Women director characteristics and earnings quality: evidence from banking industry in Indonesia

Abstract

Recognizing the growing importance of gender diversity in corporate governance, this study delves into the impact of women directors’ attributes on earnings management. Focusing on listed banks in the Indonesia Stock Exchange, we hand-collect women directors’ profiles from 2017 to 2021. Notably, our findings reveal that having women on the board correlates with reduced earnings management. Specifically, while higher educational attainment among women directors positively affects earnings management, this effect turns negative when considered in conjunction with their educational backgrounds. Additionally, the tenure of women directors appears to inversely impact earnings management. Crucially, our study breaks new ground by exploring the moderating roles of educational levels and backgrounds, thereby enriching existing literature and shedding light on nuanced effects across different levels of earnings management.

Keywords:

• Banking
• characteristics of women directors
• earnings quality
• education levels and backgrounds

REVIEWING EDITOR:

• Collins Ntim, University of Southampton, United Kingdom of Great Britain and Northern Ireland

JEL CLASSIFICATIONS:

• M4

1. Introduction

Drawing from agency theory and the upper echelon theory, this study seeks to investigate the influence of women directors’ attributes on earnings management within Indonesia’s banking industry. Earnings management refers to the deliberate use of discretionary accounting methods to present desired financial results (Bajra & Cadez, 2018; Gao et al., 2020; Gavious et al., 2012; Orazalin, 2020). In such scenarios, managers have an incentive to manipulate financial statements that potentially affect a company’s financial performance (Beneish, 2001; Christie & Zimmerman, 1994; Orazalin, 2020; Pham et al., 2017). A clear example of earnings management practice, as reported by Kompas.com (2019), is the case of Garuda Indonesia, which has treated revenue contrary to the Statement of Financial Accounting Standards in its financial statement. Another case was reported by detikFinance (2018) related to Bank Bukopin, which revised its 2016 financial statements in 2018. Excessive earnings management practices may reduce the quality of earnings as information to make decisions. In this case, the board of directors can be important in monitoring earnings reporting.

According to agency theory, the board monitoring function has an important role in reducing agency problems and monitoring managerial decisions to protect shareholder interests and ensure that financial statements are of high quality (Fama & Jensen, 1983; Lara et al., 2017; Orazalin, 2020; Xie et al., 2003). Kyaw et al. (2015) and Ntim (2015) argue that more diversification of the board of directors enhances corporate governance and mitigates the practice of earnings management in a company. One aspect of board composition that has garnered significant attention in previous studies is the presence of women on the board of directors (Al-Absy, 2022; Elzahar et al., 2022; Githaiga et al., 2022; Mensah & Boachie, 2023; Zalata, Ntim, Aboud, et al. 2019; Zalata, Ntim, Choudhry, et al. 2019). This has heightened scrutiny regarding the influence of women on the board on earnings management. The rationale stems from the perception that women directors exhibit greater independence, astute observation skills, and less tolerance for opportunistic behavior, making them effective overseers of companies (Adams & Ferreira, 2009; Fan et al., 2019). Furthermore, women frequently demonstrate a collaborative and consensus-driven approach compared to men, contributing to a more inclusive and effective board environment (Klein, 2002).

However, the previous studies of women on the board yield mixed evidence. The presence of women on the board of directors is noteworthy, as they can potentially mitigate earnings management practices within a company (Al-Absy, 2022; Elzahar et al., 2022; Fan et al., 2019; Gavious et al., 2012; Gull et al., 2018; Harakeh et al., 2019; Kim et al., 2017; Lara et al., 2017; Orazalin, 2020; Sana, 2018; Zalata, Ntim, Choudhry, et al., 2019). On the contrary, the presence of women was identified with a positive effect on earnings management (Arun et al., 2015; Buniamin et al., 2012; Kontesa et al., 2020; Zalata, Ntim, Aboud, et al. 2019). Meanwhile, according to Abdullah and Ismail (2016) and Ye et al. (2010), gender has no significant effect on earnings quality. Furthermore, Owen and Temesvary (2018) found that a positive relationship between higher gender diversity on bank boards and bank performance is only manifested after a certain threshold of female participation was reached. Pathan and Faff (2013) showed that gender diversity positively affects bank performance; however, that positive influence was only evident in the pre-Sarbanes-Oxley Act (SOX) period (1997–2002). After SOX, the link between gender diversity and performance weakened.

The debatable results of previous research regarding the impact of women on boards on earnings management raise a question about the role of women on boards, especially in Indonesia. This research focuses on Indonesia, a country with a male-dominated, aggressive masculine gender culture where people are driven to compete, achieve, and succeed from the start of their school till their organization’s life (Hofstede, 2001). The lack of normative regulations concerning women’s participation on corporate boards in Indonesia prompts inquiries into their role, particularly in mitigating earnings management and ultimately enhancing earnings quality. Furthermore, we studied this issue in the banking industry as this industry is crucial for Indonesia’s economy, and the existence of earnings management cases in the banking industry will have a crucial impact on making economic decisions.

Within the Indonesian context, the studies have suggested that the presence of women may diminish firm performance (Aryani et al., 2018; Darmadi, 2013; Nugrahani & Yuniarti, 2021). However, in the case of earnings quality, Purwa and Setiawan (2021) found that gender positively affects accounting conservatism, indicating that companies with low masculinity report earnings conservatively. Furthermore, Sudarman and Hidayat (2020) found that women’s audit committees negatively reduce earnings management. Anggraini and Gustivani (2022) and Nurmayanti et al. (2022), found that women CEO do not significantly influence accrual earnings management. While Karina (2021) found that board diversity moderates the positive relationship between board independence and earnings management.

Therefore, research on the existence and role of women in companies, especially in mitigating earnings management, is still inconclusive. The inconclusive evidence of the effect of women on board on earnings management is possible due to the missing moderating effect of other variables. Namazi and Namazi (2016) argue that a moderating relationship is better in examining the relationship between two variables in a complex business rather than the direct effect.

Upper echelon theory suggests that the top management team’s characteristics influence the organization’s decisions and actions. This theory posits that the board of directors is composed of individuals with diverse backgrounds, experiences, and perspectives. The board diversity leads to a more innovative and effective organization (Adams & Ferreira, 2009; Klein, 2002). Therefore, we include women’s characteristics on board, which are educational level, educational background, and tenure, as moderating variables.

Based on an accounting perspective, Kontesa et al. (2020) stated that director competency is important in governance efficacy. The board of directors must possess competence to effectively oversee and monitor the financial reporting process as well as enhance earnings quality (Reeb & Zhao, 2013). The greater a person’s level of education, the more inclined they are to make informed decisions (Cheng et al., 2010; Papadakis & Barwise, 2002). Directors with backgrounds in economics, finance, and accounting are better equipped to manage profits than those lacking such relevant expertise (Xie et al., 2003). However, directors with an accounting background have more incentives to engage earnings management in financial reports (Hu et al., 2017; Ngo & Nguyen, 2022). Another aspect discussed in this study is the tenure of women directors in office, which has been reported negatively correlated to earnings management in earlier studies (Bouaziz et al., 2020; Fan et al., 2019). Nonetheless, Hu et al. (2015) stated that there was no linear relationship between the tenure of women directors and earnings management.

However, while the insights from previous literature are valuable, they primarily examined women’s characteristics solely as independent variables. To conduct earnings management, a person needs an accounting background. The better decision of a person with higher education will change with an accounting background in the case of earnings management, as well as with experience that is measured with the tenure of women directors in office. Consequently, the current paper seeks to make the following contribution to the existing literature. First, our research design introduces the moderating effects of educational levels and background. Second, using sub-sample analysis, we provide evidence of how women directors relate to the conditional level of earnings management. Third, this study uses banking companies as the focus observation, whereas banking companies are often not included in the research, especially those that discuss earnings management. Finally, we provide evidence of the role of women on the board of directors in countries with male-dominated cultures.

Our analysis reveals that the presence of women on the board of directors has a negative and significant effect on earnings management. The educational level of women directors has a positive and significant effect on earnings management. However, when their educational backgrounds moderate the effect of the women directors’ educational levels on earning management, the result shifts to a negative effect. Meanwhile, the term office for women directors has a negative effect on earnings management. In addition, the findings indicate that the governance attribute of women directors is conditional to the level of earnings management.

The rest of the article is organized as follows. Section 2 presented the context and background of the study. Section 3 reviews the theoretical framework. Section 4 reviews related prior literature, followed by the development of hypotheses. Section 5 demonstrates the research design. Section 6 reports and discusses the empirical findings. Section 7 reports the additional analysis. Section 8 reports the robustness test. Finally, section 9 concludes the article and suggests avenues for future research.

2. Background

In recent years, emerging markets have gained significant prominence in the global economy, and only a handful of researchers overlook them. There’s limited understanding of the implications of corporate governance policies, specifically in financial markets of emerging economies, for example, Indonesia. This research focuses on Indonesia for several reasons. First, in Indonesia, gender equality in work is regulated and guaranteed by the state through the Indonesia Constitution 45 Article 28 D paragraph 2 (Indonesia Constitution, 1945), which affirms that every Indonesian citizen has the right to work and get fair and decent remuneration and treatment in employment relations. In this case, the state guarantees fair treatment of workers in terms of the type of work, placement of positions in work, and the provision of wages. Although normatively, there are equal rights between female and male workers, women’s labor conditions are still unequal in quantity and quality. Gender inequality in employment can be identified by looking at the labor force participation rate of women and men. The latest data from Statistics Indonesia (Badan Pusat Statistik (BPS), 2023) shows that there is still a high gap of the labor force participation rate by gender. The male labor force participation rate in 2022 was 83.87%, an increase compared to 2021 at 82.27%. Meanwhile, women’s labor force participation rate in 2022 was only 53.41%, up slightly compared to 2021, which was 53.34%. Women working as professionals are also only 48.65%, down from 49.99% in 2021. Compared to other countries, data on women’s empowerment in Indonesia is also still relatively low. According to the United Nations Development Programme, although Indonesia is included in the high human development group, it ranks 114 in the world, with no data for the female share of employment in senior and middle management (United Nations Development Programme (UNDP), 2023).

Second, unlike developed countries such as Canada, the UE, France, Italy, and the UK, which have adopted various gender-based quotas, either mandatory or voluntary, for positions on corporate boards (Zalata, Ntim, Aboud, et al., 2019), there is a lack of normative regulations in Indonesia that mandate the presence of women on both boards of directors. In the absence of regulation, the inclusion of women in top executive positions remains voluntary. This was confirmed by the statement of Agus Saptarina, director of the Indonesian Financial Services Authority, who stated that the contribution of women in the board of commissioners and directors was still below 50% (Otoritas Jasa Keuangan (OJK), 2021). Furthermore, Agus Saptarina stated that the number of companies with a female board of commissioners in 2019 was 40% and decreased to 38% in 2020. Meanwhile, the number of companies with female board directors in 2019 was 45% and rose slightly to 46% in 2020. Meanwhile, the number of commissioners and directors based on gender in 2019–2020 is still very small. The number of female members of the board of commissioners was only 13% in 2019–2020, while female directors were recorded at 15% in 2019 and rose to 16% in 2020. Therefore, the absence of normative regulations has resulted in an unnoticeable rise in the number of women represented on corporate boards in Indonesia.

This phenomenon makes a perfect context for examining the role of women on a company’s board, especially with the phenomenon of earnings management. First, earnings management as one of the determinants of earnings quality, as proposed by Healy and Wahlen (1999), also exists in Indonesia. Suhardianto and Harymawan (2011) found that 18% of accounting research in Indonesia examines earnings quality. It stated that corporate governance factors dominate the studies on earnings quality in Indonesia, and the conclusion is still debatable. Furthermore, Kontesa et al. (2020) argue that, similar to other developing countries, Indonesia still has a gap regarding education among firm executive members, providing an excellent platform to study this issue.

This study focuses on the banking industry due to its significant contribution to economic development through financial instruments. This industry is regarded as a mirror of economic growth (Purwa & Setiawan, 2021). Moreover, in the context of regulatory compliance, the banking industry differs from other industries, as mentioned by Nasution and Setiawan (2007). For example, in the Bank of Indonesia (2013) Regulation No. 15/2/PBI/2013, related to Status Determination and Follow-Up Supervision of Conventional Commercial Banks, the banks must comply to avoid inclusion in the category of unhealthy banks. Similarly, Financial Services Authority Regulation No. 37/POJK.03/2019 (OJK, 2019) concerns Transparency and Publication of Bank Reports related to financial statements. Furthermore, banks play a crucial role in the financial system, and their stability is vital for economic growth and individual well-being (Owen & Temesvary, 2018). Therefore, studying factors that influence earnings quality in the banking industry is critical.

Despite existing regulations for banks in Indonesia, a considerable number of financial scandals continue to emerge. Megascandal cases in Indonesia include the BLBI case, Century Bank, Panama Papers, and Fincen Files. Notably, the 2002 earnings management scandal within Indonesia’s banking sector that garnered significant attention involved Lippo Bank, which issued dual financial reports on 27 December 2002. Atqiyaa and Triastuti (2022) argue that Lippo Bank manipulated the financial statements presented in comparative form between 30 September 2002 (audited) and 30 September 32001 (unaudited) by stating that the company’s total assets were IDR 24.185 trillion and net profit amounted to IDR 98.77 billion. The report also stated that the value of collateral taken over amounted to IDR 2.393 trillion, and the Capital Adequacy Ratio (CAR) amounted to 24.77% on 27 December 2002. However, what the Jakarta Stock Exchange published on 27 December 2002, differed from the financial statements published by Lippo Bank to the public. In the financial statements presented in comparative form as of 30 September 2002, 31 December 2001 and 31 December 2000, it is written that the company’s total assets amounted to IDR 22.8 billion, the value of collateral taken over as of 30 September 2002 amounted to IDR 1.43 trillion and net loss as of 30 September 2002, amounted to IDR 1.273 trillion and CAR amounted to 4.23% (Atqiyaa & Triastuti, 2022).

The most recent earnings management case occurred at Bukopin Bank in 2018. Bukopin Bank revised its financial statements from 2015, 2016 and 2017 due to manipulating credit card data spanned over five years. Bukopin Bank adjusted its 2016 net profit from IDR 1.08 trillion to IDR 183.56 billion. The decline was caused by credit card provision and commission income, from IDR 1.06 trillion to IDR 317.88 billion. The revisions also occurred in the financing of Bukopin’s Syariah Bank (BSB) subsidiary related to the addition of reserve balances for impairment losses for certain debtors so that the allowance for impairment losses on financial assets increased from IDR 649.05 billion to IDR 797.65 billion. At the end of 2016, Bukopin Bank revised the decline in equity owned from IDR 9.53 trillion to IDR 6.91 trillion due to a downward revision of the profit balance from IDR 2.62 trillion to IDR 5.52 trillion. The revision occurred because the previously reported profit was incorrect. The effect of declining equity can also decrease the Capital Adequacy Ratio from 15.03 to 11.62%, which also increases Bukopin Bank’s Non-Performing Loan (detikFinance, 2018).

Based on the above description, this study analyzes whether women influence the board of directors in reducing earnings management in the banking industry in Indonesia. In addition, it also examines whether the characteristics of women directors, such as education level, educational background, and tenure in the banking industry of Indonesia, influence earnings management.

3. Theoretical framework

In order to maximize shareholders’ wealth, prior studies indicate that effective corporate governance is a critical mechanism to mitigate manager opportunistic behavior, such as earnings management (Fama & Jensen, 1983). As part of corporate governance, the board of directors can do a task to monitor manager behavior or hold a function as an advisor. Prior research proves that board of directors’ diversity, including gender diversity, significantly affects manager behavior positively or negatively (Githaiga et al., 2022; Musa et al., 2023). Based on the evidence provided by prior studies, women directors are the best to do the monitoring role to mitigate earnings management (Zalata, Ntim, Choudhry, et al., 2019). This raises the motivation to study more deeply about the characteristics of women directors to monitor the manager behavior in earnings management effectively. On the other hand, there are many different theories (e.g. economic, governance, sociological, and psychological) underlying the interpretation of the outcome of the investigation around the effect of gender diversity on earnings management (Elzahar et al., 2022). In this current study, we employ two theories to explain the effect of female director characteristics on earnings management: agency theory and upper echelon theory. Agency and upper-echelon theories address managers’ behavior and its implications for organizations, though they emphasize distinct aspects.

3.1. Agency theory

Agency theory is a framework to understand the relationship between principals (owners) and agents (managers) in a company (Jensen & Meckling, 1976). It indicates a potential conflict of interest between principals and agents, as managers might not consistently prioritize the shareholders’ best interests. Fama and Jensen (1983) argue that the separation of ownership and control in a company can lead to agency problems, which arise when managers pursue their own interests rather than the interests of shareholders. In this case, managers have more access to the firm’s internal information compared to shareholders, resulting in information asymmetry (Pham et al., 2017). Information asymmetry between managers and shareholders allows managers to carry out earnings management to achieve their desired outcomes. Managers are motivated to conduct earnings management for three reasons: debt covenants, managerial compensation, and stock price motivations (Fama & Jensen, 1983; Jensen & Meckling, 1976). To mitigate earnings management, the monitoring function of the board becomes critical for a company (Elzahar et al., 2022; Fama & Jensen, 1983; Jensen & Meckling, 1976). According to the strand of gender diversity studies, females are better monitors (Nguyen et al., 2020; Zalata, Ntim, Aboud, et al., 2019; Zalata et al., 2022). It is expected that the female director will provide better monitoring of the financial statements. In line with agency theory, exploring the specific attributes that empower female directors to excel as corporate monitors, safeguard shareholder interests, and uphold the integrity of financial reporting is crucial. Thus, women directors have minimized earnings management and enhanced accounting quality information (Setiawan et al., 2020).

3.2. Upper echelon theory

Upper echelon theory (UET) is a theory of organizational behavior that suggests that the characteristics of the top management team (such as the board of directors) influence the organization’s decisions and actions (Adams & Ferreira, 2009). Upper echelon theory suggests that characteristics of top executives, such as their education, experience, and values, influence organizations’ strategic decisions and performance (Adams & Ferreira, 2009; Klein, 2002). Accordingly, upper echelon theory suggests that the composition of a corporate board plays a crucial role in shaping decision-making strategies, as board members’ collective experiences and knowledge influence the decisions they make (Đặng et al., 2020; Elzahar et al., 2022; Nguyen et al., 2020). In their study on gender-diverse boards, Đặng et al. (2020) acknowledge the established notion of cognitive distinctions between female and male directors. Women are considered capable of supervising companies (Adams & Ferreira, 2009) as women directors are more independent, keen observers, and less tolerant of opportunistic behavior (Fan et al., 2019). The same result was found when women were present on the board of commissioners as chief financial officers (Setiawan et al., 2020). In the case of earnings management, there is still debate regarding women’s role in mitigating earnings management. Some studies found that women directors are able to mitigate earnings management (Al-Absy, 2022; Elzahar et al., 2022; Fan et al., 2019; Orazalin, 2020; Zalata, Ntim, Choudhry, et al., 2019). On the other hand, some studies found that the presence of women directors increase the earnings management (Arun et al., 2015; Buniamin et al., 2012; Kontesa et al., 2020; Zalata, Ntim, Aboud, et al. (2019). Whereas, other studies proved no relationship between women on board and earnings management (Abdullah & Ismail, 2016; Ye et al., 2010). The debate around the effect of women on board directors on earnings management is possibly due to the inherent characteristics of women directors. Consequently, the study of women directors’ characteristics and earnings management focusing on the banking industry in Indonesia is beneficial.

4. Literature review and hypothesis development

4.1. The presence of women on the board of directors on earnings management

Based on agency theory, there is a potential conflict between the agent (management) and principal (stakeholders) (Fama & Jensen, 1983; Jensen & Meckling, 1976). In this case, management has more control over the company’s operational information than the stakeholders, which results in information asymmetry. To reduce the information asymmetry problem, management reports the company’s operational results through financial reporting. However, the quality of earnings reported in financial statements is often questionable as management is motivated to do earnings management. Earnings management is the outcome of the accrual-based accounting system used in financial reporting (Setiawati & Nai’m, 2000). The earnings management activities are related to manipulating earnings to influence financial reporting and cannot deny the role of the managerial staff (Bajra & Cadez, 2018). Therefore, management is incentivized to manipulate earnings, which can affect a company’s financial performance (Beneish, 2001; Christie & Zimmerman, 1994; Orazalin, 2020; Pham et al., 2017). To guarantee high earnings quality, the company must have a board monitoring function to avoid excessive earnings management practices that contradict accounting standards. Several studies have stated that gender diversity on the board of directors can positively impact the company because a more diverse gender can better mitigate the practice of earnings management (Kyaw et al., 2015) and, finally, positively impact organizational valuation (Ntim, 2015).

This study discusses the relationship between the presence of women on the board of directors and earnings management, as the relationship can positively impact the company, as reported by several researchers (Abdullah & Ismail, 2016; Lara et al., 2017; Orazalin, 2020). In addition, in one of the reports from the International Monetary Fund compiled by Sahay et al. (2015), the role of a woman on the board of directors was mentioned as a person who can influence the decision-making process, offer quality supervision, and significantly improve the financial stability of the bank.

Previous studies about the presence of women directors in a company identify the benefits due to a high level of independence and active supervision of women directors (Fan et al., 2019). In addition, women are intolerant of opportunistic activities. As a result, they can improve the quality of monitoring and reporting of financial statements (Al-Absy, 2022; Elzahar et al., 2022; Fan et al., 2019; Zalata, Ntim, Choudhry, et al., 2019). Therefore, women can reduce earnings management in companies, especially in the banking sector (Bouaziz et al., 2020; Fan et al., 2019; Githaiga et al., 2022; Lara et al., 2017). The board of directors works collectively as a team. In Indonesia’s context with a male-dominated culture, an interesting question arises: Will the existence of women on the board be able to influence the decisions regarding earnings management? Therefore, the following hypothesis is proposed:

H₁: The presence of women on a company’s board of directors negatively affects earnings management.

4.2. Education level of women directors on earnings management

Upper echelon theory suggests that the characteristics of the board of directors influence the organization’s decisions and actions. The theory posits that the board of directors is composed of individuals with diverse education, backgrounds, experiences and perspectives and that this diversity leads to a more innovative and effective organization. One key tenet of upper-echelon theory is that the board of directors is more likely to make decisions consistent with their own values and beliefs. This is because they have a significant amount of power and influence within the organization, and they are able to shape the organization’s culture and strategy (Adams & Ferreira, 2009; Klein, 2002). Therefore, the competency of women directors is influenced by their education level and skill.

A director is expected to be a leader who can develop relationships among companies, maintain a competitive advantage, improve performance, and represent the company’s value system (Ali et al., 2014; Fan et al., 2019; Hillman et al., 2009). Women are considered capable of being innovative in leading a company (Prabowo & Setiawan, 2021), and with the proper education and skill set, they can attain top management positions in a company (Herrmann & Datta, 2005; Nekhili & Gatfaoui, 2013; Papadakis & Barwise, 2002). The higher the level of education, the more likely it will influence a person to make better decisions. Xiong (2016) reported that the higher the education of the board of directors, the lower the level of corporate earnings management. This aligns with Cheng et al. (2010)’s finding that directors with a higher education could report better financial statements. In Indonesia’s context, although there is gender equality in education, women who sit in higher-level positions are still rare. This was confirmed by the statement of Agus Saptarina, director of the Indonesian Financial Services Authority, who stated that the contribution of women in the board of commissioners and directors was still below 50%. Thus, it is interesting to test whether the higher level of education of women on the board of directors will strengthen the impact of their existence on the board of directors to reduce earnings management. Therefore, the following hypotheses are proposed:

H₂: The presence of women with higher education levels on a company’s board of directors negatively affects earnings management.

4.3. Educational background of women directors in earnings management

According to upper echelon theory, one important characteristic of the board of directors is their expertise (Adams & Ferreira, 2009; Klein, 2002). The expertise of directors affects the monitoring of accounting policies and disclosures in financial statements; therefore, when the directors are experts, the quality of accounting in the financial statements is guaranteed (Jiang et al., 2013; Musa et al., 2023). The expertise possessed by directors can be seen in their educational backgrounds. Nekhili and Gatfaoui (2013) proposed that women directors need an educational background to reach top management positions in a company. Bédard and Gendron (2010) state that financial expertise is required to make financial reports. The experts of directors with educational backgrounds in economics, finance, and accounting are certainly more familiar with and understand earnings management. Therefore, they can understand how to manage earnings (Xie et al., 2003). However, a director with an accounting background can improve earnings management, reducing the quality of financial reports (Hu et al., 2017; Ngo & Nguyen, 2022). Ngo and Nguyen (2022) stated that a director’s financial and accounting expertise could encourage someone to manage earnings.

Meanwhile, Hu et al. (2017) mentioned that directors with accounting backgrounds act less conservatively by improving earnings management practices. It was also noted by Ason et al. (2021) and Zalata et al. (2022) that directors with an accounting background tend to be involved in earnings management practices because of their expertise.

Although previous literature proved that the educational background of board of director in economics, finance, and accounting tends to be involved in earnings management, their motivation to do so probably changes with their level of education, as Xiong (2016) reported that the higher the education of the board of directors, the lower the level of corporate earnings management. Indonesia’s context is male-dominated. Thus, it will be interesting to test whether the presence of women on the board of directors, with higher education in economics, finance, and accounting, is effectively able to reduce earnings management. Therefore, the following hypothesis is proposed:

H₃: The presence of women with higher educational levels and backgrounds in economics, finance, and accounting on a company’s board affects earnings management.

4.4. Tenure of women directors in earnings management

As discussed in upper echelon theory, the characteristics of women directors are related to their work experience. Similar to the educational history of the directors previously discussed, companies need a director to build relationships between companies; they must also build a competitive level, company performance, and a representative form of company value (Ali et al., 2014; Fan et al., 2019; Hillman et al., 2009). Therefore, companies must have competent directors. Women are considered capable of innovating in leading companies (Prabowo & Setiawan, 2021). However, to realize the company’s competitive value, the company tries to choose competent directors who contribute well (Gull et al., 2018). Xie et al. (2003) and Jiang et al. (2013) mention that directors with experience in economics, finance, and accounting understand earnings management to manage profits well. Experienced directors will certainly increase supervision over the running of the company. Deng et al. (2018) note that tenure negatively affects earnings management. This agrees with previous research, which states that directors with a much longer term of office report earnings less aggressively than those who have just started (Zhang, 2009; Al-Absy, 2022).

In a male-dominated culture like Indonesia, with the low contribution of women serving on the board of directors, as stated by the director of the Indonesian Financial Services Authority, the longer of women serving on the board of directors must significantly influence a company. The longer she serves on the board of directors, the higher her expertise in managing the organization and monitoring tasks. Thus, it is expected that women’s tenure on the board of directors can mitigate earnings management to increase earnings quality. Therefore, the following hypothesis is proposed:

H₄: The tenure of women in board directors negatively affects earnings management.

5. Research design

5.1. Sampling frame and sample selection

The sampling frame of this study is Indonesia-listed banks from 2017 to 2021, where we chose only conventional banks. We exclude Islamic banks as they have a different business model. Theoretically, Islamic finance differs significantly from conventional finance (Beck et al., 2013). Islamic banks offer Shariah-compliant products that can be separated from conventional banks in terms of complexity, agency cost, level of maturity, and development (Beck et al., 2013). Furthermore, Islamic banks forbid bank interest since it is not Shariah-compliant. In addition, Jubilee et al. (2021) also find a difference in productivity between Islamic and conventional banks. Therefore, our final sample consists of 210 observations from 42 banking companies. The sample selection process is presented in Table 1, and the list of the banks is presented in Appendix.

Table 1. Sample selection process.

Criteria	Total
Banks listed in IDX 2017–2021	46
Syariah banks listed in IDX 2017–2021	(4)
Total banks used as sample	42
Total firm-years sample observations	210

5.2. Variable measurement

5.2.1. Earnings management

Earnings management results from implementing accrual-based accounting in financial reporting (Setiawati & Nai’m, 2000). Scott (2009) argues that earnings management is natural because financial accounting standards allow management to choose accounting policies to maximize utility or firm value. The accounting of loan loss provisions significantly impacts banks’ financial performance and balance sheet figures, necessitating considerable estimation and judgment (Purwa & Setiawan, 2021). Loan loss provisions are a variable consisting of borrowing costs and the initial balance of the allowance for doubtful accounts (Chang et al., 2011). Anandarajan et al. (2007) and Collins et al. (1995) stated that loan loss provisions manage banking companies’ profits. To calculate earnings management in banking, we follow Wahlen (1994) by determining the residual value using the following equation: (1) $${\text{LLP}}_{\text{it}}=\mathbf{a}+{\mathbf{b}}_1{\text{BEGLLA}}_{\text{it}}+{\mathbf{b}}_2{\text{LCO}}_{\text{it}}+{\mathbf{b}}_3{\text{CHLOANS}}_{\text{it}}+{\mathbf{b}}_4{\text{LOANS}}_{\text{it}}+{\mathbf{b}}_5\text{NPLit}+\text{e}$$(1)

Note: α= Constant; LLP_it = Loan loss provisions divided by total initial assets for the company i in year t; BEGLLA_it = Initial Earning Asset Loss divided by total initial assets for the company i in year t; LCO _it = Loan Charge-Offs divided by total initial assets for the company i in year t; CHLOANS _it = Change in Total Loans/Credit divided by total initial assets for the company i in year t; LOANS _it = Total Loans/Credit Outstanding divided by total initial assets for the company i in year t; NPL _it = Non-Performing Loans divided by total initial assets for the company i in year t; and e = errors.

The residual value of Eq. (1)(1) $${\text{LLP}}_{\text{it}}=\mathbf{a}+{\mathbf{b}}_1{\text{BEGLLA}}_{\text{it}}+{\mathbf{b}}_2{\text{LCO}}_{\text{it}}+{\mathbf{b}}_3{\text{CHLOANS}}_{\text{it}}+{\mathbf{b}}_4{\text{LOANS}}_{\text{it}}+{\mathbf{b}}_5\text{NPLit}+\text{e}$$(1) is a discretionary LLP component, or it can be considered an abnormal LLP component (Purwa & Setiawan, 2021). Since we focus on the magnitude of earnings management, we use the absolute number of LLP abnormal components as the dependent variable in the Eq. (2)(2) $$\begin{array}{c}{\text{ALLP}}_{\text{it}}=\mathbf{a}+{\mathbf{b}}_1\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_2\text{EDU}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_3\text{BACKEDU}\_{\text{WOMEN}}_{\text{it}}\\ \hspace{1em}+{\mathbf{b}}_4\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}*\text{EDU}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_4\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}*\\ \hspace{1em}\text{BACKEDU}\_\text{WOMEN}+{\mathbf{b}}_4\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}*\text{EDU}\_{\text{WOMEN}}_{\text{it}}*\\ \hspace{1em}\text{BACKEDU}\_\text{WOMEN}+{\mathbf{b}}_4\text{TENURE}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_5{\text{ROA}}_{\text{it}}+{\mathbf{b}}_6{\text{GROWTH}}_{\text{it}}+\\ \hspace{1em}{\mathbf{b}}_7{\text{AC}}_{\text{it}}+\text{e}\end{array}$$(2) to test the hypothesis in this study.

5.2.2. Women directors

This study discusses some of the characteristics inherent in women directors. The presence of women on the board of directors is measured using the proportion of women (Fan et al., 2019). The ratio of women in the question is the number of women directors divided by the total number of directors in the company (Sila et al., 2016).

5.2.3. Education level of women directors

This study uses the average level of education with a dummy variable, namely code 0 for high school education level, code 1 for diploma education level, code 2 for undergraduate education level, code 3 for master’s education level, and code 4 for doctoral’s education level to measure the educational level of women directors (Fan et al., 2019). Following Kontesa et al. (2020), we standardized educational levels using a ranking methodology based on percentile groups. For instance, the biggest 20th percentile group will score 5, and the smallest 20th percentile group will score 1.

5.2.4. Educational background of women directors

The educational background of women directors is measured using a dummy variable, namely code 1 for women directors who have an academic background in accounting, economics, and finance, and code 0 otherwise (Ason et al., 2021; Hu et al., 2017; Ngo & Nguyen, 2022).

5.2.5. Tenure of women directors

Another characteristic of women directors who participated in this study is their tenure as women directors. Gull et al. (2018) stated that the director’s term of office should be considered to measure the director’s experience. To measure the experience of women directors, the average length of time the directors served in the banking companies was measured, as per the study by Bouaziz et al. (2020) and Fan et al. (2019).

5.2.6. Control variables

This study examines the effect of women directors’ characteristics on earnings management and analyzes several other factors that influence earnings management. Other variables influencing earnings management include audit committees, return on assets (ROA), and asset growth. Romadani and Aryani (2021) measure the audit committee by adding up all members of the audit committee in the company. ROA is the result of dividing net income by total assets (Purwa & Setiawan, 2021). Meanwhile, asset growth is calculated by dividing the total assets at the end of year t by the total assets at the end of year t − 1 (Purwa & Setiawan, 2021). In sum, the operationalization of each variable is presented in Table 2.

Table 2. Variables operationalization.

Variables	Operationalization
Dependent variable
ALLP	Absolute (Loan loss provisions divided by total initial assets)
Main independent variables
PERCENT_WOMEN	Women directors divided by the total number of directors
TENURE_WOMEN	The average length of time the directors served
Moderating Variables
EDU_WOMEN_Std	A dummy variable, namely code 0 for high school education level, code 1 for diploma education level, code 2 for undergraduate education level, code 3 for master’s education level, and code 4 for doctoral’s education level to measure the educational level of women directors (Fan et al., 2019). Following Kontesa et al. (2020), we standardized educational levels using a ranking methodology based on percentile groups. For instance, the biggest 20th percentile group will score 5, and the smallest 20th percentile group will score 1.
BACK_WOMEN	Dummy variable, namely code 1 for women directors who have an academic background in accounting, economics, and finance, and code 0 otherwise
Control Variables
ROA	Net income divided by total assets
GROWTH	The total assets at the end of year t divided by the total assets at the end of year t-1
AC	Total all members of the audit committee in the company
Variable used to do additional test
DISEXP	Total discretionary expenditures
Variable used to control for endogenity
WOMEN_D	A dummy variable, equal to one if firms have at least one female directors, and zero otherwise.

5.3. Regression method

In this section, we examine the effect of women directors’ characteristics on earnings management. The earnings management equation is estimated using Ordinary Least Square (OLS) and Generalized Least Square (GLS) Random Effect. Following (Arifin et al., 2022), we employ GLS-RE because our main independent variable, women directors’ characteristics, is relatively time-invariant. In addition to the main variables, we include several control variables that have been identified as significant factors of earnings management in prior studies. There are audit committees, return on assets (ROA), and asset growth (Purwa & Setiawan, 2021; Romadani & Aryani, 2021).

Therefore, our empirical model to test the hypotheses takes the following form: (2) $$\begin{array}{c}{\text{ALLP}}_{\text{it}}=\mathbf{a}+{\mathbf{b}}_1\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_2\text{EDU}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_3\text{BACKEDU}\_{\text{WOMEN}}_{\text{it}}\\ \hspace{1em}+{\mathbf{b}}_4\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}*\text{EDU}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_4\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}*\\ \hspace{1em}\text{BACKEDU}\_\text{WOMEN}+{\mathbf{b}}_4\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}*\text{EDU}\_{\text{WOMEN}}_{\text{it}}*\\ \hspace{1em}\text{BACKEDU}\_\text{WOMEN}+{\mathbf{b}}_4\text{TENURE}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_5{\text{ROA}}_{\text{it}}+{\mathbf{b}}_6{\text{GROWTH}}_{\text{it}}+\\ \hspace{1em}{\mathbf{b}}_7{\text{AC}}_{\text{it}}+\text{e}\end{array}$$(2)

Note: α Constants; ALLP_it = Abnormal Loan Loss Provisions divided by total initial assets for the company i in year t; PERCENT_WOMAN _it = proportion of women directors for the company i in year t; EDU_WOMEN _it = Education Level of Women Directors for the company i in year t; BACKEDU_WOMEN _it = Women Board of Directors background for the company i in year t; TENURE_WOMEN _it = Women Board of Director’s term of office for the company i in year t; ROA _it = Return on Assets for the company i in year t; GROWTH _it = asset growth for the company i in year t; AC _it = Audit Committee for the company i in year t; and e = errors.

6. Empirical findings and discussion

6.1. Empirical findings

Before testing the hypotheses, Table 3 provides the descriptive statistics of all variables.

Table 3. Descriptive statistics.

	Mean	SD	Min	Max
ALLP	0.043	0.120	0.001	1.613
PERCENT WOMEN	0.185	0.190	0	0.750
EDU std	3	1.418	1	5
BACKEDU WOMEN	0.461	0.461	0	1
TENURE WOMEN	6.316	7.670	0	38
ROA	0.004	0.024	−0.181	0.050
GROWTH	0.214	0.834	−0.398	10.863
AC	3.781	1.107	2	8

Note: ALLP = Abnormal Allowance for Earning Assets (Loan Loss Provisions) divided by total initial assets; PERCENT_WOMAN = Proportion of Women Directors; EDU_WOMEN = Education Level of Women Directors; BACKEDU_WOMEN = Women Directors Background; TENURE_WOMEN = Term of Office of Women Directors; ROA = Return On Assets; GROWTH = Asset Growth; AC = Audit Committee.

Table 3 shows that the average presence of women directors in the banking industry is 19% which is slightly higher than the average of female directors of 14.8% reported by Gavious et al. (2012). The average education level of women directors is undergraduate. Meanwhile, 49.4% of the women directors have an educational background in economics, business, accounting, and finance. The table also shows that the average tenure of women directors is 6.32 years, with a maximum term of office for women directors of 23 years. The mean value of ALLP is 0.043, which is slightly higher than the average ALLP of 0.035 reported by Mathuva and Nyangu (2022).

Table 4 shows the correlation matrix. Earnings management (ALLP) is negatively correlated with all gender variables. We also observe that audit committee (AC) and bank performance (ROA) are negatively correlated with earnings management (ALLP).

Table 4. Correlation test results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
(1) ALLP	1.000
(2) PERCENT_WOMEN	−0.067	1.000
(3) EDU_std	−0.171	0.473	1.000
(4) BACKEDU_WOMEN	−0.042	0.373	0.798	1.000
(5) TENURE_WOMEN	−0.142	0.596	0.456	0.487	1.000
(6) ROA	−0.367	0.152	0.194	0.069	0.101	1.000
(7) GROWTH	0.060	−0.064	−0.136	−0.045	−0.015	−0.008	1.000
(8) AC	−0.065	−0.120	0.116	0.080	−0.081	0.060	−0.094	1.000

Note: ALLP = Abnormal Allowance for Earning Assets (Loan Loss Provisions) divided by total initial assets; PERCENT_WOMAN = Proportion of Women Directors; EDU_WOMEN = Education Level of Women Directors; BACKEDU_WOMEN = Women Directors Background; TENURE_WOMEN = Term of Office of Women Directors; ROA = Return On Assets; GROWTH = Asset Growth; AC = Audit Committee.

Based on the results of the correlation test, no correlation coefficient exceeds 0.8. Table 5 shows that each independent variable’s variance inflation factor (VIF) is below the threshold, indicating that multicollinearity is less likely to occur.

Table 5. Multicollinearity test results.

Variable	VIF	1/VIF
EDU_std	3.32	0.301215
BACKEDU_WO ∼ N	3.06	0.326899
TENURE_WOMEN	1.79	0.559726
PERCENT_WO ∼ N	1.76	0.566986
ROA	1.07	0.933331
AC	1.07	0.934983
GROWTH	1.04	0.962463

We begin our analysis by estimating our baseline model in Eq. (2)(2) $$\begin{array}{c}{\text{ALLP}}_{\text{it}}=\mathbf{a}+{\mathbf{b}}_1\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_2\text{EDU}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_3\text{BACKEDU}\_{\text{WOMEN}}_{\text{it}}\\ \hspace{1em}+{\mathbf{b}}_4\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}*\text{EDU}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_4\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}*\\ \hspace{1em}\text{BACKEDU}\_\text{WOMEN}+{\mathbf{b}}_4\text{PERCENT}\_{\text{WOMEN}}_{\text{it}}*\text{EDU}\_{\text{WOMEN}}_{\text{it}}*\\ \hspace{1em}\text{BACKEDU}\_\text{WOMEN}+{\mathbf{b}}_4\text{TENURE}\_{\text{WOMEN}}_{\text{it}}+{\mathbf{b}}_5{\text{ROA}}_{\text{it}}+{\mathbf{b}}_6{\text{GROWTH}}_{\text{it}}+\\ \hspace{1em}{\mathbf{b}}_7{\text{AC}}_{\text{it}}+\text{e}\end{array}$$(2) . First, we estimate the regression using ordinary least squares (OLS) as our main analysis. Then, since our board characteristics are relatively time-invariant, we use a generalized least square (GLS) random effect for the robustness test (Arifin et al., 2022). Due to the limited variability of the board characteristics variable over time, the standard fixed-effect panel estimation is not appropriate for our research. Fixed-effect estimation necessitates within-firm variation in the independent variable to yield effective and reliable estimates. Independent variables that remain relatively constant over time, such as gender in directors, cannot be effectively estimated using fixed-effect estimation because they become absorbed in the time-demeaning process of variables in fixed-effect (Wooldridge, 2002). Some studies also opt for GLS-RE when dealing with independent variables that exhibit little variation over time, such as board characteristics (Jackowicz et al., 2014) and ownership structure (Pattnaik et al., 2013).

As shown in Model (1) Table 6, the OLS results reveal a negative and significant relationship between the proportion of women directors in earnings management with a significance level of 0.05; therefore, H₁ is accepted. Thus, the presence of women directors can suppress earnings management practices in a bank in Indonesia. The result remains when we use the GLS-random effect, as shown in Model (2). However, we find the coefficient of interaction variables between PERCENT_WOMEN and EDU_std in Model (1) is positive and significant at the 10% level. This contradictory result against H₂ indicates that the higher level of women’s education tends to weaken the negative effect of women directors’ presence on earnings management. Furthermore, the interaction became insignificant when we used the GLS-random effect, as shown in Model (2). Interestingly, both in Model (1) and (2), the estimated coefficients of the three-interaction-variable (PERCENT_WOMEN*EDU_std*BACKEDU_WOMEN) are significantly negative at the 1% level. One interpretation of this result is that directors with higher educational levels and economics, finance, and accounting backgrounds are more broadly enlightened about the consequences of earnings management. These results are consistent with the H₃. Meanwhile, the coefficient of TENURE_WOMEN is significantly negative, indicating that the longer the tenure of women directors, the less likely they are to manage earnings. Women directors with longer tenure gain more experience on the monitoring activity of earnings management. Women directors with longer tenure positively affect earnings quality. Therefore, our H₄ is supported in this study.

Table 6. Regression results.

	(1)	(2)
	ALLP	ALLP
PERCENT_WOMEN	−0.653**	−0.549*
	(−2.379)	(−1.843)
EDU_std	0.012	0.016
	(0.649)	(0.788)
PERCENT_WOMEN*EDU_std	0.158*	0.136
	(1.883)	(1.459)
BACKEDU_WOMEN	−0.089	−0.090
	(−0.676)	(−0.651)
PERCENT_WOMEN*BACKEDU_WOMEN	2.162***	2.194***
	(3.093)	(3.103)
EDU_std*BACKEDU_WOMEN	0.016	0.017
	(0.468)	(0.449)
PERCENT_WOMEN*EDU_std*BACKEDU_WOMEN	−0.513***	−0.519***
	(−2.840)	(−2.786)
TENURE_WOMEN	−0.002*	−0.004***
	(-1.793)	(−2.767)
ROA	−1.627***	−1.578***
	(-5.014)	(-4.69)
GROWTH	0.0057	0.004
	(0.630)	(0.505)
AC	−0.005	−0.004
	(−0.639)	(−0.546)
Constant	0.054	0.046
	(1.344)	(0.977)
Observations	210	210
Overall R^2	–	0.2371
Adj R^2	0.2036	–

Note: *, **, and *** are significant at 10%, 5%, and 1% levels, respectively.

Concerning the control variables, the only significant relationship with earnings management is the ROA. ROA is negatively and significantly related to earnings management, while audit committee and asset growth are not significantly related. Accounting practices improve earnings management because it can do so.

6.2. Discussion

As previously mentioned, the presence of women on the board of directors is negatively and significantly related to earnings management, with a significance level of 0.05; thus, H₁ is accepted. This finding supported the agency theory that strong corporate governance is essential to reduce the potential conflict between management and shareholders (Jensen & Meckling, 1976). Furthermore, this finding also supports the upper echelon theory that suggests women’s characteristics will have greater ethical decisions (Klein, 2002). In this study, we provide evidence that the presence of women on the board of directors effectively reduces earnings management and increases earnings quality. Previous studies, have stated that women in the role of directors in a company could reduce earnings management (Bouaziz et al., 2020; Fan et al., 2019; Gavious et al., 2012; Githaiga et al., 2022; Gull et al., 2018; Harakeh et al., 2019; Kim et al., 2017; Lara et al., 2017; Orazalin, 2020; Sana, 2018; Zalata, Ntim, Choudhry, et al., 2019). Women can supervise a company’s financial statements (Lara et al., 2017). Women are also considered to be very independent, tend to avoid risks, and may even be intolerant of opportunistic attitudes (Adams & Ferreira, 2009; Gibbs, 2015; Sahay et al., 2015).

The findings in this study with the population of Indonesian banking companies align with previous studies. They show that the presence of women on the board of directors is negatively and significantly related to earnings management. Although Indonesia has a masculine gender culture (Hofstede, 2001), this research shows that women can suppress earnings management in an environment dominated by men. Therefore, in this case, the presence of women can be used to reduce earnings management practices in a masculine environment, which tends to encourage directors to act non-conservatively, as stated in the study of Purwa and Setiawan (2021).

However, based on the upper echelon theory, we found that the presence of women with higher education levels on the board of directors is positively and significantly related to earnings management; therefore, H₂ is rejected. The finding can be interpreted that the negative impact of women on the board of directors on earnings management, as proven in this study (H₁), is decreased when the women have higher education. Our finding aligns with Qi et al. (2018) study that shows managers with higher education tend to engage in more earnings management than those with lower education. It appears that highly educated executives have more familiarity and understanding of how to conduct earnings management through accrual.

Furthermore, we find interesting results when the women’s education level interacts with the background of education in economics, finance, and accounting. We found that women directors with a higher level of education and economic, finance, and accounting backgrounds are more likely to have less incentive to engage in earnings management. Therefore, our argument in H₃ is supported. This finding contradicts the research conducted by Ason et al. (2021), Hu et al. (2017) and Ngo and Nguyen (2022), who argued that women directors’ financial and accounting expertise could improve earnings management, thereby reducing the quality of financial reporting. However, when combined with higher education, they become alert about the consequences of earnings management. As a result, they tend to act more carefully in earnings management engagement. This finding suggests that gender diversity policies should promote the appointment of female directors with relevant skills and experience to do the monitoring roles within the boardroom effectively.

Based on the test results that have been carried out, it can be seen that the tenure of women directors has a negative effect on earnings management. This result supports H₄ and is consistent with the findings of Bouaziz et al. (2020), Deng et al. (2018) and Elzahar et al. (2022), showing that the tenure of directors is negatively and significantly related to earnings management. This is because in the initial year, the directors are in office, they tend to be more conservative in managing profits. However, this is inversely proportional if the directors have served for an extended period (Hu et al., 2015).

7. Additional analysis

7.1. Sub-sample low vs high earnings management

In addition, we perform sub-sample analysis to address the concern over the conditional behavior of women directors. This approach allows us to extract sweeping conclusions based on the dependent variable’s conditional levels. We divide our sample into ALLP-LOW and ALLP-HIGH based on the median level and re-estimate our baseline model. The estimation in Table 7 provides evidence that the effects of women directors on earnings management differ across the level of ALLP. In particular, the effect of educational level and background is positive at the low level of ALLP but negative at the high level of ALLP. Therefore, the findings indicate that the governance attribute of women directors is conditional to the level of earnings management.

Table 7. Regression results.

	(1)	(2)
	ALLP-LOW	ALLP-HIGH
PERCENT_WOMEN	0.012	−0.079
	(0.917)	(−0.111)
EDU_std	0.002**	0.006
	(2.029)	(0.188)
PERCENT_WOMEN*EDU_std	−0.007*	−0.090
	(−1.693)	(−0.385)
BACKEDU_WOMEN	0.021***	−1.342***
	(2.97)	(−4.366)
PERCENT_WOMEN*BACKEDU_WOMEN	−0.086**	7.541***
	(−2.493)	(5.259)
EDU_std*BACKEDU_WOMEN	−0.006***	0.292***
	(−3.279)	(3.786)
PERCENT_WOMEN*EDU_std*BACKEDU_WOMEN	0.024***	−1.616***
	(2.674)	(−4.422)
TENURE_WOMEN	0.001	0.002
	(1.373)	(0.907)
ROA	0.021	−0.870*
	(0.571)	(−1.894)
GROWTH	0.001	0.035
	(0.629)	(1.306)
AC	0.001	−0.012
	(1.630)	(−1.044)
Constant	0.001	0.113*
	(0.134)	(1.697)
Observations	105	105
Overall R^2	0.265	0.443
Adj R^2	0.178	0.377

Note: *, **, and *** are significant at 10%, 5%, and 1% levels, respectively.

7.2. Real earnings management

The baseline model employs accrual-based earnings management as a dependent variable. Despite the lack of prior literature regarding real earnings management in the banking industry, we tried to examine the effect of gender on real earnings management by re-estimating our baseline using real earnings management proxy as a new dependent variable. Due to the specific characteristics of the financial industry, we only employ a proxy suitable for our research: discretionary expenditures (Roychowdhury, 2006). We use the following model to estimate our real earnings management: (3) $${\text{DISEXP}}_t/{\text{TA}}_{t-1}=\hspace{0.17em}{\alpha }_0+{\alpha }_1\left(1/{\text{TA}}_{t-1}\right)+{\beta }_1\left(R_{t-1}+{\text{TA}}_{t-1}\right)\hspace{0.17em}+\epsilon$$(3)

Where DISEXP is total discretionary expenditures measured as the sum of SG&A and advertising expenditures; TA is total assets; and R is Revenue. We show the results in Table 8.

Table 8. Real earnings management: discretionary expenses.

	AREM
PERCENT_WOMEN	0.402
	(0.128)
EDU_std	0.481**
	(2.344)
PERCENT_WOMEN*EDU_std	−1.019
	(−1.060)
BACKEDU_WOMEN	−0.458
	(−0.297)
PERCENT_WOMEN*BACKEDU_WOMEN	1.936
	(0.237)
EDU_std*BACKEDU_WOMEN	−0.286
	(−0.703)
PERCENT_WOMEN*EDU_std*BACKEDU_WOMEN	0.400
	(0.190)
TENURE_WOMEN	−0.006
	(−0.397)
ROA	−3.858
	(−1.044)
GROWTH	−0.060
	(−0.581)
AC	−0.214**
	(−2.424)
Constant	0.780
	(1.636)
Observations	201
Adj R^2	0.045

Note: *, **, and *** are significant at 10%, 5%, and 1% levels, respectively.

Unlike the main result of accrual-earnings management, we find that the presence of women on the board of directors does not significantly influence real-earnings management in the banking industry. Their level of education, background, and tenure are also not able to impact real-earnings management in the banking industry. Our failure to find evidence regarding the impact of women on the board of directors is perhaps due to the lack of measurement of real-earnings management in the banking industry.

8. Robustness test

8.1. Effect of COVID-19 period

As for the robustness test, we exclude two years of observations with the year-end in 2020 and 2021 and re-estimate our baseline model. We posit that the Covid-19 pandemic is likely to influence the financial constraints of banks. We show the result in Table 9. Consistent with expectations, the results are qualitatively similar to our main findings. Based on the results of the robustness test regression, it can be seen that the results were similar to the period before the COVID-19 pandemic.

Table 9. Sub-sample analysis: Covid-19 period excluded.

	Non Covid-19
	ALLP-LOW
PERCENT_WOMEN	−.688*
	(−1.948)
EDU_std	0.036
	(1.475)
PERCENT_WOMEN*EDU_std	0.150
	(1.478)
BACKEDU_WOMEN	0.039
	(0.230)
PERCENT_WOMEN*BACKEDU_WOMEN	2.175**
	(2.501)
EDU_std*BACKEDU_WOMEN	−0.019
	(−0.436)
PERCENT_WOMEN*EDU_std*BACKEDU_WOMEN	−0.519**
	(−2.399)
TENURE_WOMEN	−0.002
	(−1.011)
ROA	−3.587***
	(−6.222)
GROWTH	0.248***
	(5.280)
AC	0.001
	(0.010)
Constant	−0.032
	(−0.547)
Observations	126
Adj R^2	0.479

Note: *, **, and *** are significant at 10%, 5%, and 1% levels, respectively.

8.2. Endogenity test

To mitigate any potential endogeneity concerns that may arise from sample selection bias, we follow Zalata, Ntim, Choudhry, et al. (2019) by using propensity score matching (PSM). We first calculate the propensity score from probit estimation and perform Gaussian Kernel to match our sample. In the matched sample (un-tabulated), we can conclude that both control and treatment groups have similar characteristics. Therefore, using bootstrap with 100 replications, we find that the coefficient of WOMEN_D are negative and statistically significant 5% level as presented in Table 10. The findings suggest that firms with female directors on average exhibit lower earnings management behaviour. Thus, this result supports our hypothesis that the presence of women on a company’s board of directors negatively affects earnings management.

Table 10. Propensity score matching.

Gaussian Kernel: Bootstrapped 100 replications
	Coeff	z-stat
WOMEN_D	−0.558**	−2.54
Control variables	Yes
Wald chi2(1)	6.46
Prob > chi^2	0.011
Replications	100
Adj. R^2	0.043

Note: *, **, and *** are significant at 10%, 5%, and 1% levels, respectively.

9. Conclusion

We present evidence that the presence of women on the board of directors within the banking industry notably influences earnings management. This suggests that even within traditionally male-dominated cultures like Indonesia, the presence of women can serve as a deterrent to earnings manipulation. Moreover, fostering gender diversity can motivate women to pursue leadership roles, thereby enriching the banking industry. Interestingly, the study indicates that a higher educational background mitigates the potential adverse effects of female board representation on earnings management. Specifically, women directors with advanced education in fields such as economics, finance, and accounting exhibit a reduced propensity for engaging in earnings management activities.

This study breaks new ground in several ways, particularly when it comes to understanding the impact of women directors on corporate governance and earnings quality, especially in male-dominated cultures like Indonesia. First, we identify how the educational level and background of women directors can influence their effectiveness in board oversight. Second, through in-depth analysis of specific groups within our data, we reveal how the relationship between women directors and earnings management varies depending on certain conditions. Third, by focusing on the often-ignored context of banking companies, we add valuable insights to the broader conversation about earnings manipulation. Finally, our findings shed light on the unique challenges and opportunities faced by women directors in male-dominated environments.

Our results carry several important implications. First, they indicate that having women on the board of directors effectively diminishes earnings management, enhancing earnings quality. This improvement in quality is pivotal as it provides relevant information for making financial decisions that influence the broader economy. Second, our results also have important implications for gender diversity roles and policy reforms for several reasons: (1) The findings suggest that the boards should consider appointing more women directors to increase the monitoring roles; (2) the findings suggest that corporate governance reforms should focus on enhancing the monitoring role of women directors; (3) the results suggest that where normative regulation is lacking, such as in Indonesia, the government should develop gender diversity policies for companies. Finally, the findings carry managerial implications, suggesting that banks would benefit from incorporating women into their boardrooms to leverage their expertise and insights. Overall, the paper provides valuable insights into the role of women directors in corporate governance with masculine culture and the potential impact of gender diversity on earnings management and earnings quality.

While our study offers valuable insights, it’s essential to acknowledge its limitations. First, our research focuses solely on banking companies within Indonesia. Second, while we shed light on the impact of women on the board of directors concerning accrual-based earnings management, we did not delve into its effects on real-earnings management. This gap may stem from challenges in measuring real-earnings management variables within the banking sector. Recognizing these constraints, our study paves the way for future investigations. First, subsequent research could broaden its scope by examining countries with similar masculine cultural contexts as Indonesia. This would offer a comprehensive understanding of how women’s influence varies across diverse cultural settings in managing earnings. Second, there’s a need to develop robust metrics for assessing real-earnings management within the banking industry.

Authors’ contribution

The list of complete authors along with the contribution to the paper are as follows. Y. Anni Aryani, Ph.D, first and corresponding author (email: [email protected]): lead the research in all stages, from the first manuscript until the final version, therefore I contributed substantially to the conception or design of the work. Alma Evangelista Mahendrastiti (email: [email protected]): data collecting and drafting under Aryani’s supervision. Prof. Doddy Setiawan, Ph.D (email: [email protected]): adding literature review and discussion results and giving final approval. Taufiq Arifin, Ph.D. (email: [email protected]): running the test analysis and discussion results. Dr. Evi Gantyowati (email: [email protected]): adding literature review and discussion results. All the authors agreed to be accountable for all aspects of the work.

Disclosure statement

We have no conflicts of interest to disclose.

Data availability statement

Data derived from public domain resources. The data supporting this study’s findings are available in the bank’s annual financial report published on its website or derived from the Indonesia Stock Exchange at https://www.idx.co.id/id. The list of banks included as data samples in this study is in the appendix attached.

Additional information

Funding

Currently, the manuscript is under the support of LPPM UNS (Institution for Research and Community Service, Universitas Sebelas Maret), Mandatory Research Grant, No. 228/UN27.22/PT.01.03/2023.

Notes on contributors

Y. Anni Aryani

Y. Anni Aryani is Associate Professor at the Faculty of Economics and Business, Universitas Sebelas Maret. She obtained a PhD degree in accounting at Victoria University, Melbourne, Australia. Her research interests focus on financial accounting, management accounting, corporate governance, corporate disclosure, and quantitative research methods.

Alma Evangelista Mahendrastiti

Alma Evangelista Mahendrastiti is Associate Auditor at Grant Thornton Indonesia. She obtained a Bachelor’s degree in accounting at Universitas Sebelas Maret, Indonesia. Her research interests focus on financial accounting, corporate disclosure, and quantitative research methods.

Doddy Setiawan

Doddy Setiawan is Professor at the Faculty of Economics and Business, Universitas Sebelas Maret. He obtained a PhD degree in accounting at Universiti Saint Malaysia. His research interests focus on financial accounting, management accounting, corporate governance, corporate disclosure, and quantitative research methods.

Taufiq Arifin

Taufiq Arifin is Associate Professor at the Faculty of Economics and Business, Universitas Sebelas Maret. He obtained a PhD degree in accounting and finance at the University of Twente, Netherlands. His research interests focus on financial accounting, corporate governance, corporate finance, and quantitative research methods.

Evi Gantyowati

Evi Gantyowati is Associate Professor at the Faculty of Economics and Business, Universitas Sebelas Maret. She obtained a PhD degree in accounting at Universitas Diponegoro, Indonesia. Her research interests focus on financial accounting, management accounting, corporate governance, corporate disclosure, and quantitative research methods.

Appendix

Table

No	CODE	BANK NAME
1.	AGRO	Bank Rakyat Indonesia Agroniaga Tbk PT
2.	AGRS	Bank IBK Indonesia Tbk PT
3.	AMAR	Bank Amar Indonesia PT
4.	ARTO	Bank Jago Tbk PT
5.	BABP	Bank MNC Internasional Tbk PT
6.	BACA	Bank Capital Indonesia Tbk PT
7.	BBCA	Bank Central Asia Tbk PT
8.	BBHI	Bank Harda Internasional Tbk PT
9.	BBKP	Bank KB Bukopin Tbk PT
10.	BBMD	Bank Mestika Dharma Tbk PT
11.	BBNI	Bank Negara Indonesia (Persero) Tbk PT
12.	BBRI	Bank Rakyat Indonesia (Persero) Tbk PT
13.	BBSI	Bank Bisnis Internasional PT
14.	BBTN	Bank Tabungan Negara (Persero) Tbk PT
15.	BBYB	PT Bank Neo Commerce Tbk PT
16.	BCIC	Bank JTrust Indonesia Tbk PT
17.	BDMN	Bank Danamon Indonesia Tbk PT
18.	BEKS	Bank Pembangunan Daerah Banten Tbk PT
19.	BGTG	Bank Ganesha Tbk PT
20.	BINA	Bank Ina Perdana Tbk PT
21.	BJBR	Bank Pembangunan Daerah Jawa Barat dan Banten Tbk PT
22.	BJTM	Bank Pembangunan Daerah Jawa Timur Tbk PT
23.	BKSW	Bank QNB Indonesia Tbk PT
24.	BMAS	Bank Maspion Indonesia Tbk PT
25.	BMRI	Bank Mandiri (Persero) Tbk PT
26.	BNBA	Bank Bumi Arta Tbk PT
27.	BNGA	Bank CIMB Niaga Tbk PT
28.	BNII	Bank Maybank Indonesia Tbk PT
29.	BNLI	Bank Permata Tbk PT
30.	BSIM	Bank Sinarmas Tbk PT
31.	BSWD	Bank Of India Indonesia Tbk PT
32.	BTPN	Bank BTPN Tbk PT
33.	BVIC	Bank Victoria International Tbk PT
34.	DNAR	Bank Oke Indonesia Tbk PT
35.	INPC	Bank Artha Graha Internasional Tbk PT
36.	MAYA	Bank Mayapada Internasional Tbk PT
37.	MCOR	Bank China Construction Bank Indonesia Tbk PT
38.	MEGA	Bank Mega Tbk PT
39.	NISP	Bank OCBC NISP Tbk PT
40.	NOBU	Bank National NOBU Tbk PT
41.	PNBN	Bank Pan Indonesia Tbk PT
42.	SDRA	Bank Woori Saudara Indonesia 1906 Tbk PT