Do board gender diversity and remuneration impact earnings quality? Evidence from Spanish firms

Earnings quality refers to the extent to which a firm’s reported earnings capture its actual earnings (Krishnan and Parsons, 2008). When the informational quality of reported financial statements is degraded through aggressive accounting practices, earnings quality is damaged (Martin et al., 2016) and agency problems arise (Jiraporn et al., 2008). These agency problems can be mitigated through effective corporate governance systems. From a corporate governance perspective, our study aims to test the effectiveness of two boards of directors’ features promoting earnings quality: the gender diversity and remuneration of the board of directors.

The OECD/G20 Principles on Corporate Governance establish that a governance framework should ensure the strategic guidance of a company, the effective monitoring of management and a board’s fiduciary accountability to the company and its shareholders [¹]. In this spirit, Spain has developed strong corporate governance regulation. The Unified Good Governance Code of Listed Companies promotes modern corporate governance policies, promulgated by entities such as the OECD, the Basel Committee on Banking Supervision and the European Commission. Jensen (1993) identifies the role of boards of directors as an internal governance mechanism that mitigates agency conflicts. But improvements to the legal framework in Spain have not prevented financial scandals, highlighting the inability of boards of directors to inhibit aggressive accounting practices (Segui-Mas et al., 2018).

Managers can dilute earnings quality by manipulating financial reporting, leading to private benefits, and misleading outside investors (Aboody et al., 2005). Boards of directors have a crucial role in constraining these practices, but they do not always provide the transparency necessary for effective decision-making (Segui-Mas et al., 2018). Two central points that have been highlighted in social and political debates in the past few years, calling for the attention of regulators, scholars, corporations and users of financial information are the demand for gender equality in the labor market and other social spheres (Arun et al., 2015) and the questioned effectiveness of compensation packages paid to executives and board members (Merino-Madrid et al., 2009). Therefore, this paper sheds light on the social and political debate by examining the association of board gender diversity and the board remuneration system on firms’ earnings quality.

We make contributions on several fronts. First, we study earnings quality through a new, more informative metric based on the StarMine algorithm, which incorporates firms’ accruals, cash flow and operating efficiency in the measure of earnings quality. This is a forward-looking metric that combines information content relevant for future decisions; which differs from the traditional, accruals-based earnings-management measure widely used in the literature which acts as a backward-looking proxy for financial transparency (Lo, 2008).

Second, most empirical studies are based on Anglo-Saxon, common-law contexts and cannot be generalized to differing legal frameworks in civil-law countries. Indeed, La Porta et al. (2000) found evidence that common-law countries have the strongest protection of outside investors, including shareholders and creditors, in comparison to civil-law countries. Spain’s civil-law system is characterized by weaker protection of shareholder’s rights, heavier government intervention and higher self-dealing problems than common-law countries (La Porta et al., 2008). Because aggressive accounting and the subsequent deterioration of earnings quality are more likely to occur in countries with less protection for minority shareholders (Bushman et al., 2004), our study articulates a pertinent discussion of board gender diversity and board remuneration in a civil-law context. This study also builds upon the scarce literature on Spain, such as García-Sánchez et al. (2017), by analyzing new features such as earnings sustainability instead of earnings management.

Third, we calculated board gender diversity using the percentage of women on boards as well as the Blau and Shanon’s indices to capture gender equality, in line with the United Nations Sustainable Development Goal 5. Hence, this research is highly pertinent for the Spanish corporate sector and the European region, as contemporary debates about implementing directives seeking gender-balanced boards of directors and fair-remuneration policies for board members have become increasingly prominent in the corporate and social arenas (Damak, 2018).

Fourth, our analysis leads us to establish a nonlinear relationship between remuneration and earnings quality that has not been thoroughly discussed, nor empirically tested, in previous literature (Laux and Laux, 2009). Finally, we applied a new robust econometric technique based on the Tobit semiparametric estimator for fixed effects that improves the precision of results. We do this by using hand-collected data from the Spanish Stock Exchange Commission that allows us to extend the literature by finding relationships between previously untested board features.

We find that gender-balanced boards improve earnings quality of listed firms. Because most Spanish firms have low female representation on their boards, companies can benefit from promoting women to their board of directors to achieve gender-balanced board composition. Exhibited in Figure A1, boards are clearly male biased, with female participation being about half the average of the countries in the European Economic Area.

We also provide empirical evidence for Spain on the unexplored nonlinear relationship between board remuneration and firms’ earnings quality. This relationship has been theoretically suggested but not tested empirically. Earnings quality does not appear to monotonically increase with board compensation; rather, it is optimized when remuneration per board member is roughly €460,000, or 3.90% of a firm’s annual net income. While some Spanish listed firms are overpaying their board members, many could improve earnings quality by increasing board compensation.

Our findings are robust across three alternative measures of gender diversity and four metrics for board remuneration. We shed light on how board gender diversity and board remuneration policies can be used as drivers to improve firm financial transparency, as a response to stakeholder pressures like the United Nations with the Sustainable Development Goals call for action to all countries. Improving financial information quality is key to creating a healthy financial system in which corporations play a fundamental role (Pucheta-Martínez et al., 2016). Transparent and high-quality financial information is linked to good corporate governance practices (Segui-Mas et al., 2018).

The agency perspective emphasizes that the major role of the board of directors is to monitor managerial decisions to protect the interest of shareholders and ensure the integrity and transparency of corporate financial information (Jensen and Meckling, 1976). Relevant characteristics of boards include their gender balance and their compensation schemes, which serve as monitoring mechanisms to enhance governance and the quality of financial reporting (Baixauli-Soler et al., 2016).

The agency approach suggests that gender balance on a board reduces agency costs (Orazalin, 2020) because it increases monitoring of managers and in turn aligns the interests of shareholders and managers (Campbell and Mínguez-Vera, 2008). Along the same lines, Dalton and Dalton (2010) suggest that diversity of board members improves the effectiveness of the board of directors and enables it to provide better oversight of a firm’s disclosures and reports. As further support, Adams and Ferreira (2009) and Gull et al. (2018) indicate that female directors are more cautious in their decision-making, demand greater accountability from managers for weak corporate performance and are more willing to report fraudulent financial disclosures. When men participate in decision-making, they aim mostly at improving the performance of the company. This could be because men are more confident in what they do in financial matters (Barber and Odean, 2001; Krishnan and Parsons, 2008). When women join the decision-making group, their vision, attitudes and character complement those of men, enhancing decisions, corporate performance and earnings quality (Kogut et al., 2014).

Men and women take different approaches to money and investment (Barber and Odean, 2001). Female executives are more conservative in their financial decisions than male executives, which contributes to less aggressive accounting practices (García-Sánchez et al., 2017). Similarly, gender inequality in leading corporate positions results in poor corporate governance (García-Izquierdo et al., 2018). According to Saeed et al. (2019), a gender-balanced board of directors benefits from the skills, experiences and backgrounds of its varied members and engages in a high-quality decision-making process that promotes effective monitoring of earnings management and subsequently improves earnings quality. Companies that balance their board of directors by increasing the number of female members constrain earnings-manipulation practices and enhance the quality of financial reporting (Pucheta-Martínez et al., 2016). Such gender diversity contributes to the quality of the board’s deliberations and discussions of difficult issues that are often unaddressed by all-male boards (Gul et al., 2011). Similarly, as Krishnan and Parsons (2008) emphasize, high gender diversity in higher levels of management is positively related to earnings quality.

Equality among men and women is a universal legal principle that was approved by the UN General Assembly in 1979 and endorsed by Spain in 1983. Gender parity is a fundamental principle in the European Union which is also included in the United Nations’ Sustainable Development Goal agenda [²], and Spain has incorporated voluntary and compulsory regulations regarding board gender diversity [³]. The Unified Good Governance Code of Listed Companies set Spain on a new path toward achieving gender-balanced boardrooms. The code initiated a long process of reform and more recently was amended to propose a minimum of 40% female representation [⁴]. In addition, Spain approved the Organic Law 03/2007 which gives women and men equal access to employment and equal working conditions, in addition to promoting parity of women and men on the boards of directors of large corporations by mandating that no gender can represent less than 40% nor more than 60% of the seats. Moreover, the Spanish Constitution proclaims the right to equality and nondiscrimination on the basis of gender. In addition, Spain became the third European country to establish not binding gender quota, where the absence of sanctions has hindered the achievement of the law’s objectives. [⁵] Therefore, based on the theoretical, empirical and contextual arguments on the Spanish corporate sector, we hypothesize H1:

Board compensation plans are designed to encourage directors to wisely advise managers, monitor their performance and ultimately maximize shareholders’ wealth (Jensen and Meckling, 1976). Merino-Madrid et al. (2009) state that there is no scholarly consensus about the optimal board-remuneration structure and that companies should offer compensation packages with an optimal combination of fixed and variable components to align interests and reduce agency costs. Designing such packages has been a priority on the governance-reform agenda (López et al., 2015) because high compensation compromises directors’ objectivity in their governance role (Brick et al., 2006). Prior research provides significant insight on how executive compensation affects earnings quality and corporate decisions (Amewu and Alagidede, 2019), but the relationship between board compensation and earnings quality is underexplored.

Spanish companies are subject to two regulations regarding board remuneration. The first is the Unified Good Governance Code of Listed Companies of 2006. The code recommends disclosing board members’ remuneration and making this information transparent. It is characterized by the voluntary “comply and explain” principle. The second regulation is the Spanish Companies Act (or Royal Legislative Decree 1/2010), which requires companies to disclose directors’ pay, among other obligations. Remuneration policies must specify the maximum amount that companies can distribute to directors. For Spanish listed firms, Merino-Madrid et al. (2009) highlight the lack of independence on board-remuneration committees for optimizing value creation, suggesting that an optimal compensation scheme should exist.

Two competing hypotheses attempt to explain the board compensation-earnings quality relationship. The incentive hypothesis suggests that higher pay provides directors with an incentive to monitor managers, which prevents earnings manipulation and subsequently improves earnings quality and earnings informativeness (Adams and Ferreira, 2008). High compensation attracts talented directors, and such compensation aligns the interests of directors and the owners they represent (Vafeas, 2000). Empirically, this hypothesis is confirmed by Kim et al. (2013), who suggest that cash-based compensation establishes a better agency relationship between directors and shareholders than equity-based compensation.

An alternative explanation is the perverse-incentive hypothesis, which posits that excessive cash compensation may encourage directors to pay managers back by reducing their oversight of the managers, triggering higher earnings manipulation and in turn reducing earnings quality. This quid pro quo is associated with excessive compensation and weak monitoring (Brick et al., 2006). Overcompensated directors may be less likely to criticize managers and more likely to give managers the discretion to misreport earnings for personal benefit. Excessive compensation may create an environment in which managers and directors can connive to enjoy private benefits at the expense of shareholders. It reduces the pressure for directors to engage in internal monitoring, which lets managers misrepresent earnings and causes earnings quality to deteriorate. Empirically, Ye (2014) documents a negative association between earnings quality and cash compensation for independent directors. This suggests that rewarding independent directors with high cash pay compromises their independence and reduces their effectiveness at overseeing managers’ financial reporting. Similarly, by closely tying directors’ wealth to equity prices, schemes that compensate directors with an excessive amount of stock may incentivize managers to misreport financial information. As emphasized by Ye, if the market cannot see through overstated earnings and if it misvalues a company, directors compensated with stock may benefit from altering financial reports by engaging in insider trading.

To our knowledge, these hypotheses have yet to be examined empirically. For Spanish firms, the nearest study is García-Izquierdo et al. (2018), which analyzes how CEO pay correlates to the ratio of females on a board. It does not analyze earnings quality, board remuneration nor gender-balanced boards. Ours does. Because the incentive hypothesis is expected to be confirmed at relatively low compensation levels (implying a positive impact of board compensation on earnings quality), while the perverse-incentive hypothesis is expected to be confirmed at excessive compensation levels (implying a negative impact of board compensation on earnings quality), we hypothesize that:

Our sample comprises 105 nonfinancial Spanish listed firms from 2013 to 2018 in an unbalanced panel with 491 firm-year observations, with an average of 4.7 continuous observations per company. Financial firms were excluded because of their unique industry features [⁶]. Information on financial statements was obtained from Thomson Reuters Refinitiv Eikon. Data on board features were manually collected from the Annual Report of Corporate Governance published by the Spanish Stock Exchange Commission.

Our methodology uses a Tobit semiparametric estimator with firm- and industry-fixed effects because of the censored nature of our alternative dependent variables as described below. To our knowledge, this is the first paper applying this econometric technique to include these fixed effects to study earnings quality [⁷]. In addition, to address the common endogeneity problems in corporate governance studies (Barros et al., 2019) as well as the individual, time-invariant heterogeneity problem, we apply the Generalized Method of Moments system estimator (GMM-SE) following Blundell and Bond (1998). The GMM in first differencing (Arellano and Bond, 1991) is used as a second-order method to deal with endogeneity issues. Further details on the Tobit and GMM methods are discussed in  Appendix 1.

The primary dependent variables are measures of earnings quality developed by StarMine [⁸] and range from 0 to 1, with 1 being the highest. StarMine defines earnings quality as the degree to which past earnings are reliable and persistent. High-quality earnings accurately reflect a company’s current and past operating performance, indicate future operating performance and represent a reliable valuation measure for the company for all earnings levels. StarMine states that poor earnings quality indicates a likelihood of deteriorating fundamentals and low financial transparency, which makes it more difficult for investors to make informed investment decisions.

StarMine’s earnings-quality formula includes accruals, cash flow and operating efficiency. Our first measure of earnings quality is an overall measure that includes all three components (EQ1); our alternative metrics correspond to accruals (EQ2), cash flow (EQ3) and operating-efficiency (EQ4) measures of earnings quality. To our knowledge, this is the first study to use these metrics. Further details about the construction of these metrics are supplied in  Appendix 2.

To mitigate bias in estimating the earnings-quality metric, we test the consistency of our findings with a popular measure based on discretionary accruals. Following the modified Jones model (Dechow et al., 1995), the discretionary components of total accruals (TAcc) correspond to the residuals of the model to account for annual changes in operating revenues (ΔREV), annual changes in accounts receivable (ΔAR) and property, plant and equipment (PPE) as follows:

This complementary measure of earnings quality is represented by an earnings-management metric (EM) estimated as the residuals of equation (1). Given that managers can engage in income-increasing or income-decreasing manipulation of financial information, we follow the literature in considering the absolute value of EM to improve precision (Dechow et al., 1995). To standardize the earning-quality variable, we multiply EM by −1. Hence, higher (lower) values of EM indicate better (worse) earnings quality, consistently with the StarMine-based metrics.

The first explanatory variable is board gender diversity, defined as the percentage of women on the board (GD1). Given that a board composed only of male or female members exhibits a lack of diversity, we follow Saona et al. (2019) in using two additional indexed metrics to measure the construct of a gender-balanced board. The Blau (1977) diversification Index (GD2) is computed as $GD2=1−∑i=1nPi2$⁠, where P_i corresponds to the proportion of directors in each of the n = 2 gender categories (male and female). GD2 ranges between 0 (for no gender diversity) and 0.5 (equal proportions of male and female board members). The last metric is the Shannon (1948) diversification index (GD3), computed as $GD3=|∑i=1nLnPiPi|$⁠. GD3 takes values between 0 (no gender diversification) and 0.693 (equal proportions of each gender) [⁹].

Directors’ compensation schemes are measured with the log transformation of the average total remuneration per director (BRem1) [¹⁰], the fixed remuneration per director (BRem2) [¹¹] the board’s total remuneration (BRem3) and the board’s variable remuneration (BRem4), all as shares of the company’s net income. We compute the squared transformation of these variables to test their nonlinear relationship (posited in H2) with the earnings-quality measures.

We include several control variables. Following Gull et al. (2018), board-feature controls include the percentage of equity capital owned by the board (BOwn), the percentage of independent board members (BIndep), the number of board members (BSize) and the percentage of executive directors (BExec). Company characteristics include company size (Size) measured as the logarithm of total assets, a measure of default risk (Risk) based on Altman’s (1968) Z-score, the return on assets (ROA) computed as net income over total assets, the leverage ratio defined as total debt over total assets (Lev), a metric of growth opportunities (GO) defined as total assets less total common equity plus the company’s market capitalization. All controls are scaled by total assets [¹²].

Moreover, we introduce some dummy variables summarizing additional information on the degree to which Spanish companies follow Unified Good Governance Code. These variables refer to:

• if there are requirements for chairmanship;

• whether the board secretary monitors good-governance recommendations;

• the existence of supermajorities;

• whether tenure of 12 years or more of independent directors is required;

• the existence of a secretary board member;

• the existence of external advice for directors; and

• the existence of policies concerning time to prepare for board meetings.

Because these dummy variables are highly correlated, we applied the Cluster-Focused Principal Component Factoring technique, which produced three factors (named Governance1, Governance2 and Governance3 that capture an accumulated 62.06% of the variability of the corresponding seven covariates.

This results in the following partial-identification specification:

Here EQ_it represents earnings quality based on the StarMine metrics and the discretionary-accruals measure for robustness purposes for company i in period t; δ1_it represents one of the alternative measures of board gender diversity and δ2_it represents one of our measures of board remuneration; the vector $Xit’$ includes the set of controls previously defined; IndDummy represents the industry dummies capturing systematic differences across industry sectors; µ_i represents the individual effect; and ε_it is the mean-zero error term.

The firms in the sample are representative of the Spanish market [¹³].

As shown in Table A2 of  Appendix 3, the earnings-quality measures indicate that the capacity of earnings to represent Spanish companies’ actual economic performance is average. The overall measure of earnings quality (EQ1) is less than 0.5 while the other three metrics are slightly greater than 0.5. As for board gender diversity, we observe that about 16% of board seats are occupied by women, well below the recommendations of the Organic Law 03/2007 and the 40% quota imposed in other European nations. The mean values of the alternative measures of gender diversity (GD2 and GD3) are also well below gender parity.

For board remuneration (BRem1), we estimate that the average annual amount earned per director is about €183,000. Of this amount, approximately €115,000 come from fixed components (BRem2) and the remaining euros (less than 40%) come from variable components. The total remuneration equals 3.9% of the company’s net annual income (BRem3), whereas the variable remuneration amounts to 1.5% of net income (BRem4).

Examining other board features, a remarkably high proportion (20.6%) of equity capital is owned by boards (BOwn). Additionally, the average board size is 10.3 members (BSize), of whom about 41.3% are independent directors (BIndep) and 16.5% are executive directors (BExec).

Looking at firm-level control variables, the average ROA is 3.2% and the debt ratio is 28.9% of assets (Lev). Regarding the firm’s risk, the Altman (1968) Z-score metric (Risk) indicates that companies are almost in the safety zone [¹⁴]. The average growth-opportunity value (GO) indicates that Spanish companies do account for future growth opportunities because the metric is greater than 1, in line with results in previous empirical works (Adam and Goyal, 2008) [¹⁵].

Table 1 shows a consistent and statistically significant positive impact of gender diversity (GD1) on earnings quality, implying that the more gender-balanced the board is, the more transparent the financial reports disclosed by the company and the more informative its earnings. Previous empirical research finds that female participation on boards of directors is positively associated with good governance (Adams and Ferreira, 2009; Erhardt et al., 2003). Our results lead us to postulate that, in general, a gender-balanced board takes advantage of the characteristics of women and men while enhancing transparency and mitigating unethical behavior by the board. Thus, although Spanish boards’ gender diversity is still far from the European-wide average recommendations, gender diversity is value relevant. Hence, as shown by Orazalin (2020), gender-balanced boards improve the transparency and earnings quality.

Figure 1 plots the point estimates of the alternative versions of board gender diversity (GD1, GD2 and GD3) against the various measures of earnings quality (EQ1, EQ2, EQ3, EQ4 and EM). The positive relationship between gender diversity and earnings quality remains statistically significant and consistent in magnitude for all versions of earnings quality against all forms of gender diversity, except for EQ3 (earnings quality measure based on cash flows) which is not statistically significant at a 5% level for GD1, GD2 or GD3. This provides robust evidence in support of H1. This result reinforces Orazalin’s (2020) finding that gender parity in board composition improves the quality of accounting information.

The results concerning board compensation indicate a nonmonotonic relationship between compensation and earnings quality. For instance, Model (1) of Table 1 – analyzing remuneration per board member (BRem1) – shows that the estimated coefficients of BRem1 and BRem1² are positive and negative, respectively. This implies a diminishing marginal effect of total remuneration per director on earnings quality, suggesting a global maximum exists. The Lind and Mehlum (2010) test, displayed below the estimated extrema points in Table 1, does not support a monotonic relationship between BRem1 and EQ1, suggesting a threshold exists near €459,000 per director [¹⁶]. Because the mean value of BRem1 is less than the threshold, this finding suggests that the average Spanish company behaves consistently with the incentive hypothesis, meaning that board members closely monitor managerial behavior as compensation increases. This finding also indicates that several firms operate consistently with the perverse-incentive hypothesis.

Figure 2 displays the optimization of earnings quality with respect to compensation per director (BRem1), as derived in Model (1) of Table 1. The figure supports the notion that board remuneration has an inverse U-shaped relationship with earning quality. At low remuneration, increases in director compensation improve the quality of earnings. But starting near €459,000, firms (21.19% of them) begin to operate consistently with the perverse-incentive hypothesis.

Table A4 in  Appendix 4 expands on Table 1 by alternating the gender-diversity measure and the board-remuneration measure used in each specification. The positive influence of gender parity on earnings quality (EQ1) does not appear to be sensitive to the variable definition. In addition, the concave relationship between board remuneration and earnings quality does not depend on the measure of board remuneration. Indeed, the four metrics of board remuneration used in the estimations indicate that EQ1 increases until it achieves certain maximum level, after which the earnings quality deteriorates as the remuneration keeps increasing, as supported by the Lind–Melhum test. For instance, the average extrema point that represents the level at which additional compensation damages financial transparency in the various model specifications for BRem1 remains comparable to the one estimated in Table 1. In addition, it is observed that earnings quality improves as long as the log transformation of fixed board remuneration (BRem2) is not beyond the average of about €420,000. Models (3), (4), (7) and (8) of Table A4 also support the concave relationship as tested with the Lind–Melhum contrast located below the estimated extrema points [¹⁷]. Findings indicate that increasing board compensation results in improved financial transparency until it reaches an average of 13.10% of the firm’s net income (BRem3), at which point additional board remuneration begins to damage financial transparency. However, the average board-compensation package of a Spanish company is about 3.90% of net annual income [¹⁸]. This corroborates the above finding that most of the Spanish corporate sector operates consistently with the incentive hypothesis. A similar relationship is observed when we consider BRem4.

To test the robustness of our main findings, we reestimate the previous results by alternating the measures of earnings quality (EQ2, EQ3 and EQ4) and varying the metrics for gender parity (GD1, GD2 and GD3) and board remuneration (BRem1, BRem2, BRem3 and BRem4) in the specified model. The results are shown in  Appendix 4.

Robustness results displayed in Table A5 use as dependent variables the accrual measure (EQ2), cash flow measure (EQ3) and operating-efficiency measure (EQ4) of earnings quality. Again, we observe that board gender diversity increases the quality of earnings, but only when considering the EQ2 measure.

In addition, we still observe a concave relationship between board compensation (BRem1) and earnings quality when we consider EQ2 or EQ3 as the outcome measure. This relationship shows little sensitivity to our previous findings’ specification.

To address the potential for measurement errors, simultaneity problems and endogeneity problems caused by omitted variables, we test the sensitivity of our model specification by using a two-step GMM-SE with adjusted standard errors for potential heteroskedasticity as proposed by Blundell and Bond (1998) and as reported in Table A6 of  Appendix 4 [¹⁹]. This table examines EQ1 to EQ4 and the three measures of board gender diversity (GD1, GD2 and GD3) but uses only BRem1 as a measure of board remuneration; we find no differences from the results of the regressions in Tables A4 and A5. Board gender diversity appears consistent with the previously reported findings in Tables A4 and A5. The remuneration variable in this table exhibits the hypothesized relationship when we use two of the alternative metrics of earnings quality (EQ1 and EQ3). We also observe that BRem1 is not statistically significant in the EQ4 model. This sensitivity check indicates a positive impact on quality of financial reporting (EQ1 and EQ3) as total remuneration per director (BRem1) increases, ceteris paribus. This relationship continues up to a certain maximum point, after which earnings quality deteriorates as compensation continues to increase.

Because our proxy measures of earnings quality have not been applied in other contexts, the seminal application of these novel measures (EQ1 through EQ4) is reinforced with a variable that is widely used to measure earnings management (EM) as described by Dechow et al. (1995). These findings are tabulated in Table A7; the results are consistent with our initial results. We observe that the three measures of gender diversity exhibit a positive and statistically significant influence on EM. The estimations associated with board remuneration generally show a nonlinear influence: higher compensation improves earnings quality, but above a certain compensation level, earnings quality deteriorates. These results validate our findings.

Finally, as a last check for endogeneity issues, we apply the GMM in first differencing (GMM-Dif) according to Arellano and Bond (1991). Table A8 – which replicates the outputs obtained in Table A6 – shows that regardless of the metric for board gender diversity, the statistical significance and sign of the estimated parameters is consistent with the previous findings. Similarly, the board-remuneration variable (BRem1) still displays a nonmonotonic relationship with earnings quality. We conclude that while we lack a valid instrumental variable to deal with endogeneity issues for the covariates that are not strictly exogenous, GMM is an efficient technique to deal with this econometric problem (Antonakis et al., 2021; Barros et al., 2019).