Do benevolent directors influence corporate earnings persistence?

Corporate financial reporting quality is a primary concern for shareholders and capital markets worldwide, given its crucial role in assessing firms’ financial stability. Earnings play a pivotal role in firm value; thus, managers have incentives to manipulate earnings to enhance firm value and stock prices (Badertscher, 2011; Yang and Abeysekera, 2019). Earnings management is a significant approach used by managers to achieve their economic objectives, and it is more severe than corporate fraud because shareholders are unlikely to discover intentionally manipulated earnings (Badertscher, 2011; Yang and Abeysekera, 2019). However, earnings persistence (EP) is considered a decision-useful indicator of earnings quality (EQ) ^[1] as investors require useful information to forecast future cash flows (Jia and Li, 2022; Simoni et al., 2022). Investors generally prefer higher EP, as it is more sustainable, permanent and less transitory (Li, 2019). Prior studies indicate that higher accrual quality (consistency between accruals and future cash flows) is positively associated with EP (Dechow et al., 2010). In contrast, real earnings management (through understating discretionary expenses) and book-tax conformity are negatively associated with persistence (Atwood et al., 2010; Li, 2019). These results resonate with the notion that persistent earnings indicate high-quality financial reporting and valuable information on future cash flows.

Due to the importance of EQ for the integrity of capital markets, it is useful to examine the potential corporate governance characteristics that continuously influence higher EQ. Prior research in the international context provides evidence of the link between various board characteristics and EQ (Abdelsalam et al., 2021; Alves, 2023; Assenso-Okofo et al., 2021). The link between board characteristics and financial reporting quality is less understood in the Australian context. This thin literature in the Australian context is relatively older, dating back to samples in the early 2000s. Of the few notable studies, Baxter and Cotter (2009) find a negative influence of audit committee existence and expertise on intentional earnings management, Davidson et al. (2005) reveal significant positive associations between the number of nonexecutive directors on board and audit committees and EQ, Mather and Ramsay (2006) document a positive influence of board size and independence on EQ during CEO turnovers, Kent et al. (2010) find the positive associations between audit committee size, independence, the board size, independence and accrual quality, in the forms of discretionary and innate accruals. The existing international and Australian literature commonly focuses on boards’ characteristics related to directors’ demographics, structure, qualifications and expertise, with less emphasis given to personal moral characteristics in driving financial reporting quality. We aim to fill this research gap by investigating whether the moral disposition of directors, as reflected in their personal life choices of serving not-for-profit (NFPs), influences a firm’s financial reporting environment.

This paper primarily examines whether board benevolence is associated with EQ, followed by the moderating roles of CEO benevolence and gender, as well as board and audit committee independence, on the association between board benevolence and EQ. Theories of upper-echelon and ethical leadership inform the research design. Director benevolence is proxied through their “involvement in NFP leadership positions while holding corporate directorships.” Board benevolence is measured based on the proportion of directors on boards with NFP positions. Benevolence data is hand-collected from corporate annual reports and verified through public information sources such as the Australian Charities and NFP Commission (ACNC) register. EQ has several indicators, including EP, accruals, smoothness, loss recognition, restatements and investor responsiveness (Al-Enzy et al., 2023; Dechow et al., 2010). Following Atwood et al. (2010), Li (2019) and Simoni et al. (2022), EQ is measured through EP, which evaluates the sustainability of past earnings to future earnings (Simoni et al., 2022). EP data is obtained from EIKON/Refinitiv. Using a sample of Australian capital market firms from 2010 to 2019, this study finds a positive association between board benevolence and EP. Benevolent CEOs, women CEOs and the presence of a higher proportion of independent board and audit committee members strengthen this association.

Prior literature suggests a connection between board and committee composition and CEO-level drivers of EQ in the Australian context; however, board composition is typically evaluated in terms of expertise, independence and gender diversity. To the best of our knowledge, no prior research has examined the influence of directors’ moral characteristics on financial reporting quality. We argue that director benevolence is a significant channel through which board composition can be viewed and potentially influence EQ. Thus, we encourage corporate governance researchers to reflect upon corporate leaders’ personal ethics and moral standing in understanding the influence of high-quality financial reporting practices. Furthermore, prior research in Australia has commonly viewed EQ through traditional, backwards-looking measures of EQ based on accrual models. In contrast, we consider EP as a forward-looking, decision-useful EQ indicator for capital market participants (Saona et al., 2024).

The Australian context is appropriate for the study for four main reasons. First, most EQ studies focus on the US or other international contexts; however, the Australian context differs from those of other countries in terms of firm characteristics, market competition and corporate governance practices (Bugeja et al., 2015). Second, Australia has a balance-sheet-oriented framework for financial reporting, higher shareholder protection mechanisms and less conformity between tax and financial accounting (Cheung et al., 2010; Yang, 2019). Third, unlike the USA, Australia’s regulatory environment is “principle-based,” including compliance with accounting standards (Cheung et al., 2010), allowing for higher managerial discretion in financial reporting. Finally, the extent of external monitoring, such as the focus of analysts, is less intensive in the context (Habib and Hossain, 2008; Yang, 2019). Thus, earning management incentives and practices are likely to differ in Australia compared to international contexts and require separate inquiries.

This research makes significant contributions to scholarly knowledge in multiple ways. First, the originality of the study lies in the significance of examining the moral characteristics of corporate leaders in evaluating EQ. To date, although extensive research has demonstrated the influence of board characteristics on EQ, most studies have primarily focused on two aspects of board composition. Prior research is typically focused on the impact of board technical and demographic characteristics on EQ. However, contemporary research indicates that the personal characteristics of leaders play an instrumental role in shaping corporate behavior (Jiang and Ye, 2024; Ikeda et al., 2021). Thus, it is insightful to ask ourselves how the moral values of corporate leaders influence their choice of reporting consistent and quality earnings information to the market. The findings of the study do not merely contribute to upper-echelon literature as one more study but rather signal the importance of viewing board composition through different useful perspectives. Thus, the most critical and marginal contribution of this study is to show that board benevolence plays a significant role in the EP of firms, extending our prior understanding of the board-level determinants of financial reporting quality (Kent et al., 2010; Z. Li et al., 2023; Strydom et al., 2017). Prior literature has examined the concept of directors serving on NFPs, primarily focusing on charity directors; however, our study is different from these in two ways. First, prior studies consider the role of directors with charity experience from a resource capital perspective (Gertsberg et al., 2024; Chen et al., 2022), unlike ours, which focuses on the moral indication of such personal choices. Second, research focuses on charity directors, predominantly examining their influence on CSR activities (Weerasinghe et al., 2024a, 2024b; Chen et al., 2022) or the strategic use of these directors as an ESG response (Gertsberg et al., 2024). We focus on the connection between the personal choice of serving on NFPs as an indication of moral disposition and financial reporting quality, extending prior knowledge. Third, the results contribute to the upper-echelon literature that investigates the link between managerial/board characteristics and corporate outcomes (Feng et al., 2023; Liu et al., 2023), particularly financial reporting quality (Francoeur et al., 2022; Z. Li et al., 2023; Zalata et al., 2022). Fourth, the findings align with the ethical leadership view (Brown et al., 2005; Ho et al., 2015), highlighting the impact of Benevolent directors on the moral aspects of firms, specifically transparency and honesty in financial reporting. Finally, the findings reaffirm the importance of corporate governance mechanisms in enhancing EQ through persistent earnings, as demonstrated in a recent Australian sample (Davidson et al., 2005; Kent et al., 2010, 2016).

There are three significant implications for practice. First, to the best of the authors’ knowledge, this study is the first to investigate the impact of directors’ moral characteristics on EP. In the quest to enhance the quality of financial information in capital markets, our findings suggest that policymakers should consider alternative perspectives on board composition when setting corporate governance recommendations. Second, the results are value-relevant for potential investors, lenders, auditors and other stakeholders in evaluating the financial information of firms. For instance, financial reporting is more reliable in firms with a higher proportion of Benevolent directors on their boards; thus, information users can evaluate the firm’s financials accordingly. Finally, the findings reaffirm the importance of corporate governance mechanisms in promoting transparent financial reporting, which is insightful for firms in structuring their internal governance.

Higher EQ is defined by Dechow et al. (2010, p. 344) as “earnings provide more information about the features of a firm’s financial performance that are relevant to a specific decision made by a specific decision-maker”. Dechow et al. (2010) identify three categories of proxies of EQ in the literature as properties of earnings (e.g. persistence, accruals, smoothness and loss recognition), investor responsiveness to earnings (e.g. earnings response coefficient) and external indicators of earnings misstatements (e.g. restatements). Earning persistence is examined under two broad research themes: the influence of various firm and governance characteristics on persistence, and whether persistence is decision-useful for equity valuation. A simple model specification to estimate EP is:

This suggests that if a firm exhibits higher EP, its current earnings are a more reliable indicator of future performance, and annuitizing current earnings is likely to result in a smaller valuation error compared to a firm with less persistent earnings (Dechow et al., 2010). However, Dechow et al. (2010) suggest that researchers should evaluate persistence as a measure of EQ in conjunction with the firm’s fundamental performance and accounting systems. This is because the persistence is likely driven by the industries within which the firms operate. For example, a firm’s cost strategy may depend on the industry, growth, competition and the proportion of fixed costs (Dechow et al., 2010; Nissim and Penman, 2001; Soliman, 2008). Thus, it is useful to consider these in evaluating EQ through EP. Combining these extensions, StarMine has developed an EQ measure of persistence by defining it as “the degree to which past earnings are reliable and are likely to persist” ^[2] (Atas, 2005, p. 3). Evolving research in EQ adopts the StarMine measure, acknowledging the reliability (Abdelsalam et al., 2021; Fassas et al., 2023; Saona et al., 2024). Thus, we adopt this measure for EP in this research.

The board of directors has a fiduciary duty to monitor management activities to ensure the integrity of financial reporting (Cohen et al., 2004; Ranasinghe et al., 2020). Existing research suggests a connection between board characteristics and EQ (Dissanayake et al., 2023). The relationships between board characteristics and EQ are nuanced, indicating mixed results. Board gender diversity has a negative influence on earnings management practices in UK firms (Arun et al., 2015). Similar evidence exists from the USA. Srinidhi et al. (2011) and Oyotode-Adebile et al. (2022) reveal that the EQ is higher in firms with more women on boards. Australian evidence indicates similar results. Australian evidence also supports the positive influence of women on boards on EQ (Strydom et al., 2017). In contrast, some evidence suggests that board gender diversity is not a significant factor in determining EQ or is even negatively related. A study from the US indicates that female directors with prior financial failures are more likely to engage in real earnings management (Bhuiyan et al., 2020). In the UK, female directors and independent female directors are positively related to earnings management in firms with low debt levels (Arun et al., 2015).

Other internal corporate governance mechanisms, such as board and audit committee independence, show negative associations with opportunistic earnings management practices (Davidson et al., 2005). Kent et al. (2010) decomposed accruals into innate and discretionary accruals using an Australian sample from 2000 to 2006 in examining the link with corporate governance. This study reveals that firms audited by the Big 4 and those with large audit committees are associated with high-quality discretionary accruals. In contrast, board and audit committee independence is related to innate accrual quality. Kent et al. (2016) report the link between internal governance choices and EQ, specifically finding that smaller firms are likely to have effective governance without implementing the best practices recommendations of Australian securities exchange (ASX). To a lesser extent, studies indicate opposite or neutral evidence concerning the influence of board and audit committee independence on EQ. For example, Chen et al. (2015) show that independent directors on boards and audit committees are only effective in reducing earnings management in firms with rich information environments. While the evidence is mixed, the consensus in prior literature is that board independence, audit committee independence and board gender diversity enhance internal governance and provide strong oversight of management decisions (Katmon and Farooque, 2017).

While board composition is typically viewed and examined through directors’ observable characteristics, such as independence and gender, it is also possible that their personal life ethics are reflected in the governance of firms. Building on this comprehension, the boards of directors’ personal life choice to serve on NFPs concurrently with corporate leadership is perceived as Benevolent in this study, following the approaches of Chapple et al. (2025) and Weerasinghe et al. (2024a, 2024b). From a normative perspective, benevolence refers to the concern for other’s good, well-being and development (Karakas and Sarigollu, 2012). Karakas and Sarigollu (2012)’s theory of benevolence highlights several aspects of benevolent leaders, including morality (focusing on leaders’ ethics and values), spiritual actions, vitality (how leaders create positive change) and community, which emphasizes their contribution to society. Authors suggest that benevolent leaders display affective commitment and organizational citizenship behavior (Karakas and Sarigollu, 2012). Corporate leaders who choose to serve on NFPs indicate a personal moral disposition to accept their social responsibility, which resonates with the Benevolent leadership theory, particularly in terms of morality, vitality and community paradigms.

We argue that the distinction between Benevolent directors and other directors has several reasons that positively influence the financial reporting environment. First, the simultaneous holding of leadership positions in the corporate and NFP sectors may indicate a personal virtue of directors, such as altruism and a sense of moral obligation (Bekkers and Wiepking, 2011; Feng et al., 2023). Second, the objectives of NFP (for-purpose) and corporate sectors (for-profit) are distinct; thus, serving on NFPs may bring different accountabilities and exposures to these directors (Ward and Miller-Stevens, 2021). Third, by serving on NFPs, directors are likely to develop broader perspectives than their counterparts (Chen et al., 2022; Feng et al., 2023; White et al., 2014). Finally, directors who serve on NFPs likely develop positive personal values by being exposed to leaders (social cohesion effect) who primarily serve in NFP sectors (Fredette and Sessler Bernstein, 2021). Benevolent leaders’ ethical stance, diverse perspectives and exposure to higher accountabilities suggest the possibility of moral leadership.

Upper-echelon and ethical leadership theories further inform the link between board benevolence and EP. The upper-echelon theory posits that corporate outcomes reflect the perspectives and characteristics of top echelons (Hambrick and Mason, 1984). Research confirms this view, finding associations between numerous managerial personal ethics and corporate outcomes (Chapple et al., 2020; Dissanayake et al., 2025; Weerasinghe and Dissanayake, 2025). The connections between board technical or demographic characteristics, such as skills, expertise, gender and culture (Gray and Nowland, 2017; A. Weerasinghe et al., 2023; A. P. Weerasinghe et al., 2023), and firm outcomes are evident. Similarly, board composition is viewed through the lens of members’ ethics, such as altruistic behavior and empathy (Bekkers and Ottoni-Wilhelm, 2016; Haynes et al., 2017; Osei Bonsu et al., 2024). CEO characteristics also reveal connections between various personal aspects (e.g. greed and narcissism) (Osei Bonsu et al., 2024) and firm outcomes (e.g. CSR and Misconduct) (Kim et al., 2022; Sajko et al., 2021). From the perspective of upper-echelon theory, the benevolence of leaders should be reflected in corporate decision-making, including financial reporting strategies. Relatedly, the ethical leadership view suggests that ethical leaders demonstrate appropriate conduct within organizations and create an ethical work environment that encourages ethical behavior (Brown et al., 2005; Ho et al., 2015). EQ literature uses the ethical leadership view to inform the connection between women corporate leaders and EQ, arguing they are more ethical, less assertive, less aggressive and less overconfident (Ho et al., 2015; Huang and Kisgen, 2013; Martin et al., 2009; Strydom et al., 2017). Similarly, the ethical leadership view has been applied to ex-military executives, finding a negative association between ex-military leaders and adverse earnings management practices (Z. Li et al., 2023). Benevolent leadership, upper-echelon and ethical leadership views together suggest that directors who serve on NFPs likely view their corporate duties through moral, ethical and organizational citizenship obligations. Given that financial reporting is a key mechanism for informing shareholders and other stakeholders of firms’ strategic decisions, Benevolent directors likely uphold transparency in financial reporting, leading to the development of the following hypothesis:

While we adopt this narrative, directors serving on NFPs could be motivated by other reasons, such as reputation enhancement, professional networking, peer expectations and skills development (Brekke et al., 2018; Weerasinghe et al., 2024a, 2024b). For instance, service in industry-linked associations or clubs is more aligned with strategic goals than ethical intentions (Marquis and Tilcsik, 2016). We acknowledge the existence of the alternative narrative for our hypothesis. In addition, we conduct further tests in subsection 4.2.3 to validate specific potential explanations.

Prior literature has predominantly focused on board-level determinants of EQ, with limited research examining the influence of CEO characteristics. In this literature, certain CEO characteristics are found to have positive links with opportunistic earnings management. CEO social capital is positively associated with real earnings management in US firms (Griffin et al., 2021). Evidence from European firms suggests that CEO overconfidence, experience and cognitive ability are associated with increased corporate real earnings management (Kouaib and Jarboui, 2016). Narcissistic CEOs engage in accrual-based earnings management (Buchholz et al., 2020; Lin et al., 2020). In contrast, evidence suggests that women executives, including CEOs, are more ethical and risk-averse and less likely to engage in earnings management practices than their male counterparts, thereby improving EQ (Duong and Evans, 2016; X. Li et al., 2023). Prior literature indicates that board and CEO dynamic interactions are crucial in setting the tone at the top, which upholds transparency. Relatedly, leadership “homophily” refers to the phenomenon where similarities among directors (e.g. social or demographic) can positively influence corporate outcomes (Glass et al., 2016). Given the consensus that women CEOs are more likely to be ethical and transparent in their decision-making within firms, we expect these CEOs to facilitate the board’s monitoring and governance functions, particularly those of benevolent directors. Similar arguments apply to Benevolent CEOs. First, CEO benevolence is the same as director benevolence, resonating with leadership homophily. CEOs who serve across corporate and NFP sectors are likely to have distinct, broad ethical perspectives compared to CEOs who do not serve in such positions (Weerasinghe et al., 2024a, 2024b). They are less likely to engage in treacherous financial reporting practices and more likely to be transparent and honest in managing their firms. Second, Benevolent CEOs and directors share the same in-group perspective on ethical activities; thus, CEOs are likely to assist directors in improving financial reporting transparency. Therefore, we propose that Benevolent and women CEOs are likely to play an auxiliary role with Benevolent boards in enhancing the quality of financial reporting:

It is insightful to consider the role of corporate governance in the board benevolence–EP association. Governance mechanisms such as board and audit committee independence indicate strong links with EQ in the Australian context (Davidson et al., 2005; Kent et al., 2010) and in international contexts (Busirin et al., 2015; Srinidhi et al., 2011). The conceptual reason is that having more independent directors on boards and audit committees is likely to constrain opportunism and enhance the monitoring function of firms (Kent et al., 2016; Niu, 2006; Srinidhi et al., 2011), resulting in high-quality financial reporting. The board of directors is the primary internal corporate governance mechanism influencing management’s reporting decisions (Srinidhi et al., 2011). Independent directors are likely to exercise more vigilance in monitoring management actions (Alves, 2023). Prior research suggests that these corporate governance functions, along with other board and CEO characteristics, play an auxiliary role in enhancing EQ (Arun et al., 2015; Mnif and Cherif, 2021; Musa et al., 2023; Okaily et al., 2019). Thus, we propose that the presence of more independent directors in the board and audit committees is likely to assist Benevolent directors in adopting transparent reporting, leading to the following hypotheses:

The sample is ASX 300-listed firms from 2010 to 2019. The final sample is driven by data-hand-collection of benevolence and consists of 918 firm-year observations after merging the data of all variables and excluding financial firms and missing data. The sample is classified using the Global Industry Classification Standard (GICS) sectors. Table 1 reports the sample distribution by industry (Panel A) and year (Panel B).

EQ can be measured using many aspects, although research predominantly uses accrual-based models (Dechow et al., 2010; Strydom et al., 2017; Yang and Abeysekera, 2019). Following Abdelsalam et al. (2021) and Simoni et al. (2022), we proxy EQ by EP. The data are sourced from the EIKON/Refinitiv database (linked to StarMine EP data).

Benevolence is proxied by corporate leaders’ involvement in NFP leadership positions, simultaneously to corporate leadership. Board benevolence is measured by the proportion of directors on the board serving in NFP leadership positions in a year. CEO benevolence is the number of NFP positions a CEO holds each year. We hand-collected benevolence data from corporate annual reports by running keywords through reports to recognize potential NFP positions of directors and CEOs ^[3]. First, a list of ASX 300 firms is chosen in Connect4 using ASX codes. A keyword search ^[4] is conducted through annual reports using the database to refine the reports with potential NFP names under the director/CEO biographies. Second, the resulting annual reports are manually read to recognize potential NFPs. Firms with names such as Pty Ltd, Ltd, NL, Corp and PLC are disregarded (it is required by firms to disclose these under Australian Corporations Legislation). Third, the recognized list of potential NFPs is verified through four primary public information sources to confirm the NFP status: the Australian Business Number, the ACNC, the Australian Securities and Investments Commission registers and the official NFP websites. Finally, verified NFPs are listed in an Excel sheet along with director/CEO details: name, ASX firm, Industry, positions and year. The CEO gender (CEOGEN) variable equals 1 if the CEO is a woman and 0 otherwise within a given financial year. Gender is not necessarily a binary variable; however, this sample comprises only men and women, as disclosed in director biographies.

Model 1 examines the association between board benevolence and EP:

In the model, i represents the firm-year observations, and t represents the period from 2010 to 2019. All independent variables (IVs) are lagged by one year to address concerns about reverse causality. The baseline model is revised by introducing interaction terms to examine the moderating roles of Benevolent CEOs, women CEOs and internal corporate governance mechanisms on the association between board benevolence and EP (H2–H5).

EP_i,t is the dependent variable representing the EQ overall score, ranging from 0 to 100. Board benevolence is the primary explanatory variable, representing the proportion of directors on boards serving on NFPs, ranging from 0 to 100%. CEO benevolence is the number of NFP positions held by CEOs, and gender represents a dummy variable equal to 1 if the CEO is a woman and 0 otherwise. Control variables are included at three levels: firm, board and CEO characteristics, following relevant prior literature (Dissanayake et al., 2022; Kent et al., 2010; Strydom et al., 2017; Yang and Abeysekera, 2019). Firm-level variables include financial performance, measured using return on equity(ROE), size, using the natural log of total assets (LNTA), market-to-book value (MBV), age (FAGE), based on the number of years listed on ASX, capital expenditure (LNCAPEX),the natural log of proportion of capital expenditure on operating revenue, leverage (LEV), the percentage of total liability to total capital. These data are sourced from Morningstar and EIKON/Refinitiv databases. Board characteristics include board independence (BDIND) is the percentage of independent directors on the board, size (BDSIZE) is the number of directors on the board, meetings (BDMEET) is the number of meetings during the year, board gender diversity (BDDIV) is the percentage of women directors on the board. Board control data are obtained from the EIKON/Refinitiv database. CEO duality (CEODUA) variable is equal to 1 if the CEO is also the Chairman and 0 otherwise. CEO tenure (CEOTEN) refers to the number of years the CEO has held the position of CEO in the firm. CEO data are sourced primarily from the Connect4 and EIKON/Refinitiv databases, supplemented with hand collection. All the variable definitions are provided in  Appendix.

Key statistics are provided in Table 2. All control variables are winsorized at the 1^st and 99^th percentiles to adjust for outliers. The total assets and capital expenditure variables are presented as natural log values. All IVs are lagged by one year (t − 1) to address concerns about reverse causality. EP has values ranging from 1 to 100, with an average of 63.38, and exhibits a negative skew, indicating a non-normal distribution. We use a robust regression method (Hashmi et al., 2018; Katmon et al., 2019) to estimate associations, which is well-suited to handle the nature of the EP variable. Robust regression uses iterated re-weighted least squares (M-estimators) and can handle nonnormality and outliers (Huber, 1992; Pek et al., 2018). The average ROE is 11, LNCAPEX is −2.61, LEV is 0.46, FAGE is 22 years and LNTA is 14. The averages of board measures include BDIND with 67%, BDSIZE with 7 directors, BDMEET with 11, BDDIV with 17% and AUDIND with 88%. The average CEOTEN is five years, and CEODUA is a binary variable. The Pearson correlations are reported in Table 3. Most variables indicate low to moderate associations, ranging from 0 to 0.4 ^[5], indicating no threat of multicollinearity (Tabachnick et al., 2013).

Board benevolence ranges from 0 to 100%, with an average of 39%, indicating that, on average, 39% of directors on boards serve on NFPs. CEO benevolence ranges from 0 to 9, with an average of 0.51, and gender is a binary variable. Table 4 presents additional descriptive data on the board of benevolence. Panel A reports the percentage of board Benevolence of 50% or less and above 50% in different GICS industry sectors. Board benevolence above 50% accounted for 27% of the total sample, and the materials (6.32%) and industrials (5.01%) sectors indicated a higher representation of boards with above 50% Benevolent directors. Panel B reports statistical mean differences in firm size, board independence, size and audit committee independence between these two groups of board benevolence. Board benevolence above 50% is evident in relatively large firms with higher board independence, gender diversity and audit committee independence ^[6].

Table 5 presents the regression findings of H1, along with model fit statistics. Fit statistics of a null model and a model with all the variables compared, indicating an R-squared of 22% in the full model. The results support H1, indicating a significant positive association between board benevolence and EP (β = 8.161, p < 0.05). The influence of board benevolence on EP is economically significant, and a one-standard-deviation increase (Table 2, 0.26) in benevolence is associated with a 3.35% increase in the EP score of firms ^[7]. These estimates and t-values are consistent with prior literature on board characteristics and EQ (Arun et al., 2015; Strydom et al., 2017; Godigbe et al., 2018). Of the covariates, the only significant indicators are CEO gender (β = 13.922, p < 0.01), capital expenditure (β = −7.504, p < 0.01), firm age (β = 0.125, p < 0.05) and market-to-book value (β = 0.861, p < 0.01). Interestingly, firms investing more in capital expenditure are likelier to have lower EP. The findings do not indicate board independence, audit committee independence and board diversity as significant predictors of EP, thus, inconsistent with certain prior Australian research, such as Davidson et al. (2005), Kent et al. (2010) and Strydom et al. (2017). This may be due to the differences in EQ measures and the inclusion of benevolence in this study. Thus, in the presence of board benevolence, the individual influence of independence and diversity is likely to be less pronounced in predicting EP.

Table 6 reports the findings of H2 and H3. The interaction terms of CEO benevolence and gender are presented in two separate models, finding support for both hypotheses. The results indicate that the association between board benevolence and EP is strengthened in the presence of benevolent ^[8] and women CEOs. The interaction terms indicate large coefficients, particularly the interaction between board benevolence and CEO gender. This is possibly due to different scales of the independent and dependent variables (i.e. while EP has a mean of 63.38, the mean of board benevolence is 0.39). In unreported results, we tested variation inflation factor (VIF) values, finding no threat of multi-collinearity. In addition, we tested this using a standardized EP measure, and the results indicate small coefficients, implying that the scale of the EP variable drives the large coefficients. Figures 1 and 2 illustrate the nature of these interactions. Figure 1 presents the moderation role of CEO benevolence at several levels: 0, 1, 3, 5 and 7, representing the number of NFP positions by CEOs. When firms are not managed by Benevolent CEOs (blue line), the association between EP and board benevolence indicates a stable line. When CEO benevolence increases, the association between board benevolence and EP is more pronounced with significant positive slopes (all other lines). In firms managed by Benevolent CEOs, EP indicates lower starting amounts with lower board benevolence. This suggests that EP is poor in firms managed by Benevolent CEOs, in the absence of a certain proportion of Benevolent directors on boards (40%), with an increasing EP slope from the beginning. Thus, the real benefit of this interaction on EP is when firms managed by Benevolent CEOs reach 40% or more board benevolence in their firms. CEO gender acts similarly (Figure 2), with a low starting point of EP when women CEOs manage firms; however, with an increasing slope (red) surpassing the stable line (blue) when board benevolence reaches around 70%. The results confirm that CEO benevolence and gender strengthen the association between board benevolence and EP, supporting hypotheses.

The results of the moderating roles of internal governance mechanisms are reported in Table 7. A higher proportion of independent directors on boards and audit committees strengthens the association between board benevolence and EP^[9]. Figures 3 and 4 illustrate the nature of these interactions. The association between board benevolence and EP indicates a negative slope when the board independence is below 47% and the audit committee independence is below 70% (blue lines). However, board and audit committee independence significantly strengthened the association between board benevolence and EP after the above thresholds (red and green lines). Board benevolence of 30–40% (surpassing the blue lines) is the optimal point where these interactions seem to generate real benefits of financial reporting quality. The results confirm the positive moderation roles of internal corporate governance mechanisms on the board benevolence and EP association and indicate that the influence of Benevolent directors on EP is more pronounced when they are independent and serve on audit committees.

While EP represents an integral portion of financial reporting quality, prior literature commonly uses discretionary accruals as a proxy for EQ. To ensure the validity of our results regarding the influence of benevolence on EQ, we conduct an additional analysis using discretionary accruals as the outcome variable based on the modified Jones model (Dechow et al., 1995). We first obtain the total accruals of firms in year t by subtracting net cash flow from operations from net income in year t. Total accruals are then estimated using changes in total revenue (from t − 1), changes in receivables (from t − 1), gross property, plant and equipment in year t. All variables are deflated by lagged total assets, and the residuals of the regression represent discretionary accruals. These values are then used to perform robust OLS regression analysis with all IVs, including board benevolence. The results are reported in Table 8 and are consistent with the initial analysis, indicating that board benevolence is positively associated with EQ (less discretionary accruals), with a coefficient of −0.346 (p < 0.05).

The hypothesis was informed by the main narrative, with the existence of the alternative narrative that directors are motivated to serve on NFPs for multiple reasons, such as increasing networks and reputation. While we cannot thoroughly test all potential motivations of directors serving on NFPs, we investigate whether benevolent directors form a distinct group from other directors involved in multiple corporate directorships. We obtained board member corporate affiliations data from the Refinitiv database and used it as a proxy for “board networks.” The measure represents the average number of board affiliations with other corporate leadership (Ahmed et al., 2024; Remo‐Diez et al., 2025). The correlation between board benevolence and board networks is examined, finding a weak correlation (0.327). This suggests that benevolent directors and most corporate networked directors are distinct groups of directors. Then, we include board networks in the regression analysis to ensure such networks do not drive the results. The results indicate that board networks have a marginal negative association with EQ (β = −3.641, p < 0.10). The influence of board benevolence remains consistent with previous analyses despite the inclusion of board networks (β = 9.104 p < 0.05).

Several robustness tests are performed to minimize endogeneity concerns: entropy balancing (EB), accounting for the error term of board benevolence (residual inclusion method), and two-stage least squares (2SLS) regression. First, the results may be biased due to differences in firm, governance and CEO characteristics between firms with higher and lower board benevolence. For example, key differences in firms with higher board benevolence may drive results rather than benevolence. To alleviate this concern, the EB method is used to generate a matched sample by eliminating statistical differences between firms (Francoeur et al., 2022) with two levels of board benevolence. Second, board benevolence could be a proxy for firm and governance characteristics if certain firms are more likely to recruit benevolent directors than others. To ensure that a unique portion of the board benevolence variable influences EQ, the residual inclusion method is performed. This method is used to isolate the error term of the outcome variable (board benevolence) in a predicted regression model, including firm and governance characteristics as predictors (Gul et al., 2011). Finally, we use the 2SLS method, using lagged values of board benevolence as an instrumental variable. 2SLS regression is used to address potential simultaneity, omitted variable bias and reverse causality (Li et al., 2018), particularly if unobserved firm attributes influence board benevolence and EQ.

EB is a relatively new and robust sample matching method (Bhandari and Golden, 2021; Francoeur et al., 2022). Research refers to EB as a “doubly robust” technique (Zhao and Percival, 2017) and has several advantages over commonly used propensity score matching (PSM). Primarily, unlike PSM, EB retains the original sample size, as matching is done through weighting observations and does not allow the researcher discretion in generating the weights (Bhandari and Golden, 2021; McMullin and Schonberger, 2020). In this process, the sample is divided into a treatment and control group based on benevolence (BBENBIN), with a score of> 50% indicating High (treated) and <50% indicating Low (control). The BBENBIN binary variable was used as the DV in creating the matched sample, and all other covariates were predictors. The final sample includes 918 observations, and the statistical mean differences of all covariates between treatment and control groups are zero and non-significant. The regressions are re-performed using the matched observations (Table 9). The results of the EB sample validate the baseline results for H1, H2, H4 and H5; however, they do not support H3. Board benevolence is positively associated with EP (β = 12.359, p < 0.01) and CEO benevolence (β = 10.122, p < 0.05), and internal governance functions strengthen this association (board independence: β = 85.840, p < 0.01 and audit committee independence: β = 52.396, p < 0.05). Thus, the EB sample regression is consistent with the baseline results.

We use an alternative approach to address endogeneity, following Gul et al. (2011). First, we built a predictor model of board benevolence using firm, board and CEO characteristics from the baseline regression model (variables were chosen based on univariate correlations). The rationale here is to build a model that predicts firms with higher board benevolence and obtain the residual values. The predicted model indicates that board benevolence is a linear combination of firm, board and CEO characteristics; thus, if these characteristics explain the most variation in EP, board benevolence is only an aggregate proxy of those characteristics (Gul et al., 2011). In contrast, if the unexplained variance of board benevolence in the predicted model, which is the error term (or residuals), explains the variance of EP, board benevolence is more likely to be linked with EP independently of the control variables. The obtained residuals are used next, replacing board benevolence (new variable BBENRES) in examining the association with EP, revising the baseline regression model:

Table 10 reports the findings of the predictor model (Step 1), and the results of Step 2 are reported in Table 11. The predictive model shows an R-squared of 0.355, thus, explaining 36% of the variance in board benevolence by the predictors. The remaining 64% represents the unexplained variance, also known as the residuals (BBENRES). The Step 2 regression results are consistent with the previous findings concerning H1, H2, H4 and H5; however, they do not support H3, like EB sample regressions. The unexplained variance of board benevolence is significantly and positively associated with EP (β = 7.065, p < 0.10). This association is strengthened in the presence of Benevolent CEOs (β = 5.866, p < 0.05), a higher number of independent members on boards (β = 86.752, p < 0.01) and audit committees (β = 62.305, p < 0.01). The model fit R-squared statistic is also similar to baseline regressions, with a value of around 0.22, indicating the validity of the baseline and other analyses.

To further address endogeneity concerns, we followed the 2SLS regression, using lagged values of the main IV as an instrument (Harjoto et al., 2015; Li et al., 2018; Srinidhi et al., 2011). However, there is no feasible external IV available for board benevolence, given the nature of this construct; thus, we construct an IV using the lagged values (t − 1) of the board benevolence measure (InstrumentBBEN). Using lagged values of the primary endogenous variable is consistent with accounting and finance literature (Le et al., 2024). A valid instrument must satisfy two criteria: relevance (strong correlation with the endogenous variable) and exclusion (no direct correlation with the outcome variable without the endogenous variable (Knyazeva et al., 2013; Wang and Bellemare, 2019). We first test the relevance condition by estimating board benevolence with the instrument and other covariates. Column 1 of Table 12 reports the results of the first-stage OLS regression, indicating that the instrument satisfies the relevance condition with a coefficient of 0.740 at the 1% significance level and an F-statistic of 66.05, which exceeds the critical value of 10 (Le et al., 2024). Predicted values of first-stage regression (PredictedBoardBEN) are used in the second-stage regression, replacing the endogenous variable (Column 2). The results indicate a significant positive association between the predicted board benevolence and EP, validating our results. Wang and Bellemare (2019) show that lagged endogenous variables are less likely to have direct connections with the outcome variable. To confirm, we test the exclusion criteria by regressing the residuals from the second stage regression (error term of EP) on the instrument and other covariates. Column 3 reports the results, showing that there is no significant association between the instrument and the error term of EP, which satisfies the exclusion criteria. Thus, our results are consistent with the IV approach, indicating robustness.

This paper examines the association between board benevolence and EP (H1). It further explores the moderating roles of CEO benevolence (H2), CEO gender (H3) and board and audit committee independence (H4 and H5) on the association between board benevolence and EP. Theories of upper-echelon and ethical leadership inform the research design. Using a sample of ASX 300 firms from 2010 to 2019, the analysis reveals a significant positive association between board benevolence and EP. This association is strengthened in the presence of benevolent CEOs, women CEOs and a higher number of independent members on the board and audit committees. The results are robust to a matched sample using the EB approach, residual inclusion of board benevolence and instrumental variable regressions. The findings are consistent across the additional tests for H1, H2, H4 and H5, indicating greater validity of these associations. H3 was supported in the baseline regression; however, it is insignificant in the EB sample and when accounting for the error term in board benevolence.

The findings align with the notion that Benevolent directors who serve on NFPs, while simultaneously holding corporate leadership positions, differ from their counterparts in their influence on financial reporting. This is likely to be driven by their ethical stance, personal values and exposure to higher ethical accountabilities and perspectives (Bekkers and Ottoni-Wilhelm, 2016; Feng et al., 2023; Ward and Miller-Stevens, 2021; White et al., 2014). Benevolent directors are likely to be ethical leaders in managing their organizations (Brown et al., 2005; Ho et al., 2015), and less likely to engage in unethical financial reporting practices. Their values are reflected in the firms they manage (Chapple et al., 2020; Hambrick and Mason, 1984; Haynes et al., 2017) through transparency and honesty in publicizing actual financial status. The same reasoning applies to the moderation roles of CEO benevolence and gender, indicating that ethical CEOs support Benevolent directors in enhancing EP. This finding is consistent with the literature, which suggests that women or ethical CEOs can serve as influencers for transparent financial reporting (Arun et al., 2015; Srinidhi et al., 2011; Zalata et al., 2022). The results indicate the enhanced monitoring function of two central internal governance mechanisms, the board and the audit committee independence. Early Australian research provides evidence of the positive influence of these functions on EQ (Davidson et al., 2005; Kent et al., 2010), the findings reinstate this evidence with a recent sample of Australian capital market firms from 2010 to 2019. The association between board benevolence and EP is strengthened in the presence of higher proportions of board and audit committee independence.

The study has several limitations. First, the sample is data-driven, and the results are only generalizable to the Australian capital market context. However, the findings will be insightful for allowing researchers to develop studies that examine the influence of different board characteristics on EP. Future research can extend this study to other contexts and determine whether the impact of board benevolence is significant to EP. Second, we used two CEO and governance characteristics as moderators, and future studies can investigate more characteristics that may interplay with board benevolence in influencing EP. Third, we recognize directors serving on NFPs as benevolent in this study; however, directors are incentivized to hold such roles for multiple reasons (e.g. reputation and professional networks). While we conduct additional tests to show that benevolent directors are a distinct group of directors with multiple ASX directorships, and our measure may be valid, it is not possible to fully rule out this alternative explanation. Thus, we acknowledge this as a limitation of the study. Future research can examine the influence of board benevolence on different accrual management methods by extending the prior work of Kent et al. (2010) in the Australian context. Fourth, future research can extend our study by exploring the stock market response to the appointment or departure of Benevolent directors. Finally, COVID-19 significantly impacted the governance and performance of firms, including EQ (Landier and Thesmar, 2020); thus, it will be insightful to investigate the role of Benevolent directors during the crisis period.

This paper is developed based on Dr Ashesha Weerasinghe’s Doctor of Philosophy research project conducted under QUT and the Australian Government’s Research Training Program scholarship.