Exploring the Next Level of Boardroom Independence: Are Boards and Committees Driving Firm Performance or Risk in Western Europe?

Abstract

This research responds to recent calls to explore the independence conditions under which boards’ leadership becomes economically meaningful for performance and risk in continental European governance systems. Using an unbalanced panel dataset of 223 non-financial publicly listed companies from Western Europe over 10 years between 2015 and 2024, this research examines how boards and their committee independence influence return on assets and return on equity, return volatility, indebtedness and liquidity volatility. Econometric methods include OLS regressions, industry fixed effects, linear and nonlinear models, including alternative specifications. The results highlight a U-shaped relationship between board and audit committee independence and operational efficiency, consistent with the critical mass interpretation. Board, audit and nomination committee independence reduce return volatility, reflected in a linear relationship. Audit committee independence is likely to reduce indebtedness beyond a balanced level, while the relationship of nomination committee independence with debt level is linear and negative across specifications. All governance mechanisms related to independence exhibit nonlinear relationships with liquidity volatility, with an immediate negative effect, while excessive independence oversight reduces their marginal effect. The findings suggest the existence of an optimal level of board and committee independence that is economically meaningful, providing practical guidance for shaping board and committee composition, to enhance performance and control risk.

1. Introduction

Corporate governance represents a fundamental mechanism through which firms align managerial decisions with shareholders’ and stakeholders’ interests. This perspective is well established in the corporate governance literature, which emphasizes the role of boards in mitigating agency conflicts and aligning managerial incentives, according to Jensen and Meckling (1976) and Fama and Jensen (1983). In the context of increasing regulatory pressure and the growing attention paid to accountability, the role of the board of directors and its committees is enhanced by the independence mechanisms, and becomes central to discussions on corporate performance and risk management. Prior studies highlight that board structures, independence, and committee oversight play a crucial role in shaping firm performance and risk outcomes, according to Adams et al. (2010) and Hermalin and Weisbach (2003). The effectiveness of governance mechanisms is particularly relevant in continental European economies, where ownership concentration, stakeholder orientation, and institutional frameworks differ significantly from the Anglo-Saxon model.

Several studies investigate the impact of board characteristics on firm outcomes, yet with mixed results (Bhagat & Black, 1998; Hermalin & Weisbach, 2003; Fauver et al., 2017). While board independence may be associated with improved monitoring and reduced agency conflicts, other studies document negative or insignificant effects on performance, suggesting that excessive independence may constrain strategic flexibility and reduce access to firm-specific knowledge (James et al., 2022). Similarly, the role of board committees, such as audit, nomination, and compensation committees, may vary across institutional settings, which is often attributed to differences in legal systems, ownership structures, and governance traditions across countries according to La Porta et al. (1998) and Aguilera and Jackson (2003), who documented mixed implications for profitability and risk. Overall, as supported by the studies mentioned above, the evidence shows that the effects of board independence and board committee structures on firm outcomes are context dependent, with mixed results across different environments, depending on the conditions under which firms operate and the markets in which they are embedded. However, the differences in ownership structures, governance traditions, and regulatory environments discussed above, may only partially explain the implications of board’s independence mechanisms for firm outcomes, therefore supporting our approach to enlarge the empirical specifications to better capture the complexity of governance mechanisms.

The empirical literature review highlights two major gaps. First, the prior literature traditionally focuses on linear specifications when examining the relationship between governance mechanisms and firm outcomes, such as Loukil and Yousfi (2015), although more nuanced relationships may also exist, according to Liu and Du (2026). Second, relatively limited attention has been devoted to the joint analysis of performance and risk (Mohsni et al., 2021), particularly when considering the triangular intersection between structural and cognitive diversity proposed by Behlau et al. (2024) or task-oriented attributes discussed by Harjoto et al. (2018), with compensation, oversight mechanisms and board’s policies, framed in the continental European governance system. Against this background, the present study aims to examine whether board and committee independence shape the corporate outcomes in Western Europe. By simultaneously analyzing profitability and risk indicators, this study provides a comprehensive assessment of firm performance. Moreover, by employing nonlinear specifications, it explicitly tests for threshold effects in governance mechanisms, addressing recent calls in the literature for more nuanced empirical approaches.

The analysis focuses on Western European countries belonging to the unified corporate governance continental regime, having consistent reporting for both financial data and corporate governance disclosures. Using pooled OLS, industry fixed effects models, and random effects models, this study evaluates the impact of board mechanisms related to independence and board structure, as well as the board’s knowledge level, remuneration, oversight mechanisms and policies, along with the board members’ knowledge and skills and the influence of the compensation level on firm performance and risk, while controlling firm-specific factors such as firm size, age, total sales, receivables turnover, cash flow, including liquidity, leverage, and operational efficiency.

By focusing on a continental European sample and integrating nonlinear econometric techniques, this paper contributes to the corporate governance literature in three ways. First, it provides empirical evidence on the performance–risk implications of board and committees’ independence, complemented by several new board diversity mechanisms, such as remuneration, oversight structures and board’s policies, proposed by the authors. Second, it highlights the existence of a governance threshold for the efficacy of the independence mechanisms, rather than uniform effects. Third, it offers policy-relevant insights for regulators and practitioners seeking to balance effective monitoring with strategic decision-making related to performance and risk. Unlike much of the existing literature, this study jointly examines performance and risk within a unified nonlinear framework.

2. Literature Review and Hypotheses Development

Despite the extensive empirical literature on corporate governance, important gaps in explaining corporate outcomes remain. Prior studies have often examined governance mechanisms separately, focusing on individual board attributes, such as size and independence, as discussed by Rudyanto and Kusnadi (2025) and Ahrens et al. (2025). Comparatively less attention has been devoted to how different governance mechanisms, such as board structure, cognitive diversity, directors’ incentives, oversight practices or board policies, may interact with committee-level independence within the broader boardroom framework. This fragmented approach limits understanding of how multiple governance dimensions jointly shape firms’ outcomes.

Moreover, empirical findings are often context-dependent and mixed; therefore a new perspective aims to bring more clarity. Board size, board tenure, CEO duality, and board independence exhibit heterogeneous and sometimes nonlinear effects across institutional environments, suggesting the need for integrated empirical frameworks that allow governance mechanisms to operate simultaneously (Ahrens et al., 2025; Tang, 2017; Adams et al., 2005; Hermalin & Weisbach, 2012). In addition, several practices increasingly emphasized by governance codes—such as board effectiveness reviews, formal governance and diversity policies, and sustainability committees—remain underexplored in large-sample empirical studies and are often captured through limited proxies.

Finally, much of the existing evidence relies on single-country or sector-specific samples, constraining generalizability as documented in prior studies focusing on individual national settings, such as the UK (Guest, 2009), the US (Yermack, 1996), or specific industries (Adams & Ferreira, 2009), which limits the external validity of their findings. This study addresses these gaps by providing a holistic empirical analysis of board governance mechanisms and their association with firm performance, explicitly accounting for complementarities, nonlinearities, and emerging governance practices within a unified empirical framework.

Building on the gaps identified in the literature, this study adopts an integrated empirical approach to examine how multiple board governance mechanisms jointly influence firm performance and risk outcomes. The following section outlines the existing evidence on the related topic and proposes the hypotheses to be tested using a comprehensive set of variables and econometric methodology.

2.1. Board Structural Characteristics and Corporate Outcomes

Board size is a core structural attribute of corporate governance, reflecting a trade-off between enhanced monitoring capacity and coordination costs. Larger boards may improve oversight and provide access to diverse expertise and external resources; however, they may also reduce decision-making efficiency due to communication difficulties and diluted responsibility. Recent empirical evidence highlights these competing effects. Msomi and Nzuza (2025), using panel data from financial firms listed on the Johannesburg Stock Exchange, find that board size positively influences market-based performance measures such as Tobin’s Q, while exerting a negative effect on accounting-based performance, particularly return on assets. This suggests that larger boards may strengthen external legitimacy but impair operational efficiency. Similarly, a large-scale meta-analysis by Ahrens et al. (2025), covering 346 studies across 110 countries, shows that the impact of board size on performance is highly context dependent, with stronger negative effects observed for market-based performance in certain institutional settings. Overall, the recent literature suggests that board size does not have a uniform effect on firm outcomes, with its influence depending on the balance between monitoring benefits and coordination costs, as well as on the broader governance environment.

The role of the chief executive officer within the board of directors has received considerable attention in the corporate governance literature, as it directly affects the balance between managerial authority and board oversight, according to Bel-Oms and Grau Grau (2025) and Tang (2017). Two central governance mechanisms associated with CEO power are CEO–chairman of the board duality and CEO presence on the board as an executive member. CEO–chairman duality concentrates leadership authority by assigning both executive and supervisory roles to a single individual. Agency theory predicts that such concentration weakens board independence and reduces monitoring effectiveness. Empirical evidence largely supports this view. Krause et al. (2014) document that CEO duality is associated with weaker monitoring structures, although its performance implications depend on firm-specific conditions. Similarly, Tang (2017) finds that CEO duality increases firm risk-taking, consistent with reduced internal control when leadership roles are combined. Bel-Oms and Grau Grau (2025) show that CEO duality moderates the effectiveness of audit, compensation, and nomination committees, influencing overall firm performance through governance interaction effects.

CEO board membership represents a more nuanced governance mechanism. On the one hand, CEOs serving as board members improve information sharing and strategic coordination between management and the board. On the other hand, excessive CEO influence may compromise board independence. Adams et al. (2005) show that powerful CEOs with strong board presence can significantly shape firm outcomes, increasing both performance potential and risk. Complementing this evidence, Hermalin and Weisbach (2012) argue that CEO participation on the board enhances advisory efficiency but may gradually erode the board’s monitoring role. Liu and Du (2026) argue that CEO power affects capital structure decisions more than leverage levels, with CEO board presence potentially leading to lower leverage to maintain control. Overall, the literature highlights a fundamental trade-off in CEO-related governance mechanisms. While CEO involvement can enhance strategic effectiveness through superior information access, excessive concentration of power may undermine board independence and oversight quality.

Based on the empirical evidence reviewed above, we formulate the following hypotheses:

H1a. 

Board size is associated with reduced performance and risks outcomes.

H1b. 

CEO duality and CEO board membership have a positive impact on performance and are associated with reduced risks.

2.2. Board Cognitive Characteristics and Corporate Outcomes

Board tenure captures the extent of directors’ firm-specific experience and stability, which can enhance monitoring and strategic decision-making. At the same time, prolonged tenure may reduce board independence and increase the risk of entrenchment. Recent studies provide evidence of these opposing effects. Mielcarz and Osiichuk (2025) investigate the effects of average board tenure—operationalized as excess tenure beyond a standard term—on board oversight effectiveness. Using an international panel dataset of over 3000 publicly listed firms from 59 countries (2003–2018) and static panel regression analysis, they examine whether longer director tenures weaken the board’s ability to monitor management effectively. The study finds that excessive board tenure is positively associated with higher executive compensation and lower sensitivity of pay to negative performance surprises, suggesting that long-serving boards may become complacent or aligned with management interests. Additionally, firms with longer tenured boards have a lower likelihood of managerial turnover following poor performance, indicating reduced governance responsiveness. These results highlight a tenure–independence tradeoff, whereby long average board tenure may undermine independent oversight even when formal independence criteria are met. In addition, Lun et al. (2025) document a nonlinear, inverted U-shaped relationship between board tenure and firm performance in U.S. hospitality firms, indicating that moderate tenure improves performance, whereas excessive tenure leads to diminishing returns. In summary, the literature suggests that average board tenure entails a trade-off between experience gains and entrenchment risks, supporting the inclusion of both linear and nonlinear specifications in empirical analyses of board structure and firm outcomes.

Regarding board member affiliations, Aggarwal et al. (2019) examine listed Indian firms and test how business group affiliation, a structured form of affiliation network, conditions the diversity–performance link. Using panel regressions and performance proxied by market-based measures such as Tobin’s Q, they find that board diversity is generally beneficial, but its positive effect is weakened, or can turn adverse, in group-affiliated firms, consistent with the idea that affiliation networks may dilute independence and monitoring effectiveness. Additionally, Latif et al. (2020) analyze non-financial firms listed on the Pakistan Stock Exchange and examine directors’ multiple directorships. Using empirical panel models, they show that busy directors can be associated with weaker monitoring or lower effectiveness, highlighting a fiduciary trade-off: external affiliations may bring reputation and expertise but can also reduce attention and oversight quality when excessive.

Based on these arguments, we formulate our second hypotheses:

H2a. 

Board tenure and higher affiliation are associated with improved performance.

H2b. 

Board tenure and higher affiliation are associated with reduced risk outcomes.

2.3. Board Compensation and Firm Performance

Using newly available individual-level remuneration data, Acero and Alcalde (2020) analyze 1110 director–year observations from 90 Spanish listed firms. Their results show no robust pay–performance sensitivity and limited evidence that governance mechanisms effectively discipline director compensation, suggesting that incentive structures may not consistently enhance firm performance. By contrast, Fang and Huang (2024), using U.S. data and a difference-in-differences design around Delaware court rulings, find that stronger governance reduces director pay and is associated with positive stock market reactions. This supports the view that well-designed incentive mechanisms can improve firm value. Overall, the mixed evidence highlights the need to further examine the role of incentive mechanisms in shaping both firm performance and risk-taking. In addition, Adams et al. (2005) show that greater CEO influence involves a trade-off between control and expertise, improving performance while increasing risk, thereby supporting the role of incentive mechanisms and providing a foundation for H3.

Accordingly, we propose the third hypotheses:

H3a. 

Board incentive mechanisms are positively associated with firm’s performance.

H3b. 

CEO compensation linked to risk, is associated with higher corporate risks.

2.4. Board Oversight: Effectiveness Reviews and Meeting Attendance

Board effectiveness reviews represent a formalized mechanism through which firms assess board performance and strengthen oversight. These practices are increasingly promoted by corporate governance codes as tools to enhance accountability and decision-making (OECD (2015) and Financial Reporting Council (2018)). Although direct evidence remains limited, prior studies suggest that structured evaluations can improve monitoring quality and board effectiveness, proposed by Adams et al. (2010). In parallel, more active boards—proxied by meeting frequency and attendance—are associated with stronger monitoring and improved firm outcomes, including performance and risk, as argued by Vafeas (1999) and Adams and Ferreira (2008).

Board meeting attendance represents one of the most widely used quantitative proxies for board oversight and director engagement. High attendance rates reflect directors’ commitment to monitoring responsibilities and active participation in strategic and controlling processes. Unlike board effectiveness reviews, meeting attendance is an observable and measurable characteristic, which has been extensively examined in empirical literature. Barros and Sarmento (2020) analyze United Kingdom listed firms and show that higher board meeting attendance is associated with lower corporate tax avoidance, indicating stronger board oversight. Using panel regressions with firm fixed effects, the authors find that active director participation constrains opportunistic managerial behavior. Although not directly linked to profitability, the results support board meeting attendance as a reliable proxy for oversight quality. Similarly, Sahoo (2023), using evidence from listed firms, finds that higher board meeting attendance is positively associated with firm performance, reinforcing the view that director engagement plays a critical role in effective oversight. However, several researchers (Simionescu et al., 2021; Vintilă et al., 2025) provided empirical evidence of negative influence of the number of board meetings on performance (ROA and ROE), arguing that increased costs and time burden have an impact on performance, while higher meeting frequency may signal underlying problems or distress within the company. Overall, the literature provides empirical support for board meeting attendance as a key oversight mechanism, while evidence on the mechanism of board effectiveness reviews remains largely conceptual and exploratory. Together, these mechanisms capture both the behavioral and procedural dimensions of board oversight, justifying their joint inclusion in empirical analyses of corporate outcomes.

Considering the above argumentation related to oversight practices, we propose to test the following hypotheses:

H4a. 

Oversight practices, such as board effectiveness reviews and board meeting attendance, are associated with reduced performance.

H4b. 

Oversight practices, such as board effectiveness reviews and board meeting attendance, are associated with reduced risks.

2.5. Board Policy and Firm Performance

Hussain et al. (2024) study full takeover M&A deals (2003–2015) where both bidder and target are covered by Refinitiv ASSET4 ESG. Using OLS regressions, they test whether governance transfers occur post-merger. Their board-structure pillar explicitly includes a balanced board structure policy (BSP) implementation indicator. Results show that when the bidder’s governance quality exceeds the target’s pre-deal, the combined firm’s post-deal governance improves, consistent with governance convergence toward higher standards. Additionally, Fernandes et al. (2023) use Thomson Reuters Eikon, Refinitiv data (2005–2019) for firms in the logistics and transportation sector, estimating fixed effects panel regressions to examine social sustainability out-comes. They model board structure policy (BSP) as a governance-policy construct and show that BSP strengthens (positively moderates) the relationship between board gender diversity and social sustainability, while the moderation is not supported for cultural diversity in the same way.

Considering policy board diversity (PBD), Wawryszuk-Misztal (2021) analyzes 268 non-financial domestic firms listed on the Warsaw Stock Exchange (2016–2018) and determines whether firms declare implementation of a board diversity policy in corporate governance statements (comply-or-explain setting). Using logistic regression, the study finds that larger firms, firms with larger management boards, and firms with women as presidents of supervisory boards are more likely to implement a board diversity policy. Additionally, Ben-Amar et al. (2022) examine initial board gender diversity disclosures of 506 publicly traded firms on the Toronto Stock Exchange. Using textual analysis (tone and clarity metrics) linked to subsequent appointments, they show that firms that appear committed to enhancing board diversity provide clearer disclosures and a more optimistic tone, while reluctant firms exhibit obfuscation and certainty language consistent with impression management. This is directly useful when policy board diversity is proxied through disclosed diversity-policy commitments. Considering these findings, we propose the fifth hypothesis:

H5. 

Board policies are associated with higher firm performance and lower risk.

2.6. Board Independence and Monitoring Effectiveness

Chairperson’s independence refers to the separation of leadership roles between the chief executive officer and the chair of the board, or to the appointment of a non-executive chair who is independent from management. From an agency theory perspective, an independent chair improves the board’s ability to supervise management by minimizing role conflicts and facilitating objective performance evaluation. Early empirical evidence by Daily et al. (1997) suggests that separating the CEO and chair roles strengthens board monitoring and is associated with improved long-term firm outcomes, particularly in firms characterized by high growth opportunities or information asymmetry. More recent evidence supports these conclusions in broader institutional contexts. Omenihu (2025), using a dynamic panel framework, finds that independent board leadership is positively related to firm performance, especially in environments where external governance mechanisms are weaker. These findings indicate that chairperson independence enhances oversight effectiveness by reinforcing the board’s authority relative to management. Coles and Hesterly (2000) show that chairman independence strengthens governance effectiveness, especially when combined with a balanced board structure, leading to improved firm performance.

The presence of independent, non-executive directors constitutes another central mechanism through which boards enhance monitoring quality. Independent directors are expected to bring objectivity, reduce managerial entrenchment, and protect shareholder interests by exercising unbiased judgment. Empirical evidence largely supports these expectations. Mihail and Micu (2021), analyzing firms listed on the Bucharest Stock Exchange, document a positive association between the proportion of independent board members and return on equity, suggesting that independent directors contribute to improved financial performance through stronger oversight. Complementing firm-level studies, a meta-analytic regression analysis by Zubeltzu-Jaka et al. (2019) synthesizes evidence from over one hundred studies and finds that board independence is generally positively related to accounting-based performance measures, although the strength and direction of the relationship vary across institutional contexts. So, taken together, these findings indicate that governance mechanisms may generate both beneficial and adverse effects, depending on institutional context and intensity as well as firms’ strategic objectives. Tai (2025) further argues that board independence positively affects performance, especially when independent directors maintain decision-making objectivity supported by adequate incentives.

Building on these arguments, we propose the sixth hypothesis, including testing for possible nonlinearities:

H6. 

The proportion of independent board members, including chairperson independence, is associated with higher firm performance and lower risk, reflecting their role in enhancing board oversight.

2.7. Board’ Specialized Committee’s Independent Oversight

Audit committee independence is widely recognized as a critical governance mechanism for ensuring the integrity of financial reporting and strengthening internal monitoring. Prior empirical research consistently documents that independent audit committees play a central role in constraining managerial opportunism. Klein (2002) provides early but foundational evidence that firms with fully independent audit committees exhibit significantly lower levels of abnormal accruals, indicating reduced earnings management and higher reporting quality. Complementing this evidence, Bédard et al. (2004) show that audit committee independence, particularly when combined with financial expertise and active committee engagement, is associated with a lower likelihood of aggressive earnings management practices. Together, these studies establish audit committee independence as a key determinant of effective financial oversight and accounting transparency. Alhababsah and Azzam (2024) argue that audit committee independence is particularly relevant in developing countries, where other monitoring mechanisms are weaker, although its direct impact on firm performance is limited.

Nomination committees are designed to safeguard the board appointment process from managerial influence and to promote objective director selection. Empirical evidence suggests that independent nomination committees contribute to stronger board structures by facilitating the appointment of independent and diverse directors. Ruigrok et al. (2007) demonstrate that independent nomination committees are positively associated with higher levels of board independence in European firms, supporting the view that such committees enhance governance quality through more transparent selection processes. Similarly, Vafeas (2003) shows that independent committee structures play a crucial role in improving overall board monitoring capacity, including director appointments. Taken together, these findings suggest that nomination committee independence strengthens governance by reinforcing the board’s autonomy and monitoring effectiveness. Mans-Kemp and Viviers (2019) emphasize that nomination committee independence is a key driver of board diversity and decision-making quality.

The establishment of dedicated CSR or sustainability committees reflects the growing importance of non-financial oversight within corporate governance frameworks. Empirical evidence indicates that such committees contribute positively to sustainability-related outcomes, particularly when they operate independently from management. Liao et al. (2015) find that firms with environmental or CSR committees provide higher-quality environmental disclosures, suggesting improved oversight of sustainability reporting. Similarly, Fuente et al. (2017) show that sustainability committees are associated with greater adoption of international sustainability reporting standards and improved CSR performance, suggesting that CSR and sustainability committees enhance the board’s capacity to oversee non-financial risks and long-term value creation.

Taken together these findings offer support for our seventh hypothesis, which also tests possible nonlinearities, as documented in prior research:

H7. 

Specialized governance independence, such as independence of audit and nomination committees, contribute to reducing risks, while they may reduce also performance.

3. Research Methods

3.1. Description of the Database and Variables

The sample consists of an unbalanced panel of 223 listed non-financial companies, primarily included in the European Index Stoxx600, from seven Western European countries—Austria, Belgium, France, Germany, Luxembourg, the Netherlands and Switzerland—and ten industrial sectors to which they belong, over a 10-years period between 2015 and 2024.

Details about the sample structure are presented in

Table 1. Description of the sample, by countries and industries.

Country	Frequency	Percent	Cumulated
Panel A: by country
Austria	5	2.24	2.24
Belgium	12	5.38	7.63
France	60	26.91	34.53
Germany	60	26.91	61.43
Luxembourg	6	2.69	64.13
Netherlands	28	12.56	76.68
Switzerland	52	23.32	100
Panel B: by industry
Communication Services	11	4.93	4.93
Consumer Discretionary	32	14.35	19.28
Consumer Staples	17	7.62	26.91
Energy	8	3.59	30.49
Health Care	30	13.45	43.95
Industrials	62	27.80	71.75
Information Technology	15	6.73	78.48
Materials	26	11.66	90.131
Real Estate	14	6.28	96.41
Utilities	8	3.59	100

Source: Author’s own work.

.

All variables included in the study, their definitions and calculation formulas, as well as notation and measurement are presented in

Table 2. Description of variables, definitions and measurements.

No.	Variable Name	Variable Code	Variable Category	Variable Definition/Measurement
Dependent Variables
1	Return on Assets	ROA	performance	Total Return/Total Assets (%)
2	Return On Equity	ROE	performance	Net Income/shareholders’ equity (%)
3	ROA Volatility	sdROA	risk	Standard deviation over 5Y window of Total Return/Total Assets (%)
4	Financial Leverage	LevDE	risk	Total Debt/Total Equityx100 (%)
5	Quick Ratio Volatility	sdQR	risk	Standard deviation over 5Y window of Quick Ratio, measured as Current Assets less Inventory divided by Total Current Liabilities (%)
Independent Variables
6	Chairperson Independent	ChairInd	independence	Is the company chairperson independent (not affiliated with the company’s management or major shareholders)? Considered if independent status of the chairperson is explicitly disclosed (0/1)
7	Independent Board Members	IndBM	independence	Percentage of independent board members as reported by the company (%)
8	Audit Committee Independence	CAuditInd	independence	Percentage of independent board members on the audit committee as stipulated by the company (%)
9	Nomination Committee Independence	CNomInd	independence	Percentage of independent board members on the nomination committee as stipulated by the company (%)
10	Compensation Committee Independence	CCompInd	independence	Percentage of independent board members on the compensation committee as stipulated by the company (%)
11	Board Size	BSize	structural	Total number of Board Directors (number)
12	CEO Chairman Duality	CEODual	structural	CEO–Chairperson dual role. A dummy variable assigned a value of 1 for the same person holding the roles of CEO and Chairperson and 0 when there is a separation in the roles) (0/1)
13	CEO Board Member	CEOBM	structural	The CEO is a board member (0/1)
14	CSR Sustainability Committee	CSRCom	structural	A dummy variable taking the value of 1 if the company has a CSR committee or team (0/1)
15	Board Tenure	BTenure	cognitive	Average number of years each board member has been on the board (years)
16	Board Member Affiliations	BMAfill	cognitive	Average number of other corporate affiliations for the board member (number)
17	Board Member Compensation	lnBMComp	compensation	Total compensation of the board members in US dollars (EUR equivalent)
18	CEO Compensation Link to TSR	CEOCompRisk	compensation	Is the CEO’s compensation linked to total shareholder return (TSR)? (0/1)
19	Board Meeting Attendance Average	BMeet	oversight	Board meeting average is the attendance average of members versus the total number of board meetings held (%)
20	Board Effectiveness Review	BER	oversight	A dummy variable taking the value of 1 if the company reports existence of the process of assessing and analyzing the effectiveness and efficiency of a company’s board of directors in fulfilling their responsibilities and achieving the business goals and a clear time frame for this specific process (0/1)
21	Board Structure Policy	BSP	policy	A dummy variable taking the value of 1 if the company has a policy for maintaining a well-balanced membership of the board (0/1)
22	Policy Board Diversity	PBD	policy	A dummy variable taking the value of 1 if the company has a policy to maintain a well-balanced board regarding the gender diversity and/or intercultural (race, religion, culture) representation on the board (0/1)
Control Variables
23	Date of Incorporation	Age	firm	Years of establishment (years)
24	Firm size	lnTA	firm	Total Assets EUR (natural logarithm of total assets of listed companies) (EUR)
25	A/R Turnover	ART	firm	This item measures the number of times Receivables are cycled through in a given period. It is calculated as Primary Revenue for the fiscal period divided by the Average Total Net Receivables for the same period (number)
26	Free Cash Flow	lnFCF	firm	This item represents Cash Flow excluding Capital Expenditures and Total Cash Dividends Paid for the fiscal year (EUR, natural logarithm)
27	Net Sales	lnSALES	firm	Net Sales (natural logarithm), EUR

Source: All variables and definitions were retrieved from LSEG Workspace on 6th of September 2025. ROA, ROE, LevDE, QR have been included in estimations both as dependent variables and control variables, to account for performance and risk in case of estimations of risk and performance, respectively.

.

Building on prior research estimating firms’ performance, such as Adams et al. (2005), Simionescu et al. (2021) and Ahrens et al. (2025), financial performance is proxied by operational efficiency and profitability ratios, while risk is measured through the standard deviation of returns, current ratio, and leverage ratio, as proposed by Cheng (2008), Jane Lenard et al. (2014) and Loukil and Yousfi (2015). Volatility of short-term liquidity is alternatively included in this study, measured as the standard deviation of quick ratio. To estimate these corporate outcomes, we regress these five dependent variables on the corporate governance independent variables, having board and committees’ independence as main independent variables of interest. Accounting-based measures ROA, ROE, LevDE, and QR are used both as dependent variables and, in alternative specifications, as control variables. This modeling choice is intentional and reflects the joint relationship between performance and risk. Specifically, when performance measures (ROA, ROE) are used as dependent variables, risk indicators (e.g., LevDE, QR, sdROA) are included as controls; conversely, when risk measures are the dependent variables, performance indicators are introduced as controls. Board structure, knowledge, compensation level, oversight mechanisms and board policies are also discussed along with main results related to independent mechanisms, as they are expected to influence the corporate outcomes. A set of control variables for firm’s characteristics have been included in our analysis to account for time-invariant firm heterogeneity, such as company’s age, firm size (total assets), accounts receivable turnover, free cash flow, and net sales. All variables are measured according to financial-accounting standards, in line with prior studies, and are available as such in the data source, except variables measuring volatility, which are built by the authors using the standard deviation method.

3.2. Quantitative Methods

The empirical analysis is based on an unbalanced panel dataset consisting of 223 non-financial companies from Western Europe, belonging to the continental European corporate governance system, over a ten-year period, namely 2015–2024, selected from the European index Stoxx600, consisting of 2164 observations. The sample includes companies established in countries that benefit from sound corporate governance mechanisms: Austria, Belgium, France, Germany, Luxembourg, Netherlands and Switzerland.

The corporate outcomes analyzed in this study as dependent variables are represented by two indicators of performance and five of risk: return on assets (ROA) and return on equity (ROE) as performance indicators, and the standard deviation of ROA and QR on a 5-year rolling window (sdROA, sdQR) and the leverage ratio (LevDE), as risk proxies. Building on Harjoto et al. (2018) and the literature review study of Behlau et al. (2024), the board diversity included in this study is clustered in relation-oriented and task-oriented variables and structural, demographic or cognitive dimensions, respectively. Adding to this evidence, several board mechanisms, such as remuneration, oversight and policies, are proposed originally by the authors to be included in the empirical design to estimate firms’ performance and risk, alongside variables related to board independence mechanisms. The source of all variables is LSEG Workspace, and their definitions and measurements are presented in Table 2. This approach responds to recent calls to explore the governance independence conditions under which board diversity becomes economically meaningful, highlighting the results after controlling sectoral heterogeneity, relevant to the composition of our sample.

To ensure the validity of the results, variables were organized in several models based on the board diversity conceptual categories. To account for possible nonlinearities of governance independence in line with critical mass interpretation, quadratic terms for board and committees’ independence have been considered in our models.

Estimations were performed in Stata MP 15 (64-bit), using relevant results from prior research from official sources, specifically ScienceDirect for the theoretical foundation, and data from LSEG Workspace, accessed on 6 September and 23 November 2025. Standard deviation has been computed by authors using input data collected from our data source.

Selected variables have been tested for collinearity using the Pearson correlation matrix, being considered in different models if correlation > 0.7. To ensure the reliability of our parametric estimations, all continuous variables were winsorized at the 5th and 95th percentiles to limit the influence of outliers. To address skewness and reduce the influence of extreme values, selected financial variables were transformed using natural logarithms. This proactive treatment effectively mitigated the influence of extreme outliers and reduced the distributional tails. Post-winsorization, skewness coefficients for all continuous variables ranged between −1.75 and 1.78, which falls within the widely accepted threshold of +/−2 for large-sample research (N = 814), following the guidelines of Bulmer (1979) and Hair et al. (2010), while formal normality tests show a significant level. Although formal normality tests remained significant, the lack of substantial asymmetry justifies the use of mean-based estimation. Accordingly, we employ ordinary least squares (OLS) as our baseline model. To further validate our findings and account for unobserved heterogeneity, we follow the OLS baseline estimation with a fixed Effects (FE) specification as a rigorous robustness check. Considering the panel structure (sub-sample of a market index), industry fixed effects have been preferred by authors to account for industrial sector peculiarities and time-invariant unobserved heterogeneity such as between-sectors corporate culture, ownership structure, or historical governance practices. Mean-centering was applied to quadratic terms to correct the inherent collinearity with the term itself, in case it causes high VIFs that make models hard to interpret.

Several panel regression models were estimated, including linear and nonlinear specifications, as well as fixed effects and random effects models, with the optimal selection made using the Hausman test. The nonlinear models include quadratic terms for independence mechanisms, and for statistically significant relationships, the turning points were determined where applicable.

The empirical models are designed to test the research hypotheses developed in Section 2, which examine the impact of board independence and specialized committee independence on corporate performance and risk (H6 and H7), while incorporating board diversity dimensions related to board structure (H1), knowledge (H2), remuneration (H3), oversight (H4) and policies (H5).

The baseline specifications of the econometric models for corporate outcomes’ regression on board independence and disaggregated committees’ independence mechanisms, supported by board policies and practices, are presented in Equations (1) and (2) below:

$$\begin{gathered} {CorporateOutcome}_{it}={\beta }_0+{\beta }_1{BoardIndependence}_{it}+{\beta }_2{BoardStructure}_{it}+{\beta }_3{BoardCognitiv}_{it}+ \\ {\beta }_4{BoardRemuneration}_{it}+{\beta }_5{BoardOversight}_{it}+{\beta }_6{BoardPolicy}_{it}+\sum {\beta }_{it}{Firm}_{it}+{\epsilon }_{it} \end{gathered}$$

(1)

$$\begin{gathered} {CorporateOutcome}_{it}={\beta }_0+{\beta }_1{AuditIndependence}_{it}+{{\beta }_2{NominationIndependence}_{it}+\beta }_3{BoardStructure}_{it}+ \\ {\beta }_4{BoardCognitiv}_{it}+{\beta }_5{BoardRemuneration}_{it}+{\beta }_6{BoardOversight}_{it}+\sum {\beta }_{it}{Firm}_{it}+{\epsilon }_{it} \end{gathered}$$

(2)

Equation (3) presents the regression model used to estimate corporate outcomes in case of a possible nonlinear influence of board independence:

$$\begin{gathered} {CorporateOutcome}_{it}={\beta }_0+{\beta }_1{BoardIndependence}_{it}+{{\beta }_2{sqBoardIndependence}_{it}+\beta }_3{BoardStructure}_{it}+ \\ {\beta }_4{BoardCognitiv}_{it}+{\beta }_5{BoardRemuneration}_{it}+{\beta }_6{BoardOversight}_{it}+{{\beta }_{7}BoardPolicy}_{it}+\sum {\beta }_{it}{Firm}_{it}+{\epsilon }_{it} \end{gathered}$$

(3)

Equation (4) presents the regression model used to estimate corporate outcomes in case of a possible nonlinear influence of audit committee independence:

$$\begin{gathered} {CorporateOutcome}_{it}={\beta }_0+{\beta }_1{AuditIndependence}_{it}+{{\beta }_2{sqAuditIndependence}_{it}+\beta }_3{BoardStructure}_{it}+ \\ {\beta }_4{BoardCognitiv}_{it}+{\beta }_5{BoardRemuneration}_{it}+{\beta }_6{BoardOversight}_{it}+\sum {\beta }_{it}{Firm}_{it}+{\epsilon }_{it} \end{gathered}$$

(4)

Equation (5) presents the regression model used to estimate corporate outcomes in case a possible nonlinear influence of nomination committee independence is present:

$$\begin{gathered} {CorporateOutcome}_{it}={\beta }_0+{\beta }_1{NominationIndependence}_{it}+{{\beta }_2{sqNominationIndependence}_{it}+ \\ \beta }_3{BoardStructure}_{it}+{\beta }_4{BoardCognitiv}_{it}+{\beta }_5{BoardRemuneration}_{it}+{\beta }_6{BoardOversight}_{it}+\sum {\beta }_{it}{Firm}_{it}+{\epsilon }_{it} \end{gathered}$$

(5)

where:

• ${CorporateOutcome}_{it}$ denotes the measures for performance and risk outcomes;
• ${BoardIndependence}_{it}$ measures the percentage of independent members in board;
• ${AuditIndependence}_{it}$ measures the percentage of independent members in audit committee;
• ${NominationIndependence}_{it}$ measures the percentage of independent members in nomination committee;
• ${sqBoardIndependence}_{it}$ is the quadratic term for board independence;
• ${sqAuditIndependence}_{it}$ is the quadratic term for audit committee independence;
• ${sqNominationIndependence}_{it}$ is the quadratic term for nomination committee independence;
• ${BoardStructure}_{it}$ includes variables measuring board size, CEO duality and CEO board member;
• ${BoardCognitive}_{it}$ measures board knowledge level expressed by tenure and number of affiliations;
• ${BoardRemuneration}_{it}$ includes variables measuring board member’s and CEO’s risk related remuneration level;
• ${BoardOversight}_{it}$ includes variables measuring board effectiveness review and meeting attendance;
• ${BoardPolicy}_{it}$ includes variables measuring the existence of policies (BSP, PBD) at board level;
• ${\beta }_{it}$ are coefficients corresponding to firm-related attributes control variable;
• ${\epsilon }_{it}$ is the error term.

3.3. Limitations and Endogeneity Bias

A limitation of this study is the restriction of the sample to a limited number of developed European countries, which may reduce the generalizability of the results. However, the selection is representative for the objective of the analysis, which focuses exclusively on this Western region and particularly the continental European corporate governance regime. Despite this limitation, the use of panel data techniques, multiple performance and risk indicators, and nonlinear specifications offer methodological support for the robustness of the analysis and enhance the explanatory power of the empirical results.

Although our empirical design incorporates industry fixed effects and an extensive set of controls, endogeneity concerns—particularly dynamic endogeneity, simultaneity, and reverse causality—cannot be fully addressed. Similarly, two-way fixed effects models (firm and year fixed effects estimators) might complement industry fixed effects and provide new insights. Endogeneity is especially relevant in our setting, where both profitability (ROA, ROE) and risk outcomes (sdROA, LevDA, sdQR) may be jointly shaped by and interact with governance characteristics over time. Firm performance and risk profiles can influence subsequent governance adjustments, while governance structures may simultaneously shape future performance and risk, giving rise to feedback effects that fixed effects models do not fully capture.

While fixed effects estimation mitigates bias from time-invariant unobserved heterogeneity (Hermalin & Weisbach, 2003; Wintoki et al., 2012), it does not resolve dynamic endogeneity arising from such bidirectional relationships. Standard causal identification approaches in the governance literature—such as dynamic panel GMM or instrumental variable techniques—require strong and credible instruments, which are difficult to establish for the broad set of board diversity and corporate governance mechanisms examined in multi-dimensional frameworks.

Accordingly, and consistent with prior literature (Adams, 2016; Yang et al., 2019), we deliberately interpret our results as associational rather than causal. The analysis focuses on the direction, magnitude, and statistical significance of relationships, without imposing restrictive identification strategies that could limit model specification and interpretability. We acknowledge that observed governance configurations—focusing in our case on independence mechanisms—may partly reflect unobserved firm characteristics, and that reverse causality remains plausible—for instance, firms experiencing changes in profitability or risk may adjust board or committee independence, in response to prior outcomes.

To mitigate sample selection bias, the analysis is restricted to firms operating within mature governance, continental-like environments, characterized by well-established institutional frameworks and strong enforcement. This setting supports more consistent cross-firm comparisons and reduces institutional heterogeneity. Additionally, sourcing our sample from the Stoxx600 index—covering large-, mid-, and small-cap companies across 18 European countries—combined with extensive firm-level controls and industry fixed effects, provides the grounding, enhances comparability and helps account for residual cross-sectional heterogeneity.

4. Results and Discussion

To examine the impact of corporate governance independence on companies’ profitability and risk outcomes, an empirical analysis was conducted. The data analyzed are provided by the LSEG Workspace database and cover the period 2015–2024; however, data were collected from 2011 for standard deviation computation of the risk proxies. This study provides a rigorous assessment of the relationship between the explanatory variables and the profitability–risk duo of the firms in the sample.

4.1. Descriptive Statistics and Correlation Matrix

Table 3. Descriptive statistics.

	Count	Mean	SD	Min	Max
ROA	2120	0.0537431	0.0441572	−0.0235373	0.158346
ROE	2023	0.1555333	0.092199	0.01899	0.381
sdROA	2089	0.0228402	0.018319	0.0041878	0.0721006
LevDE	2140	0.8340296	0.6794755	0.015914	2.577278
sdQR	2117	0.2090851	0.2069	0.0310694	0.8378708
Age	1981	36.67491	29.61162	2	109
lnTA	2152	23.20628	1.339567	20.74979	25.54272
BSize	1903	11.63426	4.172758	6	20
BMeet	1606	95.89739	3.692158	87	100
BMAfill	1906	0.9723089	0.6121666	0.0909091	2.181818
ChairInd	1744	0.5017202	0.5001405	0	1
CEODual	1907	0.3476665	0.4763544	0	1
CEOCompRisk	1904	0.5803571	0.4936301	0	1
BTenure	1888	6.901904	2.475369	2.9375	12.48077
IndBM	1906	61.33814	28.70212	0	100
CEOBM	1751	0.3974872	0.4895181	0	1
CAuditInd	1890	71.34634	30.00728	0	100
CNomInd	1725	96.03748	8.153201	75	100
CCompInd	1446	74.61548	27.32049	0	100
lnBMComp	1874	13.967	0.8545743	12.40901	15.49366
BER	849	0.7055359	0.4560701	0	1
BSP	1908	0.9732704	0.1613342	0	1
PBD	1906	0.7681007	0.4221558	0	1
CSRCom	1906	0.8609654	0.3460735	0	1
lnSALES	2052	22.54697	1.442508	19.80788	25.00201
QR	2148	1.107554	0.5797379	0.38943	2.70007
ART	2109	6.199418	3.62671	2.097866	16.96226
lnFCF	1713	19.63122	1.521866	10.83958	22.03941
N	2164

Source: Author’s own work.

presents the descriptive statistics of the variables, including the number of available observations, their mean, standard deviation, and extreme (minimum and maximum) values.

The analyzed firms have, on average, a positive level of financial performance, with ROA of 5.3% and ROE of 15.5% on average. The leverage ratio (LevDE) has an average of 8.34%, with significant differences between companies in terms of financing structure. Risk and liquidity indicators show moderate to positive values; for example, QR has an average of 110.7%, while its volatility has an average of 20.9%, with large variability between the firms in the sample. While the sample is similar in terms of total assets, the working history suggests that both traditional and newly established firms are included.

Regarding corporate governance, the average size of the board of directors is approximately 12 members, having an average tenor of 7 years, spanning from almost 3 to 12. A share of 50% of the companies appointed an independent chairman, while in 35% of the cases, the CEO is also board member. A share of 38% of CEOs have remuneration linked to risks, while the level of compensation between board members is similar in the sample. In the case of board independence mechanisms, over 60% of board members are independent, while the committee independence exceeds 70%, and independence is 96% in the case of nomination committees. The majority of firms have implemented board policies, that is, between 70 and 96% of firms’ have BER, BSP or PBD.

Overall, the descriptive statistics suggest that the sample is diverse and can be used as a basis for conducting econometric analyses.

In

Table 4. Pearson correlation matrix.

	ROA	ROE	sdROA	LevDE	sdQR	Age	lnTA	BSize	BMeet
ROA	1
ROE	0.682 ***	1
sdROA	−0.00532	−0.0236	1
LevDE	−0.331 ***	0.0542 *	−0.132 ***	1
sdQR	0.116 ***	−0.0747 ***	0.320 ***	−0.219 ***	1
Age	0.0494 *	0.0134	−0.0587 **	−0.0525 *	−0.130 ***	1
lnTA	−0.306 ***	−0.186 ***	−0.106 ***	0.267 ***	−0.312 ***	0.105 ***	1
BSize	−0.256 ***	−0.0937 ***	−0.0683 **	0.197 ***	−0.276 ***	0.0455	0.594 ***	1
BMeet	0.0272	0.0288	0.042	−0.0740 **	−0.000249	−0.0576 *	−0.0650 **	−0.112 ***	1
BMAfill	−0.0137	0.0838 ***	0.0464 *	0.0229	−0.113 ***	0.138 ***	0.401 ***	0.117 ***	−0.0412
ChairInd	−0.0127	0.0146	0.0639 **	−0.00339	0.0468	−0.0995 ***	−0.0162	−0.0447	0.115 ***
CEODual	0.0542 *	0.0361	−0.102 ***	−0.0364	−0.114 ***	0.110 ***	0.0768 ***	0.0931 ***	−0.147 ***
CEOCompRisk	−0.0762 ***	0.0464 *	0.0571 *	0.0991 ***	−0.0644 **	0.0129	0.159 ***	−0.0152	0.200 ***
BTenure	0.176 ***	0.00937	−0.127 ***	−0.0986 ***	−0.0742 **	0.274 ***	0.0542 *	−0.0516 *	−0.0146
IndBM	0.0658 **	0.109 ***	0.00439	0.0211	0.00782	−0.0651 **	0.0708 **	−0.178 ***	0.109 ***
CEOBM	−0.0308	−0.0437	−0.0268	0.00919	−0.152 ***	0.038	0.213 ***	0.179 ***	−0.143 ***
CAuditInd	0.0417	0.0548 *	−0.0158	0.0386	−0.0556 *	−0.0540 *	0.126 ***	−0.136 ***	0.0822 **
CNomInd	−0.0173	−0.0303	0.0812 ***	0.0124	0.0542 *	−0.106 ***	0.0376	0.0292	0.0700 **
CCompInd	0.0314	0.0691 **	0.039	−0.0259	−0.000263	−0.0605 *	0.0610 *	−0.0993 ***	0.0742 **
lnBMComp	−0.0344	0.0740 **	−0.000667	0.0322	−0.115 ***	0.0481 *	0.441 ***	0.364 ***	0.129 ***
BER	−0.00718	0.0462	−0.0515	0.0784 *	−0.168 ***	0.044	0.256 ***	0.235 ***	0.028
BSP	−0.00232	0.0297	0.0580 *	0.0324	−0.164 ***	0.00228	0.152 ***	0.0907 ***	−0.00685
PBD	−0.160 ***	−0.0543 *	0.0892 ***	0.182 ***	−0.104 ***	−0.0747 **	0.283 ***	0.291 ***	0.0205
CSRCom	−0.0999 ***	0.00011	−0.00607	0.142 ***	−0.167 ***	0.0905 ***	0.332 ***	0.294 ***	−0.0408
lnSALES	−0.188 ***	−0.000641	−0.156 ***	0.168 ***	−0.403 ***	0.151 ***	0.841 ***	0.571 ***	−0.0186
QR	0.289 ***	0.0738 ***	0.240 ***	−0.319 ***	0.568 ***	−0.0514 *	−0.349 ***	−0.215 ***	0.0276
ART	0.0498 *	−0.0196	0.191 ***	−0.0738 ***	0.178 ***	−0.0242	−0.0981 ***	−0.159 ***	−0.0104
lnFCF	−0.0467	−0.0148	−0.00727	0.169 ***	−0.203 ***	0.0519 *	0.789 ***	0.462 ***	−0.031
	BMAfill	ChairInd	CEODual	CEOCompRisk	BTenure	IndBM	CEOBM	CAuditInd	CNomInd
BMAfill	1
ChairInd	0.022	1
CEODual	0.198 ***	−0.459 ***	1
CEOCompRisk	0.146 ***	0.0859 ***	−0.0134	1
BTenure	0.0451	−0.226 ***	0.196 ***	−0.231 ***	1
IndBM	0.185 ***	0.526 ***	−0.137 ***	0.227 ***	−0.165 ***	1
CEOBM	0.155 ***	−0.456 ***	0.510 ***	0.0249	0.146 ***	−0.202 ***	1
CAuditInd	0.240 ***	0.272 ***	0.0556 *	0.209 ***	−0.0457 *	0.797 ***	0.0839 ***	1
CNomInd	−0.00905	0.271 ***	−0.213 ***	0.0324	−0.148 ***	0.114 ***	−0.194 ***	0.0416	1
CCompInd	0.168 ***	0.366 ***	−0.0769 **	0.120 ***	−0.178 ***	0.786 ***	−0.0839 **	0.809 ***	0.153 ***
lnBMComp	0.216 ***	0.125 ***	−0.0745 **	0.126 ***	0.0643 **	0.125 ***	−0.0849 ***	0.0584 *	−0.0393
BER	0.284 ***	−0.0633	0.0921 **	0.121 ***	−0.00546	0.0344	0.194 ***	0.145 ***	0.0431
BSP	0.149 ***	0.00054	0.0801 ***	0.127 ***	−0.0458 *	0.0962 ***	0.0655 **	0.105 ***	0.134 ***
PBD	0.0789 ***	0.0783 **	−0.0868 ***	0.162 ***	−0.201 ***	0.0708 **	0.0915 ***	0.0980 ***	0.198 ***
CSRCom	0.189 ***	0.0522 *	0.125 ***	0.247 ***	−0.106 ***	0.0710 **	0.0677 **	0.0895 ***	0.0566 *
lnSALES	0.373 ***	−0.0311	0.0733 **	0.181 ***	0.0547 *	0.0815 ***	0.144 ***	0.100 ***	0.00713
QR	−0.0673 **	−0.0555 *	−0.0395	−0.132 ***	0.0762 ***	−0.0369	−0.0434	−0.0629 **	0.042
ART	−0.130 ***	−0.0186	−0.123 ***	−0.0653 **	−0.0523 *	−0.0223	0.0139	0.0231	−0.0169
lnFCF	0.351 ***	−0.0356	0.0443	0.166 ***	0.0868 ***	0.0568 *	0.160 ***	0.115 ***	0.0669 *
	CCompInd	lnBMComp	BER	BSP	PBD	CSRCom	lnSALES	QR	ART
CCompInd	1
lnBMComp	0.0446	1
BER	0.108 **	0.110 **	1
BSP	0.156 ***	0.0426	0.151 ***	1
PBD	0.176 ***	−0.0991 ***	0.191 ***	0.302 ***	1
CSRCom	0.0965 ***	0.174 ***	0.172 ***	0.159 ***	0.214 ***	1
lnSALES	0.107 ***	0.474 ***	0.254 ***	0.188 ***	0.252 ***	0.307 ***	1
QR	−0.0106	−0.0757 **	−0.138 ***	−0.0166	−0.153 ***	−0.164 ***	−0.284 ***	1
ART	0.0670 *	0.0189	−0.0444	0.0254	0.0163	−0.0634 **	−0.0317	−0.0507*	1
lnFCF	0.0814 **	0.352 ***	0.223 ***	0.146 ***	0.284 ***	0.294 ***	0.715 ***	−0.243 ***	0.0108
	lnFCF
lnFCF	1

Source: Author’s own work. Notes: *, **, and *** represent significance at 10%, 5%, and 1%, respectively.

the correlations between the variables can be observed.

The correlation analysis highlights several strong correlations (above +/−0.7) between board and committee independence variables; therefore, in the empirical methods the authors carefully consider these relations when selected the variables and included them in different specifications.

4.2. Estimation Results for Performance Using Pooled OLS Regressions

To conduct the econometric analysis, pooled OLS regression models have been estimated as baseline models, and fixed effects and random effects models were estimated for the robustness of the results. The optimal specification between fixed effects and random effects models was selected using the Hausman test, applying a 5% significance level. Accordingly, when the p-value is below 5%, the fixed effects (FE) model is preferred, whereas when the p-value exceeds this threshold, the random effects (RE) model is selected. Results are presented separately for each dependent variable, including the relevant explanatory variables for each model. The analysis highlights the existence of a notable relationship among variables; therefore, the detailed examination of each set of results allows for the identification of practical implications for investors, managers, and policymakers.

Table 5. Estimation results for ROA using pooled OLS regression.

ROA	(1)	(2)	(3)	(4)	(5)
IndBM	0.000108	−0.000593 **
	(1.4903)	(−2.1947)
sqIndBM		0.000005 ***
		(2.8663)
CAuditInd			−0.000040	−0.000362	−0.000030
			(−0.4904)	(−1.1889)	(−0.3673)
sqCAuditInd				0.000003
				(1.0981)
CNomInd			0.000283	0.000291
			(1.1832)	(1.2157)
c_CNomInd					0.000986
					(1.3679)
c_sqCNomInd					0.000049
					(1.0335)
BSize	−0.003006 ***	−0.002683 ***	−0.000630	−0.000483	−0.000515
	(−5.3780)	(−4.7165)	(−0.9790)	(−0.7345)	(−0.7878)
BTenure	0.000291	0.000557	0.003255 ***	0.003380 ***	0.003216 ***
	(0.3708)	(0.7077)	(3.4202)	(3.5269)	(3.3776)
BMeet	−0.000130	−0.000208	−0.001275 **	−0.001328 **	−0.001329 **
	(−0.2371)	(−0.3811)	(−2.3780)	(−2.4665)	(−2.4669)
BMAfill	−0.001538	−0.002023	0.005725	0.005652	0.005892
	(−0.4835)	(−0.6385)	(1.5661)	(1.5460)	(1.6104)
CEODual	−0.000163	0.002432
	(−0.0446)	(0.6468)
lnBMComp	0.001164	0.000024	0.000226	−0.000265	0.000426
	(0.4543)	(0.0094)	(0.0809)	(−0.0938)	(0.1524)
BER	−0.003921	−0.003099
	(−0.9819)	(−0.7779)
BSP	0.033357	0.038931 *
	(1.6136)	(1.8839)
PBD	−0.002872	−0.002383
	(−0.5127)	(−0.4274)
ChairInd			−0.001593	−0.000909	−0.002136
			(−0.3448)	(−0.1950)	(−0.4595)
CEOBM			0.007797 *	0.008911 *	0.007679 *
			(1.6878)	(1.8841)	(1.6617)
CEOCompRisk			0.009644 **	0.009999 **	0.010015 **
			(2.1666)	(2.2406)	(2.2427)
CSRCom			0.003273	0.003609	0.002792
			(0.4011)	(0.4420)	(0.3416)
lnAge	0.004801 **	0.004100 *	0.008466 ***	0.008496 ***	0.008357 ***
	(2.2908)	(1.9514)	(3.8233)	(3.8369)	(3.7700)
ART	0.002789 ***	0.002809 ***	−0.000826	−0.000874	−0.000798
	(5.7263)	(5.7981)	(−1.4408)	(−1.5203)	(−1.3920)
lnFCF	0.005580 ***	0.005782 ***
	(3.5753)	(3.7206)
LevDE	−0.019990 ***	−0.019940 ***	−0.021314 ***	−0.020980 ***	−0.021088 ***
	(−7.5931)	(−7.6148)	(−7.1960)	(−7.0469)	(−7.1008)
QR	0.032184 ***	0.031553 ***	0.001015	0.000462	0.001020
	(8.9237)	(8.7774)	(0.2478)	(0.1119)	(0.2489)
lnTA			−0.017282 ***	−0.017612 ***	−0.017697 ***
			(−5.7282)	(−5.8094)	(−5.8146)
lnSALES			0.007091 ***	0.007269 ***	0.007106 ***
			(2.6782)	(2.7403)	(2.6838)
_cons	−0.097542	−0.072028	0.352424 ***	0.370991 ***	0.386628 ***
	(−1.4127)	(−1.0390)	(4.7618)	(4.8872)	(5.3501)
F statistic	22.395420 ***	21.676662 ***	11.824877 ***	11.237048 ***	11.227980 ***
R-sq	0.368373	0.376238	0.155950	0.156885	0.156778
Obs	592	592	1106	1106	1106
Mean VIF	1.33	3.67	1.63	3.71	2.61

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

and

Table 6. Estimation results for ROE using pooled OLS regression.

ROE	(1)	(2)	(3)	(4)	(5)
IndBM	0.000040	−0.001014 *
	(0.2437)	(−1.6609)
sqIndBM		0.000009 *
		(1.7900)
CAuditInd			−0.000029	−0.000913 **	−0.000027
			(−0.2575)	(−2.1739)	(−0.2360)
sqCAuditInd				0.000007 **
				(2.1838)
CNomInd			0.000144	0.000163
			(0.4341)	(0.4935)
c_CNomInd					0.000307
					(0.3093)
c_sqCNomInd					0.000011
					(0.1742)
BSize	−0.006486 ***	−0.006005 ***	−0.001747 *	−0.001305	−0.001720 *
	(−5.1114)	(−4.6380)	(−1.9368)	(−1.4147)	(−1.8796)
BTenure	−0.002107	−0.001661	0.000418	0.000761	0.000408
	(−1.1971)	(−0.9359)	(0.3163)	(0.5727)	(0.3084)
BMeet	−0.000800	−0.000892	−0.000517	−0.000642	−0.000530
	(−0.6412)	(−0.7157)	(−0.6917)	(−0.8574)	(−0.7050)
BMAfill	0.001488	0.000836	0.020529 ***	0.020197 ***	0.020571 ***
	(0.2108)	(0.1185)	(4.0654)	(4.0049)	(4.0673)
CEODual	0.003665	0.007300
	(0.4496)	(0.8705)
lnBMComp	0.010614 *	0.008805	0.015829 ***	0.014349 ***	0.015877 ***
	(1.8389)	(1.5054)	(4.0771)	(3.6471)	(4.0774)
BER	−0.006644	−0.005202
	(−0.7413)	(−0.5792)
BSP	0.096964 **	0.105365 **
	(2.1191)	(2.2950)
PBD	0.005772	0.006443
	(0.4633)	(0.5179)
ChairInd			0.010285	0.012118 *	0.010165
			(1.6020)	(1.8748)	(1.5735)
CEOBM			0.020319 ***	0.023190 ***	0.020299 ***
			(3.1526)	(3.5313)	(3.1474)
CEOCompRisk			0.002678	0.003848	0.002759
			(0.4340)	(0.6223)	(0.4457)
CSRCom			0.008084	0.009178	0.007971
			(0.7144)	(0.8116)	(0.7030)
lnAge	0.003775	0.002777	0.005981 *	0.006094 **	0.005953 *
	(0.8089)	(0.5920)	(1.9547)	(1.9949)	(1.9421)
ART	0.003647 ***	0.003677 ***	−0.001682 **	−0.001819 **	−0.001675 **
	(3.3721)	(3.4060)	(−2.1151)	(−2.2844)	(−2.1039)
lnFCF	0.007490 **	0.007776 **
	(2.1623)	(2.2468)
LevDE	0.018541 ***	0.018705 ***	0.022456 ***	0.023318 ***	0.022507 ***
	(3.1463)	(3.1800)	(5.4866)	(5.6809)	(5.4826)
QR	0.020137 **	0.019151 **	−0.005561	−0.007186	−0.005560
	(2.5187)	(2.3943)	(−0.9791)	(−1.2565)	(−0.9784)
lnTA			−0.055587 ***	−0.056491 ***	−0.055682 ***
			(−13.3814)	(−13.5555)	(−13.2826)
lnSALES			0.035408 ***	0.035852 ***	0.035410 ***
			(9.6203)	(9.7429)	(9.6162)
_cons	−0.125876	−0.088727	0.418448 ***	0.469887 ***	0.433946 ***
	(−0.8209)	(−0.5745)	(4.0830)	(4.4759)	(4.3348)
F statistic	4.419979 ***	4.360002 ***	14.199587 ***	13.722733 ***	13.400298 ***
R-sq	0.104196	0.109212	0.183514	0.187127	0.183537
Obs	586	586.	1092	1092	1092
Mean VIF	1.34	3.73	1.63	3.73	2.60

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

present corporate performance estimates using regression models without effects, proxied by ROA and ROE respectively. Five models were developed to examine the impact of governance mechanisms related to independence on firms’ profitability related to their asset-based operational efficiency (ROA) and engaged equity (ROE).

The estimation results for corporate performance proxied by ROA using pooled OLS regression models are presented in Table 5. The coefficient of determination (R²) indicates that, on average, up to 37% of the variation in performance measured by ROA is explained by the independent variables included in the analysis, depending on the specification.

The share of independent board members is associated with a negative effect on ROA (−0.000593, p < 0.05), indicating that, at lower levels, board independence is likely to reduce performance, proxied by operational efficiency (ROA). We find that the relation of board independence and performance is nonlinear, given that the coefficient of the quadratic term exhibits an opposite sign, which is statistically significant (0.000005, p < 0.01), suggesting that when the share of independent members on a board is above 59%, the higher the governance independence, the higher the ROA.

Similarly, results related to audit committee independence suggest a nonlinear relation with ROA; however, the results are not statistically significant (the linear coefficient of CAuditInd is −0.000362, while the quadratic term coefficient flips to positive 0.000003), suggesting that audit committee independence is likely to improve ROA above a threshold level. These results contradict Alhababsah and Azzam (2024) who show that audit committee independence represents a corporate governance mechanism with a particularly important impact in developing countries, where other control mechanisms are less effective, but does not affect performance.

Discussing other board-related variables, we find complementary results. In the case of large European firms governed under a continental corporate governance regime, board size and number of board meetings are likely to erode ROA (p < 0.01, p < 0.05), supporting the literature strands related to larger boards’ inefficiency due to difficulties in reaching consensus or making more compromises in final decisions, and therefore less extreme or opportunistic decisions than the ones made by smaller groups (Cheng, 2008). Board tenure positively influences ROA (p < 0.01), indicating that the accumulated experience of board members helps ensure adequate use of assets, which in turn leads to an increase in performance. Lun et al. (2025) contribute with similar results through their study investigating the role of board tenure in influencing firms’ capabilities. The number of memberships (board member affiliations) improves the performance of large listed firms (ROA), and the coefficients are statistically significant, between p < 0.1 and p < 0.05 levels. This result adds to the resource dependence theory, arguing that board members’ external connections bring advantages to the firm through access to resources, additional information that helps the management body to make the correct decisions, and multiple opportunities that improve firm performance. However, the existence of a board structure policy has a significant positive effect (p < 0.1) on performance, suggesting that controlling excessive membership at board level leads to increased profitability, due to better time allocation and active involvement in firm strategy and the taking of responsibilities. Additionally, the presence of the CEO as a board member positively influences ROA (p < 0.05), indicating that CEO involvement on the board can be beneficial by facilitating strategic decision-making and enhancing within-firm knowledge to boost profitability. Results show that CEO compensation related to firm’s risk also has a positive and significant effect on ROA (p < 0.05), showing that risk- and performance-based incentives can motivate management to improve firm outcomes. Coles and Hesterly (2000) argue that chairman independence enhances the effectiveness of corporate governance, especially when combined with a balanced board structure, which leads to improved performance.

Discussing firm attributes, this study adds to the literature by providing evidence that well-established firms, with higher sales, higher receivable turnover, and higher cash-flow and short-term liquidity, improve performance, while higher indebtedness erodes the firm’s revenue. The results provide additional evidence that larger firms with a larger asset base tend to have lower ROA (lnTA have negative and statistically significant coefficients, with 0.05 < p < 0.01), confirming the inverse relationship between asset size and efficiency of asset utilization, since the capital-intensive allocation for a larger asset base may erode profitability, therefore requiring more careful management.

The estimation results for ROA using polled OLS regression support the theoretical arguments employed in this study, confirming that board independence supports efficiency of asset utilization, offering support for H6 if the share of independent directors is relevant, adding to the arguments of critical mass theory. Related to larger firms with larger board sizes, our finding is consistent with studies such as Msomi and Nzuza (2025), who document a negative relationship between firm size and ROA, thus confirming H1a. Cognitive attributes of board members, proxied by board tenure and number of affiliations, have a positive impact on ROA, confirming the hypothesis H2a that accumulated experience contributes to more efficient decision-making and aligning with the conclusions of Lun et al. (2025), who emphasize the benefits of board experience on performance. Adams et al. (2005) discuss the balance between control and managerial expertise, further corroborating our findings, as we find that CEO presence on the board (H1b) and its compensation related to risks (H3a) positively effects ROA through inside-firm expertise and better risk management in large and complex settings, confirming these hypotheses.

Table 6 presents the estimation results for performance, alternatively measured by capital efficiency (ROE), using pooled OLS regression models. These results are largely in line with our estimation results for ROA, suggesting that our results are robust and not influenced by various accounting measures for performance in the proposed models.

The proportion of independent directors exhibits a nonlinear effect on ROE (−0.001014, p < 0.1 and 0.000009, p < 0.1), suggesting that governance independence above 56% improves ROE, offering support for critical mass arguments which suggest that to effectively influence decision-making, the share of independent directors on the board must be relevant. Similarly, audit committee independence (−0.000913, p < 0.05 and 0.000007, p < 0.05) improves capital efficiency above a 65% share of independent members in the audit committee. Board member compensation has a significant positive impact on ROE, with financial incentives playing an important role in aligning the interests of the board of directors and shareholders. Tai (2025) argues that director independence has a positive effect on performance, and that performance is maximized when independent directors can maintain decision-making objectivity while also benefiting from appropriate compensation.

In case of other board-related variables, larger boards are likely to erode ROE, while board member affiliations have a positive impact on ROE, indicating that professional networks lead to improved capital efficiency and return for shareholders. Board structural policies (BSPs) also have a significant positive impact on ROE, indicating that corporate governance policies to control excessive memberships facilitate a more efficient decision-making process, thereby increasing the value of shareholders’ equity. Regarding CEO-related roles, chairman independence has a positive effect on performance (p < 0.1), indicating that the separation of leadership roles creates more responsible oversight. The presence of the CEO as a board member positively and strongly influences return on equity, suggesting that inside-firm knowledge matters in complex settings, supporting a stronger link between strategic decisions and shareholder objectives. In their study, Bel-Oms and Grau Grau (2025) argue that CEO duality moderates the effectiveness of the audit, compensation, and nomination subcommittees, as well as firm performance, such that the overall impact on performance is positive. In contrast, audit committee independence exerts a significant negative effect, indicating that excessive monitoring and a higher level of audit control may lead to additional future costs, with an unfavorable impact on return on equity.

Firms’ attributes have been discussed alongside independent variables as their characteristics are expected to influence performance. Firm size (lnTA) is inversely correlated with ROE (p < 0.01) indicating that larger firms may encounter higher operational costs related to business expansion, thus reducing the capital return. Asset-intensive firms’ lower ROE can be explained by the sectoral capital-intensive peculiarities, meaning they may need higher equity to support long-term investments, thus eroding the efficiency of equity. Firm age positively influences performance, as more mature firms benefit from greater experience and their stable market. Operational efficiency, reflected in accounts receivable turnover, contributes significantly to increased capital return, as does free cash flow, which supports performance through improved liquidity.

Markedly, the higher indebtedness improves capital return efficiency, highlighting the role of financial leverage in improving the return on the invested capital of shareholders. High liquidity reduces financial risk and contributes to improved return of equity. Lastly, higher net sales have a significant positive impact on ROE, confirming the importance of sound commercial policies in enhancing shareholder profitability.

Accordingly, the negative and statistically significant effect of board size on ROE is aligned with the existing literature and supports hypothesis H1a regarding the importance of board structural characteristics in shaping the value for shareholders. With respect to board member affiliations, Aggarwal et al. (2019) confirm that such affiliations can contribute positively to firm performance, like our results, thereby supporting hypothesis H2a that the cross-experience benefit of affiliations enhances capital efficiency. In line with Hussain et al. (2024) and Fernandes et al. (2023), who highlight that the existence of formal governance policies improves decision-making processes, our findings suggest that the existence of a policy of maintaining well-balanced memberships (BSP) improves performance, offering support for hypothesis H5. The influence of the board and audit committee’s independence on ROE exhibits a positive influence on ROE, starting with an independence share threshold of 55–65%, supporting hypothesis H6. These results are in line with critical mass arguments and corroborated by the substantial body of literature that documents a positive effect of audit oversight on financial reporting quality, as shown by Klein (2002) and Bédard et al. (2004).

4.3. Estimation Results for Risk Outcome Using Pooled OLS Regressions

Table 7. Estimation results for sdROA using pooled OLS regression.

sdROA	(1)	(2)	(3)	(4)	(5)
IndBM	−0.000053 *	−0.000202 *
	(−1.7746)	(−1.7951)
sqIndBM		0.000001
		(1.3706)
CAuditInd			−0.000083 ***	−0.000097	−0.000090 ***
			(−3.6841)	(−1.1383)	(−3.9624)
sqCAuditInd				0.000000
				(0.1676)
CNomInd			0.000068	0.000069
			(1.0201)	(1.0248)
c_CNomInd					−0.000411 **
					(−2.0543)
c_sqCNomInd					−0.000034 **
					(−2.5414)
BSize	−0.000431 *	−0.000363	0.000269	0.000275	0.000189
	(−1.8647)	(−1.5346)	(1.4990)	(1.5012)	(1.0401)
BTenure	−0.000810 **	−0.000754 **	−0.001173 ***	−0.001167 ***	−0.001149 ***
	(−2.4967)	(−2.3069)	(−4.4170)	(−4.3660)	(−4.3362)
BMeet	−0.000351	−0.000368	−0.000050	−0.000053	−0.000014
	(−1.5473)	(−1.6191)	(−0.3372)	(−0.3508)	(−0.0928)
BMAfill	0.002102	0.001999	0.004208 ***	0.004205 ***	0.004095 ***
	(1.5978)	(1.5185)	(4.1355)	(4.1301)	(4.0313)
CEODual	−0.003583 **	−0.003034 *
	(−2.3703)	(−1.9416)
lnBMComp	0.000764	0.000523	0.000314	0.000293	0.000176
	(0.7204)	(0.4866)	(0.4038)	(0.3720)	(0.2269)
BER	0.001312	0.001486
	(0.7940)	(0.8973)
BSP	−0.003588	−0.002409
	(−0.4195)	(−0.2804)
PBD	0.004991 **	0.005095 **
	(2.1538)	(2.1990)
ChairInd			0.002258 *	0.002287 *	0.002623 **
			(1.7553)	(1.7613)	(2.0312)
CEOBM			0.000684	0.000732	0.000769
			(0.5318)	(0.5552)	(0.5990)
CEOCompRisk			0.004350 ***	0.004366 ***	0.004101 ***
			(3.5105)	(3.5117)	(3.3075)
CSRCom			0.002817	0.002831	0.003140
			(1.2402)	(1.2450)	(1.3835)
lnAge	0.000600	0.000452	0.000152	0.000153	0.000235
	(0.6924)	(0.5176)	(0.2446)	(0.2473)	(0.3793)
ART	0.000736 ***	0.000741 ***	0.001480 ***	0.001478 ***	0.001461 ***
	(3.6550)	(3.6785)	(9.2788)	(9.2353)	(9.1774)
lnFCF	0.000153	0.000196
	(0.2376)	(0.3038)
LevDE	−0.002782 **	−0.002772 **	−0.001502 *	−0.001487 *	−0.001652 **
	(−2.5549)	(−2.5470)	(−1.8207)	(−1.7927)	(−2.0031)
QR	0.006355 ***	0.006221 ***	0.009376 ***	0.009353 ***	0.009369 ***
	(4.2593)	(4.1641)	(8.2212)	(8.1341)	(8.2355)
lnTA			0.002977 ***	0.002963 ***	0.003257 ***
			(3.5441)	(3.5077)	(3.8539)
lnSALES			−0.004539 ***	−0.004532 ***	−0.004548 ***
			(−6.1592)	(−6.1343)	(−6.1857)
_cons	0.043699	0.049097 *	0.032619	0.033409	0.034479 *
	(1.5299)	(1.7041)	(1.5837)	(1.5805)	(1.7181)
F statistic	5.098940 ***	4.904952 ***	14.668043 ***	13.842327 ***	14.281522 ***
R-sq	0.117220	0.120094	0.186594	0.186615	0.191403
Obs	592	592	1105	1105	1105
Mean VIF	1.33	3.67	1.63	3.71	2.60

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

,

Table 8. Estimation results for LevDE using pooled OLS regression.

LevDE	(1)	(2)	(3)	(4)	(5)
IndBM	0.000936	−0.002083
	(0.8603)	(−0.5097)
sqIndBM		0.000026
		(0.7664)
CAuditInd			−0.000676	0.006779 **	−0.000866
			(−0.8403)	(2.2382)	(−1.0719)
sqCAuditInd				−0.000062 **
				(−2.5522)
CNomInd			−0.002779	−0.002960
			(−1.1670)	(−1.2457)
c_CNomInd					−0.016783 **
					(−2.3486)
c_sqCNomInd					−0.000983 **
					(−2.0781)
BSize	0.012224	0.013412	0.002559	−0.000754	0.000277
	(1.4179)	(1.5306)	(0.3987)	(−0.1154)	(0.0426)
BTenure	−0.010375	−0.009185	−0.003313	−0.006570	−0.002656
	(−0.8834)	(−0.7750)	(−0.3453)	(−0.6806)	(−0.2772)
BMeet	−0.012793	−0.013118	−0.016686 ***	−0.015327 ***	−0.015531 ***
	(−1.5603)	(−1.5973)	(−3.1417)	(−2.8787)	(−2.9129)
BMAfill	0.022921	0.020690	−0.049441	−0.047987	−0.052746
	(0.4813)	(0.4335)	(−1.3586)	(−1.3219)	(−1.4503)
CEODual	−0.100632 *	−0.089285
	(−1.8445)	(−1.5789)
lnBMComp	−0.040729	−0.045514	−0.112046 ***	−0.100107 ***	−0.115622 ***
	(−1.0622)	(−1.1711)	(−4.0660)	(−3.5903)	(−4.1940)
BER	−0.006461	−0.003020
	(−0.1080)	(−0.0503)
BSP	0.163750	0.189508
	(0.5280)	(0.6073)
PBD	0.031605	0.033565
	(0.3768)	(0.3998)
ChairInd			−0.020839	−0.037056	−0.010088
			(−0.4531)	(−0.8001)	(−0.2183)
CEOBM			−0.098559 **	−0.124390 ***	−0.096025 **
			(−2.1454)	(−2.6507)	(−2.0927)
CEOCompRisk			0.091917 **	0.082615 *	0.083982 *
			(2.0776)	(1.8657)	(1.8941)
CSRCom			0.032279	0.024378	0.041755
			(0.3978)	(0.3009)	(0.5145)
lnAge	0.038774	0.035979	0.016095	0.014744	0.018028
	(1.2318)	(1.1350)	(0.7270)	(0.6675)	(0.8149)
ART	0.005735	0.005977	−0.001663	−0.000627	−0.002223
	(0.7649)	(0.7961)	(−0.2914)	(−0.1098)	(−0.3896)
lnFCF	0.032838	0.034076
	(1.3837)	(1.4321)
ROA	−5.830148 ***	−5.892858 ***	−3.829512 ***	−3.686397 ***	−3.777075 ***
	(−8.2210)	(−8.2517)	(−8.0249)	(−7.6914)	(−7.9160)
QR	−0.171342 ***	−0.172166 ***	−0.178127 ***	−0.165286 ***	−0.177824 ***
	(−3.0200)	(−3.0329)	(−4.3882)	(−4.0510)	(−4.3874)
lnTA			0.099178 ***	0.107703 ***	0.107520 ***
			(3.2551)	(3.5226)	(3.5039)
lnSALES			−0.006494	−0.011265	−0.006983
			(−0.2451)	(−0.4252)	(−0.2640)
_cons	2.074044 **	2.174735 **	2.559401 ***	2.092872 ***	2.136619 ***
	(2.0073)	(2.0872)	(3.4695)	(2.7603)	(2.9605)
F statistic	10.875431 ***	10.225118 ***	14.082944 ***	13.729852 ***	13.581049 ***
R-sq	0.220707	0.221502	0.180359	0.185241	0.183602
Obs	592	592	1106	1106	1106
Mean VIF	1.37	3.74	1.65	3.73	2.62

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

and

Table 9. Estimation results for sdQR using pooled OLS regression.

sdQR	(1)	(2)	(3)	(4)	(5)
IndBM	−0.000108	−0.001951 *
	(−0.3498)	(−1.6804)
sqIndBM		0.000016
		(1.6465)
CAuditInd			−0.000503 **	−0.003575 ***	−0.000472 **
			(−2.1914)	(−4.1835)	(−2.0438)
sqCAuditInd				0.000026 ***
				(3.7304)
CNomInd			0.002215 ***	0.002239 ***
			(3.2829)	(3.3382)
c_CNomInd					0.004482 **
					(2.2037)
c_sqCNomInd					0.000159
					(1.1815)
BSize	−0.004312 *	−0.003537	−0.003962 **	−0.002474	−0.003591 *
	(−1.7504)	(−1.4125)	(−2.1342)	(−1.3104)	(−1.9080)
BTenure	−0.003961	−0.003211	−0.006390 **	−0.005185 *	−0.006527 **
	(−1.1843)	(−0.9528)	(−2.3510)	(−1.9055)	(−2.3998)
BMeet	0.001398	0.001240	−0.002179	−0.002600 *	−0.002356
	(0.5905)	(0.5239)	(−1.4188)	(−1.6989)	(−1.5271)
BMAfill	0.010910	0.009844	0.045448 ***	0.043938 ***	0.046042 ***
	(0.8149)	(0.7355)	(4.3844)	(4.2608)	(4.4373)
CEODual	−0.038560 **	−0.031979 **
	(−2.5095)	(−2.0171)
lnBMComp	0.007731	0.004756	0.007972	0.002849	0.008645
	(0.7055)	(0.4289)	(0.9932)	(0.3519)	(1.0746)
BER	−0.005730	−0.003248
	(−0.3366)	(−0.1903)
BSP	−0.365875 ***	−0.350347 ***
	(−4.1954)	(−4.0000)
PBD	0.025719	0.026907
	(1.0872)	(1.1386)
ChairInd			−0.013655	−0.007033	−0.015325
			(−1.0339)	(−0.5309)	(−1.1540)
CEOBM			−0.052731 ***	−0.042508 ***	−0.053011 ***
			(−3.9631)	(−3.1471)	(−3.9843)
CEOCompRisk			0.000419	0.004780	0.001555
			(0.0331)	(0.3779)	(0.1224)
CSRCom			−0.066130 ***	−0.061685 ***	−0.067699 ***
			(−2.8461)	(−2.6672)	(−2.9094)
lnAge	0.005311	0.003568	−0.000518	−0.000075	−0.000905
	(0.5995)	(0.4005)	(−0.0823)	(−0.0119)	(−0.1435)
ART	0.005472 ***	0.005582 ***	0.008231 ***	0.007862 ***	0.008319 ***
	(2.6533)	(2.7092)	(5.0973)	(4.8885)	(5.1473)
lnFCF	−0.022706 ***	−0.022099 ***
	(−3.4427)	(−3.3505)
LevDE	−0.075665 ***	−0.074898 ***	−0.038512 ***	−0.034525 ***	−0.037795 ***
	(−6.9514)	(−6.8850)	(−4.5713)	(−4.0894)	(−4.4754)
ROE	−0.046653	−0.057280	−0.099474	−0.113921 *	−0.099870
	(−0.5899)	(−0.7230)	(−1.5871)	(−1.8250)	(−1.5937)
lnTA			0.032613 ***	0.029674 ***	0.031254 ***
			(3.6377)	(3.3168)	(3.4584)
lnSALES			−0.078683 ***	−0.076642 ***	−0.078653 ***
			(−9.9825)	(−9.7581)	(−9.9805)
_cons	0.855638 ***	0.914961 ***	1.294552 ***	1.454328 ***	1.530934 ***
	(2.9716)	(3.1577)	(6.2882)	(6.9559)	(7.7182)
F statistic	9.131718 ***	8.756106 ***	20.202414 ***	20.082594 ***	19.164645 ***
R-sq	0.193749	0.197572	0.242297	0.251997	0.243281
Obs	586	586	1092	1092	1092
Mean VIF	1.33	3.67	1.63	3.71	2.61

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

present corporate risk outcome estimates proxied by return volatility (sdROA), leverage (LevDE) and short-term volatility (sdQR), using regression models without effects. Standard deviation is computed by the authors using a 5-year rolling window, thus including data back to 2011. Five models were developed to examine the impact of governance independence on risk outcome.

Table 7 presents the estimation results for ROA volatility, proxied by the standard deviation over 5 years (sdROA), as a frequently used measure for operational risk (see Cheng, 2008 or Jane Lenard et al., 2014), and using pooled OLS regression as baseline. Higher volatility indicates that the ability of the firm to consistently generate profit from its assets is unstable. This instability occurs in the firm’s core business, such as managing sales and demand, high operating leverage or poor asset exploitation.

The results highlight that board independence (H6) and nomination committee independence (H7) reduce operational risk by preventing earning instability, as governance independence strengthens the monitoring function and limits opportunistic managerial decisions. Audit committee independence (H7) contributes to risk reduction through well-established internal control mechanisms and high-quality financial reporting, reducing errors and fraud. Mans-Kemp and Viviers (2019) emphasize that nomination committee independence is a determining factor for board diversity and quality, with positive implications for the decision-making process. An independent nomination committee tends to select directors based on broad and professional criteria, which is reflected in lower volatility of returns (sdROA), that is, reduced operational risk. Accordingly, the results related to board and committee independence reflect their influence on reducing risk, providing support for hypotheses H6 and H7.

Boards’ structural and cognitive diversity, proxied by board size and board tenure, have a significant negative effect on return volatility, suggesting that firms with larger boards have lower variability of corporate performance (Cheng, 2008). However, despite increasing the cognitive level of the board members, the number of affiliations may contribute to increasing return volatility, since the time burden may affect oversight responsibilities. Confirming its positive influence on corporate outcomes discussed in Section 4.2, CEO duality is associated with lower variability of outcomes, building on firm-specific knowledge that may lead to faster and more coherent decision-making, which may reduce uncertainty. The existence of a policy for board diversity (PBD) brings different perspectives and supports different interests, which may lead to more innovative but also more unpredictable decisions, given the diversity of opinions and viewpoints, contrary to our expectations in H5. Similarly, the influence of risk-based CEO incentives highlights the propensity for boards to take higher risk and assume higher return volatility, following the principle that “if you do not take risks, you do not gain”, thus supporting H3b. Higher short-term liquidity (QR) may lead to higher risk-taking and performance volatility, indicating that cash buffers may influence the board to explore riskier opportunities that may lead to higher fluctuations of returns.

The results show that high-leverage firms exhibit lower operating volatility (sdROA), suggesting that these firms may face certain constraints from the lenders, preventing them from taking riskier decisions. Moreover, the results highlight that less volatility of return may not capture higher indebtedness. In the case of shrinking net sales, this is most likely to be reflected directly in fluctuating returns, while a higher volume of commercial activity stabilizes earnings.

Discussing possible nonlinear influences of our predictors on operational risk, we find that the quadratic terms of board independence and audit committee independence, despite suggesting a turning point (the coefficient changes sign) the results are not significant. In the case of nomination committee independence, the quadratic term coefficient remains negative and statistically significant (−0.000411, p < 0.05 and −0.000034, p < 0.05) suggesting that the relation between nomination committee independence and operational risk is linear and negative.

The negative influence of board size and board tenure on operational risk is in line with Lun et al. (2025), and supports hypothesis H1a and H2b. The negative relationship between CEO duality and sdROA is consistent with the arguments advanced by Krause et al. (2014) and Tang (2017) and confirms hypothesis H1b. Board independence and nomination committee independence are associated with lower operational risk (sdROA), supporting hypotheses H6 and H7.

Table 8 presents the baseline OLS regression results for financial leverage (LevDE). Audit committee independence exhibits a nonlinear relation with debt level (0.006779, p < 0.05 and −0.000062, p < 0.05), indicating that stricter monitoring is associated with lower debts if the share of independent members is above the 55% threshold level. Nomination committee independence has a linear negative influence on leverage (−0.016783, p < 0.05 and −0.000983, p < 0.05), indicating that stronger independence mechanisms within this committee reduce debts in the capital structure, in turn suggesting that members nominated by a more highly independent committee may promote more prudent financing policies. These results related to the independence level of boards and committees, and leverage, are in line with Zubeltzu-Jaka et al. (2019) and Fuente et al. (2017), offering support for H6 and H7.

Further results show CEO duality (H1b), memberships (H2b), board member compensation (H3a) and board meeting attendance (H4b) reduce the leverage level, offering support for these hypotheses. Other results for returns (ROA) and high short-term liquidity (QR) indicate that these factors reduce the need for higher financial leverage. These results suggest there is a need for responsible governance mechanisms adapted to firms’ peculiarities, which, together with better returns and high liquidity, allow firms to strategically finance their activities from internal resources or leveraging debts. Our results are in line with Liu and Du (2026), who argue that CEO power influences debt structure more than its level, and that CEO presence on the board may lead to a reduction in leverage as a strategy to maintain control.

In the case that the CEO compensation is related to risk, firms’ indebtedness is likely to be higher, indicating that risk-based CEO incentives encourage firms engaging in higher debts to finance high-return strategies, aligning with the arguments of Acero and Alcalde (2020) and Fang and Huang (2024) and supporting H3b.

Table 9 reports estimation results for the volatility of the short-term liquidity ratio (sdQR) using pooled OLS baseline regression.

The effect of board independence on liquidity stability suggests the existence of a nonlinear relation with liquidity volatility, although the quadratic term coefficient flips its sign to positive and it is still not statistically significant. Audit committee independence exhibits a nonlinear relation with liquidity volatility (−0.003575, p < 0.01 and 0.000026, p < 0.01), suggesting that the share of independent directors in audit committee is associated with stable liquidity only below a threshold level (68.75%). These results support the excessive monitoring arguments that lead to limited marginal improvement of corporate outcome beyond this threshold; however, they offer support for H6. The independence of the nomination committee is likely to increase liquidity volatility in our sample. These results may be associated with the nomination committee’s independence director’s preference for the selection of candidates that are more willing to take risks, potentially leading to unstable liquidity.

The results for liquidity volatility show that larger boards (H1a) and CEO–Chair duality and CEO presence on the board (H1b), as well as extended tenure of board members (H2b) constrain liquidity stability, suggesting that diversity of perspective and inside-firm knowledge prevent the company from liquidity shortage in the short-term. These findings are consistent with agency theory predictions and prior evidence on coordination costs and CEO power concentration (Krause et al. (2014), Tang (2017), and Msomi and Nzuza (2025)). Similarly, a more participative board (meeting attendance) delivers better oversight, while board structure policies (BSPs) control excessive memberships, thus contribute to reducing liquidity risk (H4a and H5). The results suggest that board independence contributes to stricter monitoring, having a negative influence on short-term liquidity volatility (−0.001951, p < 0.1), and suggesting that independent mechanisms preserve liquidity, thus supporting H6. This is consistent with agency theory predictions and empirical findings in Zubeltzu-Jaka et al. (2019). Markedly, firms that established a CSR committee are likely to encounter reduced cash fluctuations (by 6.7% compared to firms without a CSR), since its influence is negative and statistically significant (−0.067699, p < 0.01), and these results are risk dependent (in the case of return volatility and leverage, the influence is positive but not statistically significant). Conversely, excessive affiliations may disrupt decision-making and lead to riskier and unstable liquidity due to insufficient time allocation of board members to oversee the firms’ risk exposure, in line with the positive influence presented above for return volatility (sdROA), and contrary to H2b.

Several firm attributes, such as return on equity (p < 0.1) and high net sales (p < 0.01), are negatively associated with fluctuating liquidity. Firms with higher receivables turnover (ART) are more likely to support their commercial intensive activity with existing liquidity, at the expense of cash fluctuations over financial exercises, while free cash flow buffer (FCF) supports stable liquidity in large non-financial listed firms, results that are largely accepted in corporate finance.

Discussing potential nonlinearities related to board’s committees, we find that the relationship of nomination committee independence with liquidity fluctuation is linear and positive, and statistically significant, but has limited economic impact. In contrast, the audit committee’s independence relation with liquidity volatility is nonlinear, contributing to more stable liquidity at the initial stage, below a threshold level of independent members. Thus, concluding about the nexus between independence mechanisms and liquidity volatility, we provide notable results relating to the audit committee’s independence, highlighting that the marginal effect of audit independence on liquidity stability starts diminishing beyond the 68.75% level, supporting the excessive monitoring arguments discussed by critical mass theories, while offering support for H7.

The results exhibit a significant negative effect of board size (H1a), CEO presence on board and CEO duality (H1b), and board tenure and affiliations (H2b) on liquidity stability, offering support for these hypotheses. Similarly, active oversight, reflected in meeting attendance, contributes to stricter monitoring and is likely to reduce liquidity volatility, as proposed by H4a This supports the documented trade-off between experience and entrenchment effects in Mielcarz and Osiichuk (2025) and Lun et al. (2025), as well as monitoring dilution effects from affiliations shown in Aggarwal et al. (2019) and Latif et al. (2020).

4.4. Fixed Effects/Random Effects Models for Performance and Risk Outcomes

Table 10. Estimation results for ROA using regression with industry fixed effects.

ROA	(1)	(2)	(3)	(4)	(5)
IndBM	0.000027	−0.000551 **
	(0.3640)	(−2.0699)
sqIndBM		0.000005 **
		(2.2580)
CAuditInd			−0.000131	−0.000496 *	−0.000123
			(−1.6415)	(−1.6614)	(−1.5265)
sqCAuditInd				0.000003
				(1.2678)
CNomInd			0.000222	0.000233
			(0.9463)	(0.9932)
c_CNomInd					0.000868
					(1.2321)
c_sqCNomInd					0.000045
					(0.9723)
BSize	−0.003003 ***	−0.002722 ***	−0.000677	−0.000514	−0.000567
	(−5.3180)	(−4.7223)	(−1.0529)	(−0.7835)	(−0.8691)
BTenure	0.000338	0.000604	0.004318 ***	0.004485 ***	0.004272 ***
	(0.4132)	(0.7344)	(4.5117)	(4.6438)	(4.4574)
BMeet	−0.000022	−0.000085	−0.000997 *	−0.001050 **	−0.001048 **
	(−0.0416)	(−0.1581)	(−1.9018)	(−1.9977)	(−1.9893)
BMAfill	−0.002339	−0.002746	0.004121	0.004024	0.004249
	(−0.7259)	(−0.8538)	(1.1312)	(1.1048)	(1.1655)
CEODual	−0.000773	0.001321
	(−0.2138)	(0.3550)
lnBMComp	0.003340	0.002272	0.003059	0.002496	0.003232
	(1.2964)	(0.8704)	(1.1130)	(0.8970)	(1.1735)
BER	−0.002524	−0.001923
	(−0.6391)	(−0.4876)
BSP	0.029205	0.033639 *
	(1.4439)	(1.6612)
PBD	−0.002444	−0.002288
	(−0.4376)	(−0.4112)
ChairInd			−0.002958	−0.002065	−0.003519
			(−0.6384)	(−0.4407)	(−0.7538)
CEOBM			0.011171 **	0.012338 ***	0.011102 **
			(2.4562)	(2.6596)	(2.4406)
CEOCompRisk			0.009712 **	0.010119 **	0.010084 **
			(2.2241)	(2.3118)	(2.3006)
CSRCom			−0.000233	0.000208	−0.000706
			(−0.0292)	(0.0260)	(−0.0882)
lnAge	0.007074 ***	0.006373 ***	0.006959 ***	0.006977 ***	0.006905 ***
	(3.3393)	(2.9873)	(3.1118)	(3.1208)	(3.0869)
ART	0.003812 ***	0.003793 ***	0.000683	0.000632	0.000713
	(6.8708)	(6.8605)	(1.0280)	(0.9503)	(1.0720)
lnFCF	0.004508 ***	0.004641 ***
	(2.8860)	(2.9802)
LevDE	−0.019389 ***	−0.019199 ***	−0.019329 ***	−0.018989 ***	−0.019107 ***
	(−6.8999)	(−6.8540)	(−6.2610)	(−6.1293)	(−6.1719)
QR	0.024115 ***	0.023623 ***	−0.005085	−0.005914	−0.005123
	(6.2828)	(6.1667)	(−1.1572)	(−1.3315)	(−1.1658)
lnTA			−0.016088 ***	−0.016195 ***	−0.016490 ***
			(−3.8694)	(−3.8954)	(−3.9464)
lnSALES			0.008070 *	0.007925 *	0.008131 *
			(1.9415)	(1.9063)	(1.9558)
_cons	−0.114520 *	−0.090854	0.246433 ***	0.268492 ***	0.273873 ***
	(−1.6707)	(−1.3148)	(3.3268)	(3.5295)	(3.7835)
F statistic	18.867403 ***	18.134708 ***	9.820204 ***	9.369157 ***	9.326688 ***
R-sq overall	0.3620	0.3603	0.1353	0.1363	0.1359
Obs	592	592	1106	1106	1106
N Sectors	10	10	10	10	10
FE/RE	FE	FE	FE	FE	FE
Chi2	0.0000	0.0000	0.0000	0.0000	0.0000

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

,

Table 11. Estimation results for ROE using regression with industry fixed effects.

ROE	(1)	(2)	(3)	(4)	(5)
IndBM	−0.000167	−0.000740
	(−1.0783)	(−1.3049)
sqIndBM		0.000005
		(1.0499)
CAuditInd			−0.000191 *	−0.001014 **	−0.000187 *
			(−1.7460)	(−2.4953)	(−1.7009)
sqCAuditInd				0.000007 **
				(2.1029)
CNomInd			−0.000078	−0.000055
			(−0.2427)	(−0.1714)
c_CNomInd					0.000215
					(0.2247)
c_sqCNomInd					0.000021
					(0.3249)
BSize	−0.005894 ***	−0.005616 ***	−0.001834 **	−0.001432	−0.001784 **
	(−4.8757)	(−4.5387)	(−2.0507)	(−1.5688)	(−1.9658)
BTenure	−0.000236	0.000060	0.001446	0.001825	0.001423
	(−0.1371)	(0.0342)	(1.1034)	(1.3814)	(1.0839)
BMeet	−0.000771	−0.000821	−0.000569	−0.000671	−0.000592
	(−0.6695)	(−0.7118)	(−0.7878)	(−0.9288)	(−0.8162)
BMAfill	0.000434	0.000049	0.015997 ***	0.015665 ***	0.016061 ***
	(0.0644)	(0.0073)	(3.2163)	(3.1531)	(3.2253)
CEODual	0.000222	0.002146
	(0.0292)	(0.2746)
lnBMComp	0.015639 ***	0.014525 ***	0.017725 ***	0.016342 ***	0.017806 ***
	(2.8606)	(2.6084)	(4.6878)	(4.2646)	(4.6970)
BER	−0.002367	−0.001667
	(−0.2836)	(−0.1991)
BSP	0.078822 *	0.083230 *
	(1.8691)	(1.9641)
PBD	0.008778	0.008883
	(0.7505)	(0.7595)
ChairInd			0.011301 *	0.013257 **	0.011055 *
			(1.7832)	(2.0729)	(1.7313)
CEOBM			0.025169 ***	0.027640 ***	0.025148 ***
			(4.0142)	(4.3396)	(4.0090)
CEOCompRisk			0.005723	0.006812	0.005883
			(0.9566)	(1.1362)	(0.9797)
CSRCom			0.005781	0.006920	0.005566
			(0.5287)	(0.6331)	(0.5079)
lnAge	0.011309 **	0.010624 **	0.007645 **	0.007705 **	0.007615 **
	(2.5458)	(2.3667)	(2.5056)	(2.5294)	(2.4938)
ART	0.007965 ***	0.007945 ***	0.001313	0.001192	0.001327
	(6.8703)	(6.8527)	(1.4450)	(1.3115)	(1.4581)
lnFCF	0.003820	0.003947
	(1.1708)	(1.2091)
LevDE	0.020033 ***	0.020298 ***	0.028168 ***	0.028890 ***	0.028269 ***
	(3.3803)	(3.4223)	(6.6915)	(6.8511)	(6.6945)
QR	−0.007315	−0.007794	−0.018041 ***	−0.019968 ***	−0.018058 ***
	(−0.9119)	(−0.9701)	(−3.0027)	(−3.2905)	(−3.0040)
lnTA			−0.044240 ***	−0.044509 ***	−0.044424 ***
			(−7.7726)	(−7.8304)	(−7.7635)
lnSALES			0.027062 ***	0.026735 ***	0.027087 ***
			(4.7813)	(4.7294)	(4.7833)
_cons	−0.142331	−0.119461	0.335787 ***	0.385675 ***	0.331199 ***
	(−0.9915)	(−0.8228)	(3.3130)	(3.7107)	(3.3458)
F statistic	7.265476 ***	6.881518 ***	10.495666 ***	10.190108 ***	9.910115 ***
R-sq overall	0.0649	0.0679	0.1481	0.1508	0.1480
Obs	586	586	1092	1092	1092
N Sectors	10	10	10	10	10
FE/RE	FE	FE	FE	FE	FE
Chi2	0.0000	0.0000	0.0000	0.0000	0.0000

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

,

Table 12. Estimation results for sdROA using regression with industry fixed effects.

sdROA	(1)	(2)	(3)	(4)	(5)
IndBM	−0.000053 *	−0.000254 **
	(−1.7746)	(−2.3408)
sqIndBM		0.000001
		(1.4047)
CAuditInd			−0.000109 ***	−0.000097	−0.000113 ***
			(−4.9352)	(−1.1383)	(−5.1113)
sqCAuditInd				0.000000
				(0.1676)
CNomInd			0.000049	0.000069
			(0.7612)	(1.0248)
c_CNomInd					−0.000287
					(−1.4826)
c_sqCNomInd					−0.000024 *
					(−1.8442)
BSize	−0.000431 *	−0.000468 **	0.000124	0.000275	0.000066
	(−1.8647)	(−1.9879)	(0.7000)	(1.5012)	(0.3670)
BTenure	−0.000810 **	−0.000819 **	−0.001290 ***	−0.001167 ***	−0.001269 ***
	(−2.4967)	(−2.4393)	(−4.8786)	(−4.3660)	(−4.7967)
BMeet	−0.000351	−0.000204	0.000103	−0.000053	0.000130
	(−1.5473)	(−0.9315)	(0.7149)	(−0.3508)	(0.8942)
BMAfill	0.002102	0.003667 ***	0.004878 ***	0.004205 ***	0.004812 ***
	(1.5978)	(2.7935)	(4.8654)	(4.1301)	(4.8017)
CEODual	−0.003583 **	−0.002764 *
	(−2.3703)	(−1.8203)
lnBMComp	0.000764	0.001351	0.000768	0.000293	0.000678
	(0.7204)	(1.2683)	(1.0152)	(0.3720)	(0.8950)
BER	0.001312	0.003082 *
	(0.7940)	(1.9145)
BSP	−0.003588	−0.003556
	(−0.4195)	(−0.4302)
PBD	0.004991 **	0.006390 ***
	(2.1538)	(2.8138)
ChairInd			0.002417 *	0.002287 *	0.002705 **
			(1.8942)	(1.7613)	(2.1061)
CEOBM			0.001081	0.000732	0.001121
			(0.8636)	(0.5552)	(0.8963)
CEOCompRisk			0.003844 ***	0.004366 ***	0.003655 ***
			(3.1967)	(3.5117)	(3.0313)
CSRCom			0.000846	0.002831	0.001090
			(0.3844)	(1.2450)	(0.4950)
lnAge	0.000600	0.001011	0.000441	0.000153	0.000477
	(0.6924)	(1.1609)	(0.7111)	(0.2473)	(0.7693)
ART	0.000736 ***	0.000803 ***	0.001303 ***	0.001478 ***	0.001288 ***
	(3.6550)	(3.5591)	(7.1232)	(9.2353)	(7.0397)
lnFCF	0.000153	−0.000595
	(0.2376)	(−0.9369)
LevDE	−0.002782 **	−0.002458 **	−0.001400 *	−0.001487 *	−0.001513 *
	(−2.5549)	(−2.1499)	(−1.6467)	(−1.7927)	(−1.7774)
QR	0.006355 ***	0.003007 *	0.005756 ***	0.009353 ***	0.005775 ***
	(4.2593)	(1.9232)	(4.7597)	(8.1341)	(4.7805)
lnTA			0.004632 ***	0.002963 ***	0.004835 ***
			(4.0424)	(3.5077)	(4.2050)
lnSALES			−0.005887 ***	−0.004532 ***	−0.005912 ***
			(−5.1393)	(−6.1343)	(−5.1671)
_cons	0.043699	0.040460	0.014935	0.033409	0.016536
	(1.5299)	(1.4347)	(0.7326)	(1.5805)	(0.8310)
F statistic		4.780931 ***	12.306158 ***		11.837309 ***
Wald Chi2	76.48			249.16
R-sq overall	0.1172	0.0987	0.1688	0.1866	0.1734
Obs	592	592	1105	1105	1105
N Sectors	10	110	10	10	10
FE/RE	RE	FE	FE	RE	FE
Chi2	0.1231	0.0000	0.0080	0.4729	0.0000

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

,

Table 13. Estimation results for LevDE using regression with industry fixed effects.

LevDE	(1)	(2)	(3)	(4)	(5)
IndBM	0.000774	0.000772
	(0.7353)	(0.2010)
sqIndBM		0.000000
		(0.0007)
CAuditInd			−0.000435	0.005704 **	−0.000609
			(−0.5608)	(1.9695)	(−0.7830)
sqCAuditInd				−0.000051 **
				(−2.1997)
CNomInd			−0.001891	−0.002084
			(−0.8321)	(−0.9176)
c_CNomInd					−0.015668 **
					(−2.3045)
c_sqCNomInd					−0.000967 **
					(−2.1495)
BSize	0.001588	0.001589	−0.004254	−0.006870	−0.006532
	(0.1906)	(0.1872)	(−0.6813)	(−1.0828)	(−1.0331)
BTenure	−0.019756 *	−0.019754 *	−0.000260	−0.003403	0.000606
	(−1.6931)	(−1.6731)	(−0.0277)	(−0.3588)	(0.0646)
BMeet	−0.009665	−0.009666	−0.011212 **	−0.010225 **	−0.010059 **
	(−1.2557)	(−1.2530)	(−2.2132)	(−2.0141)	(−1.9777)
BMAfill	0.036382	0.036380	−0.058807 *	−0.057110	−0.061355 *
	(0.7892)	(0.7870)	(−1.6697)	(−1.6240)	(−1.7440)
CEODual	−0.129946 **	−0.129937 **
	(−2.5262)	(−2.4420)
lnBMComp	−0.028889	−0.028893	−0.091530 ***	−0.081864 ***	−0.094903 ***
	(−0.7828)	(−0.7700)	(−3.4582)	(−3.0565)	(−3.5854)
BER	−0.001680	−0.001677
	(−0.0297)	(−0.0296)
BSP	0.216726	0.216746
	(0.7483)	(0.7437)
PBD	0.050327	0.050327
	(0.6303)	(0.6297)
ChairInd			−0.024388	−0.039864	−0.012521
			(−0.5430)	(−0.8784)	(−0.2772)
CEOBM			−0.088675 **	−0.108785 **	−0.087158 **
			(−2.0095)	(−2.4181)	(−1.9782)
CEOCompRisk			0.078934 *	0.071060 *	0.070390 *
			(1.8657)	(1.6766)	(1.6592)
CSRCom			0.097678	0.089987	0.107405
			(1.2638)	(1.1652)	(1.3896)
lnAge	0.057740 *	0.057737 *	0.020920	0.019960	0.021760
	(1.8908)	(1.8728)	(0.9601)	(0.9175)	(1.0002)
ART	0.004563	0.004563	0.009375	0.010025	0.008633
	(0.5507)	(0.5502)	(1.4495)	(1.5511)	(1.3351)
lnFCF	0.038739 *	0.038740 *
	(1.7161)	(1.7122)
ROA	−5.009128 ***	−5.009176 ***	−3.013328 ***	−2.897001 ***	−2.961358 ***
	(−7.2146)	(−7.1696)	(−6.4130)	(−6.1375)	(−6.3046)
QR	−0.196848 ***	−0.196849 ***	−0.178237 ***	−0.163993 ***	−0.176846 ***
	(−3.5213)	(−3.5164)	(−4.2196)	(−3.8443)	(−4.1932)
lnTA			0.211484 ***	0.213422 ***	0.219554 ***
			(5.2820)	(5.3385)	(5.4686)
lnSALES			−0.122083 ***	−0.119647 ***	−0.123069 ***
			(−3.0415)	(−2.9850)	(−3.0710)
_cons	1.534398	1.534494	1.540792 **	1.148758	1.217379 *
	(1.5625)	(1.5450)	(2.1445)	(1.5544)	(1.7331)
F statistic	9.014829 ***	8.436497 ***	10.317707 ***	10.047984 ***	10.033896 ***
R-sq overall	0.2135	0.2135	0.1570	0.1634	0.1601
Obs	592	592	1106	1106	1106
N Sectors	10	10	10	10	10
FE/RE	FE	FE	FE	FE	FE
Chi2	0.0000	0.0000	0.0000	0.0000	0.0000

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

and

Table 14. Estimation results for sdQR using regression with industry fixed effects.

sdQR	(1)	(2)	(3)	(4)	(5)
IndBM	−0.000108	−0.001951 *
	(−0.3498)	(−1.6804)
sqIndBM		0.000016 *
		(1.6465)
CAuditInd			−0.000797 ***	−0.004268 ***	−0.000761 ***
			(−3.5648)	(−5.2296)	(−3.3901)
sqCAuditInd				0.000029 ***
				(4.4195)
CNomInd			0.002120 ***	0.002157 ***
			(3.2536)	(3.3391)
c_CNomInd					0.004827 **
					(2.4698)
c_sqCNomInd					0.000190
					(1.4691)
BSize	−0.004312 *	−0.003537	−0.006254 ***	−0.004536 **	−0.005796 ***
	(−1.7504)	(−1.4125)	(−3.4167)	(−2.4438)	(−3.1229)
BTenure	−0.003961	−0.003211	−0.004277	−0.002739	−0.004488 *
	(−1.1843)	(−0.9528)	(−1.5965)	(−1.0224)	(−1.6739)
BMeet	0.001398	0.001240	−0.001596	−0.002037	−0.001816
	(0.5905)	(0.5239)	(−1.0811)	(−1.3888)	(−1.2246)
BMAfill	0.010910	0.009844	0.035003 ***	0.033216 ***	0.035595 ***
	(0.8149)	(0.7355)	(3.4401)	(3.2901)	(3.4974)
CEODual	−0.038560 **	−0.031979 **
	(−2.5095)	(−2.0171)
lnBMComp	0.007731	0.004756	0.006527	0.000563	0.007277
	(0.7055)	(0.4289)	(0.8382)	(0.0718)	(0.9329)
BER	−0.005730	−0.003248
	(−0.3366)	(−0.1903)
BSP	−0.365875 ***	−0.350347 ***
	(−4.1954)	(−4.0000)
PBD	0.025719	0.026907
	(1.0872)	(1.1386)
ChairInd			0.011971	0.019985	0.009699
			(0.9234)	(1.5397)	(0.7432)
CEOBM			−0.057170 ***	−0.046439 ***	−0.057342 ***
			(−4.4256)	(−3.5626)	(−4.4412)
CEOCompRisk			0.003932	0.008904	0.005426
			(0.3214)	(0.7310)	(0.4422)
CSRCom			−0.068605 ***	−0.062807 ***	−0.070571 ***
			(−3.0727)	(−2.8324)	(−3.1568)
lnAge	0.005311	0.003568	0.001728	0.001965	0.001461
	(0.5995)	(0.4005)	(0.2764)	(0.3170)	(0.2336)
ART	0.005472 ***	0.005582 ***	0.007749 ***	0.007370 ***	0.007881 ***
	(2.6533)	(2.7092)	(4.1810)	(4.0067)	(4.2496)
lnFCF	−0.022706 ***	−0.022099 ***
	(−3.4427)	(−3.3505)
LevDE	−0.075665 ***	−0.074898 ***	−0.028087 ***	−0.023844 ***	−0.027120 ***
	(−6.9514)	(−6.8850)	(−3.2179)	(−2.7387)	(−3.1000)
ROE	−0.046653	−0.057280	−0.116918 *	−0.130469 **	−0.117757 *
	(−0.5899)	(−0.7230)	(−1.8735)	(−2.1062)	(−1.8879)
lnTA			0.100129 ***	0.097659 ***	0.098383 ***
			(8.3872)	(8.2420)	(8.2049)
lnSALES			−0.146728 ***	−0.145729 ***	−0.146433 ***
			(−12.9116)	(−12.9322)	(−12.8906)
_cons	0.855638 ***	0.914961 ***	1.264544 ***	1.451109 ***	1.494510 ***
	(2.9716)	(3.1577)	(6.2084)	(7.0340)	(7.6057)
F statistic			23.848604 ***	24.000746 ***	22.668074 ***
Wald Chi2	136.98	140.1
R-sq overall	0.1937	0.1976	0.2066	0.2154	0.2076
Obs	586	586	1092	1092	1092
N Sectors	10	10	10	10	10
FE/RE	RE	RE	FE	FE	FE
Chi2	0.1281	0.0518	0.0000	0.0000	0.0000

Source: Own estimates. t statistics in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

present the estimates of corporate outcomes using regression models with fixed effects and random effects used for robustness checks to estimate companies’ financial performance and risk, while controlling for industry fixed effects. While the sample selection is sourced from the Stoxx600 index, which, unlike the tech-heavy S&P 500, is widely diversified, the authors proposed in this study to estimate the performance and risk, accounting for this sectoral heterogeneity and acknowledging that the two-way effects models (including firm and year) might complement and strengthen the empirical results. These fixed effects models were constructed to test the robustness of the initial OLS estimates, complementing the econometric techniques that implied using alternative specifications for performance (ROA, ROE) and risk (sdROA, LevDE, sdQR) to confirm the stability of the results. The selection of the appropriate model is based on the Hausman test. If the test result was lower than 0.05, the fixed effects model was selected.

The results show that the industry fixed effects are consistent with the baseline OLS regression, indicating that our results are robust to alternative specifications and consistent across models. Most variables retain the same statistical significance as in the models without effects, suggesting that the relationships identified between the independent and dependent variables do not depend on the model specification used and are not sensitive to the time-invariant sectoral attributes.

The industry fixed effects regressions confirm the consistency of the conclusions, indicating that the observed effects are not random outcomes (presumably sectoral-specific), but rather reflect consistent relationships between corporate outcomes related to performance and risk and governance mechanisms related to independence, including board diversity related to structure, expertise and policies, while controlling for various firms’ characteristics.

After controlling sectoral fixed effects, we see that the overall results estimating ROA using a regression model with effects remain consistent with the OLS model, as presented in Table 10.

The results using the fixed effects method are consistent with OLS specification, indicating the presence of a nonlinear relationship between governance independence and ROA, estimating the turning point at the level of about 60% for board independence (versus 50% when using pooled OLS) and 67.5% for audit committee (versus 60% when using pooled OLS). Starting with these thresholds, the influence of board and audit committees’ independence improves ROA, offering support for H6 and H7.

Most of the significance levels and signs of the coefficients are preserved when controlling for sectoral fixed effects, suggesting that a minor influence is attributable to industry peculiarities in our sample, supporting our baseline results.

In Table 11, most of the results for ROE using the fixed effects method remain consistent with the baseline models without effects.

After controlling sectoral time invariant attributes, the results show that nonlinearity between board independence and ROE is not consistent across models, suggesting that initial effects on ROE are between-sector effects rather than within-firm mechanisms. However, the results for audit committee independence are consistent across models, supporting the baseline results, and providing empirical evidence of the existence of a turning point at the level of 72.4% (versus 65% before controlling for industry fixed effects). This shows that audit committee independence significantly improves capital efficiency (ROE). These results related to the nexus between independence mechanisms and ROE offer evidence for H7.

Table 12 presents the fixed effects estimation results for the volatility of returns (sdROA), which are largely similar to the baseline estimates using pooled OLS regression. Overall, governance independence mechanisms (board, audit, nomination) exhibit a negative influence on return volatility, highlighting the nonlinearity in the case of board independence, while confirming the negative influence of board independence on return volatility and the limited impact of diminishing marginal effect due to the higher threshold level of the turning point. These results offer support for H6 and H7.

Table 13 presents regression results for financial leverage (LevDE) after controlling for industry fixed effects.

Financial leverage estimates using fixed effects models are consistent with the baseline OLS results, confirming the audit committee independence nonlinearity present at the level of 56% (compared to 55% identified before when controlling for industry fixed effects), suggesting that stricter monitoring is likely to reduce indebtedness above this threshold level. Nomination committee independence exhibits a negative influence on indebtedness, and the results are consistent across OLS and fixed effects models, indicating that nomination committee independence is a powerful governance mechanism in reducing excessive indebtedness. These results might be explained by the selection procedures this committee undertakes, which support the promotion of prudent financing policies when higher levels of independence occur.

Table 14 presents regression results for short-term liquidity volatility (sdQR) after controlling for industry fixed effects.

The estimation results using the fixed effects model show consistency with the baseline OLS estimations. In the case of board independence, the results show that the turning point is present at the level of 61% (p < 0.1) in our sample, starting when the relationship turns to positive, meaning a diminishing marginal effect of board independence on liquidity volatility above this threshold, supporting H6. Similarly, audit committee independence holds its negative and nonlinear relation with liquidity volatility (p < 0.01), showing a turning point at the level of 73.6% (versus 68.75% when using OLS models), offering support for H7. These results are consistent with prior literature related to excessive monitoring arguments, according to which some may face limited marginal improvement of risk outcomes beyond a threshold, according to Zubeltzu-Jaka et al. (2019) and Fuente et al. (2017). Results for short-term liquidity underline also that independence of the nomination committee holds its positive influence on liquidity volatility in industry fixed effects models, which contradicts H7. Taken together, these results suggest that a larger number of independent directors in the nomination committee is associated with lower volatility of returns (see Table 12), but also with higher volatility of short-term liquidity. These results suggest that the influence of nomination committee independence, as a governance mechanism, on corporate risk is risk dependent.

Table 15. Summary of results and statistical significance of predictors coefficients based on pooled OLS regression.

Variables	ROA	ROE	sdROA	LevDE	sdQR
IndBM	− **	− *	− *		− *
sqIndBM	+ ***	+ *
CAuditInd		− **	− ***	+ **	− ***
sqCAuditInd		+ **		− **	+ ***
CNomInd					+ ***
c_CNomInd			− **	− **	+ **
c_sqCNomInd			− **	− **
BSize	− ***	− ***	− *		− **
BTenure	+ ***		− ***		− **
BMeet	− **			− ***	− *
BMAfill		+ ***	+ ***		+ ***
CEODual			− **	− *	− **
lnBMComp		+ ***		− ***
BER
BSP	+ *	+ **			− ***
PBD			+ **
ChairInd		+ *	+ **
CEOBM	+ *	+ ***		− **	− ***
CEOCompRisk	+ **		+ ***	+ **
CSRCom					− ***
lnAge	+ ***	+ **
ART	+ ***	+ ***	+ ***		+ ***
lnFCF	+ ***	+ **			− ***
LevDE	− ***	+ ***	− **		− ***
ROA				− ***
ROE					− *
QR	+ ***	+ **	+ ***	− ***
lnTA	− ***	− ***		+ ***	+ ***
lnSales	+ ***	+ ***	− ***		− ***

Source: Own estimates. * p < 0.1, ** p < 0.05, *** p < 0.01. “+” indicates a positive influence and “‒” indicates a negative influence.

presents more clearly, as a straightforward conclusion of the findings, a summary of the signs and statistical significance of the relationship between our predictors and corporate outcomes, proxied by performance variables (ROA and ROE) and risk outcome accounting-based measures (sdROA, LevDE, sdQR), using the OLS baseline model.

5. Conclusions

This study investigates the impact of board and committee independence on corporate performance and risk in continental European governance regimes, using a sample of Western European countries over 10 years between 2015 and 2024. Using panel data techniques and econometric analysis, this study provides empirical evidence on how governance independence mechanisms, as well as board structure, oversight mechanisms and board policies, shape firm outcomes.

The nonlinear estimations reveal a U-shaped relationship between board independence and performance, indicating the existence of a threshold of a minimum independence level of 50% for a meaningful influence. Beyond this point, the benefits of stronger monitoring and reduced agency conflicts outweigh the coordination and information costs associated with independent boards. This finding reconciles conflicting evidence in the literature and underscores the importance of considering possible nonlinearities when evaluating governance mechanisms. Board independence is likely to reduce return volatility, and the relation is linear. Relevant to this study is that the influence of board independence on reducing short term liquidity volatility is noticed right from the initial levels of independence, while the marginal effect is less relevant as the threshold rises above 60%.

Specialized board committees emerge as key determinants of both performance and risk. The influence of audit committee independence on corporate outcome is nonlinear, and is associated with higher performance if beyond a 72% threshold and likely to lower indebtedness risks if beyond 55%. Depending on the corporate risk, we find that audit independence consistently preserves return volatility, as its influence is linear and negative across models. Likewise, its influence diminishes liquidity volatility if the independence level is below about 70%, while above this level, the audit independence mechanism less effectively stabilizes short-term liquidity.

The contribution of nomination committee independence appears to be risk dependent. However, the influence of nomination committees is noticeable when considered for risk-reduction strategies. Thus, we find that nomination committee independence contributes to reducing return volatility and the results are consistent across specifications. These results suggest that selection strategies may lead to more prudent financial policies, highlighting the nomination committee role in ensuring board leadership and long-term stability and consistency of returns. Similarly, nomination committee independence is likely to prevent excessive indebtedness, since its influence is linear and negative across specifications. Markedly for the contribution of this study, nomination committee independence is likely to contribute to short-term liquidity volatility, as selection procedure may lead to candidates willing to sacrifice short-term liquidity stability for consistent returns, as our findings suggest.

The study results emphasize that board and committee-level governance independence mechanisms play an outstanding role in shaping performance and risk oversight.

This study provides empirical results suggesting that larger boards are associated with reduced operational efficiency, and on the contrary preserve the stability of returns, thus reducing associated risks. CEO involvement on the board also contributes positively to performance and preserves stability of short-term liquidity, supporting the view that managerial insight and inside-firm knowledge can complement the board’s monitoring role. CEO duality consistently contributes to reducing outcome volatility and indebtedness.

Board tenure is consistently associated with higher performance and lower volatility of firms’ outcomes, benefits that might be considered for a growing strategy by larger and complex companies. Affiliations are positively associated with firm performance, suggesting that experience and external professional networks enhance decision-making quality and resource access. Conversely, excessive affiliation may limit the oversight capabilities of board members, leading to higher volatility of corporate outcomes.

Taking the results together, the findings indicate that independent boards and committee oversight contribute nonlinearly to improving performance and reducing indebtedness, supporting critical mass arguments, while consistently reducing outcome volatility, supporting the marginal effects arguments due to excess independent oversight. This evidence confirms the trade-off between performance enhancement and risk management inherent in corporate governance design.

Overall, this study demonstrates that corporate governance effectiveness depends not only on the presence of independent directors in board or specialized committees, but also on the benchmarks for intensity. The results suggest that governance codes should avoid rigid independence requirements and instead promote flexible frameworks that account for firms’ strategic objectives, firms’ characteristics, governance regimes and institutional contexts. For practitioners, the findings highlight the importance of balancing advisories with monitoring roles or board policies with strategic roles of the CEO to achieve sustainable corporate performance.

Despite its contributions, this study is subject to several limitations. First, the analysis is restricted to large listed firms from Western Europe, which may limit the generalizability of the findings to other institutional settings. Second, some governance mechanisms are captured through available disclosure-based proxies, and from the firms that reported, which may limit the observations to this sub-sample. Nevertheless, the presence of dynamic endogeneity remains an inherent limitation and motivates future research explicitly designed to identify causal mechanisms.

Future research could extend this analysis by incorporating ownership structure, more ESG dimensions, or cross-country comparisons between continental European and Anglo-Saxon governance systems, thereby further enriching the understanding of governance–corporate outcome dynamics.