Board Structure and Firm Performance: The Moderating Role of National Governance Quality

Abstract

This study empirically investigates the relationship between board composition, focusing on board size and board independence, and firm performance. It further examines how national governance quality moderates this relationship. Using a panel dataset of 1604 firms from 41 developed and emerging economies, the study employs pooled ordinary least squares (OLS) as the baseline regression method, alongside two-stage instrumental variable regression and system generalised method of moments (GMM) to address potential endogeneity concerns. Firm performance is measured using return on equity (ROE) and Tobin’s Q. Board size is captured by the number of directors on the board, while board independence is measured by the proportion of non-executive directors. The findings indicate that while board size and independence are positively associated with firm performance, the strength of these relationships weakens in countries with high governance quality. Our findings remain robust after controlling for dynamic endogeneity and unobserved time-invariant heterogeneity inherent in the corporate governance–performance nexus.

1. Introduction

The board of directors is considered a fundamental element of corporate governance, tasked with the dual roles of overseeing managerial conduct and offering strategic guidance (Githaiga et al., 2022). Consequently, it has garnered significant academic interest, especially about its structure and performance (Jabbouri & Almustafa, 2021; Uribe-Bohorquez et al., 2018). Among the principal structural characteristics of the board, the size of the board and the independence of its members have surfaced as prominent topics in discussions of governance. Established perspectives, bolstered by agency theory and a significant number of empirical findings, assert that smaller boards exhibit decisiveness and oversight (Guest, 2009), while larger boards face coordination challenges and diluted accountability (Jensen, 1993; Altass, 2022). Concurrently, the emergence of the norm advocating for board independence, alongside the prioritisation of a majority of non-executive, independent directors, has become a global standard in governance codes (MacNeil & Esser, 2022). Independent boards are regarded as essential protection against corporate entrenchment and the expropriation of resources by powerful shareholders, facilitating the more impartial evaluation of strategic decisions (Zattoni & Cuomo, 2010).

However, these assumptions are increasingly being challenged by scholars who emphasise the role of national governance quality (D. Li et al., 2006). In countries with weak legal institutions, inadequate regulatory oversight, and pervasive corruption, even well-designed board structures may serve merely as symbolic compliance mechanisms with limited practical effect (Sharma et al., 2024; Raza et al., 2020). In environments characterised by weak governance, high transaction costs, disparities in information, and self-serving behaviours diminish the effectiveness of internal governance mechanisms (North, 1991). Insufficient governance structures hinder investment by restricting enforceable property rights and diminishing the dependability of market transactions. On the other hand, high-governance settings are characterised by robust legal enforcement, regulatory transparency, and diminished information costs, which mitigate opportunism and augment enterprises’ capacity to utilise governance frameworks for enhanced performance (Chari & Banalieva, 2015). Formal institutions and qualities such as the rule of law, political stability, regulatory quality, and government effectiveness significantly moderate the effects of board structures, thereby enhancing firm performance (Oliveira et al., 2024).

Despite extensive research on board size and independence (e.g., Uribe-Bohorquez et al., 2018; Hu et al., 2023; Qadorah & Fadzil, 2018), the moderating influence of national governance quality remains underexplored. This limits our understanding of how institutional environments shape the effectiveness of board structures (P. Kumar & Zattoni, 2013). Although previous studies have thoroughly explored the separate impacts of board size and board independence on firm outcomes (e.g., Uribe-Bohorquez et al., 2018; Al-Saidi, 2021; Hu et al., 2023; Qadorah & Fadzil, 2018), many have overlooked the potential moderating role of national governance quality in these relationships. This oversight constrains our understanding of the interaction between board membership and institutional contexts in influencing business performance (P. Kumar & Zattoni, 2013). Although recent studies have started to investigate institutional moderators, including the research by Atugeba and Acquah-Sam (2024); Sharma et al. (2024) on corporate governance and national institutions, T. Nguyen et al. (2021) on board gender diversity; and Jabbouri and Almustafa (2021) on cash holdings, few have focused specifically on board composition, particularly board size and independence across varying governance contexts. Moreover, there is a lack of comprehensive cross-country comparative studies that assess whether these dynamics differ systematically between developed and emerging markets.

This study addresses these gaps by examining how board size and board independence influence firm performance and how national governance quality moderates these relationships. Drawing on institutional theory, we argue that formal governance mechanisms such as the rule of law, regulatory quality, governmental effectiveness, political stability, control of corruption, and accountability can either enhance or diminish the efficacy of board structures. Our analysis is based on a dataset of 1604 firms from 41 developed and emerging markets. The sample includes countries selected for the availability of firm-level and governance data, ensuring comparability and relevance. This cross-national diversity enables an evaluation of institutional differences and their effects on corporate governance. Our findings show that although larger boards typically are associated with better firm performance, the positive effect lessens in nations with robust governance quality. In a similar vein, the advantages of board independence are favourable yet diminish as the quality of national governance improves, indicating a substitution effect between external institutional enforcement and internal governance structure. The findings also indicate that improving national governance quality in emerging nations can substantially enhance the beneficial impacts of board structures on corporate performance. To ensure robustness, we address endogeneity using instrumental variable two-stage least squares (IV-2SLS) and dynamic system GMM estimation.

This study makes several key contributions. First, it extends institutional theory by demonstrating how national governance quality interacts with board characteristics to affect firm outcomes. Second, by integrating a diverse cross-national sample of firms from both developed and emerging economies, we enhance the existing literature on comparative corporate governance. We undertake a detailed analysis that contrasts the moderating influences of national governance quality between emerging and non-emerging markets, uncovering significant institutional disparities. Finally, our findings have practical implications for policymakers, investors, and boards aiming to customise governance methods to their own institutional circumstances. The remainder of the paper is structured as follows: Section 2 reviews the relevant literature and outlines our hypotheses. Section 3 details the methodology, followed by results in Section 4 and the conclusion and policy implications in Section 5.

2. Literature Review and Hypothesis Development

This section explores two key dimensions of the study: the first is the impact of board composition, specifically board size and board independence on firm performance, and the second is the moderating role of national governance quality in shaping the relationship between board composition and firm performance.

2.1. Impact of Board Structure on Firm Performance

The size of a board is a crucial aspect of corporate governance structure, supported by a substantial yet unresolved body of literature discussing its influence on company performance. From the perspectives of resource dependence and stakeholder theory, larger boards are considered beneficial as they typically include a wider range of skills, experiences, and representation of stakeholders. The presence of diversity is thought to enrich board discussions and elevate strategic decision-making, particularly in intricate or evolving environments (Mishra, 2023; S. Kumar et al., 2022). Larger boards might possess a greater capacity to obtain essential external resources, enhance legitimacy, and provide specialised guidance (Rashid, 2018).

However, agency theory presents a more measured perspective. It recognises that a larger board could potentially lower agency costs by enhancing managerial oversight, yet it also cautions against issues such as coordination challenges, diffusion of responsibility, and diminished oversight in boards that are excessively large (Jensen, 1993; Dalton et al., 2007). Indeed, several empirical studies provide evidence that supports these concerns. In a study conducted by Guest (2009) involving a substantial panel of UK firms, a significant negative correlation was found between board size and firm performance, as assessed through profitability, Tobin’s Q, and share returns, particularly in the context of larger firms. P. Nguyen et al. (2016) also observe that in Australia, larger boards correlate with diminished business value, increased costs, and CEO compensation more closely tied to firm size rather than performance, indicating inefficiencies and inefficient incentives. Yan et al. (2021) with Altass (2022) provide corroborative evidence of an inverse correlation between board size and business performance in the United States and other developed markets.

In contrast, results from emerging markets add complexity to this narrative. Investigations conducted in Indonesia (Fahlevi et al., 2023), China (Zhu et al., 2022), and East Africa (Githaiga et al., 2022) demonstrate positive associations between board size and firm value, earnings management, or sustainability performance. The presence of larger boards could serve as a compensatory function by improving legitimacy, facilitating access to external resources, and reinforcing managerial discipline. Rashid (2018) illustrates that in these contexts; larger boards may effectively mitigate agency costs by broadening the distribution of monitoring responsibilities. Collectively, these findings underscore certain aspects of the effects of board size. Potharla and Amirishetty (2021) suggest that the relationship could be non-linear, indicating that moderate increases in board size may lead to positive governance outcomes until a certain threshold is reached, after which the drawbacks of scale may start to surpass the advantages. Considering this detailed evidence, the current study proposes the following hypothesis:

H1. 

There is a positive relationship between board size and firm performance.

Board independence is defined as the proportion of non-executive, outside directors. It represents another fundamental principle of efficient governance, especially within corporate governance frameworks. Independent directors are viewed as objective overseers who focus on the interests of shareholders and mitigate the potential for managerial self-serving behaviour. Consequently, regulatory authorities worldwide have incorporated board independence into governance codes, grounded in the conviction that it improves organisational accountability and transparency (Hillman & Dalziel, 2003).

The agency theory emphasises the importance of an independent board for effective management supervision and evaluation (Arora & Soni, 2023). This theory posits that the board must comprise a significant number of independent members who supervise management to aid shareholders in preserving ownership by mitigating agency problems. Fama and Jensen (1983) have proposed the notion of maintaining most non-executive directors, as they are likely to perform their roles as overseers with greater efficiency. The presence of independent directors on corporate boards is further substantiated by the resource dependency theory framework, given their access to external resources (Hillman et al., 2000). Conversely, stewardship theory posits that insiders, rather than outsiders, are more effectively aligned with the objectives of the organisation. Stewardship theorists contend that managerial behaviour is fundamentally aligned with organisational interests, suggesting that an excessive focus on independence could hinder cohesion and diminish the flow of information, particularly when boards are devoid of business environment-specific ideas (Donaldson & Davis, 1991). Shleifer and Vishny (1997) posited that managers who generate favourable financial outcomes for investors build a solid track record, enabling them to re-enter financial markets for the firm’s future requirements. From this viewpoint, enhanced firm performance was associated with a predominance of inside directors on the board, as they possess a greater knowledge of the business, are more adept at governance than outside directors, and consequently can make more informed decisions.

Nevertheless, existing empirical literature offers a variety of perspectives on this viewpoint. Several studies indicate that a greater degree of board independence corresponds positively with firm performance (Liu et al., 2015; Uribe-Bohorquez et al., 2018; Al-Saidi, 2021; Hu et al., 2023). Furthermore, certain empirical studies, including those by Rashid (2018), Mishra (2023) and Martín and Herrero (2018), documented an insignificant relationship between board independence and firm performance. It remains uncertain whether the board members are not operating autonomously or if the absence of performance enhancement stems from insufficient board independence. Both Al-Gamrh et al. (2020) and Khan et al. (2024) also identified a negative link between board independence and business growth. The divergent viewpoints indicate that board independence may not have a consistent effect in different scenarios. The effectiveness may be contingent upon firm size, industry complexity, structure of ownership, or the overall institutional context. Nevertheless, considering the substantial evidence indicating performance-enhancing effects under specific settings, we propose the following:

H2. 

There is a positive association between board independence and firm performance.

2.2. Board Structure and Firm Performance: The Role of National Governance Quality

Institutional theory offers a compelling perspective for analysing the ways in which wider societal and regulatory environments shape organisational conduct and governance frameworks. It suggests that organisations are not isolated entities; rather, they exist within an institutional context that includes both formal mechanisms like regulations, laws, and judiciary systems, as well as informal components such as norms, culture, and shared beliefs (Scott, 1995; North, 1991). Specifically, institutional theory indicates that the effects of corporate governance mechanisms on performance vary by country and are influenced by the quality of the national institutional context (Filatotchev et al., 2013).

Firms act as economic units that react to the regulations and conditions set by the environment in which they operate. A practice deemed efficient or legitimate in one national setting may be perceived as ineffectual or illegitimate in another. Consequently, the quality of national governance, defined as the degree to which public institutions preserve the rule of law, enforce contracts, mitigate corruption, and ensure regulatory openness, serves as a vital mediator in the links between governance and performance (T. Nguyen et al., 2015). Robust governance frameworks mitigate knowledge asymmetries and agency dilemmas by guaranteeing legal enforcement, safeguarding investor rights, and bolstering the credibility of managerial oversight systems (Aslan & Kumar, 2012). On the other hand, in nations marked by fragile institutions, significant corruption, or unclear regulations, even thoughtfully crafted corporate governance frameworks like independent boards or significant large boards can become ineffective (Kaufmann et al., 2011; Bello et al., 2020). Corruption distorts resource allocation and raises the cost of doing business while also weakening the enforcement of formal governance mechanisms. In these contexts, companies might find themselves driven to pursue rent-seeking activities, which can undermine internal controls and diminish investor protection.

This institutional variation is effectively demonstrated in recent empirical studies. Zattoni et al. (2017) contend that although board independence might have a restricted direct impact on performance, its efficacy is notably amplified in nations characterised by high institutional quality. In a similar vein, Jabbouri and Almustafa (2021) analysed data from twelve MENA countries and discovered that cash holdings have a positive impact on firm value, particularly in environments with robust institutional frameworks. Their findings suggest that national governance enhances internal governance processes by reducing agency conflicts. Similarly, Atugeba and Acquah-Sam (2024) indicate that adherence to national governance norms mitigates the typically adverse impacts of corporate governance actions on company efficiency, reinforcing the perspective that the institutional context influences governance results.

This study extends the analysis to the connection between board structure, specifically board size and board independence, and business performance. We contend that the quality of national governance acts as a moderating factor that either amplifies or diminishes the impact of board composition on organisational results. In nations with robust governance quality, formal institutional processes like judicial enforcement and regulatory supervision can augment board responsibilities, thereby improving their performance.

H3a. 

In countries with low national governance quality, the positive association between board size and firm performance is stronger.

H3b. 

In countries with high national governance quality, the positive association between board size and firm performance is weaker.

On the other hand, in weak institutional settings, similar board characteristics might not lead to performance improvements because of the lack of strong external support systems and dependable legal enforcement. In these environments, investors, including both shareholders and bondholders, face heightened risks of agency conflicts, especially in situations where legal protections are weak and enforcement mechanisms lack effectiveness (La Porta et al., 2002). This enhances dependence on internal governance structures like board independence to protect investor interests. Moreover, the characteristics of ownership concentration and the identities of predominant shareholders such as families, states, or business groups exhibit considerable variation across different institutional settings, influencing the nature and intensity of agency issues (Aguilera & Crespi-Cladera, 2015).

H4a. 

In countries with low national governance quality, the positive association between board independence and firm performance is stronger.

H4b. 

In countries with high national governance quality, the positive association between board independence and firm performance is weaker.

In emerging markets, the presence of institutional voids obstructs the efficient operation of markets, leading to increased transaction costs (Khanna & Palepu, 2000). Firms in these environments face differing levels of institutional frictions due to variations in transaction and transformation costs, affecting their competitiveness and profitability (Ngobo & Fouda, 2012). Wan (2005) contends that companies in emerging economies may establish institutional obstacles to entry to maintain enduring competitive advantages. However, improvements in the quality of national governance may subject these companies to increased rivalry in the product market, thereby amplifying external discipline.

Conversely, developed economies are generally defined by robust governance systems and advanced financial markets (La Porta et al., 2002). Institutional investors in these markets not only choose firms with strong internal governance but also actively participate in improving governance standards (Aggarwal et al., 2011; Chung & Zhang, 2011; Harford et al., 2018). As a result, robust national institutions in developed nations could act as alternatives to internal governance mechanisms, including board oversight. Consequently, developed nations can be seen as institutional frameworks that provide an extra dimension of governance control. This stance corresponds with our empirical evidence, indicating that the adverse moderating influence of national governance quality on the association between board structure and firm performance is more significant in developed economies. Based on these differences, we hypothesise that the following are true:

H5a. 

A positive moderating effect of national governance quality on the board structure–performance relationship is stronger in emerging markets.

H5b. 

A negative moderating effect of national governance quality on the board structure–performance relationship is stronger in developed markets.

3. Materials and Methods

3.1. Sample and Data Sources

The research employed data from an initial sample of 1650 publicly traded companies spanning 45 nations. Upon removing enterprises with incomplete data, the final sample consisted of 1604 firms from 41 countries, covering the period from 2013 to 2022. Financial data at the firm level were obtained from the Bloomberg Terminal, whereas national governance indicators were derived from the Worldwide Governance Indicators (WGI) project established by Kaufmann et al. (2011). The WGI dataset provides yearly assessments of governance quality across six key dimensions: voice and accountability, political stability, control of corruption, government effectiveness, regulatory quality, and rule of law. The sample construction procedure commenced with the extraction of firm-level data from Bloomberg, encompassing essential financial performance and governance factors. The first dataset produced 19,011 observations for Tobin’s Q, 19,246 observations for both ROE and ROA, and 18,914 observations for board size. The firm-level data were subsequently integrated with macro-level governance indicators sourced from the WGI, yielding around 17,643 matched firm-year observations.

To maintain consistency in all empirical estimations and descriptive statistics, we employed listwise deletion, preserving only those firm-year observations having comprehensive data for all variables included in the analysis. This procedure resulted in a final balanced panel comprising 14,857 firm-year observations, which served as the final sample data for all statistical analyses conducted in this paper. A step-by-step breakdown of the sample construction process is presented in Appendix A

Table A1. Sample construction.

Step	Description	Observations Remaining
1	Initial financial data (ROE, ROA, MTB, FRSZ, etc.)	19,246
2	Board structure data (e.g., board size, PCWB, PIND)	18,914–18,609
3	Governance indicators merged (from WGI)	~17,643
4	Observations with missing data removed (listwise deletion)	14,857
5	Final sample used for all analyses	14,857

. Although it is standard to exclude companies in the financial and utilities sectors because of their unique regulatory frameworks and capital structures, we faced challenges in systematically removing these firms from the dataset due to the lack of industry classification codes (e.g., SIC or NAICS). This limitation is clearly recognised, and we suggest that subsequent studies with access to more comprehensive industry identifiers explore this issue for improved sample accuracy.

3.2. Variables Measurement

Following previous research, firm performance is assessed using Tobin’s Q, Return on Assets (ROA), and Return on Equity (ROE) (Almustafa et al., 2023). Tobin’s Q is a market-based indicator employed by investors to assess their views on market dynamics and corporate governance; it is defined as the ratio of a company’s market value to the replacement cost of its assets. On the other hand, ROE is a performance metric derived from accounting data. Return on Equity (ROE) is calculated by dividing net income by total equity. Improvements in ROE indicate a more efficient distribution of resources in generating profits.

In line with Potharla and Amirishetty (2021), board size is defined as the total number of directors on a firm’s board. According to agency theory, smaller boards are typically viewed as more efficient and easier to coordinate, which improves monitoring efficiency and favourably impacts business performance (Yermack, 1996). Conversely, resource dependence theory asserts that larger boards provide a wider array of talents, experience, and external networks, which can serve as significant resources that improve organisational performance (Dalton et al., 2007). Notwithstanding these theoretical viewpoints, empirical evidence regarding the connection between board size and business performance remains ambiguous, with research yielding both positive and negative associations, reflecting a lack of consensus in the literature.

Board independence is seen as a vital component of corporate governance, especially in resolving the principal-agent problem. Fama and Jensen (1983) assert that an increased ratio of independent directors enhances the board’s supervisory function. This consequently restricts management opportunism and aligns the interests of managers with those of shareholders. Agency theory suggests that a more autonomous board is likely to positively impact corporate performance. This study defines board independence as the percentage of independent directors on the board.

CEO duality is measured using a binary variable, assigned a value of 1 when the CEO concurrently holds the position of board chairperson, and 0 in all other instances. Scholars suggest that having dual leadership can promote quick decision-making and a cohesive plan of action, potentially benefiting organisations in an unpredictable business setting, thereby improving overall performance (Sahoo et al., 2023; Saleh et al., 2024; Chiu et al., 2021). Stewardship theory asserts that the fusion of the CEO and board chair positions (CEO duality) promotes a unified and cohesive leadership structure at the highest level of the business (Donaldson & Davis, 1991). This centralised leadership can diminish uncertainty among employees, managers, and stakeholders by explicitly defining who possesses ultimate power, hence facilitating swifter and more efficient decision-making (Finkelstein & D’Aveni, 1994). On the other hand, Agency theory argues that having power centralised in one individual weakens the board’s capacity to properly oversee management, which could result in agency issues and a decrease in firm value (Duru et al., 2016; F. Li et al., 2016; Bristy et al., 2022).

Agency theory (Jensen & Meckling, 1976) and resource dependence theory (Pfeffer, 1973) suggest that diversity within boards can enhance firm performance by expanding access to essential resources and varied viewpoints. Nevertheless, the empirical evidence regarding this relationship is still not definitive, primarily because of differences in model specifications, measurement techniques, and contextual variations among the studies conducted. This study defines board diversity in terms of the percentage of female directors present on the board.

Board tenure is measured as the average duration of the tenure of independent directors on the firm’s board. The empirical literature about board tenure presents inconclusive results. For instance, both Huang and Hilary (2018); Bonini et al. (2015) contend that prolonged employment allows directors to comprehend the firm’s operational and financial details more effectively, thereby enhancing their advisory capacity. Conversely, Livnat et al. (2021) posits that, over time, long-tenured board members may cultivate intimate connections with management, resulting in more alignment with managerial interests and diminished propensity to contest executive decisions. The erosion of objectivity may impair board oversight and reduce business value.

Firm size is measured as the natural logarithm of total assets. The association between firm size and firm performance is still ambiguous in the literature. Larger enterprises may gain from economies of scale, increased access to financing, and augmented market dominance, resulting in better performance. Conversely, exceeding a specific threshold in size may result in bureaucratic inefficiencies, coordination difficulties, and protracted decision-making that might impair performance. This duality is evident in the conflicting empirical results: Palaniappan (2017) identifies a positive correlation between firm size and performance, yet studies like Manna et al. (2016) reveal a detrimental effect. The Market-to-Book ratio (MTB) serves as an indicator of growth prospects and is computed by subtracting book equity from total assets, adding market equity, and then dividing by total assets (Simoens & Vander Vennet, 2021). MTB increases when a corporation possesses important intangible assets, including market dominance, goodwill, a portfolio of patents, proficient management, and lucrative investment prospects (Tejerina-Gaite & Fernández-Temprano, 2021).

3.3. Measurement of Governance Quality

We employ the Worldwide government Indicators (WGI) created by the World Bank to evaluate the quality of national governance. The WGI is a composite indicator derived from a weighted average of six governance dimensions: government effectiveness, voice and accountability, political stability, rule of law, lack of violence/terrorism, regulatory quality, and control of corruption. These variables are quantified in standardised units, spanning from −2.5 to +2.5, with elevated values signifying enhanced governance performance. However, some studies argued that these six components are often significantly correlated, which raises issues regarding multicollinearity when they are included together in regression models (Dang et al., 2022). Consequently, depending on a mere average score may induce estimation bias or obscure the unique impacts of individual governance characteristics (Almustafa et al., 2023). As a result, we adopted Principal Component Analysis (PCA) in order to develop a unified governance quality index. This allowed us to overcome issues regarding multicollinearity and capture the underlying structure of the governance indicators.

Table 1. Description and measurements of research variables.

Name of Variable	Acronym	Definition	Sources
Outcome variables
Tobin q ratio	TOBIN Q	Tobin’s Q is calculated as the sum of the market value of equity, total liabilities, preferred equity, and minority interest, divided by the value of total assets.	Bloomberg
Return on Equity	ROE	ROE serves as a financial performance measure that evaluates a company’s capacity to make profit relative to its shareholders’ equity. It is measured by dividing net income by the equity held by shareholders.	Bloomberg
Main Independent variables
Board size	BDSZ	Board size is the total number of directors	Bloomberg
Board Independence	PIND	This is measured as percentage of independent directors on the board of selected firms.	Bloomberg
Governance quality index	GOVQ	A composite governance index constructed using Principal Component Analysis (PCA) based on six governance indicators from World Bank	World Bank—World Development Indicators
Control variables
Board Average tenure	BDAT	This is the firm average board tenure.	Bloomberg
CEO duality	CEO	A dummy variable that takes the value of 1 if the chairperson is also the CEO and zero otherwise	Bloomberg
Board diversity	PCWB	This is measured as the percentage of female directors on the board.	Bloomberg
Market to book value	MTB	The ratio of a firm’s accounting-based equity value to its market-based equity value.	Bloomberg
Firm size	FRSZ	Firm size is the natural log of the total assets of the firm	Bloomberg

provides a description and measurement details of the research variables used in this study

3.4. Empirical Models

The empirical models employed to evaluate the hypotheses of the study are outlined in Equations (1)–(4). The principal independent variables of interest are board size and board independence. Equation (1) specifies the relationship between board size and firm performance, while Equation (2) examines the effect of board independence on firm performance. To formally test these hypothesised relationships, we specify the following equations.

$$Y_{it}={\beta }_0+{\beta }_{1}BDSZ+{\sum }_{m=1}^M\mathrm{\mu }m{Controls}_{mit}+{\mathrm{ɲ}}_i+{\lambda }_i+{\mathrm{Ɛ}}_{it}$$

(1)

$$Y_{it}={\beta }_0+{\beta }_{1}PIND+{\sum }_{m=1}^M\mathrm{\mu }m{Controls}_{mit}+{\eta }_i+{\lambda }_i+{\mathrm{Ɛ}}_{it}$$

(2)

Here Y_it denotes the firm’s financial performance as proxied by ROE and Tobin. Q Moreover, “i” represents the firm (ranging from 1 to 1604), while “t” represents the year within the sample period (2013–2022), ${\mathsf{\eta }}_{\mathrm{i}}$, represents firm-specific fixed effects, ${\lambda }_i$ represents time fixed effects while Ɛ_it represents the disturbance term. To estimate models 1 and 2, we applied the pooled OLS regression. Subsequently, to mitigate potential endogeneity issues in the association between board composition and firm performance, we recognise the likelihood of reverse causality and unobserved heterogeneity (Wintoki et al., 2012). Such issues may emerge if the performance of the firm impacts the structure of the board, or if excluded variables simultaneously influence both board composition and performance. To mitigate these risks, we employ an instrumental variables (IV) methodology, which addresses endogeneity arising from reverse causality, selection bias, and measurement error. We use a two-stage least square (2SLS) regression methodology. In the 2SLS estimation, we choose instruments that have a strong correlation with the endogenous board composition variables while being plausibly exogenous to firm performance, meaning they affect firm outcomes only through their influence on board structure.

The second aspect of this paper examines the role of national governance quality in moderating the relationship between board composition, particularly board size and board independence, and firm performance. To empirically evaluate Hypotheses H3 and H4, which suggest that the influence of board composition on firm performance is more significant in nations with robust national governance frameworks, we integrate the National Governance Index (NGI) into Models 1 and 2. The empirical model for this phase is defined as follows:

$$\begin{array}{cc}\hfill Y_{it}={\beta }_0+& {\sum }_{n=1}^2\alpha n{BDSZ}_{it}+{\beta }_1{GOVQ}_{it}+{\beta }_2({BDSZ}_{it}\ast {GOVQ}_{jt})+\hfill \\ & {\sum }_{m=1}^M\mathrm{\mu }m{Controls}_{mit}+{\eta }_i+{\lambda }_i+{\sigma }_c+{\mathrm{Ɛ}}_{itc}\hfill \end{array}$$

(3)

$$\begin{array}{cc}\hfill Y_{it}={\beta }_0+& {\sum }_{n=1}^2\alpha n{PIND}_{it}+{\beta }_1{GOVQ}_{it}+{\beta }_2({PIND}_{it}\ast {GOVQ}_{jt})+\hfill \\ & {\sum }_{m=1}^M\mathrm{\mu }m{Controls}_{mit}+{\eta }_i+{\lambda }_i+{\sigma }_c+{\mathrm{Ɛ}}_{itc}\hfill \end{array}$$

(4)

Here, GOVQ_jt represents the national governance quality index, which is a composite measure encompassing six dimensions of governance at the country level specifically, voice and accountability, political stability and the absence of violence/terrorism, government effectiveness, regulatory quality, rule of law, and control of corruption, as established by Kaufmann et al. (2011). σ_c represents country fixed effects. In analysing the moderating effect of national governance quality, we adopted the pooled OLS regression as our baseline line model. We employed the two-step system generalised method of moments (System GMM) estimator, following Cave et al. (2023). This method is very reliable in terms of heteroskedasticity and autocorrelation (Arellano & Bover, 1995). The panel data technique tackles unobserved heterogeneity by applying first-difference to the data, which removes firm-specific effects and yields more consistent estimators. Moreover, System GMM adeptly addresses endogeneity by employing lagged levels and differences in the explanatory variables as internal instruments, leveraging their orthogonality conditions (Blundell & Bond, 1998). To verify the model’s validity, we conduct the Arellano–Bond test for autocorrelation in the residuals and the Hansen J-test to evaluate the suitability of the instruments.

This study used the Least Absolute Shrinkage and Selection Operator (Lasso) alongside conventional regression methods to mitigate overfitting and improve model parsimony (Bhattacharyya, 2025). The rationale for employing LASSO as the primary variable selection method is its computational efficiency and its ability to maintain the stability characteristic of Ridge regression (Vu et al., 2023). The lasso2 command in Stata 15 was employed with the Extended Bayesian Information Criterion (EBIC) to facilitate the selection of factors that most effectively predict company performance, measured by Tobin’s Q. To maintain uniformity throughout all empirical analyses, we utilised listwise deletion (commonly referred to as complete case analysis) to manage missing data (Newman, 2014; Enders, 2022). Only observations with complete data for all variables included in the regression models were retained in the final sample. This method guarantees that the number of observations stays uniform throughout all tables and estimations.

4. Results

This section provides a comprehensive analysis of the empirical findings derived from our research. The analysis begins with descriptive statistics and correlation analysis to assess the distribution, variation, and interrelationships among the selected variables. To ensure the robustness and validity of the findings, various model specifications and estimating methods are employed as part of the robustness tests.

4.1. Descriptive Statistics

Table 2. Descriptive Statistics.

Variables	Obs	Mean	Std. Dev.	Min	Max	Skew.	Kurt.
TOBIN Q	14,857	2.069	1.841	0.677	11.348	2.901	12.574
ROE	14,857	14.541	18.403	−47.502	100.614	1.434	10.993
ROA	14,857	5.802	6.929	−13.759	31.45	0.96	5.561
BDSZ	14,857	10.139	2.976	2	23	0.591	3.482
PIND	14,857	56.432	21.217	0	100	0.184	2.255
GOVQ	14,857	0.02	0.987	−1.968	1.646	0.073	1.31
PCWB	14,857	17.72	13.413	0	70	0.49	2.597
CEO	14,857	0.211	0.408	0	1	1.414	3
MTB	14,857	8.784	12.152	−2.862	33.633	1.428	3.259
FRSZ	14,857	11.318	2.812	5.406	19.254	0.633	3.251

Note: This table provides the descriptive statistics for all variables examined in this study. Observations with missing values in any significant variables were removed using listwise deletion, maintaining a uniform sample across all models.

presents the descriptive statistics for all variables used in the analysis. The average sample firm is valued at around double the book value of its assets, as indicated by Tobin’s Q (mean = 2.069, SD = 1.841), with values ranging from 0.677 to 11.348. Consequently, this demonstrates that the markets have favourable opinions of the capacity of businesses to make use of limited resources (Lewellen & Badrinath, 1997). The average ROE is 14.541%, with a standard deviation of 18.403%, and it ranges from a minimum of −47.502% to a maximum of 100.614%. The average board comprises approximately 10 directors, although some firms may have as few as 2 and others may have as many as 23. Similarly, the average firm size has a minimum of 5.406 and a maximum of 19.254. The sampled firms show an average female board representation of 17.72%, with certain firms reporting no women on their boards and others as high as 70%, indicating a significant gap in gender diversity among corporate boards. The descriptive statistics indicate that the firms have an average market-to-book ratio of 8.784, implying that, on average, the market value of these firms is much higher than their book value. Out of the 41 countries employed in this study, 18 are categorised as emerging economies, while 23 are identified as non-emerging (developed) economies, according to MSCI market classifications which is based on market accessibility, size, and liquidity criteria. Firms from emerging economies represent 51.4% of the sample (9887 firm-year observations), while those from developed economies account for 48.6% (9359 firm-year observations).

4.2. Correlation Analysis

Table 3. Pairwise correlations.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
(1) tobin Qratio	1.000
(2) roe	0.415 *	1.000
	(0.000)
(3) roa	0.566 *	0.750 *	1.000
	(0.000)	(0.000)
(4) bdsz	−0.136 *	0.005	−0.103 *	1.000
	(0.000)	(0.579)	(0.000)
(5) govq	−0.016 *	0.044 *	0.004	0.003	1.000
	(0.049)	(0.000)	(0.669)	(0.759)
(6) pcwb	0.054 *	0.126 *	0.069 *	0.100 *	0.405 *	1.000
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
(7) pind	0.049 *	0.089 *	0.023 *	−0.069 *	0.607 *	0.400 *	1.000
	(0.000)	(0.000)	(0.005)	(0.000)	(0.000)	(0.000)
(8) ceo	0.012	0.000	0.002	−0.014	−0.006	−0.074 *	0.010	1.000
	(0.158)	(0.958)	(0.846)	(0.081)	(0.448)	(0.000)	(0.240)
(9) mtb	0.264 *	0.114 *	0.131 *	−0.090 *	−0.353 *	−0.153 *	−0.259 *	0.018 *	1.000
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.030)
(10) frsz	−0.311 *	−0.137 *	−0.234 *	0.178 *	−0.275 *	−0.247 *	−0.258 *	0.074 *	−0.034 *	1.000
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)

Note: This table reports on the Pairwise correlation coefficients among the selected variables. * Shows significance at p < 0.05.

displays the correlation matrix of the employed variables. Firm performance exhibits a positive correlation with the ratio of independent directors on the board, the market-to-book ratio, and the proportion of women on the board. In contrast, board size and firm size exhibit negative associations with firm performance (Tobiin Q and ROE). All correlation coefficients among the independent and control variables are below the usual multicollinearity threshold of 0.8, indicating no significant multicollinearity issues.

Table 4. Variance inflation factor.

Variables	VIF	1/VIF
GOVQ	2.084	0.48
PIND	1.662	0.602
MTB	1.347	0.743
PCWB	1.301	0.769
FRSZ	1.173	0.853
BDSZ	1.07	0.935
CEO	1.012	0.989
Mean VIF	1.378	.

demonstrates that the highest Variance Inflation Factor (VIF) is associated with GOVQ, recording a value of 2.084, well beneath the set criterion of 10 (Kim, 2019). The VIF values for the remaining independent variables are below 10, signifying that multicollinearity is not a concern in our analysis.

In

Table 5. Variables Selected by Lasso Regression.

Dependent Variable: Tobin’s Q (Model 1-BDSZ)
Variable	Lasso Coefficient	Post-OLS Coefficient
BDSZ	−0.0432	−0.0447
GOVQ	−0.0355	−0.0435
BDAT	0.0643	0.0656
PCWB	0.0061	0.0067
CEO	0.0692	0.0808
MTB	0.0406	0.0409
FRSZ	−0.1650	−0.1661

and

Table 6. Lasso and Post-Lasso OLS Regression Results.

Dependent Variable: Tobin’s Q (Model 2-PIND)
Variable	Lasso Coefficient	Post-Lasso OLS Coefficient
PIND	0.0052	0.0056
GOVQ	−0.0861	−0.0968
BDAT	0.0638	0.0649
PCWB	0.0031	0.0034
CEO	0.0638	0.0731
MTB	0.0420	0.0423
FRSZ	−0.1738	−0.1748

, the Lasso regression employing the EBIC selection criterion identified seven principal predictors of firm performance (Tobin’s Q): governance quality (govq), board average tenure (bdat), percentage of women on board (pcwb), CEO duality (ceo), market-to-book ratio (mtb), and firm size (frsz). All chosen variables were preserved in the post-Lasso OLS regression, validating their statistical significance. The findings indicate that a limited number of governance and firm-level factors play a significant role in accounting for differences in firm performance, thereby validating the application of Lasso to prevent overfitting and enhance model simplicity (Vu et al., 2023).

Table 7. The impact of board structure on firm performance.

Panel A: POOLED OLS
	(1)	(2)	(3)	(4)
	TOBIN Q	ROE	TOBIN Q	ROE
BDSZ	0.0182	0.1684
	(0.5776)	(1.1376)
GOVQ	0.218	−1.4371	0.2254	3.2284 ***
	(1.387)	(−0.6001)	(0.9631)	(3.1164)
BDAT	0.0294 ***	0.3493 ***	0.044 ***	0.4569 ***
	(3.835)	(3.9595)	(2.7575)	(5.3721)
PCWB	0.0082 **	0.1475 ***	0.0046	0.179 ***
	(2.5938)	(3.2387)	(1.2013)	(4.1035)
CEO	−0.1231	−0.2938	−0.0597	−0.4608
	(−0.8389)	(−0.3449)	(−0.3372)	(−0.5928)
MTB	0.0586 *	0.3436	0.037 **	0.2641 *
	(1.8335)	(1.6624)	(2.1061)	(1.8471)
FRSZ	−0.3433 ***	−0.948 **	−0.2764 ***	−1.0443 ***
	(−4.3902)	(−2.2563)	(−5.5692)	(−3.3377)
PIND			0.0023	0.0292
			(0.5823)	(0.9677)
Constant	4.7284 ***	15.8254 ***	4.2039 ***	16.0992 ***
	(5.7046)	(2.8228)	(7.2686)	(3.8903)
Observations	14,857	14,857	14,857	14,857
R-squared	0.3144	0.1265	0.2429	0.1194
Year FE	YES	YES	YES	YES
Country FE	YES	YES	YES	YES
Panel B: IV-2SLS
	(1)	(2)	(3)	(4)
VARIABLES	TOBIN Q	ROE	TOBIN Q	ROE
BDSZ	0.563 ***	3.446 ***
	(13.347)	(9.272)
GOVQ	0.289 ***	3.882 ***	−0.781 ***	−3.833 ***
	(8.385)	(12.783)	(−9.563)	(−4.589)
BDAT	0.018 ***	0.293 ***	0.056 ***	0.542 ***
	(3.044)	(5.684)	(12.206)	(11.507)
PCWB	−0.010 ***	0.095 ***	−0.011 ***	0.071 ***
	(−4.764)	(5.214)	(−5.879)	(3.784)
CEO	0.337 ***	1.964 ***	−0.005	−0.063
	(5.981)	(3.956)	(−0.117)	(−0.156)
MTB	0.042 ***	0.290 ***	0.047 ***	0.334 ***
	(23.373)	(18.488)	(29.119)	(20.270)
FRSZ	−0.443 ***	−2.070 ***	−0.252 ***	−0.872 ***
	(−29.483)	(−15.636)	(−35.037)	(−11.864)
PIND			0.060 ***	0.433 ***
			(13.658)	(9.662)
Constant	0.595	−5.003	0.437	−10.333 **
	(1.217)	(−1.160)	(1.004)	(−2.322)
Observations	14,848	14,848	14,848	14,848
R-squared	−0.367	−0.064	0.085	0.041
Year FE	YES	YES	YES	YES
Country FE	YES	YES	YES	YES
Anderson canon LM stat	443	443	999.4	999.4
Anderson canon Prob	0	0	0	0
Cragg-Donald Wald F stat	454.9	454.9	1068	1068

Note: Panel A reports result from ordinary least squares (OLS) regressions, while Panel B presents estimates based on the IV-2SLS estimation. All variable definitions are provided in Table 1. z-statistics are reported in parentheses. Standard errors are clustered at country level. All models include year and country fixed effects. *** p < 0.01, ** p < 0.05, * p < 0.10.

, Panel A presents the baseline regression estimates derived from pooled OLS, applying Tobin’s Q and Return on Equity (ROE) as metrics for firm performance. Panel B displays findings from the two-stage least squares with instrumental variables (IV-2SLS) to mitigate potential endogeneity and omit variable bias. We develop a valid instrument for board size by interacting legal origin with the average board size at the regional-legal level. We create a categorical variable, region-legal, by integrating a firm’s legal origin with its geographical region. The average board size is subsequently calculated for each region-legal group, reflecting the institutional and cultural norms that influence board composition across varying legal contexts. The variable representing average board size by region and legal origin is interacted with legal origin to generate the instrument employed in the first-stage regression for board size. Model 2 employs similar technique, where we instrument board independence by interacting with the average proportion of independent directors (within each region-legal group) with legal origin. In all models, the Cragg-Donald Wald F-statistic above the standard threshold of 10, signifying the lack of weak instruments. The Sargan test for overidentifying limitations is not relevant, as the model is exactly identified.

In Panel A, board size (BDSZ) shows a positive but insignificant association with both Tobin’s Q and ROE. However, in the IV-2SLS estimation, board size (BDSZ) exhibits a positive and significant relationship with firm performance across both models. In model 1, panel B, the coefficient of 0.563 signifies that an increase of one person on the board corresponds with an average rise of 0.563 units in Tobin’s Q, assuming other variables remain constant. Likewise, an increase of one member in board size results in a 3.446 (model 1, Panel B) percentage rise in ROE. The positive correlation between BDSZ and company performance is corroborated by existing literature (Md Maniruzzaman, 2023; Sahoo et al., 2023). They asserted that BDSZ plays a crucial role in enhancing directors’ capacity to oversee and regulate managerial operations. A larger board is more likely to facilitate superior access to diverse resources compared to a smaller board. Agency theory posits that a corporate board possessing diversified expertise and knowledge is likely to exhibit enhanced learning and judicious decision-making capabilities, leading to improved firm value. Consequently, an increase in the number of directors enhances the boards’ capacity for oversight. Rashid (2018) contended that a larger board could more effectively monitor managerial actions, hence diminishing agency costs associated with the separation of management from ownership, which in turn enhances business performance. This result is contradicted by certain studies (Fatma & Chouaibi, 2021; Guest, 2009). Research by Guest (2009) and Fatma and Chouaibi (2021) provides inconsistent evidence, contending that excessively big boards may encounter coordination difficulties, diminished accountability, and prolonged decision-making procedures, hence undermining governance performance.

Furthermore, the results presented in Table 7, Panel B, indicate a positive and statistically significant relationship between board independence and firm performance, as measured by Tobin’s Q and ROE, using the IV-2SLS estimation method. Our findings are consistent with the results of Uribe-Bohorquez et al. (2018) and Al-Saidi (2021) who reported a positive and significant relationship between board independence and firm performance. However, our findings contradict that of Al-Gamrh et al. (2020) and Khan et al. (2024) who reported an inverse relationship between board independence and firm performance. Moreover, advocates of agency theory (1983) contend that external independent directors will diminish agency and monitoring expenses, while curbing managerial tendencies towards self-interest, hence enhancing corporate financial outcomes.

Panel B, which presents the IV-2SLS estimations, demonstrates a positive and significant association between governance quality and firm performance, as measured by Tobin’s Q and ROE. This finding aligns with Ngobo and Fouda (2012), who highlight the importance of strong national governance in enhancing business performance. A likely explanation is that robust governance frameworks help mitigate investment risk, foster regulatory certainty, and improve stock market efficiency, thereby boosting profitability and reducing firm level performance variability. Conversely, Model 2 reveals a negative relationship between governance quality and both Tobin’s Q and ROE. This disparity suggests that, when accounting for potential endogeneity, the effect of national governance quality may depend on the performance measure used and the specific model specification (T. Nguyen et al., 2015).

The results in Table 7 further show a significant positive relationship between board average tenure (BDAT) and firms’ performance (Tobin Q and ROE). The coefficient for board average tenure (BDAT) is consistently positive and statistically significant across all models, highlighting that longer-serving directors contribute positively to firm performance. This study corroborates Faleye et al. (2013), who assert that board tenure is positively associated with business longevity, suggesting that long-serving directors contribute stability and institutional knowledge to the organisation. Barroso et al. (2011) contend that directors with extended tenures are more inclined to depend on entrenched views and cognitive frameworks in their strategic decision-making. Musteen et al. (2006) observe that opposition to organisational change escalates with the length of directors’ employment. These studies indicate that prolonged board tenure may improve performance due to expertise and continuity, but it may also lead to more cognitive rigidity. The results in Table 7 further show a significant positive relationship between board average tenure (BDAT) and firms’ performance (Tobin Q and ROE). The coefficient for board average tenure (BDAT) is consistently positive and statistically significant across all models, highlighting that longer-serving directors contribute positively to firm performance. This study corroborates Faleye et al. (2013), who assert that board tenure is positively associated with business longevity, suggesting that long-serving directors contribute stability and institutional knowledge to the organisation. Barroso et al. (2011) contend that directors with extended tenures are more inclined to depend on entrenched views and cognitive frameworks in their strategic decision-making. Musteen et al. (2006) observe that opposition to organisational change escalates with the length of directors’ employment. These studies indicate that prolonged board tenure may improve performance due to expertise and continuity, but it may also lead to more cognitive rigidity.

Our OLS results suggest a significant and positive association between board gender diversity and firm performance, consistent with prior studies (Post & Byron, 2015; Safiullah et al., 2022). This finding supports the argument that women directors may contribute to more effective oversight and enhance the quality of boardroom decision-making through diverse perspectives. However, the IV-2SLS estimates reveal a more nuanced picture: while board gender diversity is negatively associated with firm value (as measured by Tobin’s Q), it shows a positive and significant relationship with accounting-based performance indicators. This divergence implies that the benefits of gender-diverse boards may be more evident in short-term operational efficiency than in market valuation.

The findings indicate that CEO duality has a statistically significant positive effect on firm performance in Model 1, Panel B. This result stands in contrast to the conclusions of Khan et al. (2024), who report a negative relationship between CEO duality and market-based performance measures. According to Khan et al., this adverse association may be attributed to complex ownership structures, such as layered control mechanisms, cross-shareholdings among affiliated firms, and limited corporate transparency. These factors can weaken governance effectiveness and erode investor confidence, thereby negatively impacting market valuation (Manna et al., 2016).

Likewise, firm size (FRSZ) exhibits a significant negative association with both performance indicators (Tobin Q and ROE) across both results (OLS regression and IV-2SLS regression). This implies that smaller firms tend to exhibit better performance than larger firms. Large corporations may experience diseconomies of scale, agency issues, or bureaucratic inefficiencies, which could account for the decrease in performance. This finding is consistent with previous research conducted by Liu et al. (2015) and Manna et al. (2016). On the other hand, this finding contradicts the studies by Atugeba and Acquah-Sam (2024) and Palaniappan (2017), who reported a positive relationship between firm size and financial performance.

4.3. Moderating Effect of GOVQ on the Relationship Between Board Structure and Firm Performance

Table 8. Board Structure, and Firm Performance: The Moderating Role of National Governance Quality.

Panel A: Baseline: Pooled OLS
	(1)	(2)	(3)	(4)
	TOBIN Q	ROE	TOBIN Q	ROE
BDSZ	0.016	0.148
	(0.5333)	(1.0197)
PIND			0.0037	0.0477
			(1.1188)	(1.38)
BDSZ*GOVQ	−0.024	−0.2231
	(−1.2721)	(−1.5533)
PIND*GOVQ			−0.0046	−0.0203
			(−1.5128)	(−0.6323)
GOVQ	0.4803 **	0.9988	0.4934 **	3.9782 *
	(2.0734)	(0.3137)	(2.177)	(1.9403)
BDAT	0.0296 ***	0.3511 ***	0.0417 ***	0.475 ***
	(3.8931)	(4.0048)	(3.8574)	(6.8737)
PCWB	0.0085 ***	0.1504 ***	0.0057	0.1289 **
	(2.8252)	(3.301)	(1.5743)	(2.6199)
CEO	−0.1197	−0.2624	−0.1964	−0.578
	(−0.8158)	(−0.3132)	(−1.2287)	(−0.6419)
MTB	0.0589 *	0.347 *	0.0492 *	0.2828 *
	(1.8396)	(1.6878)	(1.9509)	(1.816)
FRSZ	−0.341 ***	−0.9265 **	−0.2888 ***	−0.6533
	(−4.3797)	(−2.2607)	(−3.2388)	(−1.4248)
Constant	4.6928 ***	15.4948 ***	4.0823 ***	11.6564 **
	(5.3443)	(2.8176)	(4.1086)	(2.284)
Observations	14,857	14,857	14,857	14,857
R-squared	0.3154	0.1274	0.278	0.1046
Country FE	yes	yes	yes	yes
Year FE	Yes	Yes	Yes	Yes
Panel B: System GMM
	(1)	(2)	(3)	(4)
VARIABLES	TOBIN Q	ROE	TOBIN Q	ROE
L.TOBIN_Q_RATIO_W	0.551 ***		0.680 ***
	(6.480)		(11.887)
L.ROE		0.533 ***		0.442 ***
		(5.055)		(5.026)
BDSZ*GOVQ	−0.361 **	−3.051 **
	(−2.004)	(−2.474)
PIND*GOVQ			−0.039 ***	−0.183 **
			(−5.893)	(−2.169)
PIND			−0.028 ***	−0.118
			(−5.161)	(−1.612)
BDSZ	−0.641 **	−0.319
	(−2.030)	(−0.054)
GOVQ	3.782 **	15.231 *	1.961 ***	11.812
	(1.990)	(1.744)	(3.814)	(1.240)
BDAT	0.053 *	−0.059	−0.096 ***	−0.378
	(1.955)	(−0.216)	(−5.225)	(−1.494)
PCWB	0.054 ***	−0.286	0.042 ***	0.162
	(4.248)	(−1.465)	(4.116)	(0.866)
CEO	−0.831 **	10.675	−0.143	−1.513
	(−2.066)	(1.436)	(−1.086)	(−0.850)
MTB	−0.004	1.350 ***	0.072 ***	1.209 ***
	(−0.399)	(4.154)	(3.850)	(5.049)
FRSZ	−0.019	4.645 *	−0.104 ***	0.222
	(−0.240)	(1.750)	(−2.916)	(0.552)
Constant	6.996 ***	−71.343	3.151 ***	12.982
	(2.799)	(−1.299)	(4.023)	(0.951)
Observations	13,298	13,310	13,298	13,310
Number of cid	1590	1592	1590	1592
Country FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Hansen_test	8.049	3.574	5.646	3.555
Hansen Prob	0.154	0.467	0.227	0.470
AR(1)_test	−6.255	−6.101	−7.948	−6.714
AR(1)_p-value	3.99 × 10−10	1.06 × 10−9	0	0
AR(2)_test	−1.388	1.435	0.606	1.262
AR(2)_p-value	0.165	0.151	0.545	0.207
No. of Instruments	60	62	62	62

Note: Panel A reports results from ordinary least squares (OLS) regressions, while Panel B presents estimates based on the two-step System GMM estimator. All variable definitions are provided in Table 1. z-statistics are reported in parentheses. Standard errors are clustered at country level. All models include year and country fixed effects. *** p < 0.01, ** p < 0.05, * p < 0.10.

presents the findings on the relationship between board size, board independence, and firm performance, with a particular focus on the moderating role of national governance quality. The analysis employs two estimation techniques: Panel A details the findings from the baseline pooled OLS regressions, whereas Panel B showcases estimates derived from the System GMM model, which addresses potential endogeneity and the dynamic characteristics of firm performance. Both models include interaction terms that connect governance quality with variables related to board structure. Firm performance is assessed through two indicators: Tobin’s Q and Return on Equity (ROE).

Panel B of Table 4 presents the empirical findings from the generalised method of moments (GMM) regressions. To verify the model’s validity and the instruments’ robustness, we performed many diagnostic tests. First, the Arellano-Bond tests for serial correlation were employed to identify any autocorrelation present in the error terms. The findings validate the expected presence of first-order serial correlation [AR(1)] and the lack of second-order serial correlation [AR(2)], suggesting that the model is appropriately specified. Second, the Hansen J-statistics was utilised to assess over-identifying restrictions, and the findings indicate that the instruments applied are valid and uncorrelated with the error term. The diagnostic tests collectively validate that the assumptions foundational to the GMM estimation hold true across all four models.

The interaction term between board size and national governance quality (BDSZ × GOVQ) yields a consistently negative but statistically insignificant coefficient in Panel A (Pooled OLS). However, in Panel B (System GMM), the coefficient becomes negative and statistically significant, suggesting a more robust moderating effect of governance quality under dynamic estimation. This indicates that whereas larger boards typically enhance business performance, their efficacy declines in nations with robust governance systems. This suggests that internal governance tools, including board size, may function as compensatory measures in weaker institutional environments but become superfluous or even counterproductive when external governance structures are robust.

Similarly, the coefficient for the interaction term between board independence and governance quality (PIND × GOVQ) is negative and statistically significant in Panel A (Pooled OLS). However, in Panel B (System GMM), the coefficient becomes negative and statistically significant. This highlights that even though an independent board improved firm performance, its marginal benefit is reduced in countries with strong governance structures. In countries with effective governance, external mechanisms can serve as a replacement for internal board oversight, thereby diminishing the additional value derived from board independence. These findings do not support hypotheses 3 and 4 of the study, which posits that compliance with national governance frameworks strengthens the relationship between board structures (board size and board independence) and firm performance.

4.4. Moderating Role of National Governance Quality: Emerging vs. Developed Economies

This research investigates the extent to which the quality of national governance influences the interplay between board structure and firm performance (Tobin Q). The analysis is undertaken distinctly for emerging markets and developed economies to effectively capture institutional variation. The dynamic panel estimation is carried out using the system generalised method of moments (GMM), which addresses potential endogeneity and unobserved heterogeneity in the governance–performance relationship. The Arellano-Bond tests for serial correlation were employed to identify any autocorrelation present in the error terms. The findings validate the expected presence of first-order serial correlation and the absence of second-order serial correlation, suggesting that the model is appropriately specified. Also, the Hansen J-statistic was adopted to examine over-identifying restrictions, and the findings indicate that the instruments applied are valid and uncorrelated with the error term.

As shown in Columns (1) and (2) of

Table 9. System GMM Estimates of the Relationship Between Board Composition and Firm Performance: The Moderating Role of National Governance Quality in Emerging and Developed Economies.

VARIABLES	Emerging	Non-Emerging	Emerging	Non-Emerging
L.tobin_q_ratio	0.510 ***	0.360 ***	0.597 ***	0.775 ***
	−3.492	−3.367	−8.176	−9.418
BDSZ*GOVQ	2.348 ***	−0.327 **
	−3.353	(−2.398)
PIND*GOVQ			−0.002	−0.019 *
			(−0.375)	(−1.918)
BDSZ	3.050 ***	0.309 ***
	−3.544	−2.671
PIND			0.005	0.019
			−0.964	−1.588
GOVQ	−25.042 ***	4.267 **	1.485 *	0.428
	(−3.663)	−2.3	−1.805	−1.369
BDAT	−0.005	0.005	0.002	−0.056
	(−0.107)	−0.311	−0.052	(−0.812)
PCWB	−0.055 ***	−0.001	0.001	−0.02
	(−3.302)	(−0.231)	−0.146	(−1.464)
CEO	0.47	0.149	−2.508	−0.753
	−1.279	−0.733	(−0.878)	(−0.318)
MTB	0.313 ***	0.249 ***	0.058	0.043 **
	−2.623	−5.165	−1.047	−2.048
FRSZ	−0.145	−0.153 ***	−0.596 ***	−0.356 *
	(−1.153)	(−3.836)	(−5.220)	(−1.732)
Constant	−35.754 ***	−2.173	13.032 ***	4.963 *
	(−3.885)	(−1.463)	−3.833	−1.849
Observations	6673	6625	6661	6612
Number of cid	818	772	817	772
Country FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Hansen_test	0.816	2.502	1.183	3.983
Hansen Prob	0.846	0.644	0.757	0.263
AR(1)_test	−3.418	−5.297	−7	−5.037
AR(1)_p-value	0.000631	1.18 × 10−7	0	4.74 × 10−7
AR(2)_test	−0.977	1.18	−0.184	1.628
AR(2)_p-value	0.329	0.238	0.854	0.104
No. of Instruments	39	39	39	36

Note: Table presents estimates from the System GMM regression. All variable definitions are provided in Table 1. Dependent variable is Tobin’s Q. All models include year fixed effects. Standard errors are clustered at country level. z-statistics are reported in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.10. The asterisk (*) indicates an interaction term between board size (BDSZ) and governance quality (GOVQ).

, the estimated coefficient for the interaction between board size and national governance quality (BDSZ × GOVQ) is positive and statistically significant in emerging markets, but negative and statistically significant in non-emerging markets. These results support Hypothesis H5a, indicating that national governance quality strengthens the positive relationship between board size and firm performance in emerging markets, whereas it weakens this relationship in developed markets. This finding corroborates with that of Atugeba and Acquah-Sam (2024) who documented that governance quality moderates the corporate governance-firm performance relationship in Ghana. Likewise, T. Nguyen et al. (2015) contended that national governance quality strongly influences the association between ownership structures and performance in developing economies where enforcement is weak.

As shown in Columns (3) and (4) of Table 9, the coefficient for the interaction between board independence and national governance quality (PIND × GOVQ) is negative and statistically significant in developed markets. This supports Hypothesis H5b, which posits that the negative moderating effect of national governance quality on the board independence—performance relationship is more pronounced in developed markets, where robust national institutions and institutional investors provide additional layers of governance. This finding contradicts that of Zattoni et al. (2017) who reported that national business systems moderate the relationship between board independence and firm performance in IPO firms. The interaction between board independence and national governance quality is statistically insignificant in emerging economies, suggesting that governance quality does not meaningfully moderate the impact of board independence on firm performance in emerging markets.

4.5. Additional Analysis

We developed a composite financial performance index to comprehensively assess firm performance, utilising three well recognised indicators: Return on Equity (ROE), Return on Assets (ROA), and Tobin’s Q. All variables were initially standardised using z-scores to address variations in magnitude and distribution. The standardised values were thereafter averaged to produce a singular performance measure. This composite index encompasses both accounting-based performance (ROE and ROA) and market-based value (Tobin’s Q), providing a more balanced and robust indicator of firm performance than any singular metric. We employed the composite performance index as a proxy for firm performance in

Table 10. IV-2SLS Regression Results of Board structure on Firm Performance Across Governance Quality Levels and Full Sample.

	(1)	(2)	(3)	(4)	(5)	(6)
VARIABLES	Low-Governance	High-Governance	Low-Governance	High-Governance	Full Sample	Full Sample
BDSZ	−0.954 ***	−0.080 ***			0.193 ***
	(−4.590)	(−8.915)			(11.162)
PIND			0.041 ***	−0.018 ***		0.054 ***
			(12.952)	(−8.888)		(17.458)
GOVQ	−0.249	−0.009	0.256 ***	0.016	0.142 ***	−0.992 ***
	(−0.781)	(−0.096)	(2.905)	(0.157)	(10.058)	(−15.750)
BDAT	0.056 ***	0.016 ***	0.018 ***	−0.005	0.017 ***	0.045 ***
	(5.773)	(5.680)	(5.856)	(−1.478)	(7.123)	(16.559)
PCWB	0.003	0.010 ***	0.003 ***	0.013 ***	0.001	−0.004 ***
	(1.114)	(10.809)	(2.737)	(12.092)	(0.871)	(−3.613)
CEO	−0.259 ***	−0.160 ***	0.035	−0.128 ***	0.114 ***	−0.031
	(−2.714)	(−6.441)	(1.359)	(−5.266)	(4.946)	(−1.414)
MTB	0.007 ***	0.034 ***	0.018 ***	0.034 ***	0.016 ***	0.022 ***
	(2.665)	(28.857)	(17.930)	(28.773)	(22.308)	(23.540)
FRSZ	0.355 ***	−0.105 ***	−0.094 ***	−0.143 ***	−0.164 ***	−0.137 ***
	(3.522)	(−18.473)	(−17.596)	(−32.631)	(−26.618)	(−29.621)
Constant	4.823 ***	1.611 ***	−0.694 **	2.738 ***	−0.501 **	−2.254 ***
	(4.406)	(8.619)	(−1.979)	(12.511)	(−2.502)	(−8.620)
Observations	7372	7476	7372	7476	14,848	14,848
R-squared	−7.799	0.279	−0.105	0.275	−0.132	−0.422
Year FE	YES	YES	YES	YES	YES	YES
Country FE	YES	YES	YES	YES	YES	YES
Anderson canon LM stat	23.54	1274	530.2	741.6	443	650.3
Anderson canon Prob	1.22 × 10−6	0	0	0	0	0
Cragg-Donald Wald F stat	23.51	1526	568.8	818.5	454.9	677.6

Note: All models are estimated using IV-2SLS regressions. The dependent variable is a composite firm performance index constructed from ROE, ROA, and Tobin’s Q. Columns (1) and (2) present estimates for board size as the regressor across low- and high-governance subsamples, respectively. Columns (3) and (4) show results for board independence as the regressor across the same governance subsamples. Columns (5) and (6) report full-sample regressions for board size and board independence, respectively. Robust z-statistics are reported in parentheses. *** p < 0.01, ** p < 0.05.

below.

To further examine the impact of governance quality, we performed a sub-sample analysis by categorising the sample into high- and low-governance groups based to the median value of the composite governance quality index. Observations with a governance score at or above the median were categorised as high-governance, while those below the median were classified as low-governance. Subsequently, we re-evaluated our models individually for each group to determine if the effects of board size and independence on company performance vary across governance settings. We employed instrumental variable two-stage least squares (IV-2SLS) estimation and used a composite performance index developed from ROE, ROA, and Tobin’s Q as an indicator of firm performance. The findings indicate that board size consistently exerts a negative and statistically significant influence on firm performance in both high- and low-governance quality environments. However, the extent of the adverse impact is significantly more pronounced in countries with low governance than in those with high governance. This indicates that boards are less effective, or potentially harmful, in low-governance quality environments. due to the lack of an institutional framework that facilitates efficient board operations (T. Nguyen et al., 2015).

In the full sample, board size (BDSZ) exhibits a positive and statistically significant relationship with the composite measure of firm performance, consistent with the finding that board independence (PIND) is also positively and significantly associated with firm performance. This finding is consistent with our initial regression reported in Table 7, where Tobin’s Q and ROE were used individually as firm performance measures. In both cases, board size and board independence demonstrated a positive and statistically significant relationship with firm performance, reinforcing the robustness of the effect across alternative specifications.

To ensure robustness, we re-estimate our models by omitting observations from the two nations with the greatest sample sizes, China (20.9%) and the United States (9.5%), to evaluate if our findings are disproportionately affected by these predominant contributors. The resultant reduced sample (N = 10,398) serves as a valuable benchmark for assessing the consistency of our findings across a more equitable population. The results demonstrate that board size (BDSZ) has a positive and statistically significant relationship with Tobin’s Q, aligning with our primary findings presented in Table 7. Nonetheless, its correlation with ROE is positive albeit statistically insignificant (see

Table A2. Fixed effect model without China and United States.

	(1)	(2)	(3)	(4)
VARIABLES	FE	FE	FE	FE
BDSZ	0.0124 ***	0.1020
	(3.1260)	(0.7222)
GOVQ	0.3187	−4.1839	0.3178	−4.3389
	(1.1545)	(−1.5635)	(1.1548)	(−1.5743)
BDAT	−0.0010	0.1414	0.0001	0.1554
	(−0.1717)	(1.1703)	(0.0117)	(1.2863)
PCWB	−0.0008	0.0848 ***	−0.0007	0.0809 ***
	(−0.7796)	(3.1492)	(−0.6326)	(3.0183)
CEO	−0.0726	−0.7615	−0.0821	−1.0332
	(−1.4189)	(−0.7555)	(−1.5858)	(−1.0505)
MTB	0.1206 ***	0.4230 ***	0.1224 ***	0.4244 ***
	(6.7646)	(3.6030)	(6.8697)	(3.5911)
FRSZ	−0.1301 ***	−1.9231	−0.1313 ***	−1.7464
	(−2.9761)	(−1.3216)	(−2.9993)	(−1.2213)
2013.year	0.0060	2.7794	−0.0039	2.6650
	(0.0447)	(1.2446)	(−0.0289)	(1.2185)
2014.year	0.0501	3.0848	0.0439	2.9290
	(0.3604)	(1.4699)	(0.3164)	(1.4123)
2015.year	0.1696	2.0473	0.1623	1.8681
	(0.8373)	(0.9797)	(0.8042)	(0.9023)
2016.year	0.1513	2.0351	0.1453	1.8805
	(0.7681)	(0.9790)	(0.7414)	(0.9137)
2017.year	0.2193	2.7374	0.2161	2.5455
	(1.0855)	(1.3683)	(1.0801)	(1.2847)
2018.year	0.2731	3.4060 *	0.2714	3.2336
	(1.2278)	(1.7346)	(1.2272)	(1.6569)
2019.year	0.2506	2.2733	0.2530	2.0808
	(1.1161)	(1.0465)	(1.1360)	(0.9747)
2020.year	0.1447	−0.6762	0.1430	−0.9637
	(0.9514)	(−0.3250)	(0.9608)	(−0.4701)
2021.year	0.3418	1.7213	0.3419	1.5148
	(1.4141)	(0.8258)	(1.4509)	(0.7428)
2022.year	0.3505	4.4935 *	0.3558	4.2090 *
	(1.2910)	(1.8208)	(1.3310)	(1.7445)
pind			−0.0027 **	0.0267
			(−2.3971)	(1.0360)
Constant	2.5768 ***	28.8224 **	2.8698 ***	26.5249 *
	(4.1014)	(2.0605)	(4.4233)	(1.8258)
Observations	10,398	10,398	10,195	10,195
R-squared	0.1902	0.0260	0.1934	0.0261
Number of cid	1112	1112	1104	1104
Year FE	YES	YES	YES	YES

t-values are in parentheses. Standard errors are clustered at country level. *** p < 0.01, ** p < 0.05, * p < 0.10.

).

In a similar vein, board independence (PIND) demonstrates a negative and significant correlation with Tobin’s Q, which underscores apprehensions regarding the efficacy of independent boards in augmenting firm value in specific scenarios. Nonetheless, the relationship with ROE is positive yet insignificant, suggesting that market-based and accounting-based performance measures might reflect different dimensions of board effectiveness. The findings highlight the necessity of considering cross-country institutional differences when analysing the connections between governance and performance.

Furthermore, to evaluate the sensitivity of our findings to possible sample imbalance among nations, we performed Weighted Least Squares (WLS) regressions adopting the inverse square root of the number of observations per country as weights. Some of the findings reported in

Table A3. FIXED EFFECT—Weighted Least Squares (WLS) regressions using inverse square root.

	(1)	(2)	(3)	(4)
	TOBIN_Q	ROE	TOBIN_Q	ROE
bdsz	0.0171 ***	0.0203
	(3.2829)	(0.1809)
govq	0.0891	−3.9336	0.0853	−4.1388
	(0.6749)	(−1.593)	(0.6407)	(−1.6345)
bdat	−0.0018	0.1292	−0.001	0.1459
	(−0.309)	(0.9152)	(−0.1672)	(1.0115)
pcwb	−0.0009	0.0712 **	−0.0009	0.0689 **
	(−0.7455)	(2.4213)	(−0.6828)	(2.3444)
ceo	−0.096 ***	−2.4505 *	−0.1039 ***	−2.5681 **
	(−3.1606)	(−2.0189)	(−3.2631)	(−2.0949)
mtb	0.091 ***	0.5013 ***	0.0923 ***	0.5033 ***
	(4.7459)	(5.097)	(4.6537)	(5.0592)
frsz	−0.1642 ***	−0.1515	−0.1663 ***	0.0321
	(−3.2357)	(−0.122)	(−3.3774)	(0.025)
2012bn.year
2013.year	0.1033	4.0648 **	0.0935	4.0092 **
	(0.6439)	(2.0501)	(0.5739)	(2.0318)
2014.year	0.1474	4.1126 *	0.1426	4.026 *
	(0.8461)	(1.96)	(0.806)	(1.9194)
2015.year	0.1977	3.1848	0.1896	3.0684
	(1.1821)	(1.5216)	(1.1197)	(1.4654)
2016.year	0.166	2.8503	0.1585	2.7542
	(0.9568)	(1.3271)	(0.9002)	(1.2782)
2017.year	0.2578	3.8334 *	0.2542	3.7134 *
	(1.3782)	(1.7749)	(1.3413)	(1.7057)
2018.year	0.2392	4.821 **	0.2341	4.7281 **
	(1.266)	(2.165)	(1.2223)	(2.1096)
2019.year	0.2558	3.1924	0.2557	3.0829
	(1.3655)	(1.3691)	(1.352)	(1.318)
2020.year	0.2845	0.3692	0.2846	0.1606
	(1.3781)	(0.1658)	(1.3611)	(0.0715)
2021.year	0.3794 *	3.3312	0.382 *	3.2364
	(1.7592)	(1.331)	(1.7515)	(1.2863)
2022.year	0.3047	5.2145 **	0.3087	5.0711 *
	(1.4984)	(2.0486)	(1.5028)	(1.9766)
pind			−0.0025 **	0.0073
			(−2.6637)	(0.2683)
constant	2.7998 ***	7.744	3.1396 ***	5.5027
	(5.334)	(0.5946)	(5.6941)	(0.4204)
Observations	15,060	15,105	14,857	14,900
Within R2	0.1309	0.0314	0.1326	0.0317
country FE	NO	NO	NO	NO
Year FE	Yes	Yes	Yes	Yes

t-values are in parentheses. Standard errors are clustered at country level. *** p < 0.01, ** p < 0.05, * p < 0.1.

align closely with the results in Table 7: the majority of key variables maintain their expected signs and statistical significance. However, PIND demonstrates a significant negative association with Tobin’s Q under WLS, indicating that its relationship to firm performance may differ across countries with varying sample sizes. We also note variations in the impacts of governance quality and CEO duality, which may indicate differences in governance frameworks across countries.

5. Conclusions

This research analysed the impact of two specific board characteristics, namely board size and board independence, on financial performance, adopting ROE and Tobin’s Q as indicators of firm performance. The dataset includes 1604 firms over a span of 12 years, specifically from 2013 to 2023. We employed pooled ordinary least squares (OLS) as our baseline regression, subsequently employing instrumental variables, two-stage least squares (IV-2SLS), and dynamic panel estimation through a system generalised method of moments (System GMM). These methods collectively address concerns like reverse causality, omitted variable bias, and unobserved firm-specific heterogeneity. The research findings show a significant positive relationship between the size of the board and the performance of the firm, even after controlling firm-specific heterogeneity and endogeneity. In a similar vein, board independence is shown to boost firm performance, perhaps because independent directors can mitigate insider opportunism and promote investment efficiency (Liu et al., 2015).

Furthermore, the study contributes to the understanding of the interplay between national governance quality and the effectiveness of internal corporate governance mechanisms. The results present significant theoretical and practical implications. In contrast to our hypothesis, our findings indicate a different outcome, highlighting that a more robust national governance seems to reduce the effectiveness of these internal board mechanisms. This indicates a substitution effect, wherein robust external governance mechanisms diminish the necessity for companies to depend significantly on internal frameworks such as extensive or autonomous boards. In other words, a strong institutional environment can diminish the marginal advantages of specific governance characteristics. National governance quality is not a universal solution and what proves effective in one nation may be irrelevant or even detrimental in another. In nations with deficient institutional frameworks, internal governance systems may fulfil a compensating function, alleviating agency issues in the lack of robust external oversight (Atugeba & Acquah-Sam, 2024).

In addition, the findings contribute to the expanding body of knowledge on corporate governance by providing comparative insights from both emerging and developed markets. The findings indicate that improving the quality of national governance in emerging economies can boost the beneficial impacts of board structures on corporate performance. Based on institutional and agency theory, this study indicates that the governance outcomes at the firm level are influenced not only by the internal structures of the board but also by the calibre of the institutional environment within a country.

While this research presents significant findings, several study limitations must be acknowledged. First, firm performance is assessed exclusively through Return on Equity (ROE) and Tobin’s Q. Although widely recognised in prior literature, these metrics capture only a narrow dimension of financial outcomes. Other indicators employed such as profit before tax, profit after tax, return on assets (ROA), and return on capital reflect different aspects of organisational performance. Incorporating a broader range of measures in future research would enhance both the robustness and generalisability of findings. Second, the study focuses on only two board characteristics: board size and board independence. While these are key structural variables, they do not capture the complexity of board dynamics. Attributes such as average board tenure, age, educational diversity, functional expertise, and CEO duality may also shape firm outcomes. Third, the analysis omits environmental, social, and governance (ESG) dimensions. Given their growing importance in shaping corporate behaviour and strategic orientation, their exclusion limits the explanatory scope of the findings. Future studies would benefit from richer datasets that incorporate these diversity and ESG-related factors, and from exploring their interaction with national governance quality, to provide a more nuanced understanding of governance effectiveness across institutional contexts.