Exploring Gender Diversity, Board Heterogeneity, and Corporate Risk Outcomes: Evidence from STOXX600 Firms

Abstract

This study examines the evolving role of board heterogeneity, including gender diversity, board attributes, and governance practices, in shaping corporate risk outcomes. In mature governance settings, corporate risk management emerges from the interaction between board structure, independence, leadership arrangements, and boardroom composition, such that gender diversity in isolation may no longer fully capture board effectiveness. We argue that while gender diversity remains relevant, its explanatory power operates in conjunction with other board characteristics that condition the quality of decision-making in already well-functioning boards. Using multiple regression estimations on a sample of STOXX600 firms, our main outcomes show that in mature European boards gender diversity (1) improves the operational efficiency, conditional by model specification, (2) increase debts level to finance growth, thereby enabling more rapid expansion than would otherwise be possible, without pushing to extensive borrowing, while reduce leverage starting 33% (3) prevents corporate failure starting 40% women on board (4) gender-diverse boards increase liquidity when critical mass is met (33%). Overall, the findings suggest that gender-diverse boards contribute to a reconfiguration of firms’ risk exposure across operational, financial, liquidity, and failure dimensions, rather than a uniform reduction in risk.

1. Introduction

The fundamental role of a corporate board is to substantiate the risks awareness, good rules and practices, communication and transparency, good citizenship and ethics in community, competitiveness and resilience, and lastly, the company’s interests over any individual interest. As companies mature and the landscape of corporate governance evolves, a more nuanced understanding of board attributes and dynamics has emerged. The discussion aims to move beyond the traditional focus on gender diversity to improve corporate outcomes. We highlight the importance of extended board characteristics and interaction dynamics, like board members’ professional backgrounds or skills, tenure, independence, membership, remuneration, committees, and board policies, for shaping effective risk management and robust decision-making. These attributes, when combined, create a board environment that is better equipped to identify, assess, and mitigate risks for the long-term interests of the company and its stakeholders.

Despite the extensive literature on board gender diversity and corporate risk, several important gaps were identified, clustered on four dimensions. First, gender diversity is often analyzed in isolation, with limited attention to how its impact may depend on other board attributes. We note that characteristics like tenure, independence, committee structures, and governance practices jointly shape monitoring and decision-making processes (Harjoto et al., 2018), along with the largely discussed binding gender quota in institutional settings, looking for greater social equality. In this context, it becomes essential to systematically examine how the presence of female directors, combined with other board heterogeneity determinants, shapes corporate outcomes, observing possible suppressors or conditional effects on corporate risk outcomes in developed markets. Second, prior research focuses on commonly classified aggregate risk measures because they summarize a firm’s overall risk exposure into a single composite indicator, like market-based measures embedded in stock prices (e.g., total or idiosyncratic return volatility and beta), rather than isolating specific sources of risk. This approach may reflect investors’ interest in jointly assessing all underlying uncertainties, while it can mask the underlying drivers of risk and potentially mislead interpretation. This limitation arises because risk-aggregated measures do not distinguish whether risk arises from operational inefficiencies, financial policy, or liquidity constraints, leading to reduced explanatory power in models linking governance characteristics to aggregated risk outcomes. Third, while mature and highly regulated markets have been widely studied, existing research has focused mainly on performance outcomes rather than linking disclosed board attributes to differentiated risk exposures. Fourth, alternative risk measures employed in this study include four major accounting-based risk dimensions: operational risk, financial risk, liquidity risk, and bankruptcy risk, of which operational efficiency and liquidity soundness research is scarce in corporate finance, and these risks are rather approached by researchers for the banking sector. Nevertheless, these studies frequently find conditional evidence that female board representation improves cost efficiency, but the effect is sensitive to the share of women on the board (critical mass), bank capitalization/regulatory context, and model specification (nonlinearities, interactions). Research has shown that findings differ based on whether accounting-based or market-based risk measures are applied, noting that market-based measures are preferred by US-based research (Teodósio et al., 2021).

Questioning further about the relevance of gender diversity in the ESG era, we find that prior Environmental, Social and Governance (ESG) research often emphasizes disclosures for performance, while our study contributes by examining the internal governance mechanisms through which corporate risk is managed. Gender diversity on the board is a governance mechanism that influences decision-making and risk oversight. By linking gender diversity to differentiated risk dimensions, our analysis complements ESG-oriented research and reinforces the argument that comprehensive sustainable corporate outcomes depend on effective board-level governance structures. Without denying the obvious progress of gender diversity in Europe, the baseline of female representation on boards is 35.8% on average in the second semester of 2025, as reported by EIGE (European Institute for Gender Equality|EIGE, 2025), while these results are driven by several “more gender-equal” institutional contexts, advanced by their legally binding regulations.

The European institutional setting is central to the contribution of the study, conceptually built around gender diversity from its early stages and corporate risks. We address the identified gaps by adopting a multidimensional governance framework and examining how gender diversity, together with other board characteristics, influences specific sources of risk—operational, financial, liquidity, and bankruptcy risks—discussing the outstanding and benchmarking STOXX600 firms. The focus on STOXX600 firms allows us to examine board gender diversity in mature corporate governance environments characterized by strong investor protection, advanced disclosure standards, and varying histories of gender quota implementation. This setting is particularly suited to testing non-linear and critical-mass effects, as European firms exhibit substantial cross-country variation in regulatory intensity and the timing of gender diversity reforms. By maintaining steady market development and relatively constant disclosure quality, the European context enables a cleaner assessment of how gender diversity interacts with other board attributes to shape differentiated risk outcomes. We emphasize that the European focus is not a limitation but a deliberate design choice that strengthens internal validity and facilitates theoretical inference, while offering insights that can inform governance debates in other institutional settings too.

In addition to a gender diversity focus, this paper examines a large set of publicly available diversity categories that are supposed to make a board a “good board”. In our reasoning, we also look at evidence from conventional wisdom, such as the idea that smaller and more independent boards are better boards. We try to validate this evidence in case of large and complex companies, exposed to various risks and decision-making, or finding themselves in higher need of advisory services. We add literature on evidence related to operational efficiency and liquidity in corporate settings, while testing prior evidence related to indebtedness and bankruptcy risk in the European pan-pandemic context.

The remainder of this paper is organized as follows. Section 2 is dedicated to an extensive literature review, including a theoretical framework and main empirical results related to corporate risk and gender diversity, further discussing other individual diversity components, like board structure, independence, remuneration, committees, CEO role, board’s policies and practices, to develop our hypothesis. Section 3 describes data, research methodology, and empirical approach. Section 4 presents empirical results, while Section 5 discusses the findings and limitations. Section 6 concludes by presenting a summary of key findings and their implication in how the established theoretical framework explains board diversity-risk dynamics in the European context, implications for stakeholders, and future research.

2. Literature Review and Hypothesis

2.1. Theoretical Foundation and Hypothesis Development

Different theoretical approaches have been explored by researchers in corporate governance to ground their empirical research results on board diversity. Widely employed to articulate how corporate boards support corporate outcomes, agency theory, upper-echelon theory, intergroup contact theory, resource dependence theory, and critical mass theory have been discussed in this study to substantiate our results related to board gender diversity. While sustainability and ESG considerations have become central to contemporary corporate governance research, board composition remains foundational to ESG effectiveness. Out of it, the “Governance” pillar of ESG plays a critical role in shaping Environmental and Social strategies, risk appetite, and long-term value creation.

A broad body of evidence from psychology and experimental economics documents systematic gender-related differences in risk preferences, however, with important contextual nuances. Early contributions in social psychology suggest that group diversity promotes more moderate and balanced decisions (Wallach & Kogan, 1965). The impact of board gender diversity on corporate risk has been most frequently examined through the lens of financial choices and the differentiated corporate risk channels they affect. Empirical studies in economics generally find women to be more risk-averse than men (Croson & Gneezy, 2009; Jianakoplos & Bernasek, 1998), although subsequent replications highlight that gender differences in risk attitudes are often task-specific and contingent on individual and institutional characteristics (Filippin & Crosetto, 2016).

Within corporate settings, women bring distinct values and behavioral orientations to decision-making, including stronger ethical sensitivity, higher regulatory awareness, and greater concern for relational and long-term outcomes. These attributes are particularly relevant for board oversight functions, where decisions involve balancing risk, control, and strategic flexibility. Prior studies show that female directors tend to strengthen monitoring, demand greater accountability, and exhibit lower tolerance for opportunistic managerial behavior, consistent with agency theory predictions (Adams & Ferreira, 2009). Board gender diversity has been widely recognized as an important determinant of governance quality and risk management (Datta et al., 2021; Levi et al., 2014). Behavioral differences related to risk aversion, man overconfidence, and investors’ trust in financial decisions made by firms led by women further suggest that gender diversity may systematically shape firms’ financial and operational decisions (Huang & Kisgen, 2012).

At the same time, the effectiveness of gender diversity is shaped by institutional context and board composition. Research grounded in critical mass and tokenism theories points out that limited female representation may weaken women’s influence on board decisions (Kanter, 1977). Evidence from quota-driven reforms indicates that rapid increases in female board membership may initially be associated with unintended effects, including reduced firm value or increased internal frictions, particularly when appointments are constrained by a limited pool of candidates (Ahern & Dittmar, 2012). Quota regimes have expanded over time the director talent pool and professionalized nomination practices, potentially strengthening the governance role of gender-diverse boards. To account for institutional dynamics, we further control quota history, measured as the number of years since the introduction of legally binding gender quota laws, capturing the maturity of gender diversity practices across countries.

Empirical findings, however, remain heterogeneous and depend strongly on the type of risk considered and the measurement approach employed. Authors (Teodósio et al., 2021) show that European studies predominantly rely on accounting-based risk measures, whereas U.S.-based research emphasizes market-based indicators. Their review documents more consistent gender effects on accounting-related risks—such as failure probability, litigation exposure, and operational volatility—than on market-based risk measures. This pattern suggests that female directors’ influence may be more pronounced in areas related to internal controls, compliance, and balance-sheet management rather than in investors’ aggregate risk assessments.

Building on this literature, the present study adopts a differentiated risk perspective, examining how gender diversity influences both specific dimensions of corporate risk and aggregate measures. We focus on operational risk, financial leverage, liquidity, and bankruptcy risk management, allowing us to assess whether and how gender-diverse boards reallocate risks across operational, financial, liquidity, and failure dimensions.

Due to the complexity of interaction between board attributes and the evolution in gender quota legislations in the European context, the possible effects of “tokenism” or “critical mass” have been considered, controlling for a possible non-linear influence of gender diversity on corporate risk outcomes.

2.1.1. Operational Risk

From a corporate governance perspective, operational risk reflects the board’s ability to discipline managerial behavior, oversee cost structures, and ensure efficient resource allocation. Agency theory predicts that stronger monitoring reduces managerial slack and operational inefficiencies (Jensen & Meckling, 1976). In the context of operational efficiency, gender-diverse boards may enhance oversight quality, as female directors are often associated with greater diligence, risk awareness, and intolerance toward inefficient spending. Upper-echelon theory (Hambrick & Mason, 1984) states that organizational outcomes, both strategic choices and effectiveness, are partially predicted by top managerial background characteristics. Therefore, it is supposed that female directors’ cognitive traits and behavioral orientations may influence operational decisions, implying that gender diversity may affect day-to-day cost management through more cautious and process-oriented decision-making. Nevertheless, companies may face costs from heterogeneity due to greater communication and coordination problems arising among a group of directors with dissimilar or disparate backgrounds. Directors bringing varied perspectives to board deliberations, as well as larger boards, can increase conflict among board members and impede the decision-making process (Anderson et al., 2011).

While existing research often measures effectiveness by profitability indicators (ROA, ROE, EBIT), this approach may explain less of the linkage with operational expenses dynamics and industry-dependent context. Two firms with comparable profitability or liquidity positions may exhibit substantially different cost structures and operational risk exposures. Consequently, even when gender diversity does not materially affect performance levels, it may still influence operational risk through cost discipline and expenditure control. Frequently discussed in the banking sector, the presence of females on the board is associated with a reduced cost-to-income ratio and improved liquidity. Direct evidence on firms’ Operating Ratio, defined as operating expenses relative to net sales, remains scarce outside the banking literature, where cost-to-income ratios are more commonly examined. Prior findings from banking studies suggest that female representation on boards can improve cost efficiency, although effects are often nonlinear and sensitive to institutional settings.

By focusing on the Operating Ratio as a transparent and interpretable measure of operational risk, this study extends the literature beyond profitability-based assessments and banking-specific contexts. This approach allows us to directly examine whether gender-diverse boards are associated with tighter cost control and lower operational risk in non-financial mature firms; therefore, our first tested hypothesis was:

Hypothesis 1 (H1).

Gender-diverse boards are associated with a lower Operating Ratio.

Hypothesis 1.1 (H1.1).

The influence of gender diversity on operational efficiency changes when the presence of women on the board becomes relevant.

2.1.2. Financial Risk

Capital structure decisions represent a central strategic choice through which boards influence corporate financial risk. Classical corporate finance theory, like the “position of irrelevance” defended by Modigliani and Miller (1958) demonstrate that, under a set of idealized assumptions—including perfect capital markets, no taxes, and no bankruptcy costs—a firm’s capital structure and dividend policy are irrelevant to its market value. These propositions serve as fundamental theoretical benchmarks, providing a baseline to examine how real-world frictions such as taxes, information asymmetries, and agency conflicts influence firm value. Alternatively, the trade-off theory (Miller, 1977; Modigliani & Miller, 1958) posits that firms balance the tax benefits of debt against expected distress costs to determine an optimal leverage level and maximize company value. Conversely, the pecking order theory (Myers, 1984; Myers & Majluf, 1984) posits that firms prefer a sequential order for financing: first using internal funds (like retained earnings), then debt, and finally issuing new equity only as a last resort. From a governance perspective of the agency theory, it is likely that the disciplinary role of debts constrains managerial discretion and reduces cash flow problems. Gender-balanced boards with stronger monitoring capacity may therefore support higher leverage when debt is perceived as a governance mechanism rather than a source of excessive risk. From the resource dependence theory perspective (Barney, 1991; Pfeffer & Salancik, 1978), gender-diverse boards may improve firms’ access to debt markets and influence financing choices by providing broader expertise, legitimacy, improved monitoring capabilities, and external connections. Also, intergroup contact theory (Pettigrew & Tropp, 2006) suggests that female directors may gradually adopt the dominant decision-making norms of their male counterparts, potentially neutralizing presumed differences in risk preferences.

Along with the theoretical framework, empirical evidence on gender diversity and financial risk remains inconclusive. Several studies associate female leadership with more conservative financing policies (Faccio et al., 2016) and find that firms led by female CEOs exhibit lower leverage and more stable earnings, while García and Herrero (2021) document a negative relationship between female board representation and leverage, cost of debt, and debt maturity. Conversely, Ahern and Dittmar (2012) report that increased female representation following Norway’s quota reform is associated with higher leverage and lower cash holdings, suggesting increased financial risk. Adams and Funk (2012) challenge the assumption of universal female risk aversion, showing that women directors may exhibit higher risk tolerance than their male counterparts, while Matsa and Miller (2013) found no significant change in leverage following gender quota implementation.

Taken together, these mixed findings suggest that the effect of gender diversity on financial risk is context-dependent and shaped by board structure and intergroup dynamics, like size or tenure, monitoring incentives, and firms’ strategic financing needs. In large, mature firms, gender-diverse boards may support higher indebtedness as part of a deliberate governance and resource-allocation strategy, rather than as a sign of excessive risk-taking. This perspective motivates an empirical examination of whether gender diversity is associated with increased leverage in complex corporate settings. Based on these statements, the second tested hypothesis was:

Hypothesis 2 (H2).

Gender-diverse boards are associated with higher corporate indebtedness.

Hypothesis 2.1 (H2.1).

The influence of gender diversity on financial decisions changes when the presence of women on the board becomes relevant.

2.1.3. Bankruptcy Risk

Managing failure reflects a firm’s ability to sustain operations and meet its financial obligations under adverse conditions, making it a central outcome of board oversight and strategic governance. From an agency theory perspective, boards that provide stronger monitoring and strategic guidance reduce the likelihood of excessive risk-taking that could lead to financial distress. Gender-diverse boards may enhance this function by promoting more cautious, forward-looking decision-making and by improving the quality of internal controls and risk assessment processes. Upper-echelon theory further suggests that directors’ demographic and cognitive traits influence strategic choices related to solvency, restructuring, and crisis response, thereby affecting bankruptcy risk.

Empirical evidence on gender diversity and bankruptcy risk is mixed but generally supportive of a stabilizing role for female board representation. In a European cross-country setting, Faccio et al. (2016) find that female CEOs are associated with lower financial, operational, and failure risks. Similarly, Wilson and Altanlar (2009) report that a higher proportion of women on corporate boards reduces insolvency risk. Mittal and Lavina (2018) also document a negative relationship between female board representation and bankruptcy likelihood. These findings suggest that gender diversity may contribute to firm resilience and solvency through enhanced oversight and more prudent strategic behavior.

However, other studies emphasize that the impact of board characteristics on failure risk is highly context dependent. Fich and Slezak (2008) show that larger boards are associated with a higher probability of bankruptcy, while Darrat et al. (2016) find that larger boards reduce bankruptcy risk only in complex firms that require broader expertise. Similarly, the role of inside directors varies with firm sophistication and information needs. Santen and Donker (2009) find no significant relationship between director gender and financial distress, underscoring the lack of consensus in the literature. Other authors (García & Herrero, 2021) reconcile these findings by demonstrating that board composition—particularly the combination of smaller, more independent boards with higher female representation—significantly reduces the likelihood of financial distress in European firms.

Given these mixed results, bankruptcy risk appears to be shaped not by gender diversity alone, but by its interaction with broader board structures and governance mechanisms. To capture this dimension, we employ the Altman Z-score, a widely used composite indicator of bankruptcy risk that reflects firms’ profitability (retained earnings, EBIT), leverage (working capital), sales, and market value of equity. We examine how gender-diverse boards influence Z-scores, one of the traditional aggregated measures considered by researchers to assess financial distress. This choice allows us to conduct a comprehensive assessment of firm resilience beyond the financial ratios aggregated in the Z-score, while disaggregated risk measures are also included in our study.

Hypothesis 3 (H3).

Gender-diverse boards are associated with higher Z-scores and a lower likelihood of bankruptcy.

Hypothesis 3.1 (H3.1).

The influence of gender diversity on bankruptcy risk is changing when the presence of women on the board becomes relevant.

2.1.4. Liquidity Risk

Liquidity risk reflects a firm’s ability to meet short-term obligations and absorb unexpected cash flow shocks, making it a key outcome of board oversight over financial policies. Gender diversity may influence liquidity risk through more conservative cash management and precautionary policies.

From an agency theory perspective, cash holdings involve a fundamental trade-off: while liquidity provides insurance against distress, excess cash may exacerbate agency problems by increasing managerial discretion. Similarly, excessive prudent behavior for higher liquidity might increase agency cost while reducing the value for shareholders. Boards, therefore, play a critical role in balancing safety against efficiency considerations. Gender-diverse boards may influence this balance by shaping attitudes toward risk, control, and financial flexibility.

From the resource dependence and intergroup contact theory perspective, gender-diverse, financially skilled boards in mature firms may deliberately reduce excess liquidity when external financing is readily available, reallocating cash toward investment, debt servicing, or shareholder payouts. From a governance perspective, this may result in lower excess liquidity and more efficient working capital management, particularly in leading firms where growth constraints are strongly related to agency costs.

Empirical evidence on the relationship between gender diversity and liquidity is mixed and often context-dependent. Ahern and Dittmar (2012) show that increased female representation on Norwegian boards is associated with higher leverage and lower cash holdings, suggesting reduced liquidity buffers. These findings are consistent with the view that gender-diverse boards may favor more active capital allocation and less reliance on inactive cash. Similarly, Florackis and Sainani (2018) document that firms with strong financial CFO hold less cash, relying instead on their ability to access external financing even during periods of financial stress. García-Meca et al. (2022) reported that the relationship between female board representation and payout policy follows an inverted U-shape, consistent with critical mass theory. At low levels of female representation, women directors may support higher dividend payouts to reduce agency conflicts and enhance legitimacy when first interacting with their male counterparts, thereby reducing liquidity. Once a critical mass is reached, more traditional risk-averse and precautionary financial behaviors emerge, leading to lower payouts and improved liquidity positions. Recent research reveals no significant gender-based differences in the optimal leverage ratios set by CFOs (Krystyniak & Staneva, 2024). This suggests that lower liquidity does not necessarily reflect higher vulnerability but may instead signal confidence and financial sophistication that might be driven by board financial leaders.

These considerations indicate that liquidity outcomes depend not only on the presence of women on boards but also on their relative representation and interaction with other governance mechanisms. In mature and well-capitalized firms, gender-diverse boards may reduce excess liquidity as part of a disciplined financial strategy and growth constraints, relying on governance quality and access to external financing rather than cash storage.

Hypothesis 4 (H4).

Gender-diverse boards are associated with lower corporate liquidity.

Hypothesis 4.1 (H4.1).

The influence of gender diversity on liquidity risk changes when the presence of women on the board becomes relevant.

Taken together, the four hypotheses reflect a coherent risk management framework wherein board gender diversity influences corporate risk through multiple, interconnected channels. Gender-diverse boards are expected to enhance operational discipline by reducing inefficiencies (H1), while simultaneously shaping financial risk through more active and deliberate leverage choices (H2). At the same time, improved governance quality and confidence in external financing may reduce reliance on overly prudent liquidity buffers, resulting in lower liquidity holdings (H4). These effects contribute jointly to greater firm resilience and lower failure risk, as captured by higher bankruptcy-distance measures (H3). This multidimensional perspective recognizes that gender-diverse boards do not uniformly reduce all forms of risk but contribute to a reconfiguration of firms’ risk exposure across operational, financial, and liquidity dimensions, highlighting the importance of analyzing differentiated risk outcomes rather than aggregate measures.

2.2. Board Heterogeneity

Beyond gender diversity, the corporate governance literature increasingly conceptualizes board heterogeneity as a multidimensional construct encompassing directors’ demographic, cognitive, and structural attributes. Board heterogeneity is expected to influence corporate outcomes by shaping the board’s monitoring capacity, advisory quality, and risk oversight effectiveness.

Authors (Anderson et al., 2011) argue that heterogeneous boards bring diverse backgrounds, experiences, and skills into the boardroom, strengthening managerial monitoring and reducing CEO dominance. At the same time, heterogeneity may generate coordination costs, longer deliberation, and internal conflicts, potentially slow decision-making, and weaken efficiency. Board heterogeneity may arise across multiple dimensions, including education, professional experience, tenure, gender, ethnicity, and age. These authors further document that CEO influence—proxied by CEO ownership, tenure, and free cash flow—is negatively associated with board heterogeneity, suggesting that powerful CEOs tend to select more demographically similar boards. They distinguish between occupational heterogeneity (education, experience, profession) and social heterogeneity (gender, ethnicity, age), a classification later echoed in the broader diversity literature. While proponents view board heterogeneity as value-enhancing through richer perspectives and improved oversight, critics argue that diversity may sometimes be driven by social or ethical considerations rather than firm value maximization.

Subsequent research (Bernile et al., 2018), refines this distinction by classifying board diversity into demographic diversity (gender, age, ethnicity) and cognitive diversity (education, expertise, experience) or, alternatively (Jebran et al., 2020) into relation-oriented diversity, related to surface-level differences, such as age and gender, and task-oriented diversity, which is related to job-related differences, such as tenure and education. These propositions highlight that heterogeneity operates through multiple channels and may exert multidirectional effects on corporate risk, beyond gender-specific behavioral traits.

Building on this framework, the present study considers board heterogeneity as a complementary governance mechanism to gender diversity. In addition to gender composition, we jointly examine structural and functional board attributes, like CEO role, board independence, oversight structures, governance policies, and compensation design, given their expected interaction with gender diversity in shaping firms’ operational, financial, liquidity, and failure risks in European advanced corporate governance settings.

The corporate governance literature emphasizes that no single board configuration is universally optimal. As governance models expand to include multiple board attributes, empirical findings often suffer from endogeneity, omitted-variable bias, or suppressor effects, contributing to inconsistent results across institutional contexts and firm characteristics. Consistent with the “no one-size-fits-all” view articulated by Coles et al. (2008), complex firms with greater advisory and monitoring needs tend to operate with larger, more heterogeneous boards. Firm size further conditions the relevance of board heterogeneity. Larger firms are typically more diversified, face lower default risk, and enjoy easier access to long-term debt due to more collateralizable assets (Alves et al., 2015; Ghosh et al., 2011). Although larger firms are often perceived as less risky, higher leverage may increase the cost of debt (Pandey et al., 2020).

These characteristics suggest that board heterogeneity is likely to play a more pronounced role in European leading listed firms, where governance structures extend beyond gender composition toward advanced oversight, policy frameworks, independence, and risk-linked incentives.

2.2.1. Board Size

Board size represents a fundamental dimension of board heterogeneity and a key governance mechanism for mitigating agency problems. Adams et al. (2010) identify board size, independence, and CEO power as central governance levers, particularly in large firms, a view reinforced by Belghitar and Clark (2015).

From a resource-based perspective, larger boards expand the pool of expertise, skills, experience, and external connections available to the firm, thereby enhancing advisory capacity and legitimacy (Carter et al., 2007). This argument is consistent with the notion that complex firms benefit from larger boards capable of addressing diversified strategic and operational challenges (Dalton et al., 1999). Conversely, agency theory and organizational economics emphasize the coordination and communication costs associated with larger boards. Recent evidence, on a sample of large-listed companies that benefit from gender equality legislation in its early stages, indicates that board size increases operating ratio, thus reducing operational efficiency (Tiloiu, 2025). Similarly, empirical evidence suggests that larger boards may adopt more cautious and compromise-oriented decisions, potentially dampening firm performance and value (Cheng, 2008). Although larger boards provide resources and connect broader human capital, these benefits may be offset by slower decision-making and coordination burdens, adversely affecting profitability across both large and small firms (De Andres et al., 2005; Eisenberg et al., 1998). Discussing smaller boards, they are often considered more efficient, as fewer members reduce managerial discretion, accelerate decision-making, and enhance accountability (Jensen, 1993).

2.2.2. CEO Role

The configuration of the CEO role, particularly CEO duality, constitutes another key dimension of board heterogeneity. Agency theory posits that CEO duality concentrates power, weakens board oversight, and increases agency costs, especially in large and complex organizations where monitoring is already challenging (Jensen, 1993; Jensen & Meckling, 1976). In contrast, resource dependence theories argue that unified leadership enhances coordination, reduces information asymmetry, fosters inside-firm sectoral knowledge, and facilitates faster strategic responses, which may be advantageous in dynamic environments (Boyd, 1995; Fich & Slezak, 2008; Pfeffer, 1972).

Empirical research remains inconclusive. Adams and Ferreira (2007) suggest that more powerful CEOs may share information more openly with directors, implying that management-friendly boards can be optimal under certain conditions. Concerning the choice of CEO duality and its value implications, competing theories predict that combining the CEO and chairman roles has either net costs or net benefits (Finkelstein & D’Aveni, 1994). Goergen et al. (2020) discuss the disclosure requirements of firms and their reasoning for combining or separating the roles of CEO and chairman. These competing perspectives suggest that the CEO role configuration interacts with board composition and heterogeneity in the reconfiguration of firms’ risk exposure.

2.2.3. Board’s Independence

Board independence is widely viewed as a cornerstone of effective corporate governance, aimed at enhancing oversight quality and reducing managerial oversight opportunism. Independent directors are expected to strengthen monitoring and protect shareholder interests; however, empirical evidence on their performance implications remains mixed.

Studies indicate that the effectiveness of independent directors depends on contextual factors such as firm complexity, industry characteristics, and board processes rather than independence alone (Boyd, 1995). Terjesen et al. (2016) show that in cross-country settings, the mere presence of women directors matters more for firm outcomes than their formal independence.

Questioning whether there is an appropriate board size and board composition, Coles et al. (2008) found that larger firms with complex advisory needs have larger boards with more outsider directors. Moreover, the appointment of CEOs as outside directors generates stronger market reactions and positive long-term performance effects compared to other outside directors (Fich, 2005), highlighting the nuanced role of independence within heterogeneous boards.

2.2.4. Board Oversight, Policies, and Practices

Board heterogeneity also manifests through formal oversight structures, governance policies, and board practices. Boards are increasingly expected to assume explicit responsibility for risk oversight, and higher board engagement is associated with greater risk-management maturity and lower volatility in corporate outcomes.

McNulty et al. (2013) demonstrate that board processes—such as information flows, timeliness, and engagement—and board effectiveness matter for firms’ exposure and response to the crisis, for better risk outcomes, while stronger board processes are associated with superior liquidity and financial risk outcomes. Similarly, Carter et al. (2007) argue that diversity across gender, independence, skills, and committee composition enhances monitoring quality and decision-making effectiveness.

As board complexity increases, formalized governance frameworks—such as specialized committees and board policies—contribute to more structured and purpose-aligned decision-making, with implications for corporate risk resilience.

2.2.5. Board Compensation

Compensation design represents a final dimension discussed herein, through which board heterogeneity affects corporate risk. Grounded on agency theory and managerial context, effort is only an overall incentive problem, prompting the use of incentive mechanisms to align managerial risk-taking with firm objectives (Fama, 1980; Holmström, 1999).

Boards play a central role in determining remuneration policies that balance risk-taking and value creation. Consequently, the literature examines how corporate governance mechanisms can encourage managers to undertake value-enhancing risk. These mechanisms include external monitoring by shareholders as well as internal incentives such as risk-sensitive compensation structures (Coles et al., 2006).

The board of directors is widely regarded as having a particularly significant influence on managerial risk-taking and the firm’s overall risk exposure. One function of the board is to determine remuneration policies in a way that aligns CEO incentives with a company’s objectives. Discussing female directors’ differences in terms of risk-taking tolerance, this may also be reflected in the compensation packages that gender-diverse boards offer to the CEO (Sila et al., 2016). Given documented gender differences in risk preferences, it is supposed that gender-diverse boards may influence compensation design in ways that reflect more balanced attitudes toward risk and potentially reconfiguration of firms’ risk exposure across operational, financial, liquidity, and failure dimensions.

3. Data and Research Methodology

3.1. Database

This study focuses on the European context. We selected all the companies’ constituents of STOXX600, namely 600 companies, and collected extensive data between 2015 and 2024. We have excluded financial industry sectors (GICS Sector Name) due to specific reporting regulations and accounting measures. Our sample analysis consists of 475 companies having their headquarters in 18 European countries. Primary data have been obtained from LSEG Workspace (accessed during August–September 2025). Multivariate regression techniques, including year- and firm-fixed effects, are employed to investigate the relationship between board gender diversity, along with other board diversity variables, and corporate risk outcomes. Some data for year-end 2024 might still have been unreported in August 2025. A few variables included in our study have been calculated by the author (like Operating Ratio or alternative measures for gender diversity). To improve the reliability of the data and prevent outliers from disproportionately skewing results, continuous variables were winsorized (capping the highest values at the 95th percentile and the lowest values at the 5th percentile), making the results more robust without removing the outliers completely. For variables highly skewed, a log transformation has been applied to make the data more symmetrical and less sensitive to outliers. For robust results, we have considered different measures for corporate risks and many governance variables, well distributed across board diversity categories, like structure, independence, CEO role, remuneration, committee, policy, and board practices.

Since we focus on European companies’ constituents of STOXX600, we acknowledge that our results may be biased by sample selection, in favor of publicly listed, internationalized, resilient, and prosperous companies from the European space, having large, mid-, and small-cap size, located in 18 European countries, included in the index. Our sample choice is primarily motivated by the research topic of—mature European boards and corporate governance, undoubtedly grounded on the UK Cadbury’s code (The Cadbury Archive—Publications, n.d.), where governance, regulatory, and gender equality legislations first started. Moreover, data availability, regulatory comparability, and the relevance of Europe as an early adopter of corporate governance and gender diversity reforms have been of reference for our choice. Another limitation may arise from corporate governance self-reported data, as the data might be systematically provided by companies that report. In case of some variables, like oversight practices or board policy, the data history is limited to the most recent years (not available at the beginning of the analyzed period); therefore, results need to be treated accordingly. Similarly, data for 2024 might still be under completion at the retrieval date. Our final sample consists of 475 companies after excluding the financial sector and 4712 firm-year observations distributed in an unbalanced panel for 10 years (2015–2024).

3.2. Distribution of Data and Sample Analysis

The companies included in our study have their headquarters in 18 European countries and are acting in 10 industrial sectors. 20.63% of observations belong to the UK, where the first framework for best practices in corporate governance began in 1991 (The Cadbury Archive—Publications, n.d.). Board structure, financial reporting, and accountability, transparency, and non-mandatory “comply or explain” compliance for listed companies sets the stage for modern governance codes globally, focusing on director responsibilities, audit committees, and shareholder rights to boost investor confidence. 12.63% of observations belong equally to companies from France and Germany, while 10.95% to Switzerland. Related activity domain, the best represented sector is Industrial (27.16%), followed by Consumer Discretionary (13.26), Health Care (10.95%), and Materials (10.11%). Further details are presented in

Table 1. Distribution of companies per country and industrial sectors.

Observations/Companies-Years	Communication Services	Consumer Discretionary	Consumer Staples	Energy	Health Care	Industrials	Information Technology	Materials	Real Estate	Utilities	Total
Austria	0	0	0	10	0	10	0	20	0	10	50
Belgium	0	10	20	0	20	20	0	10	30	10	120
Cyprus	0	0	0	10	0	0	0	0	0	0	10
Denmark	0	10	20	0	80	70	0	10	0	10	200
Finland	10	0	10	10	10	60	10	40	0	10	160
France	30	110	40	40	50	190	40	30	40	30	600
Germany	40	120	20	0	80	160	40	90	30	20	600
Ireland	0	0	20	0	0	50	0	0	0	0	70
Italy	20	40	10	20	30	40	10	10	0	60	240
Luxembourg	0	20	0	10	10	10	0	10	0	0	60
Netherlands	20	30	50	20	30	80	40	10	0	0	280
Norway	10	0	40	30	0	20	0	20	0	0	120
Poland	10	10	10	10	0	0	0	10	0	0	50
Portugal	0	0	10	10	0	0	0	0	0	10	30
Spain	20	20	20	10	10	20	10	10	10	70	200
Sweden	40	30	30	0	30	190	40	50	50	0	460
Switzerland	20	30	40	0	110	150	30	90	40	10	520
United Kingdom	70	200	120	30	60	220	50	70	90	70	980
Total	290	630	460	210	520	1290	270	480	290	310	4750

Source: Author’s own work.

,

Table 2. Sample distribution per country.

Nr.	Observations/Companies-Years	Total	%
1	Austria	50	1.05
2	Belgium	120	2.53
3	Cyprus	10	0.21
4	Denmark	200	4.21
5	Finland	160	3.37
6	France	600	12.63
7	Germany	600	12.63
8	Ireland	70	1.47
9	Italy	240	5.05
10	Luxembourg	60	1.26
11	Netherlands	280	5.89
12	Norway	120	2.53
13	Poland	50	1.05
14	Portugal	30	0.63
15	Spain	200	4.21
16	Sweden	460	9.68
17	Switzerland	520	10.95
18	United Kingdom	980	20.63
	Total	4750	100

Source: Author’s own work.

and

Table 3. Sample distribution per sectors.

Nr.	Observations/Sector-Years	Total	%
1	Communication Services	290	6.11
2	Consumer Discretionary	630	13.26
3	Consumer Staples	460	9.68
4	Energy	210	4.42
5	Health Care	520	10.95
6	Industrials	1290	27.16
7	Information Technology	270	5.68
8	Materials	480	10.11
9	Real Estate	290	6.11
10	Utilities	310	6.53
	Total	4750	100

Source: Author’s own work.

.

3.3. Variables

All dependent and independent variables included in this study, clustered as risk (R), board (B), and control variables (C), their description, and measurement are presented in

Table 4. Variables’ description and measurement.

Variable	R/ B/ C	Individual Board Diversity Component (Cluster)	Simbol	Description/Measurement
Dependent Variables (Corporate Risk)
Operating Ratio *	R		OpR_w	Operating Ratio = operating expenses/net sales (%)
Debt to Equity *	R		LevDE_w	(Total Debt/Total Equity) × 100 (%)
Debt to Assets *	R		LevDA_w	(Total Debt/Total Assets) × 100 (%)
ZScore Altman *	R		ZScoreMNM_w	The Z-score is a multivariate formula that measures the financial health of a company and predicts the probability of bankruptcy within two years; combines five common business ratios using a weighting system calculated by Altman to determine the likelihood of bankruptcy. Typically, a score below 1.88 indicates that a company is likely heading for or is under the weight of bankruptcy. Conversely, companies that score above 2.99 are less likely to experience bankruptcy. Score between 2.99 and 1.88 is questionable. (MNM = manufacturing/non-manufacturing)
Current Ratio *	R		CurrR_w	Represents Total Current Assets Less Inventory divided by Total Current Liabilities (%)
Change of Long-Term Debt per Capital *	R		ChgLTDpK_w	Change in Long Term Debt/Total Capital ratio; Long Term Debt divided by Total Capital at the end of the fiscal period, and is expressed as percentage (%)
Quick Ratio *	R		QR_w	Represents Total Current Assets Less Inventory divided by Total Current Liabilities (%)
Independent Variable (Individual Board Diversity components)
Board Gender Diversity *	B	Gender	BGD	Percentage of females on the board (%)
WoB less 20% (Mere) *	B	Gender	WOBMere	New Created Binary Variable: if minimum female presence on board is <20% = 1, else = 0
WOB ≥ 20% AND < 40% (Token) *	B	Gender	WOBTK	New Created Binary Variable: if minimum female presence on board is ≥20% and <40% = 1, else = 0
WOB ≥ 40% (Balanced) *	B	Gender	WOBBal	New Created Binary Variable: if minimum female presence on board is ≥40% = 1, else = 0
Board Size *	B	Structure	BSize_w	Total number of Board Directors
Average Board Tenure *	B	Structure	BTenure_w	Average number of years each board member has been on the board (years)
Board Member Compensation *	B	Remuneration	ln_BMComp_w	Total compensation of the board members (EUR)
Total Senior Executives Compensation	B	Remuneration	ln_ExecComp_w	The total compensation (EUR) paid to all senior executives as reported by the company (natural logarithm)
Board Effectiveness Review *	B	Oversight	BER	Does the board conduct a periodic review of its effectiveness? Board effectiveness review represents the process of assessing and analyzing the effectiveness and efficiency of a company’s board of directors in fulfilling its responsibilities and achieving the business goals. The evaluation must be conducted with the objective of checking the effectiveness of the entire board. The objective of the evaluation should be to focus only on the effectiveness of the board. The company must have a clear time frame for this specific process. (0/1)
Board Meeting Attendance Average *	B	Oversight	BMeet_w	The average overall attendance percentage of board meetings as reported by the company (overall board members conduct regular meetings during the year; board meeting average is the attendance average provided details of members attended versus the total number of board meetings held) (%)
Compliance Head Report to Board *	B	Oversight	HComplB	Does the company have a head of compliance on the board, on board committees, or regularly reports directly to board or board committee meetings? This datapoint is to ascertain that executive officers responsible for legal and compliance have a direct report to the board of directors. Considered if the head of compliance on the board/board committees or regularly reports directly to board/board committee meetings. (0/1)
Code of Conduct Non-Compliance	B	Oversight	CCondNonc	Does the company have procedures to address non-compliance with the code of conduct (or equivalent)? Code of conduct non-compliance represents a breach or violation of the established rules, guidelines, and ethical standards that an organization or community has put in place to govern the behavior and interactions of its members or employees. Considered if there is an investigation in place and follow up on any non-compliance with its code of conduct (or equivalent). (0/1)
CSR Sustainability Committee *	B	Committee	CSRCom	Does the company have a CSR committee or team (board level or Senior management committee responsible for decision making on CSR strategy)? (0/1)
Board Structure Policy *	B	Policy	BSP	Does the company have a policy for maintaining a well-balanced membership of the board? (0/1)
Board Functions Policy *	B	Policy	BFP	Does the company have a policy for maintaining effective board functions? (0/1)
Policy Board Diversity *	B	Policy	PBD	Does the company have a policy regarding the gender diversity of its board? The company strives to maintain a well-balanced board through adequate female and/or intercultural (race, religion, culture) representation on the board (0/1)
Board Member Affiliations *	B	Policy	BMAfill_w	Average number of other corporate affiliations for the board member (number)
Board Individual Re-election	B	Policy	BIReel	Are all board members individually subject to re-election (not classified or staggered board structure)? (0/1)
Board Member Membership Limits	B	Policy	ln_MaxLenght_w	The maximum number of years a board member can be on board as stipulated by the company. Takes value 1 to 30 years/insufficient information/no limit; when the company has explicitly mentioned that board members will be on board only for a certain maximum number of years; when the maximum term is different for a different class of directors, use the one given for the independent/nonexecutive directors; if there is a law provision that says that directors have to retire after a number of years, then answer as per the provision (years)
Board Member Term Duration	B	Policy	BMReelect_w	The smallest interval of years in which the board members are subject to re-election. Annual re-election for board members who have served for a long time is accepted as “1” year. If data mention that about one-third of board members must retire at the AGM, then it is 3 years. (years)
Independent Board Members *	B	Independence	IndBM_w	Percentage of independent board members as reported by the company. (%)
Chairperson Independent *	B	Independence	ChairInd	Is the company chairperson independent? Chairperson independent, represents the company’s chairperson who is not affiliated with the company’s management or major shareholders, and provides impartial leadership and oversight. Considered if independent status of the chairperson is explicitly disclosed. (0/1)
Non-Executive Board Members	B	Independence	NonExBM_w	Percentage of non-executive board members. (%)
CEO Chairman Duality *	B	CEO	CEODual	Does the CEO simultaneously chair the board or has the chairman of the board been the CEO of the company? (0/1)
Chairman is ex-CEO	B	CEO	ChairExCEO	Has the chairman held the CEO position in the company prior to becoming the chairman? True: when Chairman was CEO in previous years. False: when the chairman was never the CEO of the company. False: when the chairman is currently the CEO (0/1)
Firm variables (control)
Quota History	C		QuotaHist_w	Number of years since gender quota has been implemented (if case) (years)
Date of Incorporation	C		Age_w	Number of years since incorporation (years)
Company Market Capitalization YE	C		ln_MarketCapYE_w	Company Market Capitalization at the end of the year (EUR)
Effective Tax Rate	C		ETR_w	Represents Provision for Income Taxes divided by Income Before Tax. (%)
Cash and Short-Term Investments	C		ln_CSTI_w	Cash and Short-Term Investments (EUR)
Return On Equity (ROE)	C		ROE_w	Return on Equity (%)
A/R Turnover	C		ART_w	This item measures the number of times Receivables are cycled through in each period. It is calculated as Primary Revenue for the fiscal period divided by the Average Total Net Receivables for the same period (number)
Inventory Turnover	C		InvTo_w	This is the ratio of Total Cost of Revenue for the fiscal period to the average Total Inventory for the same period. (%)
Dividend Yield	C		DivYield_w	This value is the percentage dividend yield calculated as the Dividends paid per share to the primary common shareholders for the fiscal period divided by the Historical Price Close, multiplied by 100 (%)
Free Cash Flow	C		ln_FCF_w	This item represents Cash Flow excluding Capital Expenditures and Total Cash Dividends Paid for the fiscal year. (EUR)
ROA Total Assets	C		ROA_w	Net Income/shareholders’ equity (%)
Operating Expenses	C		ln_OE_w	Operating Expenses total per year (EUR)
Net Sales	C		SALES	Net sales (EUR)
Working Capital	C		WCap_w	This item is defined as the difference between Current Assets and Current Liabilities for the fiscal period. (EUR)
EBIT	C		EBIT_w	EBIT is computed as Total Revenues for the fiscal year minus Total Operating Expenses Plus Operating Interest Expense, Unusual Expense/Income [SUIE], and Non-Recurring Items, Supplemental, Total [SUIT] for the same period. This definition excludes non-operating income and expenses. (EUR)

Source: All definitions have been retrieved from LSEG Workspace. Type R denotes a Risk variable, type B, denotes a board variable, while type C denotes firm-specific control variables. Variable marked by * denotes our main dependent and independent variable (risk dependent variables and gender diversity) accompanied by several largely discussed, in this paper, board heterogeneity variables. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

. Although this study is centered around gender diversity and corporate risk outcomes, there are several board heterogeneity dimensions that are largely discussed in the results Section, for their influence, confounding effect, or possible suppressor effect of gender diversity on corporate risk; therefore, they are marked by * in the table below. Firm-specific time-series data for all variables included in this study have been downloaded from LSEG. Following prior studies, we have analyzed 4 risk dimensions proxied by seven risk variables: operational risk (Operating Ratio), financial risk (Debt to Assets, Debt to Equity, Changes to Long Term Debt per Capital), bankruptcy risk (Z-Score), and liquidity risk (Current Ratio and Quick Ratio).

We built our study on previously proposed clusters of board heterogeneity—social and occupational, or relation-oriented and task-oriented (Anderson et al., 2011; Jebran et al., 2020)—extending analysis to most of the available individual board diversity dimensions, available in LSEG. We selected and analyzed board heterogeneity in alternative models in such a way to include at least one individual component of the following board diversity category: gender, structure, remuneration, oversight, committee, policy, independence, and CEO.

The board is a complex structure that continuously evolves. Therefore, we are interested in analyzing the individual components of board diversity, rather than a diversity index. For this purpose, we propose several models, gradually including as many variables as possible and interactions, on a best effort basis, depending on their availability in LSEG, aiming to avoid possible misspecifications or omitted variable bias.

Dependent variables are the following proxies for corporate risk outcomes: (1) Operating Ratio (OpR), (2) Debt to assets (LevDA), (3) Debt to Equity (LevDE), (4) Z-Score manufacturing/non-manufacturing (ZScoreMNM), Current Ratio (CurrR), and Quick Ratio (QR).

The main independent variable of interest is board gender diversity (BGD), measured as the percentage of women on the total board members. As an alternative measure of gender diversity, we created three new dummy variables measuring the presence of women on board (WOB), that is, mere (WOBMere), token (WOBTK), balanced (WOBBal), in line with critical mass theory’s categories—skewed, tilted, and balanced—defended by Kanter (1977).

Following the literature (Coles et al., 2008), we controlled for a possible non-linear relation between gender and corporate risks, including the squared term of BGD (sqBGD) in our specifications.

3.4. Model Specifications

Building on previous research (Bernile et al., 2018; Jebran et al., 2020) to explain the nexus between board gender diversity and other individual diversity components, and corporate risks, we propose the following main model:

The general form of the regression model used in the analysis is presented in Equation (1), below:

$$\begin{array}{c}{Risk}_{it}={\beta }_0+{\beta }_1{BGD}_{it}+{\beta }_2{BoardStructure}_{it}+{\beta }_3{BoardCEO}_{it}+{\beta }_4{BoardIndependence}_{it}+{\beta }_5{BoardOversight}_{it}+\\ {\beta }_6{BoardPolicy}_{it}+{\beta }_7{BoardCompensation}_{it}+{\beta }_8{BoardCommittee}_{it}+\sum {\beta }_{it}{Firm}_{it}+{\epsilon }_{it}\end{array}$$

(1)

Equation (2) presents the regression model used to estimate corporate risk in case a possible non-linear influence of BGD is present:

$$\begin{array}{c}{Risk}_{it}={\beta }_0+{\beta }_1{BGD}_{it}+{{\beta }^{\prime }}_1{sqBGD}_{it}+{\beta }_2{BoardStructure}_{it}+{\beta }_3{BoardCEO}_{it}+{\beta }_4{BoardIndependence}_{it}+\\ {\beta }_5{BoardOversight}_{it}+{\beta }_6{BoardPolicy}_{it}+{\beta }_7{BoardCompensation}_{it}+{\beta }_8{BoardCommittee}_{it}+\sum {\beta }_{it}{Firm}_{it}+{\epsilon }_{it}\end{array}$$

(2)

were:

• ${Risk}_{it}$ = dependent variables
• ${\beta }_0$ = constant
• ${\beta }_1,{{\beta }^{\prime }}_1$ = coefficients corresponding to board gender diversity and its squared term
• ${\beta }_{2,8}$ = coefficients corresponding to board diversity categories
• ${\beta }_{it}$ = coefficients corresponding to firm-related attributes control variable
• ${\epsilon }_{it}$ = error term
• i = [1, 475]
• t = [2015, 2024]

Our empirical results are grounded in a series of regression models constructed specifically for this research. These include linear models without effects and fixed and random effects regressions.

3.5. Regression Models, Tests, and Alternative Specifications

We have followed a systematic approach in selection of our models to test our hypothesis: (1) correlation analysis to select the variables included in each specification, (2) R-sq to assess the proportion of corporate risks variance explained by board diversity components, (3) VIF to control for multicollinearity concerns of independent variables, (4) comparing OLS regression results with fixed and random firm-year effects models, accounting for unobserved firm’s characteristics by removing the within-firm variation over time and controlling for unobserved time-invariant effects that might affect all companies in the same year, (5) Hausman test for appropriateness and efficiency of models, testing if fixed or random effects better fits for our panel data, (6) alternative specifications for dependent variables and main independent variable, (7) interaction variable to test whether the effect of main independent variable on dependent variables changes depending on the level of other observed variables and their possible synergistic effect.

4. Results

4.1. Descriptive Statistics

Three econometric models have been gradually developed, starting with the initial set of variables, expected to be the most relevant to our research. To better discuss the hypothesis, more variables have been gradually included in our analysis, considering adding value and improving the explanatory power of our models. During the model’s development, all board heterogeneity dimensions (described in Table 4) were considered for specifications.

Descriptive statistics examine the mean, minimum, maximum, and standard deviation for all variables included in models constructed for this research, as presented in

Table 5. Descriptive statistics.

	Count	Mean	sd	Min	Max
OpR_w	4392	0.8252633	0.1496791	0.392071	0.983615
LevDE_w	4609	0.814859	0.6711203	0.0200268	2.512473
LevDA_w	4642	0.2582427	0.1418102	0.0135369	0.5225217
ZScoreMNM_w	4458	3.988961	3.270041	0.7513725	13.39517
CurrR_w	4642	1.486761	0.7676492	0.52786	3.52528
ChgLTDpK_w	4600	0.0016562	0.0596656	−0.1079096	0.1392491
QR_w	4642	1.072931	0.5342007	0.38479	2.51293
BGD	4164	33.19426	10.71386	12.5	50
WOBMere	4164	0.1256004	0.3314382	0	1
WOBTK	4750	0.4766316	0.4995062	0	1
WOBBal	4164	0.3306916	0.4705187	0	1
BSize_w	4159	10.76052	3.276113	6	18
MaxLenght_w	449	11.60134	1.971697	9	15
ln_BMComp_w	4055	13.81337	0.7968898	12.34583	15.39418
BER	1749	0.7564322	0.4293575	0	1
BSP	4166	0.987278	0.1120857	0	1
BMeet_w	3627	96.77816	3.202349	88.89	100
BFP	4166	0.2405185	0.4274496	0	1
PBD	4162	0.8399808	0.3666679	0	1
BMAfill_w	4164	1.028875	0.6301427	0.125	2.266667
BIReel	4165	0.7030012	0.4569909	0	1
ChairInd	3797	0.5001317	0.5000658	0	1
CEODual	4165	0.2309724	0.4215054	0	1
ChairExCEO	4163	0.2310834	0.4215762	0	1
BTenure_w	4135	6.445258	2.267536	3.021739	11.5
IndBM_w	4164	63.81465	21.15755	23.07692	100
NonExBM_w	4164	86.37368	12.11066	62.5	100
CCondNonc	1749	0.5054317	0.5001135	0	1
HComplB	3792	0.4382911	0.4962428	0	1
HLegalB	3792	0.0184599	0.1346251	0	1
BMReelect_w	4065	2.330908	1.461583	1	5
CSRCom	4162	0.8565593	0.3505638	0	1
ln_ExecComp_w	4054	15.86856	0.9771402	13.96393	17.52642
QuotaHist_w	3468	7.369089	3.832804	1	14
Age_w	4264	40.41581	32.4293	4	116
ln_MarketCapYE_w	4505	22.91506	1.06628	21.2435	25.05894
ETR_w	4237	0.2285837	0.1036782	0.0101414	0.4643
ART_w	4584	7.220032	5.544977	2.155132	25.10094
ln_CSTI_w	4644	20.27202	1.63226	17.17894	23.05687
ln_FCF_w	3551	19.34823	1.43159	10.83958	21.67028
InvTo_w	4124	10.53497	15.16843	0.9478096	62.51196
DivYield_w	4348	0.0265603	0.0183346	0	0.0679214
ROA_w	4602	0.0609126	0.0505735	−0.0283803	0.1760975
ROE_w	4428	0.1692251	0.1078325	0.0112	0.4415
ln_OE_w	4641	22.07173	1.578651	18.70932	24.64526
ln_WCap_w	3348	20.40505	1.439911	13.75257	22.64712
ln_EBIT_w	4641	20.19271	1.38449	17.30289	22.73456

Source: Author’s own work. For definitions of variables, please see Table 4. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

.

Based on this initial analysis, there are several conclusions relevant to our testing hypothesis. While the companies exhibit an average of 33% presence of women on boards, there is still a disparity in gender balance strategies in the case of large companies’ boards, ranging from merely having women present (12.5%) to already gender-balanced boards (50%). This reality has been intuitively used for our constructed gender diversity variables (dummy), WOBMere (<20%), WOBTK (20–40%), and WOBBal (>40%), used to alternatively measure the inference of gender diversity on corporate risks. In case of large companies in our sample, 12.6% of companies barely have a woman on board, 47.7% appointed between 20% and less than 40% women on their boards, while only 33% reported a gender balanced board (considered for our study if equal to or greater than 40%). Board size spans between 6 and 18 members, correlated with the large firm’s complexity, while the maximum length of board service is 15 years, as stipulated in the companies’ governance policies. In more than 75% of cases (1749), companies that have reported about its effectiveness practices stated that the board conducts a periodic review of its effectiveness, 84% of them have a policy regarding the gender diversity of their board, and almost all large companies (99%) have a policy for maintaining a well-balanced membership of the board. Only 24% of cases (4166) confirmed that companies have a policy for maintaining effective board functions. Most of the board members have one mandate on average, with a maximum of 2.3 mandates. More than 70% of cases are board members who are individually subject to re-election (they do not belong to a classified or staggered board structure). Related to the independence a good board is expected to provide, 50% of chairmen of the board are independent, while 23% of CEO are also presidents of the board, and in 23% of cases, the president is the ex-CEO. Average board tenure is 6.4 years, with the longest tenure of 11.5 years, which reflects the preference of large and complex companies toward insider directors, with company-specific knowledge, while the board member reelection policy is, on average, within less than 3 years, and between 1 and 5 years. Nevertheless, on average, 63% of directors of large companies’ boards are independent, ranging from a minimum of 23% to even 100%, while 86% of observed cases are non-executive directors, supporting agency theory concerns and the good practices of external oversight. In 50% of cases (1749), the companies have clear procedures to report breaches of the code of conduct. Code of conduct non-compliance represents a breach or violation of the established rules, guidelines, and ethical standards that an organization or community has put in place to govern the behavior and interactions of its members or employees (LSEG Workspace definition). In 43% of cases, the head of compliance is either on the board, on board committees, or regularly reports directly to board or board committee meetings, while only 1.8% of legal managers in our sample validate the same oversight level. Lastly, 87% of companies in our sample reported a CSR committee in place between 2015 and 2024.

Analyzing the institutional context and the possible impact of a legally binding gender quota, we created and measured the quota history, defined as the number of years since the country where the company is headquartered implemented quota legislation. The average quota history of our sample is 7.4 years, while the longest history is 14 years. Therefore, we expect that the presence of women on the board depends on gender-equality institutional maturity, not only on company-specific evolutions.

While the companies from our sample are similar in terms of market capitalization, their working history (company age) is on average 40 years, the youngest being in the market for 4 years. Related performance, average return on assets is 6%, and capital rentability is almost 17%, with a dividend yield of 2.6%.

4.2. Correlation Matrix

Considering the multiple variables we selected for our study, there is a need to identify possible correlations between our independent variables and dependent ones, as highly correlated variables might influence the interpretation of our results. Considering that the strongest correlation between variables exceeds a 0.7 value of correlation coefficients, we proposed our three estimation models for corporate risks, including corporate governance variables that do not show stronger correlation. Strong correlation between dependent variables or gender diversity variables is not considered, since they are included in different specification models.

The correlation matrix provides initial insights relevant to our hypothesis-testing framework. Looking at the influence of our main independent variable, board gender diversity (BGD), on corporate risk, the correlation matrix shows that BGD negatively influences operating ratio, Z-score, and liquidity ratio, whereas BGD positively influences the level of debt and has no influence on long-term debts. Despite these initial results not providing strong evidence of the level of impact BGD has on corporate risks, it is consistent with our predicted hypothesis about the sense of the relations, except for the Z-score, for which the prediction is positive (H3)

The correlation analysis between dependent variables and corporate governance independent variables is presented in

Table 6. Pearson Correlation Matrix.

	OpR_w	LevDE_w	LevDA_w	ZScoreMNM_w	CurrR_w	ChgLTDpK_w	QR_w	BGD	BSize_w	ln_BMComp_w
OpR_w	1
LevDE_w	0.0255	1
LevDA_w	−0.156 ***	0.826 ***	1
ZScoreMNM_w	−0.193 ***	−0.498 ***	−0.511 ***	1
CurrR_w	0.0274	−0.401 ***	−0.400 ***	0.513 ***	1
ChgLTDpK_w	0.0506 ***	0.0939 ***	0.0936 ***	−0.0264	0.0389 **	1
QR_w	−0.0409 **	−0.275 ***	−0.283 ***	0.490 ***	0.830 ***	0.0308 *	1
BGD	−0.0206	0.121 ***	0.133 ***	−0.120 ***	−0.119 ***	0.00315	−0.0895 ***	1
BSize_w	0.144 ***	0.198 ***	0.107 ***	−0.347 ***	−0.207 ***	−0.0115	−0.134 ***	0.127 ***	1
ln_BMComp_w	0.144 ***	0.141 ***	0.0581 ***	−0.200 ***	−0.0948 ***	0.0114	−0.0459 **	−0.0319 *	0.418 ***	1
BER	0.0682 **	0.110 ***	0.0704 **	−0.123 ***	−0.137 ***	0.00147	−0.102 ***	0.136 ***	0.209 ***	0.164 ***
BSP	0.0208	0.00769	−0.0257	−0.0000652	−0.00368	0.00334	−0.0128	0.0872 ***	0.0493 **	0.00908
BMeet_w	−0.0104	−0.0672 ***	−0.0428 **	0.0965 ***	0.0726 ***	−0.0131	0.0106	−0.00121	−0.180 ***	0.0458 **
BFP	−0.0150	0.101 ***	0.0929 ***	−0.152 ***	−0.0913 ***	0.0168	−0.0400 **	0.144 ***	0.150 ***	0.190 ***
PBD	0.00000329	0.111 ***	0.0747 ***	−0.112 ***	−0.111 ***	0.0210	−0.0999 ***	0.233 ***	0.168 ***	0.00126
BMAfill_w	0.134 ***	0.0454 **	0.0190	−0.115 ***	−0.0769 ***	0.00532	−0.0606 ***	0.109 ***	0.0704 ***	0.222 ***
MaxLenght_w	−0.0805	0.0620	0.0626	0.0404	0.0434	−0.0430	0.153 **	0.0885	0.261 ***	0.00920
BIReel	0.0000590	−0.111 ***	−0.104 ***	0.165 ***	0.158 ***	0.0239	0.0691 ***	−0.159 ***	−0.228 ***	0.0560 ***
ChairInd	−0.00379	−0.0485 **	−0.00862	−0.0199	0.0232	−0.0250	−0.0178	0.0389 *	−0.0762 ***	0.0432 **
CEODual	−0.00440	0.0254	−0.000339	0.00238	−0.0560 ***	−0.00883	−0.00847	0.0989 ***	0.158 ***	0.0312 *
BTenure_w	−0.00830	−0.126 ***	−0.125 ***	0.148 ***	0.0665 ***	−0.0119	0.0450 **	−0.140 ***	−0.00550	−0.00193
IndBM_w	−0.00568	0.0106	0.0732 ***	−0.0349 *	−0.0146	0.0121	−0.0140	0.0637 ***	−0.232 ***	0.104 ***
NonExBM_w	0.0972 ***	0.0263	0.0243	−0.0846 ***	−0.0515 ***	−0.0139	−0.00880	0.0538 ***	0.152 ***	0.119 ***
CCondNonc	0.0288	0.0722 **	0.0149	−0.224 ***	−0.114 ***	−0.0346	−0.142 ***	0.0420	0.133 ***	0.119 ***
HComplB	−0.0799 ***	0.110 ***	0.105 ***	−0.00605	0.0175	−0.000256	0.0321 *	−0.0544 ***	−0.0206	0.124 ***
BMReelect_w	0.0680 ***	0.151 ***	0.0864 ***	−0.167 ***	−0.0931 ***	−0.0204	−0.0316 *	0.101 ***	0.378 ***	0.0102
CSRCom	0.0685 ***	0.137 ***	0.0974 ***	−0.220 ***	−0.0999 ***	0.0150	−0.107 ***	0.238 ***	0.238 ***	0.215 ***
ln_ExecComp_w	0.157 ***	0.0854 ***	0.0187	−0.132 ***	−0.108 ***	0.0172	−0.0759 ***	0.0709 ***	0.279 ***	0.510 ***
QuotaHist_w	−0.0836 ***	0.102 ***	0.118 ***	−0.110 ***	−0.0286	−0.00175	−0.0378 *	0.391 ***	0.0214	−0.113 ***
Age_w	0.112 ***	−0.0523 ***	−0.0669 ***	−0.0437 **	−0.00833	0.00829	−0.0810 ***	0.0600 ***	0.0552 ***	0.0789 ***
ln_MarketCapYE_w	0.00757	0.133 ***	0.0885 ***	−0.0934 ***	−0.156 ***	0.00304	−0.107 ***	0.172 ***	0.406 ***	0.475 ***
ETR_w	0.297 ***	0.118 ***	0.0224	−0.156 ***	−0.0713 ***	0.0655 ***	−0.0724 ***	0.125 ***	0.140 ***	0.0548 ***
ART_w	−0.0398 **	−0.107 ***	−0.0747 ***	0.113 ***	0.0569 ***	0.00651	−0.154 ***	−0.0902 ***	−0.165 ***	−0.0980 ***
InvTo_w	0.0697 ***	0.177 ***	0.112 ***	−0.125 ***	−0.234 ***	−0.0128	−0.00120	0.000486	0.0381 *	−0.0640 ***
DivYield_w	−0.106 ***	0.152 ***	0.129 ***	−0.307 ***	−0.114 ***	−0.0147	−0.112 ***	0.0743 ***	0.0856 ***	0.153 ***
ROA_w	−0.319 ***	−0.325 ***	−0.313 ***	0.649 ***	0.274 ***	−0.0869 ***	0.208 ***	−0.0548 ***	−0.250 ***	−0.138 ***
ln_OE_w	0.584 ***	0.215 ***	0.0272	−0.427 ***	−0.236 ***	0.0182	−0.246 ***	0.158 ***	0.475 ***	0.484 ***
ROE_w	−0.102 ***	0.103 ***	−0.0665 ***	0.370 ***	0.00488	−0.0478 **	0.0107	0.0125	−0.107 ***	0.0182
ln_WCap_w	0.181 ***	0.0172	−0.0421 *	−0.250 ***	0.170 ***	0.00159	0.0626 ***	0.109 ***	0.357 ***	0.404 ***
ln_EBIT_w	−0.0515 ***	0.191 ***	0.128 ***	−0.298 ***	−0.240 ***	−0.0440 **	−0.232 ***	0.209 ***	0.414 ***	0.440 ***
	BER	BSP	BMeet_w	BFP	PBD	BMAfill_w	MaxLenght_w	BIReel	ChairInd	CEODual
BER	1
BSP	0.119 ***	1
BMeet_w	0.0607 *	0.0127	1
BFP	0.158 ***	0.0488 **	−0.0643 ***	1
PBD	0.186 ***	0.260 ***	0.112 ***	0.0842 ***	1
BMAfill_w	0.216 ***	0.108 ***	0.0207	0.147 ***	0.0743 ***	1
MaxLenght_w	−0.308 ***	0.189 ***	−0.186 ***	−0.145 **	0.195 ***	−0.256 ***	1
BIReel	−0.0978 ***	−0.0363 *	0.214 ***	−0.180 ***	−0.0946 ***	−0.0176	−0.218 ***	1
ChairInd	−0.0516 *	0.00495	0.0857 ***	−0.0192	0.0461 **	0.0588 ***	0.104 *	0.134 ***	1
CEODual	0.0228	0.0317 *	−0.159 ***	0.181 ***	−0.0921 ***	0.0102	−0.0451	−0.277 ***	−0.364 ***	1
BTenure_w	−0.0453	−0.0431 **	−0.0500 **	−0.0801 ***	−0.165 ***	−0.0568 ***	0.116 *	−0.0844 ***	−0.183 ***	0.238 ***
IndBM_w	0.0448	0.0630 ***	0.104 ***	0.0929 ***	0.0423 **	0.180 ***	−0.124 **	0.0802 ***	0.437 ***	−0.173 ***
NonExBM_w	−0.124 ***	−0.0424 **	−0.134 ***	0.0369 *	−0.0946 ***	−0.0737 ***	0.558 ***	−0.157 ***	0.274 ***	−0.0425 **
CCondNonc	0.129 ***	0.0177	0.0661 **	0.0863 ***	0.126 ***	0.0517 *	−0.198 **	0.0281	0.0151	−0.0858 ***
HComplB	0.214 ***	0.0312	0.159 ***	0.122 ***	0.0993 ***	0.112 ***	−0.549 ***	0.152 ***	−0.0961 ***	−0.0369 *
BMReelect_w	−0.0127	0.0142	−0.200 ***	0.0913 ***	0.171 ***	−0.188 ***	0.768 ***	−0.532 ***	−0.0582 ***	0.187 ***
CSRCom	0.163 ***	0.100 ***	0.00576	0.155 ***	0.190 ***	0.127 ***	−0.123 **	−0.0396 *	0.0560 ***	0.0485 **
ln_ExecComp_w	0.138 ***	0.0702 ***	0.0719 ***	0.105 ***	0.117 ***	0.268 ***	−0.0837	0.0693 ***	0.106 ***	−0.0394 *
QuotaHist_w	0.196 ***	0.0974 ***	0.143 ***	0.101 ***	0.182 ***	0.0390 *	−0.379 ***	−0.0638 ***	−0.0790 ***	0.0757 ***
Age_w	0.0746 **	0.0131	0.0427 *	0.0198	0.00582	0.147 ***	0.0474	−0.0237	0.0000940	0.0122
ln_MarketCapYE_w	0.199 ***	0.101 ***	−0.0395 *	0.201 ***	0.124 ***	0.309 ***	0.0399	−0.154 ***	−0.0189	0.133 ***
ETR_w	0.0334	0.0428 **	−0.0637 ***	0.00784	0.0238	0.0166	0.126 *	−0.0788 ***	−0.0428 *	0.0323 *
ART_w	0.00975	0.0338 *	0.0277	−0.0814 ***	0.0458 **	−0.0643 ***	−0.131 **	0.100 ***	−0.0231	−0.0783 ***
InvTo_w	0.0182	−0.0114	−0.0685 ***	0.0217	0.0990 ***	−0.0609 ***	0.0271	−0.0899 ***	−0.0363 *	−0.00811
DivYield_w	0.0638 **	−0.00422	0.0306	0.0643 ***	−0.00464	0.0880 ***	−0.270 ***	0.0412 **	0.0324	−0.0652 ***
ROA_w	−0.0287	0.0137	0.0502 **	−0.112 ***	−0.0450 **	−0.0529 ***	−0.0551	0.131 ***	−0.0416 *	0.0121
ln_OE_w	0.216 ***	0.103 ***	−0.0612 ***	0.184 ***	0.132 ***	0.308 ***	−0.0486	−0.163 ***	−0.0146	0.0864 ***
ROE_w	0.0369	0.0295	0.0226	−0.0389 *	0.0265	0.0649 ***	−0.166 ***	0.127 ***	−0.0254	0.00132
ln_WCap_w	0.0969 ***	0.0564 **	−0.0154	0.177 ***	0.0516 **	0.231 ***	0.0971	−0.116 ***	0.0338	0.0788 ***
ln_EBIT_w	0.204 ***	0.0950 ***	−0.0755 ***	0.213 ***	0.124 ***	0.271 ***	0.00815	−0.191 ***	−0.0167	0.108 ***
	BTenure_w	IndBM_w	NonExBM_w	CCondNonc	HComplB	BMReelect_w	CSRCom	ln_ExecComp_w	QuotaHist_w	Age_w
BTenure_w	1
IndBM_w	−0.198 ***	1
NonExBM_w	−0.0413 **	0.169 ***	1
CCondNonc	−0.0804 ***	−0.00194	0.0421	1
HComplB	−0.0931 ***	−0.0494 **	−0.532 ***	0.113 ***	1
BMReelect_w	0.0221	−0.146 ***	0.354 ***	0.0208	−0.235 ***	1
CSRCom	−0.133 ***	0.0700 ***	0.0770 ***	0.137 ***	0.0458 **	0.0621 ***	1
ln_ExecComp_w	−0.0168	0.124 ***	0.0392 *	0.110 ***	0.111 ***	−0.0156	0.204 ***	1
QuotaHist_w	−0.0874 ***	−0.0578 **	−0.219 ***	0.0479	0.257 ***	−0.0630 ***	0.123 ***	−0.0542 **	1
Age_w	0.161 ***	−0.0616 ***	−0.0301	0.0954 ***	−0.0395 *	−0.0978 ***	0.0656 ***	0.0661 ***	0.0560 **	1
ln_MarketCapYE_w	0.0615 ***	0.0992 ***	0.152 ***	0.122 ***	−0.0493 **	0.160 ***	0.245 ***	0.462 ***	0.000305	0.101 ***
ETR_w	−0.0522 **	−0.0653 ***	0.105 ***	0.0586 *	−0.0951 ***	0.126 ***	0.0899 ***	0.0701 ***	0.0572 **	0.0236
ART_w	0.0373 *	−0.0246	−0.148 ***	−0.0186	0.121 ***	−0.0969 ***	−0.0598 ***	−0.0325 *	−0.0213	−0.0607 ***
InvTo_w	−0.0997 ***	−0.00671	−0.0225	−0.0322	0.0548 **	0.122 ***	−0.0183	−0.0242	−0.0288	−0.133 ***
DivYield_w	−0.132 ***	0.112 ***	0.0143	0.106 ***	0.0531 **	−0.0688 ***	0.106 ***	0.0906 ***	0.137 ***	0.0653 ***
ROA_w	0.136 ***	0.0144	−0.119 ***	−0.153 ***	−0.00210	−0.200 ***	−0.158 ***	−0.0851 ***	−0.0711 ***	0.0136
ln_OE_w	−0.00631	0.0643 ***	0.180 ***	0.174 ***	−0.0772 ***	0.215 ***	0.290 ***	0.455 ***	0.000955	0.168 ***
ROE_w	−0.0220	0.0417 **	−0.0770 ***	−0.0672 **	0.0671 ***	−0.169 ***	−0.0311 *	0.0642 ***	−0.0409 *	−0.00942
ln_WCap_w	0.0482 **	0.0835 ***	0.157 ***	0.165 ***	−0.0788 ***	0.199 ***	0.259 ***	0.364 ***	0.00730	0.119 ***
	ln_MarketCapYE_w	ETR_w	ART_w	InvTo_w	DivYield_w	ROA_w	ln_OE_w	ROE_w	ln_WCap_w	ln_EBIT_w
ln_MarketCapYE_w	1
ETR_w	0.0923 ***	1
ART_w	−0.0986 ***	−0.0692 ***	1
InvTo_w	−0.0911 ***	−0.00950	0.0747 ***	1
DivYield_w	0.0965 ***	0.0130	0.0190	0.0567 ***	1
ROA_w	0.0286	−0.306 ***	0.107 ***	−0.128 ***	−0.0563 ***	1
ln_OE_w	0.653 ***	0.311 ***	−0.0510 ***	0.0341 *	0.221 ***	−0.254 ***	1
ROE_w	0.165 ***	−0.0663 ***	0.0211	−0.0258	0.0399 **	0.658 ***	0.0296 *	1
ln_WCap_w	0.609 ***	0.145 ***	−0.0182	−0.157 ***	0.212 ***	−0.160 ***	0.681 ***	−0.128 ***	1
ln_EBIT_w	0.802 ***	0.182 ***	−0.0597 ***	−0.0870 ***	0.353 ***	0.0371 *	0.714 ***	0.195 ***	0.618 ***	1

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

.

The complexity of corporate boards’ dynamics might be explained by the preliminary relation observed between BGD and the other individual board diversity attributes included in our model. Thus, we find a positive relation between BGD and board size (0.127, p < 0.01), memberships (0.109, p < 0.01), and proportion of independent board members (0.0637, p < 0.01), all three suggesting the presence of the additive approach, discussed in literature exploring the critical mass theory, within large companies. Authors (Seierstad & Opsahl, 2010) have identified an unintended consequence of gender quotas, by the rise of the “golden skirts”, a concentration of board memberships among the same independent women on numerous boards, signaling superficial compliance without genuine structural change in leadership dynamics. Recent research also suggested that the presence of women on boards may be associated with increased board size, suggesting an additive approach toward board expansion, thus explaining their direct relation (Tiloiu, 2025). The presence of women on a board (BGD) is also positively associated with CEO duality (0.0989, p < 0.01) and executive directors’ compensation (0.0583, p < 0.01), and intuitively, with quota history (0.391, p < 0.01). We further find a positive association between BGD and the existence of board policies and practices (CSR committee, board effectiveness review, board structure policy, board function policy, board diversity policy). We find that the presence of women on a board is negatively associated with board member compensation (−0.0319, p < 0.1), board tenure (−0.140, p < 0.01), and the existence of a compliance head reporting to the board (−0.0544, p < 0.01).

Those results, although preliminary, provide strong evidence of the existence of a complex relationship between board gender diversity (BGD) and corporate risks and between BGD and other individual board diversity attributes related to structure, independence, remuneration, or corporate governance policies, whose influence might be moderated by the level of BGD.

4.3. Regression Results

This chapter presents the results of our regressions without effects and with fixed and random effects, assessing the relationship between gender diversity, board heterogeneity, and multiple dimensions of corporate risk. Prior studies suggest that board gender diversity may influence strategic decision-making, monitoring quality, information processing, and board practices, yet its effects are not uniform across risk categories (e.g., operational, financial, liquidity, or bankruptcy risk). Building on this evidence, we test whether gender diversity exhibits differentiated impacts depending on the nature of the risk involved. The following analyses reveal distinct patterns in how female board representation contributes in different ways to the reconfiguration of firms’ exposure across operational volatility, financial leverage risks, liquidity constraints, and bankruptcy likelihood, and whether female presence on the board moderates company choices related to the adoption of the highest standards of board practice. These findings contribute to a more nuanced understanding of the governance–risk nexus in the context of gender-diverse boards.

4.3.1. Estimates of the Operating Ratio

Baseline results of Ordinary Least Squares (OLS) regression of corporate risk outcome, proxied by Operating Ratio, on individual board diversity components are presented in

Table 7. Estimates for Operating Ratio using OLS regression.

	OpR (M1)			OpR (M2)			OpR(M3)
	(1)	(2)	(1) _int	(1)	(2)	(1) _int	(1)	(2)	(1) _int
BGD	0.0012	−0.0028	−0.0175	−0.0007 *	0.0011	0.0023	0.0002	0.0013	−0.0014
	(1.0484)	(−0.4233)	(−1.4381)	(−1.7150)	(0.6954)	(0.1679)	(0.5838)	(1.0808)	(−0.1873)
sqBGD_w		0.0001			−0.0000			−0.0000
		(0.6117)			(−1.1656)			(−0.9630)
BSize_w	0.0156 ***	0.0161 ***	0.0357 **	0.0028 **	0.0027 *	0.0090	−0.0026 **	−0.0026 **	0.0006
	(3.4684)	(3.5152)	(2.2932)	(1.9725)	(1.9202)	(1.5192)	(−2.4415)	(−2.4543)	(0.1393)
ln_BMComp_w	−0.0405 *	−0.0403 *	−0.0361	0.0187 ***	0.0181 ***	0.0196 ***	0.0055	0.0052	0.0055
	(−1.8965)	(−1.8850)	(−0.9697)	(3.2456)	(3.1148)	(3.3138)	(1.2037)	(1.1316)	(1.1896)
BER				0.0129	0.0123	0.0108	0.0077	0.0072	0.0280
				(1.5397)	(1.4638)	(1.2759)	(1.2437)	(1.1738)	(1.0426)
BSP	−0.1514 ***	−0.1348 **	−0.5604 ***	0.2059 ***	0.2029 ***	0.1976 ***	0.0633	0.0627	0.0132
	(−2.6745)	(−2.1455)	(−2.7919)	(3.2977)	(3.2482)	(3.0953)	(1.4878)	(1.4733)	(0.0450)
BMeet_w				0.0009	0.0008	0.0051	0.0013	0.0013	0.0013
				(0.7105)	(0.6647)	(1.0912)	(1.5158)	(1.5041)	(1.4666)
BFP				−0.0285 ***	−0.0279 ***	−0.0280 ***	−0.0168 ***	−0.0164 ***	−0.0630 **
				(−3.6025)	(−3.5247)	(−3.4707)	(−3.1177)	(−3.0213)	(−2.3840)
PBD				−0.0126	−0.0122	−0.0120	−0.0123	−0.0125	−0.0108
				(−1.0406)	(−1.0138)	(−0.9642)	(−1.2860)	(−1.3059)	(−0.2615)
BMAfill_w	0.1018 ***	0.1012 ***	0.1159 *	0.0171 ***	0.0172 ***	0.0093	−0.0087 *	−0.0085 *	−0.0411 **
	(6.2569)	(6.1970)	(1.9654)	(2.9011)	(2.9236)	(0.3275)	(−1.9147)	(−1.8824)	(−2.0113)
MaxLenght_w	0.0010	0.0011	0.0000
	(0.1513)	(0.1750)	(0.0052)
BIReel				−0.0085	−0.0091	−0.0086	0.0096	0.0093	0.0093
				(−0.9893)	(−1.0592)	(−0.9959)	(1.4504)	(1.4027)	(1.3945)
ChairInd				0.0328 ***	0.0322 ***	0.0325 ***	0.0085	0.0084	0.0079
				(3.8790)	(3.8033)	(3.7377)	(1.4086)	(1.3981)	(1.2812)
ChairExCEO				0.0118	0.0116	0.0111
				(1.3049)	(1.2845)	(1.2051)
BTenure_w				0.0001	0.0002	0.0001	−0.0027 **	−0.0027 **	−0.0027 **
				(0.0875)	(0.0981)	(0.0749)	(−2.2240)	(−2.2099)	(−2.1498)
IndBM_w	−0.0014 ***	−0.0013 ***	−0.0036	−0.0002	−0.0002	0.0001	−0.0000	−0.0000	−0.0000
	(−2.9081)	(−2.7769)	(−1.5259)	(−0.9528)	(−0.9253)	(0.1358)	(−0.2129)	(−0.2218)	(−0.0754)
CEODual	−0.0837 ***	−0.0840 ***	−0.0197				0.0153 **	0.0152 **	0.0172 **
	(−2.7539)	(−2.7589)	(−0.1598)				(2.3545)	(2.3414)	(2.5735)
NonExBM_w				−0.0004	−0.0003	−0.0028 **	−0.0004	−0.0004	−0.0005
				(−0.8751)	(−0.8503)	(−2.0735)	(−1.2160)	(−1.2470)	(−0.4725)
CCondNonc				0.0008	0.0004	−0.0014
				(0.1198)	(0.0520)	(−0.2000)
HComplB				−0.0190 **	−0.0191 **	−0.0195 **	0.0144 **	0.0142 **	0.0157 **
				(−2.1130)	(−2.1176)	(−2.1060)	(2.2351)	(2.1997)	(2.3685)
HLegalB	−0.0648	−0.0667	−0.0659
	(−0.5627)	(−0.5781)	(−0.5633)
BMReelect_w							0.0028	0.0029	0.0021
							(1.1801)	(1.2198)	(0.8893)
CSRCom				−0.0352 **	−0.0356 **	0.0455	−0.0554 ***	−0.0556 ***	−0.0650
				(−1.9809)	(−2.0020)	(0.7080)	(−3.6918)	(−3.7057)	(−1.2235)
ln_ExecComp_w	0.0393 ***	0.0399 ***	0.0362 ***				−0.0025	−0.0026	−0.0026
	(3.1149)	(3.1435)	(2.8116)				(−0.7164)	(−0.7453)	(−0.7497)
QuotaHist_w	−0.0084 **	−0.0084 **	−0.0116	−0.0013	−0.0012	−0.0043	−0.0015 **	−0.0015 *	−0.0001
	(−2.5114)	(−2.5188)	(−1.1394)	(−1.1667)	(−1.1415)	(−1.0017)	(−1.9739)	(−1.9362)	(−0.0355)
Age_w/ln_Age_w	0.0001	0.0001	0.0002	0.0007	0.0012	0.0010	0.0001	0.0001	0.0001
	(0.3814)	(0.4322)	(0.6036)	(0.1748)	(0.2840)	(0.2287)	(1.4889)	(1.5501)	(1.2917)
ln_MarketCapYE_w	−0.0707 ***	−0.0705 ***	−0.0693 ***				−0.0754 ***	−0.0754 ***	−0.0754 ***
	(−5.2682)	(−5.2389)	(−5.0184)				(−22.0490)	(−22.0516)	(−21.9141)
ln_CSTI_w	0.0158	0.0157	0.0083
	(1.2952)	(1.2917)	(0.6445)
ROE_w	−0.0266	−0.0194	−0.0371	−0.2670 ***	−0.2685 ***	−0.2650 ***
	(−0.2928)	(−0.2114)	(−0.3928)	(−7.3888)	(−7.4287)	(−7.2400)
ROA_w							−0.4158 ***	−0.4105 ***	−0.4211 ***
							(−6.8259)	(−6.7108)	(−6.7949)
OE_w							0.0901 ***	0.0902 ***	0.0905 ***
							(29.3455)	(29.3531)	(29.2115)
ETR_w	0.0738	0.0736	0.0385	0.0821 **	0.0820 **	0.0827 **	−0.0427	−0.0420	−0.0440 *
	(0.7858)	(0.7823)	(0.3931)	(2.1731)	(2.1717)	(2.1873)	(−1.6265)	(−1.5999)	(−1.6656)
ART_w				0.0021 ***	0.0021 ***	0.0019 ***	−0.0018 ***	−0.0018 ***	−0.0018 ***
				(2.8335)	(2.8576)	(2.6334)	(−3.6939)	(−3.6964)	(−3.6248)
ln_InvTo_w	0.0005	0.0005	0.0004	0.0344 ***	0.0340 ***	0.0350 ***	−0.0002	−0.0002	−0.0002
	(0.7476)	(0.8287)	(0.6954)	(9.2169)	(9.1028)	(9.3003)	(−1.3963)	(−1.4275)	(−1.3277)
ln_FCF_w	−0.0166	−0.0165	−0.0145
	(−1.4023)	(−1.3943)	(−1.1780)
DivYield_w				−0.1885	−0.1939	−0.1926	−2.0396 ***	−2.0391 ***	−2.0305 ***
				(−0.8984)	(−0.9243)	(−0.9077)	(−14.6416)	(−14.6376)	(−14.4259)
ln_WCap_w				0.0113 ***	0.0111 ***	0.0114 ***
				(3.2938)	(3.2371)	(3.2823)
ln_EBIT_w				−0.0308 ***	−0.0305 ***	−0.0403 ***
				(−6.9806)	(−6.9064)	(−2.6715)
BGD × Ind			0.0001			−0.0000			0.0000
			(0.9918)			(−0.3268)			(0.0512)
BGD × NonEx						0.0001 *			0.0000
						(1.8779)			(0.2150)
BGD × CSR						−0.0028			0.0004
						(−1.3746)			(0.2192)
BGD × Afill			−0.0003			0.0002			0.0008
			(−0.1741)			(0.2741)			(1.6416)
BGD × Bsz			−0.0005			−0.0002			−0.0001
			(−1.2374)			(−1.1058)			(−0.7896)
BGD × QH			0.0001			0.0001			−0.0000
			(0.4183)			(0.7556)			(−0.3997)
BGD × PBD									−0.0001
									(−0.0477)
BGD × BER									−0.0005
									(−0.7264)
BGD × BSP			0.0181 **						0.0012
			(2.1495)						(0.1716)
BGD × BFP									0.0012 *
									(1.8121)
BGD × BMComp			0.0001
			(0.0765)
BGD × CEODual			−0.0026
			(−0.6484)
BGD × Meet						−0.0001
						(−0.8677)
BGD × EBIT						0.0003
						(0.6664)
_cons	2.3443 ***	2.3613 ***	2.7831 ***	0.7294 ***	0.7171 ***	0.5928	0.5607 ***	0.5521 ***	0.6132 *
	(7.6762)	(7.6859)	(5.6640)	(4.5375)	(4.4527)	(1.0919)	(4.7310)	(4.6449)	(1.9341)
F statistic	6.8866 ***	6.5197 ***	5.2083 ***	12.6977 ***	12.2995 ***	10.0476 ***	57.9868 ***	56.0147 ***	42.8759 ***
R-sq	0.4231	0.4244	0.4456	0.3457	0.3470	0.3543	0.6439	0.6442	0.6472
VIF	2.37	7.92		1.58	3.10		1.67	3.02
Obs	188.0000	188.0000	188.0000	677.0000	677.0000	677.0000	927.0000	927.0000	927.0000

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. The numbers in parentheses are t statistics. For definitions of variables, please see Table 4. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

. Operating Ratio (OpR) is defined as operation expenses divided by net sales and reflects the operational efficiency of the companies, or how they deal with the day-to-day cost structure and operational risk. The second specification of each model (M1, M2, and M3) includes squared terms of BGD, while the third equation includes several interaction variables.

Prior findings related to banking efficiency suggest that female board representation may enhance cost discipline, while those relations are sometimes non-linear and highly context dependent (Owen & Temesvary, 2018).

We find a negative influence of gender diversity on Operating Ratio, contingent on model specification. This influence cannot be clearly attributable to BGD since new variables added to our models changed the direction of this relation, a presumable suppression effect might be discussed.

Following our three models systematically developed to better predict Operating Ratio, we find a negative influence of gender diversity (BGD) on Operating Ratio (−0.0007, p < 0.1/M2), although this influence cannot be clearly attributable to gender diversification since new variables added to our models changed the direction of this relation and did not reach statistical significance.

The models, including the squared term of BGD (sqBGD), do not provide evidence of the existence of a non-linear effects on Operating Ratio; therefore, H1.1 is rejected.

Discussing the other individual board diversity components, board size increase appears to increase Operating Ratio (M1 and M2), while this effect is negative when a few more variables are included and more observations are counted. Explanatory power of M3 increases to almost 65% and is based on the highest number of observations. Companies having higher numbers of memberships are more likely to increase Operating Ratio (0.0171, p < 0.01), likewise an independent chairman of the board (0.0328, p < 1.001), whose company-specific knowledge might be limited. Similarly, the CEO duality role of chairman of the board is increasing Operating Ratio, supporting concerns of agency theory, when a lack of independent oversight over executive managers occurs. Conversely, the longer the tenure of the board, the lower the Operating Ratio (−0.0027, p < 0.05), suggesting that increased knowledge about the company context helps to improve operational efficiency. The companies that established a CSR committee are more likely to reduce Operating Ratio (−0.0554, p < 0.01) and improve operational efficiency; likewise, the existence of a policy for maintaining effective board functions (BFP) reduces Operating Ratio (−0.0168, p < 0.01). In the same context, the influence of a policy for maintaining a well-balanced membership (BSP) is not consistent across models, suggesting the need to be interpreted industry-contingently.

We further analyzed the possible moderation effect of gender (BGD) on other individual board diversity components’ influence on operating ratio. We find that negative influence of non-executive directors on Operating Ratio (−0.0028, p < 0.05) is slightly moderated by BGD (0.0001, p < 0.1), while negative influence of board functions policy (BFP) on OpR (−0.0630, p < 0.05) is tempered by a gender-diverse board (0.0012, p < 0.1), as well as Board Structure Policy negative influence on OpR (−0.5604, p < 0.01) is moderated by a gender-diverse board (0.0181, p < 0.05). We underline that these moderating effects of interaction variables that include BGD on OpR are conditional on model specification.

Alternative measures for gender diversity in the boardroom have been deployed in our study for more robust results in the prediction of Operating Ratio (operating expenses/net sales). Details can be seen in Supplementary File S1 (WOBMere), Supplementary File S2 (WOBTK), and Supplementary File S3 (WOBBal). The library of supplementary tables is presented in Supplementary Materials.

In case the presence of women on the board is barely visible (WOBMere meaning presence less than 20%), we find a negative influence of the mere presence of women on the board on Operating Ratio, despite statistical significance depends on model (coefficient −0.0357, p < 0.05/M3), in line with M2 OLS results, that gender diversity reduces Operating Ratio. Discussing companies with above 20% but less than 40% women on board, defined by literature as tokens (WOBTK), we find a negative influence of gender on Operating Ratio (−0.0319, p < 0.1/M1), although not consistent across models, but as model complexity increases (M2), this relation turns positive (0.0218, p < 0.01). It is to be noted that the explanatory power of M2 compared with M1 decreases for predicting Operating Ratio (R-sq of M1 is 43%, while for M2, it drops to 29%). In line with literature and econometric technics suggesting that non-necessarily a more complex model is the best explanatory one and in the context of results discussed before, we will consider that M1 is a better predictor for Operating Ratio in case of alternative specification of gender diversity (WOBTK), suggesting more likely the negative influence of gender-diverse boards on OpR, although the results are not robust. The third alternative measure of gender diversity is the balanced presence of women on the board (WOBBal), where more than 40% of board directors are women. In these cases, models with higher explanatory power (M1 and M3) show an opposite to our expectation but positive, although not significant, influence of balanced presence of women on board (WOBBal) on Operating Ratio. We explain the flip of sign to positive in case of balanced presence of women on the board, by two possible effects. First, the impact of new variables, like board member independence, CEO duality, board member compensation, presence of a board structure policy, or longer quota history, that presumably suppresses and overtakes the effect of gender diversity on Operating Ratio when board gender balance is met. Second, observing that the coefficient sign of the squared term (sqBGD) is changing, we presume there is a non-linear effect of BGD on OpR. This influence is not statistically significant in our model specifications; therefore, we cannot fully validate it.

We further tested out specifications for multicollinearity using the variation inflation factor (VIF) since many corporate governance variables are simultaneously observed. As the VIF of our models spans between 1.67 and 2.37 (2.03 and 7.97 when the squared BGD term is included), we eliminate the concern of multicollinearity between the independent variables included in all three models created to predict Operating Ratio.

To test the endogeneity of our models (or misspecification due to unobservable variables) and obtain more robust results, we used a fixed effects estimator and ran regressions with firm-year fixed effects and random effects, controlling for firm-specific or time-invariant evolutions that might affect our initial OLS results. In case of endogeneity, OLS estimators will fail, since one of the assumptions of OLS is that there is no correlation between independent variables and the error term. We have selected the most appropriate model using the Hausman test. In case the Hausman test yielded a p-value < 0.05, we have rejected the null hypothesis, that is, there are no fixed effects in our panel data, then we choose models with fixed effects.

The robustness checks through fixed and random effects models show inconsistency in relation between gender diversity and Operating Ratio in our specifications, aligning with the heterogeneous findings in existing literature. Due to the changing results between models and sometimes odd signs of some relationships (for example board size, board member compensation, membership policy, number of mandates, CEO duality, also compliance head reporting to board) we emphasize that multivariate results are conditional on all included variables, therefore cautiously, will not accept hypothesis H1, that gender-diverse boards reduce Operating Ratio.

The summary of estimates using regression models with effects is presented separately for each model specification (

Table 8. Estimates using regression with firm-year effects, Model 1.

	OpR_w	LevDA_w	LevDE_w	ChgLTDpK_w	ZScoreMNM_w	CurrR_w	QR_w
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
BGD	0.0002	0.0004	0.0004	−0.0002	−0.0041	−0.0050	−0.0008
	(0.4261)	(0.4279)	(0.4279)	(−0.3698)	(−0.3112)	(−1.0686)	(−0.2367)
BSize_w	0.0087 ***	−0.0063	−0.0063	−0.0016	−0.1123	−0.0666 **	−0.0400 **
	(3.0574)	(−1.2951)	(−1.2951)	(−0.7267)	(−1.4106)	(−2.3167)	(−2.4291)
BMAfill_w	0.0283 ***	0.0239	0.0239	−0.0078	−0.0408	−0.0434	−0.1080 *
	(3.2332)	(1.5134)	(1.5134)	(−0.9069)	(−0.1751)	(−0.5160)	(−1.9393)
MaxLenght_w	−0.0117 **	0.0033	0.0033	0.0020	0.0426	0.1280 **	0.0673 **
	(−2.4617)	(0.4185)	(0.4185)	(0.6018)	(0.3075)	(2.5582)	(2.5600)
ln_BMComp_w	0.0121	0.0072	0.0072	−0.0076	−0.4830 **	−0.0304	−0.0290
	(1.4562)	(0.4405)	(0.4405)	(−0.6889)	(−2.1119)	(−0.3685)	(−0.4938)
IndBM_w	0.0004	0.0003	0.0003	0.0006 ***	0.0030	−0.0020	−0.0008
	(1.6234)	(0.6072)	(0.6072)	(2.5786)	(0.4392)	(−0.7920)	(−0.4616)
CEODual	0.0022	0.0132	0.0132	−0.0030	−0.1440	0.1188	0.1671 **
	(0.1961)	(0.5612)	(0.5612)	(−0.1860)	(−0.4356)	(0.9946)	(1.9661)
BSP	0.0077	−0.0603 *	−0.0603 *	0.0290	0.3316	0.2102	0.0473
	(0.4828)	(−1.7492)	(−1.7492)	(0.9499)	(0.7124)	(1.2496)	(0.3712)
HLegalB	0.0000	0.2325	0.2325	0.1671 ***	0.0000	0.0000	0.2887
	(.)	(1.4156)	(1.4156)	(2.6346)	(.)	(.)	(0.5959)
ln_ExecComp_w	−0.0013	−0.0003	−0.0003	0.0032	0.1115	−0.0359	−0.0020
	(−0.3363)	(−0.0417)	(−0.0417)	(0.4683)	(1.0441)	(−0.9301)	(−0.0709)
Age_w	−0.0027	−0.0004	−0.0004	0.0002	0.3268 ***	0.0227	−0.0032
	(−0.8693)	(−0.4564)	(−0.4564)	(1.3368)	(3.6630)	(0.7053)	(−1.4291)
QuotaHist_w	−0.0027	0.0059 **	0.0059 **	−0.0009	−0.3625 ***	−0.0359	−0.0138
	(−0.8486)	(2.0430)	(2.0430)	(−0.5214)	(−4.0380)	(−1.1057)	(−1.3712)
ln_MarketCapYE_w	−0.0109	−0.0194	−0.0194	0.0011	1.8685 ***	−0.0561	−0.1357 ***
	(−1.3063)	(−1.3350)	(−1.3350)	(0.1533)	(7.9682)	(−0.6623)	(−2.6961)
ln_CSTI_w	0.0032	0.0131	0.0131	0.0101 *	−0.3708 ***	0.0638	0.1673 ***
	(0.6253)	(1.3786)	(1.3786)	(1.6782)	(−2.7599)	(1.3150)	(4.8966)
ROE_w	−0.3157 ***	0.0344	0.0344	−0.0320	3.3149 ***	−0.1219	−0.1186
	(−9.0732)	(0.5263)	(0.5263)	(−0.6809)	(3.6915)	(−0.3757)	(−0.4978)
ETR_w	−0.0077	0.0814	0.0814	0.0712	−0.7510	0.0769	−0.0579
	(−0.2974)	(1.4666)	(1.4666)	(1.3913)	(−1.0141)	(0.2875)	(−0.2826)
InvTo_w	−0.0003	−0.0000	−0.0000	−0.0001	−0.0019	−0.0025	0.0023
	(−0.8520)	(−0.0727)	(−0.0727)	(−0.2817)	(−0.1980)	(−0.7104)	(1.0304)
ln_FCF_w	0.0001	−0.0010	−0.0010	−0.0119 *	−0.0615	0.0044	0.0156
	(0.0332)	(−0.1473)	(−0.1473)	(−1.9008)	(−0.6453)	(0.1267)	(0.5944)
_cons	0.9977 ***	0.3542	0.3542	−0.0242	−36.9561 ***	1.1352	0.9145
	(3.9694)	(1.0199)	(1.0199)	(−0.1587)	(−5.2447)	(0.4459)	(0.7784)
F statistic	9.1220 ***				8.3846 ***	2.2002 ***
Wald Chi2		28.13	28.06	27.58			57.30
R-sq overall	0.0010	0.0358	0.0524	0.1304	0.0139	0.0417	0.2179
Obs	188.0000	204.0000	204.0000	203.0000	204.0000	204.0000	204.0000
N Countries	41.0000	46.0000	46.0000	46.0000	46.0000	46.0000	46.0000
Chi2	0.0000	0.0726	0.8926	0.8301	0.0000	0.0000	0.6976
FE/RE	FE	RE	RE	RE	FE	FE	RE

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. The numbers in parentheses are t statistics. FE (fixed effects), RE (random effects). For definitions of variables, please see Table 4. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

for M1,

Table 9. Estimates using regression with firm-year effects, Model 2.

	OpR_w	LevDA_w	LevDE_w	ChgLTDpK_w	ZScoreMNM_w	CurrR_w	QR_w
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
BGD	−0.0000	0.0008 **	0.0048 ***	0.0000	−0.0077	−0.0028	0.0001
	(−0.1415)	(1.9865)	(2.6695)	(0.0382)	(−0.8493)	(−1.3633)	(0.0483)
BSize_w	0.0029 **	0.0029	0.0084	−0.0013	−0.0578	−0.0166	−0.0024
	(1.9744)	(1.2511)	(0.7905)	(−0.4322)	(−1.0856)	(−1.3851)	(−0.2498)
ln_BMComp_w	0.0004	−0.0018	−0.0015	0.0001	−0.1375	−0.0314	−0.0151
	(0.0744)	(−0.2438)	(−0.0476)	(0.0159)	(−0.8576)	(−0.8613)	(−0.5055)
BER	0.0054	−0.0167 *	−0.0671	0.0202	−0.4976 **	−0.0854 *	−0.0845 **
	(0.8726)	(−1.6903)	(−1.5077)	(1.6374)	(−2.1825)	(−1.7049)	(−2.0625)
BSP	−0.0039	−0.0859	−0.1771	0.1060	0.7048	0.5839 **	0.6052 ***
	(−0.1168)	(−1.5715)	(−0.7165)	(1.5498)	(0.5831)	(2.1024)	(2.6651)
BMeet_w	−0.0002	0.0011	0.0080 *	−0.0040 ***	−0.0063	−0.0019	0.0009
	(−0.3248)	(1.1432)	(1.8781)	(−3.4652)	(−0.2961)	(−0.4073)	(0.2300)
BFP	−0.0202	−0.0135	−0.0765	−0.0613 **	0.7632 *	−0.0024	0.0090
	(−1.5943)	(−0.6485)	(−0.8131)	(−2.3499)	(1.6538)	(−0.0223)	(0.1039)
PBD	−0.0005	0.0190	0.0158	0.0445 **	0.0855	−0.0450	−0.0265
	(−0.0547)	(1.2056)	(0.2167)	(2.2606)	(0.2453)	(−0.5639)	(−0.4063)
BMAfill_w	−0.0043	−0.0108	−0.0306	−0.0223	0.5360 **	0.0249	−0.0340
	(−0.6322)	(−0.9783)	(−0.6145)	(−1.6233)	(2.0959)	(0.4454)	(−0.7452)
BIReel	0.0153	−0.0242	0.0511	−0.0524 *	0.4717	0.0712	0.0855
	(0.8137)	(−0.9629)	(0.4510)	(−1.6682)	(0.8484)	(0.5579)	(0.8193)
ChairInd	−0.0006	0.0086	0.0580	−0.0031	−0.0636	0.0163	0.0154
	(−0.1033)	(0.9359)	(1.4055)	(−0.2702)	(−0.3086)	(0.3496)	(0.4038)
ChairExCEO	−0.0070	0.0014	−0.0108	0.0072	0.4885 *	−0.0328	−0.0479
	(−0.8409)	(0.1063)	(−0.1814)	(0.4322)	(1.6510)	(−0.4869)	(−0.8703)
BTenure_w	0.0004	0.0005	0.0048	−0.0008	0.0696	0.0235 *	0.0261 **
	(0.2628)	(0.1685)	(0.3922)	(−0.2441)	(1.1525)	(1.7151)	(2.3245)
IndBM_w	0.0002	0.0002	0.0018	−0.0000	−0.0016	−0.0016	−0.0011
	(1.1447)	(0.8908)	(1.4394)	(−0.1014)	(−0.2531)	(−1.1221)	(−1.0026)
NonExBM_w	−0.0001	−0.0010 *	−0.0041 *	−0.0013 **	0.0177	0.0037	−0.0000
	(−0.1681)	(−1.8629)	(−1.7405)	(−2.0453)	(1.5098)	(1.3774)	(−0.0023)
CCondNonc	0.0030	0.0160 *	0.0827 **	−0.0077	−0.0777	−0.0250	−0.0331
	(0.5572)	(1.8499)	(2.1004)	(−0.7167)	(−0.3929)	(−0.5695)	(−0.9205)
HComplB	−0.0095	−0.0231	−0.0682	0.0222	0.5857	0.0189	0.0403
	(−0.9424)	(−1.3942)	(−0.9131)	(1.0705)	(1.3827)	(0.2247)	(0.5853)
CSRCom	−0.0077	−0.0032	−0.0110	0.0014	−0.0508	0.0566	0.0470
	(−0.6904)	(−0.1866)	(−0.1407)	(0.0646)	(−0.1323)	(0.6460)	(0.6561)
QuotaHist_w	0.0033 **	0.0031	0.0013	−0.0025	−0.1039 **	0.0134	0.0084
	(2.3653)	(1.3857)	(0.1268)	(−0.8945)	(−2.0276)	(1.1978)	(0.9156)
ln_Age_w	0.0359	−0.0537	−0.4436 **	−0.0305	1.5779	−0.3561	−0.3824 **
	(1.2846)	(−1.1978)	(−2.1904)	(−0.5437)	(1.5452)	(−1.5627)	(−2.0523)
ROE_w	−0.1665 ***	−0.0363	−0.2240	−0.0953 *	2.8726 ***	−0.1118	−0.0810
	(−6.3415)	(−0.8836)	(−1.1836)	(−1.8552)	(3.0915)	(−0.5357)	(−0.4743)
ETR_w	−0.0662 ***	0.0119	0.1348	−0.0155	−0.5866	−0.1795	−0.1833
	(−3.3486)	(0.3955)	(0.9878)	(−0.4107)	(−0.8600)	(−1.1703)	(−1.4618)
ART_w	−0.0016	−0.0025	−0.0145	−0.0051 *	0.0228	0.0018	−0.0157 *
	(−1.1937)	(−1.1297)	(−1.4713)	(−1.8496)	(0.4627)	(0.1602)	(−1.7257)
ln_InvTo_w	0.0328 ***	−0.0187	0.0572	−0.0175	0.6316 *	−0.1678 **	0.0270
	(3.0238)	(−1.1222)	(0.7606)	(−0.8392)	(1.6795)	(−1.9858)	(0.3909)
DivYield_w	−0.2379 *	−0.6151 ***	−3.7249 ***	−0.5237 *	−14.1817 ***	0.4618	0.2000
	(−1.7915)	(−2.8584)	(−3.8169)	(−1.9463)	(−2.9670)	(0.4225)	(0.2238)
ln_WCap_w	−0.0014	−0.0047	−0.0112	−0.0082	0.1702 *	0.2427 ***	0.1964 ***
	(−0.5274)	(−1.1306)	(−0.5926)	(−1.5789)	(1.8092)	(11.4819)	(11.3625)
ln_EBIT_w	−0.0757 ***	0.0020	0.0131	0.0054	0.2865 **	0.0375	0.0343
	(−18.9484)	(0.3667)	(0.5427)	(0.8038)	(2.3989)	(1.3804)	(1.5466)
_cons	2.2899 ***	0.5703 **	1.7532	0.7248 **	−11.1206 **	−2.8018 **	−2.8078 ***
	(14.5131)	(2.2910)	(1.5580)	(2.3284)	(−1.9768)	(−2.2156)	(−2.7156)
F statistic	29.1174 ***	1.7833 ***	2.2088 ***	2.3833 ***	2.7806 ***	7.0592 ***	6.7788 ***
R-sq overall	0.0863	0.0099	0.0075	0.0190	0.0108	0.0566	0.0082
Obs	677.0000	718.0000	714.0000	718.0000	697.0000	718.0000	718.0000
N Countries	233.0000	247.0000	246.0000	247.0000	243.0000	247.0000	247.0000
Chi2	0.0000	0.0000	0.0000	0.0454	0.0000	0.0000	0.0000
FE/RE	FE	FE	FE	FE	FE	FE	FE

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. The numbers in parentheses are t statistics. FE (fixed effects), RE (random effects). For definitions of variables, please see Table 4. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

for M2, and

Table 10. Estimates using regression with firm-year effects, Model 3.

	OpR_w	LevDA_w	LevDE_w	ChgLTDpK_w	ZScoreMNM_w	CurrR_w	QR_w
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
BGD	0.0000	0.0006 **	0.0053 ***	0.0001	0.0001	0.0029 *	0.0046 ***
	(0.1930)	(2.1594)	(3.3606)	(0.3208)	(0.0208)	(1.7920)	(3.4175)
BSize_w	0.0012	0.0005	−0.0082	−0.0004	−0.0301	−0.0006	0.0034
	(0.8818)	(0.2897)	(−0.8811)	(−0.5249)	(−0.9808)	(−0.0684)	(0.4407)
ln_BMComp_w	−0.0017	−0.0019	0.0042	0.0009	−0.0414	−0.0164	−0.0013
	(−0.3538)	(−0.3369)	(0.1352)	(0.2759)	(−0.4075)	(−0.5226)	(−0.0493)
BER	−0.0065	−0.0011	0.0440	−0.0002	−0.0164	−0.0507	−0.0360
	(−1.1474)	(−0.1543)	(1.1575)	(−0.0428)	(−0.1271)	(−1.3099)	(−1.1203)
BSP	−0.0355	−0.0492	−0.1120	0.0577 *	0.1956	0.4191 **	0.3515 **
	(−1.1758)	(−1.2790)	(−0.5350)	(1.7218)	(0.2842)	(1.9681)	(1.9877)
BMeet_w	−0.0006	0.0006	0.0073 *	−0.0016 **	−0.0059	0.0003	0.0011
	(−1.0692)	(0.8621)	(1.8864)	(−2.4638)	(−0.4609)	(0.0805)	(0.3542)
BFP	−0.0114	−0.0244 *	−0.2038 ***	−0.0073 *	0.6642 ***	0.0369	0.0372
	(−0.9693)	(−1.7257)	(−2.6541)	(−1.8106)	(2.6229)	(0.4715)	(0.5728)
PBD	−0.0063	0.0152	0.0136	0.0129 *	−0.1117	−0.0624	−0.0464
	(−0.6383)	(1.2588)	(0.2024)	(1.7436)	(−0.5144)	(−0.9316)	(−0.8344)
BMAfill_w	0.0026	0.0052	0.0235	−0.0052	0.1292	0.0064	−0.0153
	(0.4129)	(0.7045)	(0.5799)	(−1.5781)	(0.9549)	(0.1553)	(−0.4489)
BIReel	0.0209	−0.0498 *	−0.1730	0.0028	−0.1482	−0.0234	−0.0134
	(0.7008)	(−1.9502)	(−1.2460)	(0.5584)	(−0.3239)	(−0.1656)	(−0.1141)
ChairInd	−0.0006	0.0024	0.0293	−0.0055	−0.0347	0.0015	0.0097
	(−0.1137)	(0.3718)	(0.8475)	(−1.2244)	(−0.3023)	(0.0423)	(0.3317)
CEODual	−0.0047	0.0048	0.0103	−0.0050	0.2839	−0.1516 ***	−0.1356 ***
	(−0.5538)	(0.4678)	(0.1853)	(−1.0310)	(1.5347)	(−2.6668)	(−2.8723)
BTenure_w	−0.0003	0.0006	−0.0094	0.0001	0.0524	0.0443 ***	0.0389 ***
	(−0.1855)	(0.2727)	(−0.8432)	(0.0562)	(1.4173)	(3.8922)	(4.1191)
IndBM_w	0.0002	−0.0001	−0.0002	0.0003 **	0.0023	−0.0017	−0.0011
	(1.0619)	(−0.3632)	(−0.1719)	(2.4238)	(0.5799)	(−1.3968)	(−1.1240)
NonExBM_w	−0.0002	−0.0006	−0.0030	−0.0003	0.0046	0.0000	−0.0018
	(−0.7191)	(−1.6244)	(−1.4119)	(−1.2821)	(0.6527)	(0.0183)	(−0.9856)
HComplB	−0.0097	−0.0190	−0.0862	−0.0032	0.1753	0.0095	0.0089
	(−0.8874)	(−1.3571)	(−1.1308)	(−0.6514)	(0.6297)	(0.1228)	(0.1381)
BMReelect_w	0.0014	−0.0073	−0.0619	−0.0027	−0.0797	−0.0078	−0.0258
	(0.2138)	(−0.8863)	(−1.3889)	(−1.5256)	(−0.5421)	(−0.1728)	(−0.6847)
CSRCom	−0.0059	0.0012	−0.0056	0.0016	−0.0151	0.1512 *	0.0880
	(−0.4401)	(0.0779)	(−0.0652)	(0.1400)	(−0.0530)	(1.7222)	(1.2074)
ln_ExecComp_w	0.0020	−0.0000	0.0020	−0.0021	0.0288	−0.0360 **	−0.0146
	(0.7172)	(−0.0102)	(0.1169)	(−0.8455)	(0.5003)	(−2.0334)	(−0.9916)
QuotaHist_w	−0.0007	0.0027	−0.0116	−0.0002	0.0068	0.0378 ***	0.0227 **
	(−0.3336)	(1.0743)	(−0.8505)	(−0.4346)	(0.1454)	(2.7153)	(1.9666)
Age_w	0.0005	−0.0027	−0.0088	0.0000	0.0166	−0.0243 *	−0.0183
	(0.2398)	(−1.0379)	(−0.6300)	(0.1655)	(0.3478)	(−1.7058)	(−1.5442)
ln_MarketCapYE_w	−0.0167 **	−0.0310 ***	−0.2141 ***	0.0036	1.5057 ***	0.2473 ***	0.2113 ***
	(−2.5545)	(−3.8648)	(−4.8888)	(1.4051)	(10.3952)	(5.5601)	(5.7201)
ETR_w	−0.0901 ***	−0.0069	0.0127	0.0366 *	0.7773 **	−0.0805	−0.0837
	(−4.9432)	(−0.3264)	(0.1104)	(1.8753)	(2.0361)	(−0.6884)	(−0.8619)
ART_w	−0.0016	−0.0020	−0.0062	−0.0003	0.0287	0.0007	−0.0057
	(−1.3732)	(−1.3138)	(−0.7570)	(−0.7799)	(1.0585)	(0.0849)	(−0.8158)
InvTo_w	0.0003	−0.0008 *	−0.0012	−0.0000	0.0121	−0.0066 **	−0.0025
	(0.6070)	(−1.6576)	(−0.4533)	(−0.1994)	(1.4138)	(−2.5120)	(−1.1587)
DivYield_w	−0.6928 ***	−0.6114 ***	−3.5964 ***	−0.2400 **	−3.8125	1.3674	0.9973
	(−4.9697)	(−3.4879)	(−3.7622)	(−2.2958)	(−1.2093)	(1.4084)	(1.2368)
ROA_w	−0.7836 ***	−0.3211 ***	−1.1035 ***	−0.0467	12.3953 ***	−0.1525	−0.1808
	(−13.5431)	(−4.4630)	(−2.7963)	(−1.0019)	(9.4792)	(−0.3825)	(−0.5461)
ln_OE_w	0.0379 ***	0.0395 ***	0.2864 ***	−0.0006	−1.6384 ***	−0.3781 ***	−0.3273 ***
	(4.3515)	(3.7406)	(4.9446)	(−0.2649)	(−8.6069)	(−6.4722)	(−6.7462)
_cons	0.5364 **	0.3523	−0.0377	0.0695	3.5504	5.1512 ***	3.8515 ***
	(2.1005)	(1.1241)	(−0.0219)	(0.7754)	(0.6254)	(2.9677)	(2.6717)
F statistic	11.6767 ***	4.1277 ***	4.3831 ***		13.6552 ***	4.7419 ***	4.7394 ***
Wald Chi2				44.15
R-sq overall	0.1179	0.0110	0.0351	0.0248	0.4135	0.0248	0.0263
Obs	927.0000	1013.0000	1005.0000	1013.0000	991.0000	1013.0000	1013.0000
N Countries	285.0000	305.0000	304.0000	305.0000	300.0000	305.0000	305.0000
Chi2	0.0000	0.0000	0.0000	0.0974	0.0000	0.0000	0.0000
FE/RE	FE	FE	FE	RE	FE	FE	FE

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. The numbers in parentheses are t statistics. FE (fixed effects), RE (random effects). For definitions of variables, please see Table 4. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

for M3). The main result is that BGD estimators are rather positive across models, although not statistically relevant, suggesting that negative estimators in OLS models are fragile, and might be biased by misspecification of models, considering a linear effect or by omitted variables. Nevertheless, board size and number of memberships are positive estimators in models with effects that remain stable (0.0087, p < 0.01 and 0.0283, p < 0.01), suggesting that OLS results are robust (M1 and M2. These results confirm the existence of other individual board diversity components, partially overtaking BGD influence on OpR, in the case of large companies having mature boards, concluding that board size and memberships increase Operating Ratio.

4.3.2. Estimates of the Leverage Ratio

Baseline results of Ordinary Least Squares (OLS) regression of corporate risk outcome, proxied by Leverage (indebtedness), on individual diversity components are presented in

Table 11. Estimates for Leverage Ratio using OLS regression.

	LevDA (M1)			LevDA (M2)			LevDA (M3)
	(1)	(2)	(1) _int	(1)	(2)	(1) _int	(1)	(2)	(1) _int
BGD	0.0012	0.0087	−0.0002	0.0044 *	0.0525 ***	0.0289	0.0055 **	0.0356 ***	0.0063
	(0.9615)	(1.1666)	(−0.0138)	(1.6937)	(5.8961)	(1.5092)	(2.2019)	(3.9821)	(0.5066)
sqBGD_w		−0.0001			−0.0008 ***			−0.0005 ***
		(−1.0211)			(−5.6366)			(−3.5033)
BSize_w	−0.0014	−0.0020	−0.0327 *	0.0247 ***	0.0227 ***	−0.0034	−0.0038	−0.0047	0.0020
	(−0.3007)	(−0.4332)	(−1.8369)	(2.7759)	(2.6080)	(−0.4092)	(−0.4243)	(−0.5303)	(0.2816)
ln_BMComp_w	0.0241	0.0227	0.0456	0.0530	0.0357	0.0150 *	−0.0063	−0.0148	−0.0019
	(1.0245)	(0.9653)	(1.1051)	(1.4573)	(0.9981)	(1.8226)	(−0.1656)	(−0.3905)	(−0.2575)
BER				−0.0742	−0.0886 *	−0.0119	0.0103	0.0006	−0.0672
				(−1.3769)	(−1.6796)	(−0.9988)	(0.1986)	(0.0115)	(−1.5247)
BSP	0.0830	0.0509	0.1256	0.0611	−0.0264	−0.1548 *	0.0951	0.0720	−0.1270
	(1.2817)	(0.7081)	(0.5438)	(0.1491)	(−0.0659)	(−1.6781)	(0.2567)	(0.1955)	(−0.2545)
BMeet_w				−0.0071	−0.0084	−0.0008	−0.0134 *	−0.0138 *	−0.0031 **
				(−0.9172)	(−1.1111)	(−0.1307)	(−1.8579)	(−1.9283)	(−2.2138)
BFP				−0.0681	−0.0434	−0.0153	−0.0047	0.0125	0.0690
				(−1.3499)	(−0.8756)	(−1.3400)	(−0.1041)	(0.2790)	(1.6304)
PBD				0.0683	0.0820	0.0108	0.1166	0.1142	0.0055
				(0.8648)	(1.0607)	(0.6008)	(1.4239)	(1.4017)	(0.0812)
BMAfill_w	−0.0133	−0.0118	−0.0181	−0.0677 *	−0.0682 *	−0.0894 **	0.0082	0.0097	−0.0032
	(−0.7231)	(−0.6405)	(−0.2649)	(−1.8008)	(−1.8544)	(−2.3651)	(0.2233)	(0.2652)	(−0.1031)
MaxLenght_w	0.0104	0.0102	0.0092
	(1.4495)	(1.4260)	(1.2208)
BIReel				−0.0719	−0.0954 *	−0.0205 *	0.0086	−0.0038	−0.0220 **
				(−1.3218)	(−1.7872)	(−1.7064)	(0.1581)	(−0.0695)	(−2.0489)
ChairInd				−0.1291 **	−0.1504 ***	−0.0489 ***	−0.1204 **	−0.1242 **	−0.0404 ***
				(−2.3806)	(−2.8280)	(−4.0016)	(−2.4096)	(−2.5007)	(−4.0629)
ChairExCEO				−0.1036 *	−0.1013 *	−0.0331 ***
				(−1.8188)	(−1.8187)	(−2.5878)
BTenure_w				−0.0207 *	−0.0207 **	−0.0048 **	−0.0277 ***	−0.0269 ***	−0.0034
				(−1.9317)	(−1.9747)	(−1.9861)	(−2.6680)	(−2.6090)	(−1.6412)
IndBM_w	0.0020 ***	0.0019 ***	0.0002	0.0029 **	0.0031 **	0.0019	0.0007	0.0007	0.0003
	(3.7665)	(3.6003)	(0.0853)	(2.2085)	(2.4499)	(1.5510)	(0.5859)	(0.5600)	(0.2698)
CEODual	0.0873 **	0.0870 **	0.0439				−0.0779	−0.0792	−0.0254 **
	(2.5610)	(2.5544)	(0.3043)				(−1.4352)	(−1.4667)	(−2.3479)
NonExBM_w				−0.0013	−0.0009	0.0011	0.0000	−0.0002	0.0039 **
				(−0.5079)	(−0.3552)	(0.5992)	(0.0098)	(−0.0621)	(2.2900)
CCondNonc				−0.0016	−0.0208	−0.0125
				(−0.0352)	(−0.4735)	(−1.2468)
HComplB				0.0236	0.0153	0.0115	0.0335	0.0256	0.0078
				(0.4034)	(0.2675)	(0.8691)	(0.6097)	(0.4690)	(0.7173)
HLegalB	0.3707 ***	0.3744 ***	0.3542 **
	(2.7393)	(2.7658)	(2.5702)
BMReelect_w							−0.0068	−0.0042	−0.0046
							(−0.3517)	(−0.2210)	(−1.2028)
CSRCom				0.3270 ***	0.3257 ***	0.2845 ***	0.2585 **	0.2544 **	0.1444 *
				(2.9340)	(2.9871)	(3.1955)	(2.0633)	(2.0422)	(1.6671)
ln_ExecComp_w	−0.0325 **	−0.0338 **	−0.0308 **				−0.0218	−0.0228	0.0025
	(−2.2246)	(−2.3068)	(−2.0588)				(−0.7755)	(−0.8166)	(0.4586)
QuotaHist_w	0.0120 ***	0.0121 ***	0.0184	−0.0040	−0.0033	−0.0068	−0.0115 *	−0.0106 *	−0.0041
	(3.2689)	(3.2822)	(1.5534)	(−0.5831)	(−0.4859)	(−1.1396)	(−1.7996)	(−1.6758)	(−0.6931)
Age_w/ln_Age_w	−0.0008 **	−0.0009 **	−0.0010 **	−0.0048	0.0031	−0.0129 **	0.0004	0.0004	−0.0002
	(−2.4365)	(−2.4980)	(−2.5663)	(−0.1820)	(0.1172)	(−2.1738)	(0.6820)	(0.7131)	(−1.3133)
ln_MarketCapYE_w	0.0243	0.0239	0.0197				0.0141	0.0108	0.0225 ***
	(1.6006)	(1.5710)	(1.2590)				(0.4965)	(0.3827)	(4.0260)
ln_CSTI_w	0.0090	0.0087	0.0139
	(0.7021)	(0.6782)	(0.9984)
ROE_w	0.0733	0.0622	0.0971	1.1917 ***	1.0619 ***	0.0832
	(0.7399)	(0.6243)	(0.9421)	(5.1533)	(4.6697)	(1.6242)
ROA_w							−3.4982 ***	−3.3661 ***	−0.8353 ***
							(−6.7425)	(−6.5079)	(−8.1821)
OE_w							0.0724 ***	0.0745 ***	−0.0164 ***
							(2.7664)	(2.8643)	(−3.2223)
ETR_w	0.1423	0.1434	0.1905 *	0.6311 ***	0.6319 ***	0.0883 *	−0.0359	−0.0228	−0.0052
	(1.3287)	(1.3394)	(1.7162)	(2.6456)	(2.7077)	(1.6795)	(−0.1655)	(−0.1058)	(−0.1230)
ART_w				−0.0181 ***	−0.0166 ***	−0.0038 ***	−0.0142 ***	−0.0141 ***	−0.0018 **
				(−3.8405)	(−3.5964)	(−3.6251)	(−3.4197)	(−3.4377)	(−2.2239)
ln_InvTo_w	0.0019 ***	0.0018 ***	0.0017 **	0.1076 ***	0.1045 ***	0.0082	0.0092 ***	0.0090 ***	0.0010 ***
	(2.7523)	(2.6641)	(2.4093)	(4.5704)	(4.5361)	(1.5855)	(6.1259)	(6.0517)	(3.7321)
ln_FCF_w	−0.0059	−0.0062	−0.0061
	(−0.4451)	(−0.4689)	(−0.4441)
DivYield_w				1.6277	1.6458	0.0249	2.9516 **	3.0101 ***	0.5768 **
				(1.2017)	(1.2420)	(0.0830)	(2.5393)	(2.6043)	(2.5348)
ln_WCap_w				−0.0585 ***	−0.0672 ***	−0.0227 ***
				(−2.7040)	(−3.1670)	(−4.7060)
ln_EBIT_w				0.0471 *	0.0516 **	0.0586 ***
				(1.7640)	(1.9731)	(3.0311)
BGD × Ind			0.0000			−0.0000			0.0000
			(0.6736)			(−0.4805)			(0.5976)
BGD × NonEx						−0.0000			−0.0001 **
						(−0.6857)			(−1.9675)
BGD × CSR						−0.0056 **			−0.0014
						(−2.0662)			(−0.5367)
BGD × Afill			0.0000			0.0018 *			0.0002
			(0.0196)			(1.9410)			(0.2527)
BGD × Bsz			0.0009 *			0.0002			−0.0001
			(1.8809)			(0.9143)			(−0.2900)
BGD × QH			−0.0002			0.0002			0.0001
			(−0.7083)			(1.3245)			(0.5780)
BGD × PBD									0.0005
									(0.2681)
BGD × BER									0.0019
									(1.6436)
BGD × BSP			−0.0029						0.0005
			(−0.2965)						(0.0410)
BGD × BFP									−0.0016
									(−1.5587)
BGD × BMComp			−0.0005
			(−0.5082)
BGD × CEODual			0.0018
			(0.4001)
BGD × Meet						−0.0000
						(−0.2488)
BGD × EBIT						−0.0010 **
						(−1.9919)
_cons	−0.6613 **	−0.6717 **	−0.6142	−0.0085	−0.1573	−0.7220	0.4670	0.2818	0.1655
	(−2.0324)	(−2.0635)	(−1.1179)	(−0.0082)	(−0.1557)	(−0.9655)	(0.4696)	(0.2845)	(0.3085)
F statistic	4.2353 ***	4.0682 ***	3.2068 ***	7.0166 ***	8.2042 ***	5.6917 ***	10.2750 ***	10.4584 ***	6.7768 ***
R-sq	0.2918	0.2958	0.3105	0.2164	0.2511	0.2261	0.2277	0.2373	0.2091
VIF	2.26	7.80		1.57	2.75		1.65	2.78
Obs	204.0000	204.0000	204.0000	714.0000	714.0000	718.0000	1005.0000	1005.0000	1013.0000

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. The numbers in parentheses are t statistics. For definitions of variables, please see Table 4. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

. Leverage is defined as the total debts divided by the total assets ratio (%) and reflects the level of external funding used by the company to finance its assets. While excessive borrowing is riskier, leveraging one’s own funds only likely leads to missed growth opportunities.

Following our three models systematically developed to better predict the preference for borrowing, we find a positive and significant influence of gender diversity on financial leverage (debts/assets) in extended specifications, M2 (0.0525, p < 0.01) and M3 (0.0356, p < 0.01) when gender diversity is below a certain level. Although prediction power is almost similar between models (R-sq is 22% in M2 and 24% in M3, while 29% in M1), we count on the increased number of observations in M3 (1005), therefore consider these model’s results most appropriate and confirm our second hypothesis (H2) predicting that gender-diverse boards increase financial leverage to better exploit growing opportunities.

As expected, testing the existence of non-linear effects of gender diversity on leverage ratio, by including the squared term of BGD (sqBGD) in our specifications, we bring interesting insights and evidence of a curvilinear relationship, based on inverted squared term coefficient estimates (−0.0008, p < 0.01/M2 and −0.0005, p < 0.01/M3, see Table 11). Considering these results, the H2.1 hypothesis is accepted, predicting that there is a turning point for the level of BGD influencing the Leverage. Using the second derivative of our regression function and the coefficient estimates, we have identified that the inflection point for the relation between BGD and Leverage is 34.2% in average (as presented by M2 and M3 for Leverage), above which, higher gender-diverse boards reduce Leverage. VIF has been analyzed for multicollinearity concerns related to independent variables included in all three models employed to predict Leverage (indebtedness risk), and based on their value between 1.56 and 2.26, this concern is neglected.

While BGD partially explains the debt levels for large-listed companies from our sample, we have analyzed a comprehensive number of individual board diversity components and their interplay with BGD. We find a positive influence of board size (0.0247, p < 0.01) and board member independence (0.0029, p < 0.05) on the Leverage ratio. Significantly, we advise that companies with a CSR Committee in place have a 33% (0.3270, p < 0.01) higher debt/asset ratio than the rest. On the other side, memberships (−0.0677, p < 0.1), chair independence (−0.1291, p < 0.05), ex-CEO role of the Chair (−0.1036, p < 0.1), or board tenure (−0.0207, p < 0.1), reduce financial leverage in large companies, suggesting that knowledge outside company (more memberships and independence) prevent higher indebtedness, while, in case of companies with in-firm-knowledge, longer board member experience in such companies reduces debt exposure. These results are conditional on all included variables and might be affected by a possible sample bias, so we recommend treating them with caution.

We further analyzed the possible moderation effect of gender (BGD) along with other individual board diversity components’ influence on Leverage. We find that BGD reduces (−0.0056, p < 0.05) the positive and significant influence of the CSR committee on Leverage (0.3270, p < 0.01), as presented before in Table 11. Similarly, the number of affiliations reduces Leverage (−0.0677, p < 0.1), and, if more women are present on the board, the affiliations’ negative effect diminishes (0.0018, p < 0.1). Board size increases Leverage (0.0247, p < 0.01), and BGD is amplifying this effect (0.0009, p < 0.1). All these findings together denote that the presence of women on boards operates conditionally and jointly with other board attributes and tends to increase leverage in large, listed companies.

Alternative measures for gender diversity have been included in our models to predict the Leverage ratio (Debt/Assets). In case the presence of women on a board is barely visible (WOBMere means presence less than 20%), we find a negative influence of the mere presence of women on the board on the Leverage ratio (−0.1106, p < 0.01/M2), contrary to our OLS results that gender diversity increases Leverage (H2). Discussing companies with above 20% but less than 40% and with more than 40% women on their boards, we find no significant influence of gender on leverage ratio, across models, but the coefficient signs flip to positive, suggesting a possible non-linear effect. These results can be explained by the immediate risk-averse effect on debt level when the first females join the board, an effect that might be further diluted due to the intergroup contact of women with their male counterparts (Pettigrew & Tropp, 2006).

Fixed and random effects models are presented in Table 8, Table 9 and Table 10 for all three models, respectively. Selecting appropriate models using the Hausman test resulted in a positive influence of BGD (0.0008, p < 0.05) on the Leverage ratio (LevDA), confirming that the initial OLS results are robust. Therefore, we confidently confirm our second hypothesis (H2) that companies with more gender-diverse boards tend to leverage more external funding to finance their growth in the case of large-listed companies.

4.3.3. Estimates of the Capital Structure

Capital structure has been analyzed using Debt/Equity as the main proxy and Changes in Long-Term Debt to Capital as an alternative measure of corporate risk outcome.

Baseline results of Ordinary Least Squares (OLS) regression of Debts to Equity (LevDE), on individual board diversity components, are presented in

Table 12. Estimates for Debts/Equity Ratio using OLS regression.

	LevDE (M1)			LevDE (M2)			LevDE (M3)
	(1)	(2)	(1) _int	(1)	(2)	(1) _int	(1)	(2)	(1) _int
BGD	0.0048	0.0252	−0.0035	0.0016 ***	0.0100 ***	0.0721	0.0010 **	0.0065 ***	0.0221
	(0.8591)	(0.7472)	(−0.0542)	(2.7855)	(5.0532)	(0.8309)	(2.0949)	(3.7190)	(0.3476)
sqBGD_w		−0.0003			−0.0001 ***			−0.0001 ***
		(−0.6131)			(−4.4368)			(−3.2620)
BSize_w	0.0079	0.0061	−0.0858	0.0041 **	0.0038 *	0.0187	−0.0004	−0.0006	0.0314
	(0.3822)	(0.2921)	(−1.0735)	(2.0896)	(1.9359)	(0.5029)	(−0.2273)	(−0.3240)	(0.8932)
ln_BMComp_w	0.0970	0.0941	0.1395	0.0169 **	0.0140 *	0.0457	0.0010	−0.0004	−0.0168
	(0.9191)	(0.8892)	(0.7536)	(2.1069)	(1.7624)	(1.2251)	(0.1351)	(−0.0603)	(−0.4378)
BER				−0.0130	−0.0154	−0.0702	0.0015	−0.0003	−0.3510
				(−1.0939)	(−1.3103)	(−1.2945)	(0.1438)	(−0.0273)	(−1.5693)
BSP	0.3728	0.2885	0.2863	−0.1426	−0.1580 *	0.0388	−0.0918	−0.0959	−0.8004
	(1.2813)	(0.8950)	(0.2756)	(−1.5750)	(−1.7664)	(0.0929)	(−1.2651)	(−1.3272)	(−0.3164)
BMeet_w				−0.0023	−0.0025	−0.0134	−0.0030 **	−0.0031 **	−0.0162 **
				(−1.3484)	(−1.4876)	(−0.4564)	(−2.1332)	(−2.1966)	(−2.2335)
BFP				−0.0173	−0.0132	−0.0527	0.0000	0.0030	0.6885 ***
				(−1.5596)	(−1.1947)	(−1.0201)	(0.0022)	(0.3466)	(3.2039)
PBD				0.0104	0.0131	0.0649	0.0251	0.0249	−0.1722
				(0.5993)	(0.7646)	(0.7995)	(1.5677)	(1.5607)	(−0.4980)
BMAfill_w	0.0009	0.0048	0.0959	−0.0182 **	−0.0184 **	−0.4400 **	0.0042	0.0045	−0.1826
	(0.0115)	(0.0580)	(0.3106)	(−2.2037)	(−2.2558)	(−2.5654)	(0.5883)	(0.6309)	(−1.1513)
MaxLenght_w	0.0478	0.0474	0.0448
	(1.4955)	(1.4799)	(1.3265)
BIReel				−0.0170	−0.0210 *	−0.0866	−0.0224 **	−0.0246 **
				(−1.4247)	(−1.7722)	(−1.5884)	(−2.0955)	(−2.3056)
ChairInd				−0.0424 ***	−0.0464 ***	−0.1592 ***	−0.0345 ***	−0.0354 ***	0.0068
				(−3.5620)	(−3.9419)	(−2.8615)	(−3.5415)	(−3.6483)	(0.1250)
ChairExCEO				−0.0251 **	−0.0245 **	−0.1377 **			−0.1537 ***
				(−1.9967)	(−1.9765)	(−2.3662)			(−3.0324)
BTenure_w				−0.0038	−0.0038	−0.0257 **	−0.0034 *	−0.0033	−0.0312 ***
				(−1.6163)	(−1.6433)	(−2.3488)	(−1.6938)	(−1.6375)	(−2.9842)
IndBM_w	0.0059 **	0.0057 **	0.0006	0.0013 ***	0.0013 ***	0.0098 *	0.0008 ***	0.0008 ***	0.0096 *
	(2.4809)	(2.3764)	(0.0480)	(4.4693)	(4.6920)	(1.7253)	(3.2861)	(3.2852)	(1.8619)
CEODual	0.3269 **	0.3265 **	0.6333				−0.0220 **	−0.0221 **	−0.1058 *
	(2.1456)	(2.1393)	(0.9760)				(−2.0741)	(−2.0929)	(−1.9188)
NonExBM_w				−0.0000	0.0001	0.0054	0.0008	0.0007	0.0133
				(−0.0264)	(0.1112)	(0.6353)	(1.5317)	(1.4812)	(1.5422)
CCondNonc				−0.0155	−0.0191 *	0.0089
				(−1.5696)	(−1.9496)	(0.1954)
HComplB				0.0179	0.0164	0.0014	0.0128	0.0113	0.0168
				(1.3901)	(1.2896)	(0.0235)	(1.1893)	(1.0554)	(0.3030)
HLegalB	1.6683 ***	1.6774 ***	1.6314 ***
	(2.7573)	(2.7667)	(2.6358)
BMReelect_w							−0.0059	−0.0054	−0.0051
							(−1.5525)	(−1.4453)	(−0.2597)
CSRCom				0.1017 ***	0.1015 ***	1.1068 ***	0.0945 ***	0.0937 ***	0.1992
				(4.1287)	(4.1769)	(2.7379)	(3.8499)	(3.8359)	(0.4529)
ln_ExecComp_w	−0.1358 **	−0.1398 **	−0.1400 **				0.0022	0.0021	−0.0263
	(−2.0756)	(−2.1226)	(−2.0799)				(0.4015)	(0.3790)	(−0.9395)
QuotaHist_w	0.0445 ***	0.0445 ***	0.1011 *	0.0009	0.0010	−0.0340	−0.0011	−0.0010	−0.0023
	(2.6968)	(2.6948)	(1.8824)	(0.5720)	(0.6442)	(−1.2552)	(−0.9051)	(−0.7940)	(−0.0770)
Age_w/ln_Age_w	−0.0015	−0.0015	−0.0016	−0.0118 **	−0.0104 *	−0.0133	−0.0001	−0.0001	0.0003
	(−0.9470)	(−0.9777)	(−0.9567)	(−2.0198)	(−1.7884)	(−0.4920)	(−1.1026)	(−1.0696)	(0.4388)
ln_MarketCapYE_w	−0.0866	−0.0878	−0.1016				0.0229 ***	0.0223 ***	0.0134
	(−1.2764)	(−1.2905)	(−1.4458)				(4.1189)	(4.0240)	(0.4744)
ln_CSTI_w	0.0812	0.0806	0.0967
	(1.4171)	(1.4034)	(1.5442)
ROE_w	1.6933 ***	1.6692 ***	1.7422 ***	0.0867 *	0.0641	1.1292 ***
	(3.7732)	(3.6991)	(3.7242)	(1.7187)	(1.2813)	(4.7986)
ROA_w							−0.8293 ***	−0.8051 ***	−3.6694 ***
							(−8.2152)	(−7.9927)	(−7.0413)
OE_w							−0.0166 ***	−0.0162 ***	0.0726 ***
							(−3.2574)	(−3.2087)	(2.7850)
ETR_w	0.9959 **	1.0069 **	1.1464 **	0.0860	0.0868 *	0.6289 ***	0.0023	0.0051	−0.0876
	(2.0415)	(2.0592)	(2.2574)	(1.6323)	(1.6698)	(2.6394)	(0.0538)	(0.1216)	(−0.4064)
ART_w				−0.0042 ***	−0.0039 ***	−0.0165 ***	−0.0020 **	−0.0020 **	−0.0134 ***
				(−4.0024)	(−3.8019)	(−3.4430)	(−2.4117)	(−2.4254)	(−3.2643)
ln_InvTo_w	0.0054 *	0.0053 *	0.0050	0.0079	0.0074	0.1115 ***	0.0011 ***	0.0011 ***	0.0088 ***
	(1.7881)	(1.7434)	(1.6024)	(1.5412)	(1.4464)	(4.6883)	(4.0655)	(3.9984)	(5.8399)
ln_FCF_w	0.0124	0.0110	0.0214
	(0.2075)	(0.1844)	(0.3439)
DivYield_w				0.0899	0.0959	1.4265	0.6169 ***	0.6268 ***	2.7049 **
				(0.3020)	(0.3266)	(1.0452)	(2.7233)	(2.7802)	(2.3315)
ln_WCap_w				−0.0218 ***	−0.0234 ***	−0.0652 ***
				(−4.5838)	(−4.9565)	(−2.9749)
ln_EBIT_w				0.0222 ***	0.0228 ***	0.1388
				(3.7644)	(3.9317)	(1.5837)
BGD × Ind			0.0001			−0.0002			−0.0002 *
			(0.4454)			(−1.1906)			(−1.7626)
BGD × NonEx						−0.0002			−0.0004
						(−0.8536)			(−1.6345)
BGD × CSR						−0.0234 *			0.0037
						(−1.8992)			(0.2744)
BGD × Afill			−0.0033			0.0097 **			0.0048
			(−0.3835)			(2.2459)			(1.2291)
BGD × Bsz			0.0026			0.0001			−0.0009
			(1.2264)			(0.1370)			(−0.9421)
BGD × QH			−0.0017			0.0008			−0.0002
			(−1.1830)			(1.1672)			(−0.1965)
BGD × PBD									0.0077
									(0.7941)
BGD × BER									0.0097 *
									(1.6498)
BGD × BSP			−0.0092						0.0210
			(−0.2101)						(0.3482)
BGD × BFP									−0.0169 ***
									(−3.2199)
BGD × BMComp			−0.0003
			(−0.0604)
BGD × CEODual			−0.0095
			(−0.4580)
BGD × Meet						0.0001
						(0.1752)
BGD × EBIT						−0.0024
						(−1.0855)
_cons	−0.6776	−0.7098	−0.4442	0.1888	0.1587	−1.9601	0.2925	0.2572	0.3888
	(−0.4656)	(−0.4866)	(−0.1801)	(0.8311)	(0.7076)	(−0.5786)	(1.5048)	(1.3275)	(0.1429)
F statistic	3.9474 ***	3.7467 ***	2.9259 ***	6.8945 ***	7.5314 ***	5.7717 ***	8.6223 ***	8.7734 ***	8.4455 ***
R-sq	0.2786	0.2801	0.2924	0.2125	0.2343	0.2296	0.1970	0.2056	0.2494
VIF	2.26	7.79		1.57	2.74		1.65	2.77
Obs	203.0000	203.0000	203.0000	718.0000	718.0000	714.0000	1013.0000	1013.0000	1005.0000

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. The numbers in parentheses are t statistics. For definitions of variables, please see Table 4. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

. Capital structure ratio is defined as total debt divided by total equity and reflects the company’s preference for the mix of debt and equity structure used to finance its spending. Preference for debt is increasingly considered by many companies to finance their assets, since access to capital instruments is usually more expensive, whereas debts are less expensive and improve companies’ performance. Nevertheless, long-term indebtedness poses significant risks to equity investors, primarily by increasing a company’s financial leverage, which may either potentiate the revenues or generate bigger losses.

We find a positive and statistically significant influence of gender diversity on Debt/Equity ratio (0.0100, p < 0.01/M2 and 0.0065, p < 0.01/M3, see Table 12). The prediction power of employed models is almost similar (R-sq is 21% in M2 and 20% in M3, while 28% in M1), while the highest number of observations is provided by M3 (1013). Controlling for possible non-linear effects of BGD on debt evolution, by adding the squared term of BGD in our specifications, we identified a turning point (−0.0001, p < 0.01) in the relation between BGD and Debt/Equity ratio at the level of 41.25% in average, starting when, female presence on the board reduce debts, as presented in Table 12. These results might be explained by the intergroup contacts, which may erode initial behavior contrasts when the first women join the board (Pettigrew & Tropp, 2006). Discussions on critical mass theory suggest that the impact of women on boards increases with their number; thus, when women’s presence is relevant, the way they influence risk-taking leans toward their traditional risk-averse behavior, therefore reducing indebtedness. These findings align with prior studies by Adams and Funk (2012), who argue that women directors may, in certain contexts, make riskier decisions than male directors, and this can, in some cases, adversely affect profitability and firm value. VIF between 1.57 and 2.26 for Debt/Equity models does not suggest misspecification and multicollinearity bias in our independent variables; this concern is set aside.

As more corporate governance variables are included in our more complex specification models (see Table 12), we find that BGD remains positive and significant, although its coefficient estimates diminish. This evolution might be explained by the influence of the comprehensive set of variables included in our specifications, exerting a possible suppressor effect on the BGD influence on corporate risk outcome. Discussing, though, the influence of other individual board diversity components on Debts/Equity we find that board size (0.0041, p < 0.05), compensation (0.0169, p < 0.05), members independence (0.0059, p < 0.05) and existence of CSR committee (0.1017, p < 0.01), increase indebtedness, in case of large, listed companies in European context, included in our analysis. Contrariwise, board meeting attendance (−0.0030, p < 0.05), memberships (−0.0182, p < 0.05), board individual reelection (−0.0224, p < 0.05), chair independence (−0.0424, p < 0.01) or ex-CEO as Chair (−0.0251, p < 0.05), board tenure (−0.0034, p < 0.1) or CEO duality (−0.0220, p < 0.05), are all negative and significant predictors of indebtedness. We highlight that large companies having a CSR committee in place have a Debt/Equity ratio with 10.2% higher than the rest (0.1017, p < 0.01). These results align with prior discussed findings from specification models to predict debts/assets ratio.

Looking for possible moderation effect of BGD and other individual board diversity components, and their combined influence on Debts/Equity, we find that BGD reduces (−0.0002, p < 0.1) the positive and significant influence of board members independence (0.0059, p < 0.05) on debts level and also, reduces (−0.0234, p < 0.1) the positive effect of CSR committee (0.1017, p < 0.01) on Debts/Equity, as presented in Table 12.

In case of alternative measures for BGD if presence of woman on board is barely visible (WOBMere means presence less than 20%) (see Supplementary File S1), we find a negative influence of merely presence of woman on board on Debts/Equity ratio (−0.3211, p < 0.05/M2 and −0.0580, p < 0.1/M3), contrary our OLS results that gender diversity increases debts (H2). Discussing companies with above 20% but less than 40% (see Supplementary File S2) and with more than 40% women on the board (see Supplementary File S3), we find no significant influence of gender on debt/equity ratio, across models. These results can be explained by the risk-averse effect on debt level when the first female joins the board, an effect that is further diluting, as supported by the intergroup contact theory of Pettigrew and Tropp (2006).

Table 8, Table 9 and Table 10 present fixed and random effects models predicting the Debt/Equity ratio for all three specifications analyzed in our study. After selecting appropriate models using the Hausman test, we find a positive influence (0.0048, p < 0.01/M2, 0.0053, p < 0.01/M3) of BGD on Debt/Equity ratio (LevDE), therefore, confirming our second hypothesis that gender-diverse boards increase debts (H2). Further, we underscore the non-linear effect, arguing that gender-diverse boards increase in debt if women’s presence on the board reaches the critical mass but is still not gender-balanced (between 40% and 50%). In the case of gender-balanced boards, their influence on debt level becomes negative, predicting that the influence of gender-diverse boards on debt level changes their traditional risk-averse behavior, therefore, reducing indebtedness when female presence on boards becomes relevant, consequently, confirming our H2.1 hypothesis.

Discussing long term indebtedness, measured by Changes of Long-Term Debts to Capital as an alternative proxy for corporate risk outcome describing capital structure, we find no significant influence of BGD on such a change using OLS models, suggesting that financial decisions are mainly idiosyncratic decision, within-firms specific and industry contingent, therefore changes in long term indebtedness strategy, is less probably to be attributable to increased gender diversity on board (BGD).

We further observe the influence of other individual board diversity components on changes in long-term debts. Board policies (BSP, PBD) or independent directors are more likely to increase long-term debts, although long-term indebtedness is not necessarily a riskier decision, but might suggest the company’s preference for more stable funding, without repayment pressure on the company cashflow. Board size (−0.0015, p < 0.1), meeting attendance (−0.0014, p < 0.1), board functions policy (BFP) (−0.0116, p < 0.05), memberships (−0.0064, p < 0.1), chair independence (−0.0089, p < 0.1), code of conduct non-compliance procedures (−0.0075, p < 0.1) are all negative predictors for long term debts to capital ratio. Companies having policies to maintain well-balanced membership on the board (BSP) (0.0781, p < 0.1) and head of legal reporting to the board (0.1671, p < 0.01) lead to more risk-taking and long-term debt hikes, compared with the rest of the cases.

Detailed results related to Changes of Long-Term Debts to Capital as an alternative measure for long-term indebtedness can be found in Supplementary File S4.

4.3.4. Estimates of the Z-Score

Baseline results of Ordinary Least Squares (OLS) regression to predict corporate risk outcome, proxied by Z-score (values lower than 1.8 typically indicate the distress zone), are presented in

Table 13. Estimates for Z-Score using OLS regression.

	ZScoreMNM (M1)			ZScoreMNM (M2)			ZScoreMNM (M3)
	(1)	(2)	(1) _int	(1)	(2)	(1) _int	(1)	(2)	(1) _int
BGD	0.0068	−0.1079	−0.4530 **	0.0218 *	−0.1226 ***	−0.0140	0.0093	−0.0434 *	−0.0989
	(0.3330)	(−0.8727)	(−2.1087)	(1.9310)	(−3.0927)	(−0.0356)	(1.3024)	(−1.6913)	(−0.5516)
sqBGD_w		0.0017			0.0023 ***			0.0008 **
		(0.9410)			(3.7981)			(2.1383)
BSize_w	−0.3990 ***	−0.3894 ***	0.1617	−0.1516 ***	−0.1452 ***	0.0975	−0.0844 ***	−0.0828 ***	0.1459
	(−5.2303)	(−5.0583)	(0.5976)	(−3.9138)	(−3.7819)	(0.6054)	(−3.3334)	(−3.2729)	(1.4635)
ln_BMComp_w	−0.4178	−0.3971	−0.1851	−0.2102	−0.1593	−0.1845	−0.0247	−0.0099	0.0244
	(−1.0749)	(−1.0197)	(−0.2955)	(−1.3244)	(−1.0102)	(−1.1355)	(−0.2270)	(−0.0912)	(0.2227)
BER				−0.0383	−0.0011	−0.0547	−0.1471	−0.1312	−0.1457
				(−0.1602)	(−0.0048)	(−0.2279)	(−0.9800)	(−0.8748)	(−0.2220)
BSP	−4.2963 ***	−3.8079 ***	−12.8697 ***	2.1161	2.4021	3.2907 *	2.2768 **	2.3219 **	−4.1565
	(−4.0134)	(−3.2000)	(−3.6698)	(1.1929)	(1.3664)	(1.8316)	(2.1752)	(2.2219)	(−0.5831)
BMeet_w				0.0607 *	0.0647 *	0.0784	0.0839 ***	0.0847 ***	0.0782 ***
				(1.7935)	(1.9325)	(0.6129)	(4.1109)	(4.1553)	(3.8108)
BFP				−0.1057	−0.1672	−0.1313	−0.1752	−0.2020	−1.8853 ***
				(−0.4801)	(−0.7652)	(−0.5873)	(−1.3760)	(−1.5812)	(−3.0963)
PBD				−0.6883 *	−0.7364 **	−0.8276 **	−0.8869 ***	−0.8812 ***	−1.8369 *
				(−1.9201)	(−2.0736)	(−2.2497)	(−3.7008)	(−3.6835)	(−1.8575)
BMAfill_w	−0.7396 **	−0.7622 **	−3.4375 ***	−0.3444 **	−0.3406 **	0.9302	−0.2548 **	−0.2577 **	0.4051
	(−2.4320)	(−2.4978)	(−3.3199)	(−2.0894)	(−2.0867)	(1.2259)	(−2.4353)	(−2.4676)	(0.8959)
MaxLenght_w	0.2844 **	0.2869 **	0.1828
	(2.4033)	(2.4234)	(1.5992)
BIReel				−0.0230	0.0506	0.0079	0.0576	0.0791	0.0557
				(−0.0968)	(0.2143)	(0.0333)	(0.3720)	(0.5107)	(0.3611)
ChairInd				0.0734	0.1350	0.1720	0.2042	0.2114	0.2764 *
				(0.3097)	(0.5736)	(0.7107)	(1.4410)	(1.4945)	(1.9251)
ChairExCEO				0.2399	0.2202	0.3566
				(0.9498)	(0.8802)	(1.3914)
BTenure_w				0.1074 **	0.1069 **	0.1207 **	0.0501 *	0.0488 *	0.0514 *
				(2.2940)	(2.3048)	(2.5382)	(1.6943)	(1.6518)	(1.7261)
IndBM_w	−0.0031	−0.0021	−0.0274	−0.0092	−0.0099 *	0.0088	−0.0045	−0.0045	0.0276 *
	(−0.3594)	(−0.2327)	(−0.6843)	(−1.5908)	(−1.7291)	(0.3570)	(−1.2448)	(−1.2358)	(1.8721)
CEODual	−0.2317	−0.2282	−1.9477				0.0560	0.0544	0.1249
	(−0.4112)	(−0.4047)	(−0.8880)				(0.3605)	(0.3508)	(0.7949)
NonExBM_w				0.0231 **	0.0210 *	0.0546	−0.0023	−0.0022	−0.0141
				(2.0100)	(1.8383)	(1.4666)	(−0.3060)	(−0.3032)	(−0.5704)
CCondNonc				−0.6517 ***	−0.5994 ***	−0.6583 ***
				(−3.3075)	(−3.0651)	(−3.2918)
HComplB				0.4129	0.4141	0.5434 **	0.0718	0.0791	0.1235
				(1.5903)	(1.6107)	(2.0637)	(0.4563)	(0.5035)	(0.7790)
HLegalB	−1.8376	−1.8936	−0.7969
	(−0.8211)	(−0.8456)	(−0.3807)
BMReelect_w							0.1058 *	0.1006 *	0.0660
							(1.9349)	(1.8405)	(1.1983)
CSRCom				−0.6928	−0.6381	−5.2520 ***	−0.8977 **	−0.8729 **	−3.1909 **
				(−1.3590)	(−1.2637)	(−3.0083)	(−2.4004)	(−2.3371)	(−2.5578)
ln_ExecComp_w	0.2705	0.2908	0.1451				−0.0272	−0.0270	−0.0212
	(1.1202)	(1.1991)	(0.6379)				(−0.3415)	(−0.3397)	(−0.2674)
QuotaHist_w	−0.0326	−0.0333	0.3036 *	−0.0375	−0.0391	0.2659 **	0.0072	0.0057	0.2247 ***
	(−0.5358)	(−0.5478)	(1.6912)	(−1.2235)	(−1.2900)	(2.2530)	(0.3948)	(0.3137)	(2.5928)
Age_w/ln_Age_w	−0.0050	−0.0047	−0.0099 *	0.2713 **	0.2470 **	0.2875 **	0.0032 *	0.0031 *	0.0037 **
	(−0.8808)	(−0.8174)	(−1.7302)	(2.3325)	(2.1410)	(2.4556)	(1.7856)	(1.7576)	(2.0910)
ln_MarketCapYE_w	0.9019 ***	0.9087 ***	1.0220 ***				0.5066 ***	0.5116 ***	0.5227 ***
	(3.5924)	(3.6167)	(4.2977)				(6.2858)	(6.3566)	(6.5137)
ln_CSTI_w	−0.2775	−0.2728	−0.3793 *
	(−1.3099)	(−1.2873)	(−1.7899)
ROE_w	7.3449 ***	7.5140 ***	7.4325 ***	9.9795 ***	10.3508 ***	9.6640 ***
	(4.4811)	(4.5555)	(4.7462)	(9.9728)	(10.3967)	(9.5696)
ROA_w							36.2042 ***	35.9623 ***	35.8761 ***
							(24.5828)	(24.3917)	(24.3551)
OE_w							−0.6435 ***	−0.6465 ***	−0.6468 ***
							(−8.7599)	(−8.8156)	(−8.8438)
ETR_w	−1.9276	−1.9449	−1.9268	−1.3926	−1.3634	−1.4866	1.7977 ***	1.7782 ***	1.8292 ***
	(−1.0884)	(−1.0978)	(−1.1431)	(−1.3222)	(−1.3073)	(−1.4219)	(2.9018)	(2.8753)	(2.9757)
ART_w				0.0565 ***	0.0523 **	0.0507 **	0.0273 **	0.0273 **	0.0249 **
				(2.7561)	(2.5754)	(2.4582)	(2.3376)	(2.3364)	(2.1397)
ln_InvTo_w	−0.0054	−0.0046	−0.0074	−0.3391 ***	−0.3275 ***	−0.3598 ***	−0.0090 **	−0.0088 **	−0.0079 *
	(−0.4880)	(−0.4126)	(−0.7013)	(−3.3306)	(−3.2476)	(−3.5155)	(−2.2215)	(−2.1691)	(−1.9498)
ln_FCF_w	−0.4708 **	−0.4660 **	−0.5791 ***
	(−2.1460)	(−2.1226)	(−2.7593)
DivYield_w				−44.6737 ***	−44.8289 ***	−43.2286 ***	−30.0977 ***	−30.2481 ***	−29.6647 ***
				(−7.5960)	(−7.6983)	(−7.3237)	(−9.0866)	(−9.1468)	(−8.9860)
ln_WCap_w				0.1390	0.1604 *	0.1365
				(1.4783)	(1.7194)	(1.4445)
ln_EBIT_w				−0.3027 ***	−0.3159 ***	−0.7712 *
				(−2.6187)	(−2.7588)	(−1.8871)
BGD × Ind			0.0005			−0.0005			−0.0009 **
			(0.4795)			(−0.8308)			(−2.2995)
BGD × NonEx						−0.0009			0.0003
						(−0.9118)			(0.5586)
BGD × CSR						0.1450 ***			0.0731 *
						(2.6760)			(1.8814)
BGD × Afill			0.0808 ***			−0.0319 *			−0.0167
			(2.8101)			(−1.6788)			(−1.4828)
BGD × Bsz			−0.0130 *			−0.0068			−0.0063 **
			(−1.8364)			(−1.5963)			(−2.4413)
BGD × QH			−0.0093 *			−0.0081 ***			−0.0056 **
			(−1.9357)			(−2.6344)			(−2.5324)
BGD × PBD									0.0249
									(0.8940)
BGD × BER									−0.0007
									(−0.0389)
BGD × BSP			0.5854 ***						0.1637
			(3.9667)						(0.9618)
BGD × BFP									0.0422 ***
									(2.8326)
BGD × BMComp			−0.0035
			(−0.2451)
BGD × CEODual			0.0430
			(0.6129)
BGD × Meet						−0.0009
						(−0.2791)
BGD × EBIT						0.0130
						(1.2508)
_cons	5.2197	5.3779	11.6777	2.2825	2.8710	2.5925	−2.9191	−2.5412	0.4848
	(0.9699)	(0.9985)	(1.3995)	(0.5047)	(0.6408)	(0.1686)	(−1.0301)	(−0.8967)	(0.0631)
F statistic	12.6303 ***	12.0047 ***	12.2078 ***	21.7586 ***	21.9178 ***	17.6067 ***	70.8029 ***	68.7730 ***	54.1310 ***
R-sq	0.5513	0.5535	0.6316	0.4676	0.4788	0.4825	0.6733	0.6748	0.6836
VIF	2.26	7.80		1.57	2.74		1.65	2.77
Obs	204.0000	204.0000	204.0000	697.0000	697.0000	697.0000	991.0000	991.0000	991.0000

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. The numbers in parentheses are t statistics. For definitions of variables, please see Table 4. Z-score has been developed by Edward Altman after 2008 financial crisis and is counted based on the following formula: Altman Z-Score = 1.2A + 1.4B + 3.3C + 0.6D + 1.0E, where A = working capital/total assets, B = retained earnings/total assets, C = earnings before interest and tax/total assets, D = market value of equity/total liabilities, E = sales/total assets. For this study, the value for Z-score is retrieved from LSEG Workspace. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

. While the lower Z-scores indicate the incidence of the risk of failure, negative effects of our independent variables on Z-score denote higher bankruptcy risk.

We find that the coefficient of term BGD is positive and statistically significant in linear model M2 (0.0218, p < 0.1), initially suggesting that a gender-diverse board increases the Z-Score. Looking at the potential non-linear effect of BGD, we re-run the model including the squared term (sqBGD). We find that the coefficient sign of the linear term flips to negative (−0.1226, p < 0.01), while the squared term coefficient changes to positive (0.0023, p < 0.01/M2, 0.0008, p < 0.05/M3), signaling a curvilinear relationship that needs to be discussed. Since the sign for the sqBGD changed to positive, this indicates that the curve is convex (U-shape). This change in sign for the linear term’s coefficient, after adding the squared term, suggests that the initial linear model was likely misspecified and not able to capture the data’s curvature. Thus, we conclude that the total effect of BGD on Z-score depends on the value of BGD, meaning reduce Z-score if its level is below a certain level and increases Z-score when BGD exceeds this threshold. Using the second derivative of our regression function with squared term of BGD, we identified the turning point at the level of 40% in average (26.65%/M2 and 54.25%/M3) starting when diversity level increase Z-score, thus below 40% on average gender diversity reduces Z-score (see Table 13) Models’ prediction power (R-sq) spans between 46% (M2), 55% (M1) and 68% (M3), while the highest number of observations (991) is provided by model M3. Therefore, we consider these models appropriate for our study and partially validate our third hypothesis (H3), arguing that the positive influence of BGD on Z-score occurs when the presence of women on the board exceeds 40%. Further on, we accept H3.1, confirming the existence of a non-linear relationship, arguing that Z-score increases when the presence of women on the board becomes relevant, since the turning point is validating the level of 40% on average. VIF span between 1.57 and 2.26, suggesting there is no misspecification or multicollinearity concern related to independent variables in our models.

Fixed and random effects models predicting Z-score are presented in Table 8, Table 9 and Table 10 for all three models, respectively. After selecting appropriate models using the Hausman test, it resulted in a negative but non-significant influence of BGD on Z-Score, although the results are not consistent across models. Discussing these results in the context of non-linear regression results, we conclude that gender-diverse boards reduce Z-score if BGD lasts below 40%, while starting from this relevance level, it improves Z-score.

In case of alternative measures for BGD if presence of woman on board is barely visible (WOBMere means presence less than 20%) (see Supplementary File S1), we find a negative influence of merely presence of woman on board on Z-score (−1.2630, p < 0.01/M3), confirming our OLS results, that gender diversity reduce Z-score when women are below the relevancy (about 40%), opposite to our H4 hypothesis. Discussing companies with above 20% but less than 40% women on the board (see Supplementary File S2), we still find a negative influence of gender diversity on Z-score (−0.4933, p < 0.05), confirming that their presence is still weak to reverse the trend. In case of a balanced board (WOBBal) with more than 40% woman on board (see Supplementary File S3), we find that on average, ceteris paribus (holding other variables constant), positive influence of gender on Z-score, is 55% higher (0.5594, p < 0.01) in companies with gender balanced boards, compared to the rest of the sample. These results from models using alternative measures for gender diversity, consistently support our main variable results (BGD), along with the findings related existence of non-linear effect, meaning, we face a negative effect of gender diversity on Z-score if women presence is below 40%, while this relation is turning to positive when women held more than 40% of board seats.

Discussing other individual board diversity components (see Table 13), we find that larger boards reduce Z-score (−0.1516, p < 0.01). This is in line with Anderson et al. (2011) who argue that directors bringing varied perspectives to board deliberations, as well as larger boards, can increase conflict among board members and impede decision-making process. Similarly, existence of a policy regarding the gender diversity of its board PBD (−0.8869, p < 0.01) or compliance with its code of conduct (−0.6517, p < 0.01), memberships (−0.3444, p < 0.05), CSR committee (−0.8977, p < 0.05) are negative predictors for Z-score. In the context of the nonlinear effect of BGD discussed before, their influence might also be nonlinear and so, dependent on the variable value. Conversely, board structure policy (2.2768, p < 0.05), meeting attendance 0.0839, p < 0.01), board tenure (0.1074, p < 0.05), and non-executive board members (0.0231, p < 0.05) positively influence Z-score, thus reducing bankruptcy risk. Companies that have a board structure policy in place (BSP) have significantly higher Z-scores than the rest of the sample.

Looking for a possible moderation effect of BGD on other individual board diversity components (see Table 13), we find that jointly, BGD and CSR increase Z-score (0.0731, p < 0.1), while their individual effect is negative. The existence of both BGD and BSP increases the Z-score (0.5854, p < 0.01), likewise the interaction of BGD with board function policy (BFP) (0.0422, p < 0.01), while their individual effect is not consistent across models. When BGD intersect board size, we find a negative effect (−0.0063, p < 0.05) of their joint interaction on Z-score, amplifying their negative individual effect, suggesting that women merely present women, when combined with increasing board size, this cannot reverse their individual negative impact on Z-Score.

4.3.5. Estimates of the Liquidity Risk

Liquidity has been thoroughly analyzed using the Current Ratio as the main proxy and the Quick Ratio as an alternative measure for corporate risk outcome.

Current Ratio (CurrR) is defined as Total Assets/Total Liabilities, and Quick Ratio (QR) represents Total Current Assets minus Inventory divided by Total Current Liabilities, expressed as a percentage. Since higher liquidity ratios equal lower liquidity risk, we will consider this relation in the interpretation of our results. The baseline results of Ordinary Least Squares (OLS) regression of Current Ratio on individual board diversity components are presented in

Table 14. Estimates for Current Ratio using OLS regression.

	CurrR (M1)			CurrR (M2)			CurrR (M3)
	(1)	(2)	(1) _int	(1)	(2)	(1) _int	(1)	(2)	(1) _int
BGD	−0.0087	−0.0220	0.0481	0.0002	−0.0183 **	0.1137	−0.0005	−0.0069	0.0392
	(−1.5520)	(−0.6501)	(0.7515)	(0.1044)	(−2.1825)	(1.4203)	(−0.1929)	(−0.7785)	(0.6201)
sqBGD_w		0.0002			0.0003 **			0.0001
		(0.3985)			(2.3058)			(0.7547)
BSize_w	−0.0269	−0.0258	0.0089	−0.0152 *	−0.0144 *	−0.0456	0.0095	0.0097	0.0301
	(−1.2913)	(−1.2243)	(0.1105)	(−1.8494)	(−1.7601)	(−1.3294)	(1.0806)	(1.1020)	(0.8593)
ln_BMComp_w	−0.1540	−0.1516	−0.1597	−0.1651 ***	−0.1587 ***	−0.1621 ***	0.0080	0.0097	0.0087
	(−1.4487)	(−1.4206)	(−0.8556)	(−4.9205)	(−4.7291)	(−4.7195)	(0.2122)	(0.2571)	(0.2284)
BER				−0.0392	−0.0339	−0.0309	−0.1297 **	−0.1277 **	0.2679
				(−0.7865)	(−0.6827)	(−0.6188)	(−2.5179)	(−2.4742)	(1.2024)
BSP	0.8344 ***	0.8911 ***	1.6841	−0.1638	−0.1301	−0.0020	0.3090	0.3137	1.0049
	(2.8510)	(2.7335)	(1.6122)	(−0.4320)	(−0.3440)	(−0.0053)	(0.8431)	(0.8558)	(0.3983)
BMeet_w				0.0074	0.0078	0.0234	0.0171 **	0.0172 **	0.0148 **
				(1.0353)	(1.1014)	(0.8635)	(2.4193)	(2.4308)	(2.0547)
BFP				0.0772 *	0.0680	0.0715	0.0099	0.0064	−0.0299
				(1.6588)	(1.4604)	(1.5016)	(0.2245)	(0.1435)	(−0.1396)
PBD				−0.0703	−0.0762	−0.0921	−0.0731	−0.0728	0.1423
				(−0.9643)	(−1.0487)	(−1.2311)	(−0.9038)	(−0.9002)	(0.4131)
BMAfill_w	−0.1928 **	−0.1954 **	−0.3551	−0.1332 ***	−0.1328 ***	−0.2995 *	−0.1105 ***	−0.1108 ***	−0.3878 **
	(−2.3185)	(−2.3375)	(−1.1512)	(−3.8516)	(−3.8518)	(−1.8945)	(−3.0461)	(−3.0545)	(−2.4607)
MaxLenght_w	−0.0219	−0.0216	−0.0216
	(−0.6774)	(−0.6666)	(−0.6332)
BIReel				0.1162 **	0.1249 **	0.1143 **	0.0381	0.0406	0.0248
				(2.3200)	(2.4934)	(2.2786)	(0.7058)	(0.7514)	(0.4583)
ChairInd				−0.0454	−0.0365	−0.0448	0.0567	0.0578	0.0571
				(−0.9110)	(−0.7330)	(−0.8766)	(1.1542)	(1.1746)	(1.1360)
ChairExCEO				−0.1824 ***	−0.1837 ***	−0.1809 ***
				(−3.4692)	(−3.5042)	(−3.3781)
BTenure_w				0.0097	0.0098	0.0091	0.0269 ***	0.0267 ***	0.0244 **
				(0.9850)	(0.9908)	(0.8998)	(2.6196)	(2.6036)	(2.3374)
IndBM_w	−0.0000	0.0001	−0.0008	0.0011	0.0010	0.0029	0.0022 *	0.0022 *	0.0107 **
	(−0.0015)	(0.0508)	(−0.0686)	(0.9313)	(0.8474)	(0.5480)	(1.7461)	(1.7496)	(2.0837)
CEODual	−0.1698	−0.1694	−0.6601				−0.1153 **	−0.1152 **	−0.1075 **
	(−1.1021)	(−1.0969)	(−1.0103)				(−2.1534)	(−2.1509)	(−1.9631)
NonExBM_w				0.0031	0.0029	0.0143 *	−0.0027	−0.0026	−0.0006
				(1.2706)	(1.2023)	(1.8119)	(−1.0463)	(−1.0325)	(−0.0688)
CCondNonc				−0.0739 *	−0.0660	−0.0611
				(−1.7863)	(−1.5965)	(−1.4530)
HComplB				0.1147 **	0.1180 **	0.1303 **	0.1115 **	0.1132 **	0.1343 **
				(2.1247)	(2.1923)	(2.3638)	(2.0591)	(2.0888)	(2.4379)
HLegalB	−0.3461	−0.3526	−0.2654
	(−0.5656)	(−0.5747)	(−0.4256)
BMReelect_w							0.0423 **	0.0418 **	0.0356 *
							(2.2201)	(2.1919)	(1.8325)
CSRCom				0.1516	0.1520	−0.4097	0.0838	0.0847	−0.1572
				(1.4700)	(1.4784)	(−1.1006)	(0.6762)	(0.6836)	(−0.3589)
ln_ExecComp_w	−0.0791	−0.0767	−0.0814				−0.0631 **	−0.0630 **	−0.0683 **
	(−1.1975)	(−1.1546)	(−1.2012)				(−2.2892)	(−2.2829)	(−2.4622)
QuotaHist_w	−0.0298 *	−0.0299 *	−0.0028	0.0081	0.0079	0.0617 **	0.0068	0.0066	0.0257
	(−1.7892)	(−1.7901)	(−0.0527)	(1.2744)	(1.2447)	(2.4728)	(1.0783)	(1.0508)	(0.8500)
Age_w/ln_Age_w	0.0019	0.0020	0.0013	0.0402	0.0369	0.0356	0.0007	0.0007	0.0007
	(1.2196)	(1.2402)	(0.7691)	(1.6354)	(1.5051)	(1.4314)	(1.1446)	(1.1354)	(1.0693)
ln_MarketCapYE_w	−0.1991 ***	−0.1983 ***	−0.1814 **				−0.0729 ***	−0.0721 **	−0.0752 ***
	(−2.9001)	(−2.8809)	(−2.5615)				(−2.5968)	(−2.5686)	(−2.6661)
ln_CSTI_w	0.1661 ***	0.1666 ***	0.1694 ***
	(2.8677)	(2.8697)	(2.6830)
ROE_w	−0.8286 *	−0.8090 *	−0.8741 *	0.3335	0.3832 *	0.2740
	(−1.8491)	(−1.7905)	(−1.8739)	(1.5791)	(1.8105)	(1.2788)
ROA_w							4.1065 ***	4.0781 ***	4.0156 ***
							(8.0590)	(7.9798)	(7.7785)
OE_w							−0.0792 ***	−0.0795 ***	−0.0764 ***
							(−3.0859)	(−3.0994)	(−2.9602)
ETR_w	−1.0327 **	−1.0347 **	−1.0945 **	−0.6745 ***	−0.6763 ***	−0.6747 ***	0.5458 **	0.5424 **	0.5422 **
	(−2.1329)	(−2.1321)	(−2.1799)	(−3.0584)	(−3.0760)	(−3.0673)	(2.5569)	(2.5402)	(2.5356)
ART_w				0.0156 ***	0.0150 ***	0.0161 ***	0.0126 ***	0.0126 ***	0.0129 ***
				(3.5759)	(3.4526)	(3.6551)	(3.0862)	(3.0860)	(3.1450)
ln_InvTo_w	−0.0183 ***	−0.0183 ***	−0.0177 ***	−0.1269 ***	−0.1256 ***	−0.1315 ***	−0.0111 ***	−0.0111 ***	−0.0113 ***
	(−6.0302)	(−5.9668)	(−5.6647)	(−5.8855)	(−5.8421)	(−6.0522)	(−7.8748)	(−7.8505)	(−7.9160)
ln_FCF_w	−0.0475	−0.0469	−0.0678
	(−0.7914)	(−0.7801)	(−1.0845)
DivYield_w				1.1531	1.1398	1.3586	−0.4053	−0.4169	−0.2783
				(0.9253)	(0.9175)	(1.0834)	(−0.3544)	(−0.3645)	(−0.2418)
ln_WCap_w				0.3127 ***	0.3160 ***	0.3103 ***
				(15.6742)	(15.8494)	(15.4014)
ln_EBIT_w				−0.2455 ***	−0.2470 ***	−0.1476 *
				(−9.9639)	(−10.0528)	(−1.8252)
BGD × Ind			0.0000			−0.0001			−0.0002 *
			(0.0645)			(−0.3822)			(−1.7250)
BGD × NonEx						−0.0003			−0.0000
						(−1.5387)			(−0.1527)
BGD × CSR						0.0191 *			0.0079
						(1.6821)			(0.5935)
BGD × Afill			0.0046			0.0043			0.0071 *
			(0.5339)			(1.0943)			(1.8063)
BGD * × Bsz			−0.0011			0.0008			−0.0006
			(−0.5145)			(0.8914)			(−0.6274)
BGD × QH			−0.0006			−0.0015 **			−0.0005
			(−0.4479)			(−2.2718)			(−0.6489)
BGD × PBD									−0.0066
									(−0.6832)
BGD × BER									−0.0107 *
									(−1.8265)
BGD × BSP			−0.0507						−0.0133
			(−1.1532)						(−0.2216)
BGD × BFP									0.0008
									(0.1534)
BGD × BMComp			0.0002
			(0.0459)
BGD × CEODual			0.0159
			(0.7598)
BGD × Meet						−0.0005
						(−0.7389)
BGD × EBIT						−0.0026
						(−1.2636)
_cons	8.0812 ***	8.0996 ***	7.0144 ***	1.7609 *	1.8270 *	−2.4563	3.3041 ***	3.3455 ***	1.8805
	(5.4926)	(5.4899)	(2.8221)	(1.8519)	(1.9266)	(−0.7855)	(3.3678)	(3.4040)	(0.6932)
F statistic	7.2107 ***	6.8085 ***	5.3101 ***	23.4819 ***	22.9748 ***	18.6428 ***	15.2239 ***	14.7121 ***	11.5358 ***
R-sq	0.4123	0.4128	0.4272	0.4789	0.4828	0.4889	0.3023	0.3027	0.3104
VIF	2.26	7.80		1.57	2.74		1.65	2.77
Obs	204.0000	204.0000	204.0000	718.0000	718.0000	718.0000	1013.0000	1013.0000	1013.0000

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. The numbers in parentheses are t statistics. For definitions of variables, please see Table 4. All references to “variable” across the article correspond to notation “variable_w” or “ln_variable_w” in results tables.

.

Regression results for Quick Ratio mirror the results for Current Ratio; therefore, to avoid excessive granularity around empirical results presentation, detailed outcomes for Quick Ratio OLS regression can be consulted in Supplementary File S5.

Based on our three systematically enlarged regression models, we find initial positive coefficient signs of the linear term BGD, although not statistically significant. The best explanatory power for liquidity risk is provided by M2, for both Current Ratio (CurrR) and Quick Ratio (QR); therefore, the negative and significant influence of BGD on liquidity ratio when the squared term is included is accepted. Models’ prediction power (R-sq) spans between 30% and 49% in the case of CurrR, and 21% and 37% in the case of QR, while, despite the highest number of observations (1013) being provided by model M3, its results are not significant for BGD inference. When squared term (sqBGD) is included, the BGD inference is negative and statistically significant predicting that BGD reduce liquidity ratios (−0.0183, p < 0.01/CurrR and −0.0143, p < 0.01/QR) when female presence is marginal, while the coefficient of the squared term becomes positive (0.0003, p < 0.05, 0.0002, p < 0.05, respectively). This evolution is signaling a curvilinear relationship; therefore, we need to identify and discuss the turning point of this relation to interpret our results. Since the sign for the linear term BGD coefficient is negative, sqBGD changes to positive, which indicates that the curve is convex (U-shape). This change in sign for the linear term’s coefficient, after adding the squared term, suggests that the initial linear model was likely miss-specified, not surprising the data’s twist, concluding that the total effect of BGD on liquidity ratios depends on the value of BGD, meaning it increases the liquidity when BGD increases. Using the second derivative of our regression functions, we identified an inflection point at the level of 30.5% BGD in the case of Current Ratio and 35.75% BGD in the case of Quick Ratio. We conclude that, on average, an above 33% presence of women on the board increases the liquidity ratio and reduces the liquidity risk. These results support the theories of critical mass discussed before (Konrad & Kramer, 2006), arguing that boards need at least three women directors to achieve “critical mass”, enabling them to genuinely influence board dynamics, broaden discussions (especially regarding customers/employees), challenge male perspectives, and foster more collaborative decision-making, leading to better corporate outcomes. VIF spans between 1.57 and 2.77, suggesting there is no misspecification or multicollinearity concern related to independent variables in our models.

Discussing other individual board diversity components (see Table 14 for Current Ratio and Supplementary File S5 for Quick Ratio), we find that board size reduces Current Ratio (−0.0152, p < 0.1) and almost three times higher, Quick Ratio (−0.0444, p < 0.01). This is in line with prior results, suggesting that complex firms would rather have larger boards, since they bring more experience and knowledge and offer better advice (Dalton et al., 1999), that might be holding less cash and benefit from investment opportunities. Similarly, board members compensation (−0.1587, p < 0.01/CurrR, −0.0478, p < 0.1/QR), memberships (−0.1328, p < 0.01/CurrR, −0.0800, p < 0.01/QR), reduce liquidity as their level increases, also, when chair is ex-CEO, liquidity is lower than in the rest of the cases (−0.1837, p < 0.01/CurrR, −0.0893, p < 0.05/QR). Conversely, complementary with BGD when above 40%, meetings attendance (0.0171, p < 0.05/CurrR), board structure policy (0.8344, p < 0.01/CurrR, 0.3591, p < 0.1/QR), maintaining an effective board function policy (0.0772, p < 0.1/CurrR, 0.1039, p < 0.01/QR), board individual reelection policy (0.1249, p < 0.5/CurrR), board tenure (0.0269, p < 0.01/CurrR, 0.0180, p < 0.05/QR), board member independence (0.0022, p < 0.1/CurrR), compliance head report to board (0.1180, p < 0.05/CurrR, 0.1781, p < 0.01/QR), board member term duration (0.0423, p < 0.05/CurrR), complement the positive influence of board gender diversity on liquidity, increasing liquidity ratios and reducing liquidity risk.

Looking for the possible moderation effect of BGD on other individual board diversity components (see Table 14 and Supplementary File S5), we find that together, BGD and CSR increase liquidity (0.0191, p < 0.1/CurrR, 0.0176, p < 0.1), while the CSR committee individual effect is not statistically significant.

In case of alternative measures for BGD, if the presence of women on board is barely visible (WOBMere means presence less than 20%) (see Supplementary File S1), we find no significant influence of gender diversity on liquidity. Discussing companies with above 20% but less than 40% women on board (see Supplementary File S2), we still find a negative influence of gender diversity on Current Ratio −0.0706, p < 0.1 and Quick Ratio (−0.0724, p < 0.05), confirming that their presence is not relevant enough to reverse the women’s influence to positive. In case of a balanced board (WOBBal) with more than 40% woman on the board (see Supplementary File S3), we find the expected change of coefficient sign to positive in case of the Current ratio (0.0624), plus, statistically significant in case of the Quick Ratio (0.0604, p < 0.1). These results let us conclude that, like BGD, in the case of alternative measures, when the presence of women on board becomes relevant (WOBBal > 40%), a 1% increase in gender diversity leads to 6% increase in liquidity.

Further discussing the liquidity ratio using regression models with effects (Table 8, Table 9 and Table 10), the models M1 and M2 exhibit inconsistent and insignificant results, although suggesting a negative relationship. Model M3 regression with random effects (Hausman test p > 0.05) presents a positive influence of BGD on both Current Ratio (0.0029, p < 0.1) and Quick Ratio (0.0046, p < 0.01), consistent with our regression including the squared term, underlying the non-linear effect of BGD on liquidity. These results let us conclude that fragile gender-diverse boards reduce liquidity, supporting arguments of intergroup contact theory related to initial like-male behavior, while a positive influence of BGD on liquidity shows up when women hold above 33% of board seats, reducing liquidity risk.

Altogether, we accept our fourth hypothesis (H4), arguing that gender-diverse boards reduce liquidity, so increase liquidity risk. The influence of gender diversity on liquidity risk is changing when the presence of women on the board starts to count, and BGD improves the liquidity of large-listed firms. As a nonlinear evolution of BGD is present, we confirm H4.1 hypothesis and predict that, as BGD increases and women’s presence on the board exceeds 33%, the presence of women on board in large-listed companies increases liquidity, reducing liquidity risk.

5. Discussion

Observing many corporate governance variables, along with board gender diversity (BGD), our main independent variable, we estimate their influence on corporate risk using three model specifications. We analyzed four corporate risk dimensions, specifically—operational, financial, liquidity, and bankruptcy risk—and seven accounting-based risk outcomes, Operating Ratio, Debt/Assets, Debt/Equity (alternative Changes to Long Term Debt to Capital), Z-Score, Current Ratio (alternative Quick Ratio). BGD has been proxied by three constructed binary variables (dummies) measuring the different levels of presence of women on the board: mere (less than 20%), token (between 20% and 40%), and balanced (above 40%). This is in line with critical mass theory and social categories of (Kanter, 1977), defined as skewed, tilted, and balanced groups.

Discussing Operating Ratio (operating expenses per net sales, %), we find that the presence of women on the board (BGD) coefficient sign flips from negative and insignificant to significant, then to positive, once we increase specification complexity. This might be explained by the following factors: (1) either confounding effect of predictors—for example board size plays opposite role with BGD along models, (2) the presence of suppressors effect given by other covariates added to the models or (3) presumably a non-liner effect of BGD on OpR, due to changing the coefficient sing of squared term (without statistical significance). By adding covariates in our models, we aimed to correct the omitted variable bias; a sign change of initial estimates is not totally unexpected. While we found initial effects to be negative and significant, although still low (−0.0007, p < 0.1), this effect changes to positive but non-significant when proper controls are applied. This finding supports prior results that omitting variable bias in corporate governance is one of the critical elements that affect inference in corporate governance-related corporate outcomes. Moreover, the effect of gender-diverse boards on corporate outcomes is sometimes non-linear and highly context-dependent. Therefore, on our best effort basis, we employed a larger number of variables available in our data source, in such a way to cover most of the board heterogeneity dimensions discussed in our paper—like structure, independence, remuneration, CEO role, committees, oversight, policies, and board practices. By this approach, we intend also to address the concern of omitted variable bias. We see that large-listed firms included in our sample benefit from the effect of various independent components of board diversity on Operating Ratio, while traditional predictors like gender diversity show reduced relevance. We conclude that limited board diversity components like board size and memberships increase OpR, conditional on model specifications. The negative influence on OpR of components like board tenure, CSR committee, board independence, non-executive board members, board function policy, estimated by OLS techniques, couldn’t be validated by models with effects. The non-linear effect of gender diversity on OpR is slightly visible in our specifications, since the squared term coefficient sign is changing, but not statistically significant. We identified a positive interaction effect of BGD and regressors like BSP, BFP, and non-executive board members, confirming their inference on Operating Ratio, conditional on model specifications.

Our estimates for Operating Ratio using models without effects (−0.0007, p < 0.1/M2), does not replicate the effects observed by Ahern and Dittmar (2012) who argued that female director quotas in Norway and board gender diversity may lead to increased internal conflicts, obstruct board operational efficiency, lower decision-making quality, increase functioning costs of organization, and, subsequently result in a decrease in corporation value. Conversely, we adhere to prior research arguing that woman is cost cautious, demanding higher accountability and lower opportunistic behavior (Adams & Ferreira, 2009). Nevertheless, our results in predicting Operating Ratio recommend careful interpretation, since not only possible non-linear effects of gender diversity, but also model specification, industry-specific or other unobserved behavioral traits of board members might affect the operational expenses level, thus the level of Operating Ratio might not be so gender-sensitive.

In case of financial risk—Leverage, capital structure, and Long-Term debt to Capital—our results exhibit that gender-diverse boards increase leverage, meaning the preference for debts to finance their assets rather than for their own funds, preferably on short term. These align with the results of Ahern and Dittmar (2012) who provide evidence that women on boards increase leverage and reduce cash holdings. Similarly, García and Herrero (2021) argue that percentage of women directors is the most influential board characteristic in terms of capital structure decisions and is negatively related to leverage, cost of debt, and debt maturity, we add to this results, that, in case of large-listed companies in European context, gender-diverse board reduce leverage only if presence of women on board is relevant (34–41%). Discussing alternative measures for gender diversity (binary), we find gender diversity to be a rather non-significant regressor for leverage, conditional on all variables included in our models. Analyzing other individual components of board diversity, we provide evidence that board size, independence, compensation, and the existence of a CSR committee increase leverage. However, we find no evidence that gender-diverse boards influence long-term indebtedness. This is in line with prior literature questioning whether there is an appropriate board size and board composition (Coles et al., 2008), suggesting that larger firms with complex advisory needs may have larger boards with more outsider directors. We add to these results, suggesting that in the case of large-listed firms in the European context, female preference for borrowing funds is rather short-term, preventing long-term indebtedness, which increases financial risk for investors.

Failure risk has been analyzed in our study by estimating the board diversity components’ influence on Z-score, since the higher the Z-score, the lower the failure risk. Despite the initial positive influence of BGD on Z-score, when controlling for a possible non-linear effect, we identified that the BGD positive effect is present only above the threshold of 40%. These results add to prior literature findings that a higher proportion of female directors on a company’s board reduces its risk of insolvency (Wilson & Altanlar, 2009), suggesting that in the case of large-listed companies, the gender-diverse boards increase Z-score when the presence of women on boards exceeds 40%.

Lastly, based on our estimates’ results, we argue that firm liquidity risk is conditional on the level of gender diversity in the boardroom. Like García-Meca et al. (2022), we suggest that female directors might reduce liquidity first, taking riskier decisions, to reduce agency conflicts and improve legitimacy along with their male counterparts. After an inflection point of 33%, traditional women’s characteristics, such as risk aversion, a conservative and prudent financial attitude, starts preserve liquidity to reduce liquidity risk. We find a positive influence of interaction between BGD and CSR on liquidity, boosting the individual effect of BGD (negative below 33% and positive above). Based on our analysis, in the case of large-listed companies, we provide evidence that other individual board diversity components increase liquidity, such as meeting attendance, board functions policy, board structure policy, board members’ independence, and compliance head report to the board. Conversely, there are negative regressors that influence liquidity, like board size, compensation, membership, or the Chair role of ex-CEO.

Although our empirical design incorporates alternative estimators, firm- and year-fixed effects, and a comprehensive set of both independent and control variables, as a best effort to mitigate omitted variable bias, we acknowledge that endogeneity concerns, —particularly simultaneity and potential reverse causality—cannot be fully ruled out. It is also discussed that in corporate governance research, it is difficult to establish strict causal inference solely with fixed-effects panel models, since governance decisions and risk outcomes may be jointly determined or simultaneously influenced by unobserved factors. Endogeneity and reverse causality are persistent challenges in governance research, and many empirical studies acknowledge but do not formally correct them, which motivates careful interpretation of associational findings and future causal work (Adams, 2016; Yang et al., 2019). However, traditional fixed-effects estimation can potentially ameliorate the bias arising from unobservable heterogeneity (Hermalin & Weisbach, 2003; Wintoki et al., 2012). Moreover, mostly used causal testing techniques in corporate governance studies to address reverse causality—dynamic panel GMM estimators, instrumental variable approaches, or lag methods—require valid instruments, which are not easily feasible for the large set of board diversity attributes analyzed in our multi-risk framework.

Although we acknowledge several criticisms about many merely correlational studies, focusing on direction, size and significance of the coefficients estimates, since only few studies empirically address causality-like (Adams, 2016) who largely address all three sources of endogeneity, selection bias, omitted variables and reverse causality, we choose for this study to focus on correlational relationships of analyzed variables, observe moderation effect and discuss possible suppressor effect, rather than to discuss possible causal correction techniques that could restrict variable inclusion and make interpretation difficult. We adhere to prior conclusions that it cannot be ascertained if the large number of governance factors analyzed are merely an effect of unobservable factors. If the causality is actually reversed and risk drives governance, it is theoretically plausible that firms experiencing specific risk profiles may adjust their governance structures (e.g., by appointing additional female directors or altering board composition) in response to past performance or risk outcomes, as noted (Hermalin & Weisbach, 2003; Wintoki et al., 2012). We interpret our results correlationally rather than strictly causally and emphasize that potential reverse causality and omitted variable bias are limitations that motivate future research, particularly designed to identify causal strategies.

To mitigate sample selection bias, we positioned our analysis on examining governance–risk relationships within mature boards in an institutional setting where governance mechanisms are well established, and enforcement is strong. This approach allows for cleaner identification of governance effects and sound comparison of evolutions, reducing confounding institutional noise that may arise in less developed markets. Furthermore, we incorporate extensive firm-level controls and firm-year fixed effects, to account for cross-firm heterogeneity within European space (as built, STOXX600 includes a fixed number of 600 components representing large, mid- and small-capitalization companies among 18 European countries, although not limited to Eurozone (thus going beyond EUR currency).

6. Conclusions

The present study examines the influence and significance of board gender diversity and other board diversity components on corporate risk in the European context, analyzing the STOXX600 index, the exclusive financial industry. Unlike prior studies, we employed an extensive set of corporate governance variables defining board diversity, along with gender diversity, our main variable of interest. We have systematically developed three models to estimate four categories of corporate risks—operational, financial, failure, liquidity—proxied by seven accounting-based measures: Operating Ratio, Debts/Assets, Debts/Equity, Changes to Long Term Debts to Capital, Z-score, Current Ratio, and Quick Ratio.

This study is fundamentally contributing insights into understanding how the complex mechanisms of large numbers of individual board diversity components, their interaction, and possible non-linear evolution influence corporate risks. Echoing the literature on the relationship between gender diversity and corporate risk, we expected our findings to be risk-type dependent. Applying sequential econometric technics, we carefully analyzed possible correlations, observed R-sq to assess the proportion of corporate risks variance explained by board diversity components, used VIF to control for multicollinearity of our regressors, perform regression models with firm-year effects selecting the appropriate one using Hausmann test, moreover, we have analyzed a series of interaction between BGD and other diversity components and not the last, we validated our results using three alternative measures for gender diversity.

The findings of this study contribute to the corporate governance literature by refining how established theoretical frameworks explain the relationship between board gender diversity and multiple corporate risk dimensions in mature European firms.

First, the evidence that gender-diverse boards improve operational efficiency—conditional on model specification—provides nuanced support for agency theory and upper-echelon theory. From an agency perspective, female directors appear to strengthen board monitoring and cost discipline, thereby reducing operational inefficiencies, though this effect is sensitive to the broader governance configuration. Upper-echelon theory further explains this result by suggesting that heterogeneous boards get managerial attention and resource allocation, particularly in day-to-day operational decisions. The conditional nature of the effect underscores that gender diversity operates jointly with other board attributes rather than as a standalone governance mechanism.

Second, the finding that gender-diverse boards are associated with higher debt levels—facilitating growth financing without excessive leverage—extends both resource dependence theory and intergroup contact theory. Resource dependence theory suggests that gender-diverse boards enhance access to external resources and financing channels, enabling firms to pursue expansion opportunities more effectively. Intergroup contact theory provides additional insight by explaining why, at lower levels of female representation, women may initially conform to their male counterparts’ expectations, supporting growth-oriented financing strategies. Importantly, the observed decline in leverage beyond the 33% threshold indicates a behavioral shift once meaningful integration is achieved, reflecting their traditional behavior of more balanced risk assessment and financial discipline.

Third, the strong association between gender diversity and reduced failure risk—captured by higher Z-scores—once female representation exceeds 40%, offers robust support for critical mass theory. This result suggests that only after women reach a sufficient presence on the board does their influence translate into more effective oversight of solvency, capital structure sustainability, and long-term financial resilience. Below this threshold, women’s contributions may remain silent due to tokenism or limited decision power, whereas critical mass enables substantive participation in strategic risk governance.

Fourth, the finding that gender-diverse boards improve liquidity outcomes once the critical mass threshold of approximately 33% is reached further reinforces the relevance of critical mass and intergroup contact theories. Early-stage female participation may be associated with reduced liquidity as boards prioritize growth and male members’ objectives, but once integration deepens, traditional gender-related traits—such as prudence, cost discipline, and stronger compliance orientation—become more prominent. This transition results in more conservative liquidity management, a higher liquidity buffer, and improved short-term financial soundness.

Taken together, these results suggest that gender diversity influences corporate risk through non-linear, context-dependent mechanisms rather than uniform behavioral effects. The study advances governance theory by showing that gender-diverse boards in mature institutional settings do not uniformly reduce risk, but instead reshape firms’ exposure across operational, financial, liquidity, and failure dimensions as women’s representation reaches meaningful levels. This multidimensional and theory-consistent approach helps reconcile previously mixed empirical findings and highlights the importance of considering critical mass, board dynamics, and risk specificity when evaluating the governance role of gender diversity.

Complementary, we provide empirical results that other board diversity components influence corporate risk channels. Board size increases operating ratio and leverage, while reducing Z-score and liquidity. Board tenure negatively influences operating ratio and leverage, while having a positive effect on Z-score. Board member independence reduces operating ratio; conversely, it increases leverage and liquidity. Companies having a CSR committee exhibit a lower operating ratio—emphasizing operational efficiency—but higher leverage and lower Z-score, suggesting possible influence of the committee in financing costly projects on social responsibility or environments. Board member affiliation, meaning the number of memberships, influences corporate risks. The number of memberships increases the operating ratio, but reduces leverage, Z-score, and liquidity, suggesting a combined effect of external expertise in managing leverage, with less time for oversight of operational expenses or liquidity. Related to the CEO role, we find that CEO duality reduces operating ratio and leverage, while a chair ex-CEO is more likely to increase leverage and decrease liquidity.

We further add to the literature by observing the existence of board policies and their influence on corporate risks. We provide evidence that companies implementing board policies (board structure policy, board functions policy, policy board diversity, board effectiveness review) influence corporate risk. Companies implementing a board structure policy and a board function policy are more likely to have higher operational efficiency and higher liquidity. Policy board diversity exhibits a negative effect on Z-score, while board effectiveness review increases liquidity, and its effect is moderated by gender diversity. Related board oversight practices, we provide evidence that the head of compliance reporting to the board is more likely to reduce the operating ratio, although conditional on model specifications, increase the Z-score, and increase liquidity. Lastly, non-compliance with the code of conduct reduces (the possibility of) long-term indebtedness, while diminishing the Z-score, thus increasing the failure risk, presumably affected by the reputational risk encountered.

This study brings a comprehensive contribution to a better understanding of large-listed companies’ practices in managing their risks, employing an extended set of board diversity components, like board structure, independence, remuneration, memberships, committees, oversight practices, or board policies, alongside gender.

However, several limitations might be acknowledged. Firstly, the sample consists of the largest market capitalization listed companies included in STOXX600; therefore, the overall results might not be fully applicable to another framework. We underscore that our findings should be interpreted as conditional on developed-market institutional environments and selected sectors rather than universally generalized. Nevertheless, those European firms’ journey in corporate governance might be of inspiration for enterprises in other capital markets outside the European zone, related to their corporate risk management. For emerging and developing markets, the results provide evidence that governance reforms, especially those related to board composition, should be aligned with the targeted risk outcomes depending on institutional interest (e.g., operational stability versus financial leverage control), illustrating how board gender diversity and governance structures influence specific categories of corporate risk rather than firm risk in aggregate. While findings related to regulatory enforcements, like independence, memberships, remuneration, board policies and practices or CSR committee are likely to be context-dependent (e.g., market maturity, advanced regulatory), other mechanisms—such as board size, gender diversity, tenure, CEO duality or oversight—differentiated effects across risk types, may plausibly be extended to other institutional environments from emerging economies. This framing allows international readers and policymakers to extract transferable insights while recognizing their contextual limitations. Therefore, we explicitly state that policymakers may adapt these insights to local institutional constraints and their targeted risk outcome rather than replicate European governance models outside Europe. This distinction is particularly relevant for jurisdictions currently designing or reforming governance codes, quota systems, or disclosure requirements. For firms stepping on the maturity ladder, they may primarily observe their governance maturity and particular risk weaknesses and consider the relevant findings; however, we highlight that gender diversity can be merely symbolic but can shape internal risk-management behavior when embedded within more effective and diverse board structures.

We built this study on the nexus between gender diversity, our main variable of interest, and other board heterogeneity dimensions. For future research, we recommend complementary dimensions, like ethnicity and cultural diversity, as well as skills diversity, to be analyzed as predictors for corporate risk outcomes. Also, market-based measures as an important corporate risk channel might be considered. Various firms’ sub-samples might also be analyzed, like financial sectors, state-owned (SOE), or dual governance regimes (unitary/dual).