 and 0 otherwise,

ROA = operational income/total assets,

TOBINSQ = market capitalization / (total assets + total liabilities), TENURE = board tenure,

INDEPENDENCE = board independence, SIZE = board size,

LOSS = loss dummy which equals 1 if roa<0 and 0 otherwise, ASSETS = natural log of total assets,

GDP = natural log of gdp per capita and BRD_FIN_EXP = financial expertise of the board.

Table3 shows the results of running a fixed effects regression that estimates the model.

TABLE 3

Association between earnings management and mandated gender quota EM (modified Jones) Coef. Std. Err. z P>z [95%

Conf.

Interval]

TREATMENT .201 .060 3.33 0.001*** .083 .320

POST -.000 .013 -0.01 0.990 -.025 .025

TREATMENT * POST -.039 .017 -2.26 0.024** -.073 -.005

ROA -.157 .031 -5.00 0.000*** -.219 -.096

TOBINSQ .012 .003 3.57 0.000*** .006 .019

TENURE -.026 .009 -2.98 0.003*** -.043 -.009

INDEPENDENCE -.007 .015 -0.49 0.625 -.037 .022

SIZE -.034 .019 -1.78 0.075* -.071 .003

LOSS -.005 .012 -0.39 0.697 -.030 .019

ASSETS .000 .003 0.03 0.976 -.006 .006

GDP -.258 .104 -2.47 0.014** -.463 -.053

BRD_FIN_EXP .022 .021 1.06 0.291 -.019 .064

13

_cons 2.935 1.130 2.60 0.009*** .721 5.149

*p<.10 **p<.05

***p<0.01

The main variable of interest is treatment * post as this denotes the situation of the mandated Norwegian gender quota. It can be seen that this variable is positively correlated with the earnings management variable. This means there exists a positive relationship between a mandated board gender quota and the amount of discretionary accruals estimated by the modified Jones model, which is in line with my hypothesis that a mandated board gender quota reduces the quality of earnings. This relationship is statistically significant as the p- value is 0.024. An unexpected result is the positive relationship between the financial expertise of the board and the earnings quality variable. This result, however, is not statistically significant.

Table 4 shows the results of a fixed effects regression that estimates the model when also controlling for the year variable

TABLE 4

Association between earnings management and mandated gender quota when controlling for time

EM (modified Jones) Coef. Std. Err. z P>z [95%

Conf.

Interval]

TREATMENT -.523 .175 -2.98 0.003*** -.866 -.179

POST -.497 .066 -7.52 0.000*** -.627 -.367

TREATMENT * POST .017 .021 0.82 0.415 -.024 .057

ROA -.137 .030 -4.53 0.000*** -.196 -.078

TOBINSQ .008 .003 2.38 0.017** .001 .015

TENURE -.021 .008 -2.51 0.012** -.037 -.005

INDEPENDENCE .006 .015 0.40 0.690 -.023 .034

SIZE -.016 .018 -0.87 0.382 -.052 .020

LOSS .004 .012 0.37 0.713 -.019 .028

ASSETS -.003 .003 -1.10 0.273 -.009 .003

GDP 1.028 .310 3.32 0.001*** .421 1.635

BRD_FIN_EXP .021 .020 1.02 0.310 -.019 .061

year

2001 -.293 .034 -8.65 0.000*** -.360 -.227

2002 -.331 .031 -10.63 0.000*** -.392 -.270

2003 -.357 .032 -11.23 0.000*** -.419 -.294

2004 -.382 .037 -10.21 0.000*** -.455 -.309

14

2005 -.421 .048 -8.83 0.000*** -.515 -.328

2006 -.422 .049 -8.54 0.000*** -.519 -.325

2007 -.452 .055 -8.17 0.000*** -.560 -.343

2008 .043 .023 1.83 0.068* -.003 .089

2009 .096 .034 2.87 0.004*** .030 .162

2010 .049 .027 1.85 0.064* -.003 .101

2011 .047 .025 1.89 0.058* -.002 .095

2012 .041 .025 1.63 0.102 -.008 .090

2013 .045 .025 1.79 0.073* -.004 .094

2014 .022 .023 0.94 0.345 -.023 .067

2015 0 (omitted)

_cons -10.612 3.323 -3.19 0.001*** -17.1241 -4.100

*p<.10 **p<.05

***p<0.01

When controlling for time the relationship between the gender quota and earnings quality loses its significance, suggesting the effect is caused by time rather than the quota itself. The years after the introduction of the gender quota in Norway show a positive association with the earnings management variable, whereas the years prior to the introduction of the quota show a negative association. These results, however, do not provide evidence this turnaround is caused by the introduction of the board gender quota and for this reason I cannot reject the null hypothesis.

15 CONCLUSION

In this thesis, I examine the relationship between the Norwegian mandated board gender quota and the quality of earnings of Norwegian firms. Research on the effects of gender diversity on firm characteristics is important because governing organizations like the EU are pushing for more female board participation. Previous research found that women on boards are associated with better monitoring and as a result higher quality earnings. These results, however, were based on a voluntary setting, i.e. firms were free to choose between female and male additions to their boards. Research suggests that a setting in which firms have to

prioritize women as a result of a mandated board gender quota may yield different results because gender becomes the deciding factor, rather than competence.

My results indicate that firms subject to the mandated board gender quota do exhibit lower earnings quality. This result, however, loses its significance when controlling for time. This may suggest the relationship is explained by time rather than the quota itself. “Earnings quality is an important outcome of good governance demanded by investors and therefore its improvement constitutes an important objective of the board” (Srinidhi et al., 2011:1639).

This study finds evidence that earnings quality has eroded over the past years in the tested sample. Whether this is due to the introduction of the board gender quota or some other effect is unclear. It is definitely interesting to both investors and firms as it may suggest the level of governance has also decreased over the years.

These results, however, may not be generalizable to other time settings or countries, mainly because board gender legislation may differ from the Norwegian legislation. Cultural and demographic difference may also prevent generalization. A shortcoming of this research is the fact that, although I have controlled for several governance factors, many factors remain uncontrolled for. Also, the measurement of earnings quality is prone to measurement errors.

Despite its limitations, this study provides some valuable insights into an issue that keeps becoming more important to companies and governing organizations alike. Future research could replicate this study in an attempt to eliminate the year effect, and to find out whether the decrease in earnings quality was caused by the introduction of the gender quota.

16 REFERENCES

Adams, R. B., & Ferreira, D. 2009. Women in the boardroom and their impact on governance and performance. Journal of financial Economics, 94, 291-309.

Ahern, K. R., & Dittmar, A. K. 2012. The changing of the boards: The impact on firm

valuation of mandated female board representation. Quarterly Journal of Economics, 127(1), 137-197.

Boffey, D. 2017. EU to push for 40% quota for women on company boards. [online] The Guardian. Available at: https://www.theguardian.com/world/2017/nov/20/eu-to-push-for-40- quota-for-women-on-company-boards [Accessed 7 Jan. 2018].

Bohren, O., & Staubo, S. 2012. Does mandatory gender balance work? Changing

organizational form to avoid board upheaval. Journal of Corporate Finance, 28, 152-168.

Brabeck, M. 1984. Ethical characteristics of whistle blowers. Jounral of Research in Personality, 18(1), 41-53.

Byron, K., & Post, C. 2015. Women on boards and firm financial performance: A meta- analysis. Academy of Management Journal, 58(5), 1546-1571.

Campbell, K., & Minguez-Vera, A. 2008. Gender Diversity in the Boardroom and Firm Financial Performance. Journal of Business Ethics, 83(3), 435-451.

Catalyst. 2017. Women on Corporate Boards Globally. [online] Available at:

http://www.catalyst.org/knowledge/women-corporate-boards-globally [Accessed 7 Jan.

2018].

Dechow, P. M., Sloan, R. G., & Sweeney, A. P. 1995. Detecting earnings management. Accounting Review, 193-225.

Farrell, K. A., & Hersch, P. L. 2005. Additions to corporate boards: the effect of gender.

Journal of Corporate Finance, 11, 85-106.

Jones, J. 1991. Earnings management during import relief investigations. Journal of Accounting Research, 29, 150-180.

Kothari, S.P., Leone, A. J., & Wasley, C. E. 2004. Performance Matched Discretionary Accrual Measure. Journal of Accounting & Economics, 39(1), 163-197.

17

McInerney-Lacombe, N., Biliamoria, D., & Salipante, P. F. 2008. Championing the

discussion of tough issues: how women corporate directors contribute to board deliberations.

In Vinnicombe, S., Singh, V., Burke, R., Bilimoria, D., & Huse, M. (Eds.) Women on Corporate Boards of Directors: International Research and Practice. Cheltenham, United

Kingdom: Edward Elgar

Srinidhi, B., Gul, F. A., & Tsui, J. 2011. Female Directors and Earnings Quality.

Contemporary Accounting Research, 28(5), 1610-1644.