Do Board Characteristics Influence Leverage and Debt Maturity? Empirical Evidence from a Transitional Economy

Abstract

This study examines the impact of board characteristics on capital structure decisions in the context of a transition economy, focusing on Algeria, where governance institutions are underdeveloped and the financial market remains immature. Using the Generalized Method of Moments (GMM) on a panel dataset of 120 firms over the period 2015 to 2019, we identify a U-shaped relationship between board size and leverage, and an inverted U-shaped relationship between board size and debt maturity. Furthermore, increased nationality diversity on boards is found to significantly reduce debt maturity. These findings highlight the critical role of board composition in shaping corporate financing strategies in transition economies and provide novel insights into corporate governance dynamics in a relatively underexplored institutional context. The results are particularly relevant for national entities such as COSOB and Hawkama El Djazaïr and may guide banking sector practices by promoting the integration of board governance criteria into credit evaluation processes.

1. Introduction

The configuration of corporate governance actors and their influence on strategic decision-making remain key topics in corporate finance research. In particular, the composition of the board of directors has received considerable attention regarding its impact on financial decisions especially capital structure as emphasized in recent studies (Martín-Ugedo & Mínguez-Vera, 2023; Abdel-Wanis & Rashed, 2023). Board members’ cognitive frameworks, values, and personalities as well as the dynamics within governance mechanisms help explain variations in strategic choices, particularly those aimed at fostering long-term value creation (Kavadis et al., 2024).

In this context, the board of directors, as the cornerstone of corporate governance, plays a critical role in shaping corporate strategy, including decisions related to capital structure. The choice between debt and equity financing is not only vital for maximizing returns but also functions as a disciplinary mechanism for managerial behavior. Agency theory of free cash flow highlights this disciplinary role of debt. As argued by Jensen (1986), debt serves as a commitment device that limits managerial discretion by requiring regular interest payments. Failure to meet these obligations can lead to financial distress or bankruptcy, exposing managers to personal costs such as reduced compensation, reputational damage, and the loss of non-financial benefits. This mechanism aligns managerial incentives with shareholder interests and contributes to the reduction in agency costs.

In this study, the theoretical framework primarily focuses on agency theory to explore whether debt can serve as a governance mechanism that compensates for the absence or weakness of external governance controls. While corporate governance encompasses a broad range of theories including resource dependence theory, stewardship theory, and others, agency theory remains especially relevant in explaining how debt disciplines managerial behavior by reducing agency conflicts between managers and shareholders.

Empirical research examining the impact of board composition on capital structure has produced mixed results. For example, studies conducted across various institutional contexts including Hermassi et al. (2015) in Canada, Purag et al. (2016) in Malaysia, Sewpersadh (2019) in South Africa and Thakolwiroj and Sithipolvanichgul (2021) in Thailand, Elmoursy et al. (2025) on UK-Listed Firms, Askarany et al. (2025) in Iran have reported inconsistent findings, ranging from positive and negative associations to no significant effects. These discrepancies may stem from variations in institutional environments, firm-level characteristics, time frames, and methodological approaches.

This study seeks to address several important gaps in the literature. First, it focuses on Algeria, a country with distinctive institutional characteristics, including a shallow capital market, a state-dominated banking system, widespread family ownership, and weak corporate governance frameworks. In such a context, board effectiveness and composition are strongly influenced by institutional and legal constraints. Governance models rooted primarily in agency theory may be insufficient to explain decision-making in such settings (Grosman et al., 2019).

This raises a critical question: which board characteristics are most relevant for effective financial decision-making in Algeria, where institutional conditions diverge substantially from those in developed markets?

Second, while much of the existing literature focuses exclusively on the level of total debt, this research extends the analysis to include debt maturity structure. This allows for a more nuanced understanding of how firms manage short- versus long-term debt, and whether short-term debt serves as a disciplinary mechanism to constrain managerial discretion.

Moreover, the study explores potential nonlinear relationships such as U-shaped or inverted U-shaped effects that are often missed by linear models. This is particularly relevant in transition economies like Algeria, where the relationship between governance and financing decisions may exhibit greater complexity.

Practically, Algerian firms offer a valuable context to examine how governance mechanisms function in environments characterized by limited capital access and weak external oversight. This setting reveals how institutional and cultural specificities shape board behavior, showing that identical board structures can yield divergent outcomes depending on the national context.

This approach brings originality to the corporate governance literature by addressing a gap in transition economy research and offering insights relevant to scholars, regulators, and practitioners aiming to better understand governance dynamics in such environments.

The rest of this paper is structured as follows: Section 2 reviews the relevant literature and develops the hypotheses. Section 3 describes the research methodology and data. Section 4 reports the empirical results and discussion. Finally, Section 5 concludes the paper, highlighting the main findings, implications, limitations, and suggestions for future research.

2. Literature Review and Hypotheses Development

This section reviews the main theories and empirical studies related to board characteristics and their impact on capital structure, including total debt and debt maturity. The analysis focuses on six board features: board size, CEO duality, board independence, gender diversity, nationality diversity, and age diversity.

The hypotheses developed in this study are based on a multi-theoretical approach, with a primary focus on agency theory. This theory emphasizes the role of governance mechanisms in controlling managerial behavior and aligning managers’ interests with those of shareholders (Jensen & Meckling, 1976). From this perspective, debt serves as a disciplinary tool to reduce agency costs and limit managerial discretion (Jensen, 1986).

In addition, resource dependence theory sees the board of directors as a link between the firm and its external environment (Pfeffer, 1972; Pfeffer & Salancik, 1978). Board composition acts as a strategic response to external pressures by facilitating access to key resources and improving decision-making (Zahra & Pearce, 1989; Ward & Forker, 2017). Board diversity, for instance, can assist firms in navigating environmental uncertainty and making more informed strategic decisions.

Stewardship theory, developed by Davis et al. (1997), offers a more optimistic view of managerial behavior. It assumes that managers are collectivist and intrinsically motivated to act in the best interests of shareholders and stakeholders. Under this model, trust and empowerment serve as key governance mechanisms, replacing control and monitoring.

More recently, institutional theory has emerged as a complementary lens, especially in explaining variations in governance effectiveness across national contexts. This theory posits that governance outcomes are shaped by broader institutional factors, legal, cultural, and economic, which differ substantially across countries (Filatotchev et al., 2013). The same governance mechanisms may produce divergent effects depending on the institutional setting.

This study aims to provide a better understanding of the relationship between board characteristics and capital structure by examining these links in the specific context of Algeria. The business environment in Algeria is characterized by significant institutional fragility, heavy bureaucracy, and an underdeveloped financial market, all of which considerably hinder economic dynamism. The entrepreneurial landscape is largely dominated by family-owned firms, often of modest size, where management practices remain informal and lack professionalization. Governance structures are archaic, with decision-making power concentrated in the hands of a single leader or family circle, and a weak separation between ownership and control. This limited internal organization is accompanied by a lack of transparency and a culture of secrecy, which complicates access to external financing and erodes the confidence of investors and partners.

The regulatory framework for corporate governance remains embryonic. The Algerian Governance Code, launched in 2009 by the Enterprise Action and Reflection Center (CARE), remains a voluntary initiative lacking legal enforcement. As a result, it is poorly disseminated and rarely implemented, especially among SMEs. Boards of directors tend to play a symbolic role, with low diversity and limited strategic oversight, mainly ratifying decisions already made by the firm’s leader. They also often lack independent directors, reducing their capacity for effective governance.

Meanwhile, Algeria’s financial market remains underdeveloped, with a largely inactive stock exchange, limiting alternative financing options outside the banking system. In this context, capital structure decisions are often influenced more by relational and political considerations than by financial criteria. Bank loans, the main source of external financing, are frequently granted through opaque processes, with little emphasis on governance quality, risk, or performance.

In this context, conventional governance mechanisms struggle to take hold, while financial decisions are largely driven by family dynamics and personal relationships rather than a rigorous analysis of risk, performance, or good governance practices. Although organizations such as COSOB (Securities and Investments Organization and Monitoring Commission) and the Algerian Institute of Corporate Governance “Hawkama El Djazaïr” are working to promote better practices, their impact remains limited. Broader reforms are urgently needed, including the adoption of effective governance mechanisms and independent oversight, to improve the decision-making environment in Algeria.

2.1. Board Size

From an agency theory standpoint, particularly Jensen’s (1986) free cash flow theory, debt can act as a disciplinary governance mechanism, especially short-term debt, to limit managerial discretion over free cash flows and prevent inefficient investments. This mechanism becomes particularly relevant in contexts with underdeveloped financial markets, where external governance pressures (such as hostile takeovers or activist investors) are weak or absent. In such settings, debt can substitute for missing market-based governance, helping to align managerial and shareholder interests.

However, as board size increases, agency costs may re-emerge in new forms: communication breakdowns, diluted accountability, and the “free rider” problem among directors can undermine board effectiveness (Lipton & Lorsch, 1992). This inefficiency may prevent the board from effectively using debt as a governance tool, potentially leading to suboptimal financial risk management.

Several studies have explored the relationship between board size and capital structure, yet findings remain mixed and context dependent. In many emerging or developing economies, such as Pakistan (Ur Rehman et al., 2010; Amin et al., 2022), India (Gill et al., 2012), Egypt (Abdel-Wanis & Rashed, 2023), and Bahrain (Alabdullah & Mohamed, 2023),evidence points to a positive relationship, where larger boards are associated with higher leverage. These results support the notion that in weak institutional environments, larger boards may actively employ debt as a substitute disciplinary mechanism.

Conversely, studies from other emerging or developing markets, including Ghana (Ahmed, 2019), Nigeria (Sani et al., 2020), and South Africa (Sewpersadh, 2019) reveal a negative association, suggesting that larger boards may favor lower debt levels. These contrasting findings highlight the variability in board effectiveness across emerging markets, likely influenced by differences in institutional quality, ownership concentration, and firm-level governance practices.

In developed economies such as Austria, Belgium, China, Denmark, Finland, France, Germany, the United Kingdom, Hong Kong, Japan, Korea, Switzerland, and the United States (Kijkasiwat et al., 2022), as well as Spain (Martín-Ugedo & Mínguez-Vera, 2023), empirical evidence tends to show a negative relationship between board size and leverage. These findings are consistent with the view that in robust legal and governance environments, firms rely less on debt to discipline management. Instead, stronger and more effective boards fulfill that role, reducing the need for debt as a substitute.

Other studies, such as those by Hussainey and Aljifri (2012) in the UAE, Hermassi et al. (2015) in Canada, Purag et al. (2016) on Malaysian family firms, and Thakolwiroj and Sithipolvanichgul (2021) in Thailand, report no statistically significant relationship between board size and leverage. These findings suggest that the effect of board size on capital structure may be highly contingent on institutional settings, governance systems, and firm-specific characteristics.

In the Algerian context, marked by underdeveloped financial markets and the dominance of family-owned SMEs, short-term debt may act as a substitute for weak external governance mechanisms, as suggested by agency theory (Jensen, 1986). Smaller boards, often composed of closely affiliated or family members, may exercise strong internal control, reducing the need for debt. However, as board size increases, coordination inefficiencies and diluted accountability may hinder board performance. In such cases, firms may increasingly rely on short-term debt to discipline managers and mitigate agency costs.

These considerations lead us to expect a nonlinear relationship between board size and capital structure. Initially, increasing board size may strengthen monitoring and governance, but beyond a certain point, the costs of complexity and inefficiency may outweigh the benefits.

H1a. 

The relationship between board size and leverage in Algerian firms is nonlinear.

H1b. 

The relationship between board size and debt maturity in Algerian firms is nonlinear.

2.2. CEO Duality

CEO duality refers to the concentration of the roles of Chief Executive Officer and Chair of the Board in a single individual (Rechner & Dalton, 1991). From an agency theory perspective, this structure may weaken board independence and reduce the effectiveness of oversight (Baliga et al., 1996; Dalton et al., 1998), thereby increasing agency costs.

Empirical findings on the link between CEO duality and capital structure are mixed and context sensitive. In emerging markets such as India (Gill et al., 2012), Pakistan (Nazir et al., 2012), Nigeria (Ranti, 2013), Greece (Dasilas & Papasyriopoulos, 2015), and Bahrain (Alabdullah & Mohamed, 2023), studies often report a positive association between CEO duality and leverage. These results suggest that centralized leadership can facilitate quicker decision-making and greater risk-taking, leading firms to rely more heavily on debt.

In contrast, studies in developed economies, including the UK, France, Germany, and Scandinavian countries (García & Herrero, 2021), as well as Malaysia (Abobakr & Elgiziry, 2015; Purag et al., 2016) and Thailand (Detthamrong et al., 2017; Thakolwiroj & Sithipolvanichgul, 2021), generally find no significant relationship. These differences likely reflect variations in institutional maturity, governance systems, and investor protections. Kijkasiwat et al. (2022) further emphasize this contrast, showing a positive relationship in emerging economies and a negative or null effect in developed ones.

In the Algerian context, CEO duality is common, especially in family-owned and founder-led firms, where it is often seen as a symbol of stability, continuity, and a long-term vision, consistent with stewardship theory. These leaders tend to value family reputation and intergenerational legacy and often maintain strong political and financial networks. This facilitates easier access to long-term financing and allows them to manage capital structure in line with long-term strategic objectives.

While agency theory highlights that CEO duality may weaken board independence and reduce oversight effectiveness, in the Algerian institutional setting, it may enable managers to avoid short-term debt pressures and external constraints. This view does not contradict stewardship theory but rather complements it. Based on this theoretical integration and the specific Algerian context, we propose the following hypotheses:

H2a. 

CEO duality is positively associated with leverage in Algerian firms.

H2b. 

CEO duality is positively associated with debt maturity in Algerian firms.

2.3. Board Independence

From an agency theory perspective, board independence plays a central role in ensuring effective oversight and mitigating agency conflicts (Fama & Jensen, 1983). Independent directors, by virtue of their detachment from internal management, are expected to exercise more rigorous monitoring and better protect shareholder interests. Within this framework, debt (particularly short-term debt) can serve as a disciplinary substitute, used to control managerial discretion in the absence of effective governance (Jensen, 1986). Yet, the link between board independence and capital structure decisions remains highly sensitive to institutional context.

Empirical findings are mixed. In several emerging or developing economies, such as Ghana (Ahmed, 2019; Yakubu & Oumarou, 2023), Pakistan (Amin et al., 2022), and Saudi Arabia (Bazhair, 2023), studies report a positive relationship between board independence and leverage. These findings suggest that in weak governance environments, independent directors may support the use of debt, particularly of shorter maturity, as a compensatory mechanism to limit managerial opportunism.

In contrast, research from other emerging or developing markets, like Malaysia (Purag et al., 2016; Tahir et al., 2023), Nigeria (Sani et al., 2020), Thailand (Thakolwiroj & Sithipolvanichgul, 2021), and Egypt (Abdel-Wanis & Rashed, 2023), reports a negative association. In these cases, board independence appears to promote more prudent financial strategies, discouraging excessive reliance on debt. This suggests that effective monitoring by independent directors may reduce the firm’s need for debt as a control tool.

In developed countries, most studies find either a negative or statistically insignificant relationship. Kijkasiwat et al. (2022), for example, show that in strong institutional settings, independent boards tend to be associated with lower leverage, since other external governance mechanisms already constrain managerial behavior. Similarly, García and Herrero (2021) and Maurice et al. (2015) report no robust link between board independence and debt maturity, highlighting that firm-level governance quality often matters more than board composition alone.

Several other studies (e.g., Hermassi et al., 2015; Abobakr & Elgiziry, 2015; Detthamrong et al., 2017; Sewpersadh, 2019) also find no significant effect, reinforcing the idea that the influence of board independence depends less on its presence, and more on the broader governance environment.

In contexts like Algeria, where the number of independent directors on boards is generally low and governance practices are informal, limited board independence results in weak monitoring of management. This situation exacerbates information asymmetry between banks and firms, which restricts access to external financing, especially for companies lacking strong social capital or networks that could facilitate credit access. Consequently, firms with few independent directors often rely more heavily on debt, particularly short-term debt, as a substitute disciplinary mechanism to control managerial behavior.

However, as the proportion of independent directors increases, board oversight becomes more effective and credible, reducing the reliance on debt as a governance tool. Better monitoring lowers agency problems and enhances transparency, which also improves the firm’s relationship with lenders, facilitating access to longer-term and potentially less costly financing.

Therefore, the relationship between the proportion of independent directors and capital structure (both leverage and debt maturity) is expected to be nonlinear: initially positive due to limited oversight and high debt use, then negative as independent oversight improves and debt reliance decreases.

H3a. 

The relationship between the presence of independent directors and leverage in Algerian firms is nonlinear.

H3b. 

The relationship between the presence of independent directors and debt maturity in Algerian firms is nonlinear.

2.4. Gender Diversity

Female representation on corporate boards is widely recognized for promoting stronger oversight, higher ethical standards, and increased investor confidence. Together, these factors can improve firm performance and decision-making by fostering cognitive diversity, creativity, and innovation, consistent with resource dependence theory (Nekhili et al., 2016). According to agency theory, gender-diverse boards may better monitor managers, reduce agency costs, and influence financing choices.

However, empirical evidence on the relationship between gender diversity and leverage is mixed and depends on country context, governance quality, and market conditions. Alves et al. (2015), analyzing data from 33 developed and emerging economies, found that gender diversity enhances board independence and effectiveness, reduces information asymmetries between managers and external investors, and supports more balanced capital structures with higher equity ratios and less reliance on short-term debt.

In contrast, Elmoursy et al. (2025) find that in the UK, a developed but market-sensitive economy, gender diversity increases leverage, but mainly under stable market conditions. This suggests that the effect of female board representation on leverage varies across economic cycles, highlighting the importance of contextual factors.

Similarly, studies from developing countries like Ghana (Yakubu & Oumarou, 2023) and Nepal (Bajagai et al., 2019) report a positive link between female presence on boards and leverage. This is often explained by the stricter managerial discipline imposed by gender-diverse boards through higher indebtedness, compensating for weaker external governance in these institutional environments.

On the other hand, research from developed European countries often shows a negative relationship between female board representation and corporate leverage. García and Herrero (2021) and Martín-Ugedo and Mínguez-Vera (2023) find that more women on boards relate to lower leverage and stronger internal controls, consistent with agency theory. Likewise, Kijkasiwat et al. (2022) observe no significant effect in emerging markets but a negative association in developed economies, suggesting that the institutional environment influences this relationship.

Moreover, some studies find no significant impact of gender diversity on capital structure, particularly in countries with weak institutional frameworks. For example, Sani et al. (2020) in Nigeria, Detthamrong et al. (2017) and Thakolwiroj and Sithipolvanichgul (2021) in Thailand, and Bazhair (2023) in Saudi Arabia suggest that where governance weaknesses persist, gender diversity alone may not be enough to affect financing decisions.

The evidence on debt maturity is also mixed. Some studies (La Rocca et al., 2020; Askarany et al., 2025) indicate that female directors prefer short-term debt, valuing its flexibility and lower long-term risk. Others (Briozzo et al., 2019) find a preference for long-term debt, consistent with a strategy focused on stability. These conflicting findings point to possible moderating factors like institutional context, governance quality, and cultural norms that need further study.

In Algeria, female representation on boards remains low due to cultural barriers and the lack of gender parity laws. This limits women’s influence in strategic decisions and their access to business networks. As a result, firms with female board members may face more difficulty obtaining external financing, especially long-term loans. This often leads to conservative financial strategies relying more on internal funds or short-term debt. Additionally, where female presence remains below a critical mass, their representation risks being symbolic rather than substantially impacting governance or financial policy.

H4a. 

Female representation on the board is negatively associated with leverage in Algerian firms.

H4b. 

Female representation on the board is negatively associated with debt maturity in Algerian firms.

2.5. Nationality Diversity

Board nationality diversity is often linked to improved governance and transparency (Armstrong et al., 2010; Yousef et al., 2020). According to agency theory, debt, especially short-term debt, acts as a monitoring mechanism that reduces agency conflicts by requiring frequent refinancing and external oversight (Jensen, 1986). However, in the Algerian context, characterized by underdeveloped financial markets and family-owned private firms, access to external financing remains difficult and costly.

Given these financing constraints, Algerian private firms may rely more on long-term debt, which requires less frequent refinancing and is better suited to environments with limited credit availability. In this context, the presence of foreign directors, who are typically independent, plays a key governance role by substituting the monitoring function usually carried out by short-term debt. Their strict oversight can reduce managerial opportunism without relying heavily on debt-based control mechanisms.

As a result, foreign directors help strengthen board oversight and improve the firm’s legitimacy, which facilitates access to scarce long-term funding despite the challenging financial environment. Moreover, their international networks and expertise can support broader access to global financing options and contribute to the firm’s strategic growth, consistent with resource dependence theory (Pongelli et al., 2023). This dynamic encourages firms to prefer long-term over short-term debt, combining enhanced governance through board diversity with financing strategies adapted to local market constraints.

These observations align with Sani et al. (2020), who show that in emerging or developing economies, the presence of foreign board members is positively linked to higher corporate leverage. In contrast, Yousef et al. (2020), studying 3773 U.S. firms, find that greater nationality diversity on boards tends to reduce dependence on debt, highlighting the different dynamics between developed and developing markets.

H5a. 

Board nationality diversity is positively associated with leverage in Algerian firms.

H5b. 

Board nationality diversity is positively associated with debt maturity in Algerian firms.

2.6. Age Diversity

Board age diversity introduces a heterogeneous mix of experiences, cognitive frames, and strategic preferences that can shape financial decision-making, including capital structure choices. Younger board members tend to bring openness to innovation, higher risk tolerance, and familiarity with contemporary financial instruments, whereas older directors offer experience, prudence, and a long-term perspective, in line with resource dependence theory. This generational complementarity has the potential to enhance board deliberations and foster a more adaptive financial strategy.

Empirical studies have shown divergent findings regarding the impact of age diversity. Harris (2014), focusing on Fortune 500 firms, reported a modest but significant negative association between average board age and leverage, suggesting that older boards tend to adopt more conservative financing policies. In contrast, Nisiyama and Nakamura (2018) in Brazil, and Arioglu (2021) in Turkey, found that age diversity was positively associated with firms’ risk-taking behavior and debt levels, implying that generationally diverse boards are more inclined to support leveraged growth strategies.

Recent findings also highlight the complexity of the relationship between age diversity and board outcomes. Gardiner (2024) notes that the effect of age diversity on performance, and, by extension, on decisions such as capital structure, largely depends on internal contextual factors such as team dynamics, corporate culture, and the complexity of strategic tasks. In firms where intergenerational collaboration is encouraged or where innovation is central to the business model, age diversity can enhance strategic agility and foster more flexible financing approaches. Conversely, in more rigid or hierarchical environments, generational differences may create tensions that limit the benefits of age diversity.

In the Algerian context, older directors, often respected for their experience and seniority, may possess extensive social and professional networks that facilitate trusted relationships with financial institutions. This accumulated social capital can ease access to credit, particularly long-term debt, which aligns with their preference for financial stability and lower refinancing risk. Thus, the presence of senior board members can encourage longer debt maturities and reduce reliance on short-term, high-risk funding.

However, in Algerian firms where hierarchical culture and seniority norms prevail, younger board members may be seen as outsiders, limiting their influence and hindering collaboration. These intergenerational tensions can weaken board cohesion and compromise collective decision-making. As a result, the board’s ability to develop coherent debt maturity strategies or maintain stable financing policies may be impaired.

In this setting, firms may increasingly rely on short-term debt as a substitute governance mechanism, compensating for the absence of effective internal monitoring. Short-term debt, with its frequent refinancing and external scrutiny, serves as an informal control tool to address agency issues when internal governance is weakened by intergenerational conflict.

H6a. 

Board age diversity is negatively associated with leverage in Algerian firms.

H6b. 

Board age diversity is negatively associated with debt maturity in Algerian firms.

3. Methodology

3.1. Sample

This study is based on panel data from an initial sample of 360 Algerian firms covering the period 2015–2019. Financial information related to capital structure was collected from the “Sidjilcom” platform, maintained by the National Center for the Commercial Register, which is accessible via subscription. Data on board composition and characteristics were obtained directly from firms through internal documents and interviews with company officials.

Due to incomplete financial records for some firms and non-responses regarding board-related information, companies with missing data were excluded. As a result, the initial sample of 360 firms was reduced to a final balanced panel of 120 firms, generating 600 firm-year observations. This final sample represents a compromise between the practical limitations of manual data collection and the need to ensure sufficient data for achieving research objectives and producing a representative analysis.

A non-probability sampling method was employed, based on practical selection criteria, as used in prior corporate governance research (e.g., Wahyudi & Chairunesia, 2019). While such a method is useful in data-constrained environments, it may introduce selection bias and limit the generalizability of the findings.

In the Algerian context, restricted access to firm-level data remains a significant barrier. Consequently, previous studies have often relied on small samples or case studies. For example, Sadaoui and Khenniche (2015) examined five firms, Belmedjahed and Douah (2018) studied two insurance companies, and Cheurfi (2019) used a sample of 18 firms. In contrast, our study offers a more comprehensive and robust analysis by using a larger sample and a longer observation period.

3.2. Variables

3.2.1. Dependent Variables

Following the methodologies of Kim (2015), Sani et al. (2020), and García and Herrero (2021), this study measures capital structure using two dependent variables. Leverage, defined as the ratio of total debt to total assets, and Debt maturity, defined as the ratio of long-term debt (with a maturity of more than one year) to total debt, consistent with the approaches of Scherr and Hulburt (2001), Séverin (2012), and Kim (2015).

3.2.2. Independent Variables

Six board characteristics are examined as independent variables: board size, CEO duality, board independence, gender diversity, nationality diversity, and age diversity. The operational definitions and measurement methods for these variables are provided in

Table 1. Variables measurement.

Variables	Acronym	Measurement
Dependent Variables
Leverage	LEV	Ratio of total Debt to total Assets
Debt maturity	DM	Ratio of long-term debt to total debt
Independent Variables
Board size	BS	Natural logarithm of the number of directors on the board
CEO duality	DUA	Dummy variable taking the value 1 if the chairperson of the board is the CEO and 0 otherwise
Board Independence	IND	Ratio of independent directors to total directors on the board
Gender diversity	DGND	Ratio of women directors to total directors on the board
Nationality diversity	DNAT	Ratio of foreign directors to total directors on the board
Age diversity	DAGE	Coefficient of variation in directors’ age
Control Variables
Firm size	FS	Natural logarithm of total assets
Tangibility	TNG	Ratio of fixed assets to total assets
Growth opportunities	GO	Year-Over-Year sales growth
Liquidity	LIQ	Ratio of current assets to current liabilities
Family ownership	FAM	Dummy variable, equal to 1 when the company is family-owned, and 0 elsewhere
Firm age	AGE	Natural logarithm of the years since the creation of the company
Industry		Based on the nomenclature of economic activities adopted by the National Center for Commercial Register, four dummy variables were retained: (1) the goods production sector or artisanal production; (2) the wholesale distribution sector or retail distribution; (3) the services sector; and (4) the export sector or import for resale in the state.
Year	Y	Five dummy variables from 2015 to 2019

.

3.2.3. Control Variables

Control variables commonly recognized in the literature as determinants of capital structure are included in the analysis. These consist of firm size, firm age, ownership structure, asset tangibility (collateral), liquidity, and growth opportunities. In all regression models, we incorporate industry and year fixed effects to account for unobserved heterogeneity. Ignoring such heterogeneity may lead to biased parameter estimates.

3.3. Research Model

The analysis starts with a linear model to examine the relationship between board characteristics and capital structure. Subsequently, a nonlinear model is specified to capture potential curvilinear effects by including the squared terms of certain board attributes. This approach allows us to investigate whether these board characteristics have an optimal level that influences capital structure decisions. The empirical analysis employs the following autoregressive models:

Linear model (1):

$$\begin{array}{ll}C{S}_{it}=\alpha +\theta L.C{S}_{it}& +{\beta }_{1}B{S}_{it}+{\beta }_{2}DU{A}_{it}+{\beta }_3 IN{D}_{it}+{\beta }_{4}DGN{D}_{it}+{\beta }_{5 }DNA{T}_{it}\\ & +{\beta }_{6}DAG{E}_{it}+{\gamma }_{1}F{S}_{it}+{\gamma }_2 AG{E}_{it} +{\gamma }_{3}FA{M}_{it}+{\gamma }_{4}TN{G}_{it}\\ & +{\gamma }_{5 }LI{Q}_{it}+{\gamma }_{6 }G{O}_{it}+Indusry+Year+{\alpha }_i+{\epsilon }_{it}\end{array}$$

Nonlinear model (2):

$$\begin{array}{ll}C{S}_{it}=\alpha +\theta L.C{S}_{it}& +{\beta }_{1}B{S}_{it}+{\beta }_{2}B{S}_{it}^2+{\beta }_{3}DU{A}_{it}+{\beta }_4 IN{D}_{it}+{\beta }_{5}IN{D}_{it}^2\\ & +{\beta }_{6}DGN{D}_{it}+{\beta }_{7 }DNA{T}_{it}+{\beta }_{8}DAG{E}_{it}+{\gamma }_{1}F{S}_{it}+{\gamma }_2 AG{E}_{it} \\ & +{\gamma }_{3}FA{M}_{it}+{\gamma }_{4}TN{G}_{it}+{\gamma }_{5 }LI{Q}_{it}+{\gamma }_{6 }G{O}_{it}+Industry+Year\\ & +{\alpha }_i+{\epsilon }_{it}\end{array}$$

• $\alpha$: Intercept
• $C{S}_{it}$: The dependent variable (the capital structure) observed for individual i in period t
• $L.C{S}_{it}$: The dependent variable lagged one period ($C{S}_{it-1}$)
• θ, β, ϒ: The vectors of the coefficients of the lagged dependent variable, the independent variables, and the control variables, respectively.
• ${\alpha }_i$: The individual fixed effects.
• ${\epsilon }_{it}$: The error term.

3.4. Estimation Method

To address potential endogeneity, commonly encountered in corporate finance and governance research, we adopt the Generalized Method of Moments (GMM), which is widely regarded as the most suitable estimation technique in such contexts (Barros et al., 2020). In addition to confirming the presence of endogeneity via the Durbin–Wu–Hausman test, diagnostic checks revealed two further econometric issues in the dataset: heteroskedasticity and autocorrelation.

In line with Roodman (2009b), in the presence of heteroskedasticity, the two-step system GMM estimator, combined with robust standard errors and the Windmeijer (2005) finite sample correction, is more efficient than the one-step alternative. Therefore, this study employs the two-step system GMM estimator as the primary estimation method. For robustness and comparative purposes, we also report results from static OLS, dynamic OLS, and fixed effects models. Following Nguyen et al. (2014), such comparisons help assess the validity and robustness of the GMM specification. Moreover, as suggested by Bond (2002), the coefficient of the lagged dependent variable estimated using system GMM should lie between the OLS estimate (which is usually upward biased) and the fixed effects estimate (which is usually downward biased). This provides an additional validity check for the GMM estimator. For this reason, we report the results from these alternative estimators as a benchmark.

To further validate the GMM results, we apply the Arellano and Bond (1991) tests for first-order (AR(1)) and second-order (AR(2)) serial correlation in the differenced residuals. As per standard expectations, AR(1) should be present, while AR(2) should not. Furthermore, the number of instruments used must be lower than the number of cross-sectional units (firms). Finally, we employ the Hansen test for overidentifying restrictions to assess the validity of the instrument set. A valid model is typically indicated by a p-value between 0.05 and 0.80. However, Roodman (2009b) recommends a more conservative range of 0.10 to 0.25. Some of our results exceed this range but remain well within the broader acceptable interval, supporting the validity of the instruments in our estimations.

4. Results and Discussion

4.1. Descriptive Statistics

Table 2. Descriptive statistics.

Variables	Obs	Mean	Std. Dev.	Min	Max	Median
Leverage	600	0.532	0.255	0.01	0.971	0.54
Debt maturity	600	0.29	0.271	0	0.973	0.225
Board size	600	1.473	0.311	1.099	2.079	1.609
CEO Duality	600	0.428	0.495	0	1	0
Board independence	600	0.471	0.197	0	0.833	0.5
Gender diversity	600	0.151	0.194	0	0.667	0
Nationality diversity	600	0.129	0.238	0	0.833	0
Age diversity	600	0.13	0.052	0.008	0.303	0.125
Firm size	600	21.528	1.605	16.923	26.601	21.558
Firm age	600	2.82	0.492	0	4.22	2.833
Family ownership	600	0.75	0.433	0	1	0
Tangibility	600	0.26	0.214	0.002	0.9	0.205
Liquidity	600	3.027	3.37	0.159	24.555	1.841
Growth opportunities	600	0.116	0.416	−0.996	2.205	0.072

presents descriptive statistics for the variables related to board characteristics, capital structure, and firm-specific attributes in the sample. The average total debt ratio (LEV) is 53%, which is comparable to the 44% reported for Korean firms (Kim, 2015) and the 57% for Chinese firms (Li et al., 2009). This suggests that Algerian firms generally operate with relatively high levels of debt. A plausible explanation is the dominance of the banking sector and the underdevelopment of capital markets in Algeria, which limits firms’ access to equity financing and leads to a greater reliance on bank loans.

Regarding debt maturity (DM), the average value is 29%, closely aligning with the 31.2% reported for Korean firms by Kim (2015). This indicates that Algerian firms tend to favor short-term debt, potentially as a governance tool to reduce agency costs through frequent monitoring by creditors.

In terms of board composition, Algerian corporate boards are relatively small, averaging approximately 4.5 members. On average, 47.1% of board members are independent, reflecting a relatively high proportion of outside directors. CEO duality, where the CEO also serves as the board chair, occurs in 42.8% of firms, suggesting a moderate level of role separation.

Board diversity in Algeria remains limited. Gender diversity, nationality diversity, and age diversity average 15.1%, 12.9%, and 13%, respectively, indicating a low degree of heterogeneity in board composition.

The sample is composed of firms from various sectors, with the production sector representing 53% of the sample, followed by services at 29%, foreign trade at 13%, and distribution at 5%. We include these sectors as control variables in our estimations to account for potential sector-specific effects on the relationship between board characteristics and capital structure. Controlling for industry ensures that our results capture the economic heterogeneity across sectors, enhancing the robustness and generalizability of our findings.

4.2. Preliminary Analysis

Prior to analysis, a preliminary data processing stage was conducted. Winsorization at the 5th and 95th percentiles was applied, particularly for variables such as growth opportunities, to reduce the influence of outliers. A logarithmic transformation was applied to the liquidity variable to improve normality and linearity. Additionally, we conducted standard diagnostic tests related to classical linear regression assumptions, including stationarity, normality, multicollinearity (VIF), heteroskedasticity (Breusch-Pagan/Cook-Weisberg), autocorrelation (Wooldridge), and endogeneity (Durbin-Wu-Hausman).

Table 3. Correlation matrix.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)
(1) LEV	1.000
(2) DM	0.037	1.000
(3) BS	−0.068 *	0.192 ***	1.000
(4) DUA	0.082 **	0.121 ***	0.043	1.000
(5) IND	0.043	0.059	0.486 ***	−0.055	1.000
(6) DGND	−0.013	−0.092 **	−0.042	−0.194 ***	0.024	1.000
(7) DNAT	0.091 **	−0.259 ***	−0.088 **	−0.232 ***	0.220 ***	−0.050	1.000
(8) DAGE	−0.125 ***	0.002	0.185 ***	0.003	−0.011	−0.019	−0.061	1.000
(9) FS	0.090 **	0.214 ***	0.233 ***	−0.044	0.156 ***	−0.110 ***	−0.002	0.102 **	1.000
(10) AGE	−0.070 *	0.095 **	0.097 **	0.083 **	0.107 ***	−0.115 ***	−0.104 **	0.126 ***	0.124 ***	1.000
(11) FAM	0.070 *	0.139 ***	0.011	0.146 ***	0.083 **	−0.181 ***	−0.129 ***	−0.106 ***	−0.019	0.043	1.000
(12) TNG	−0.068 *	0.403 ***	0.052	0.081 **	−0.058	−0.051	−0.094 **	0.009	0.147 ***	−0.072 *	0.023	1.000
(13) lnLIQ	−0.613 ***	0.320 ***	0.024	−0.008	0.058	−0.039	−0.117 ***	−0.008	−0.083 **	0.170 ***	−0.009	−0.133 ***	1.000
(14) GO_w	0.213 ***	0.048	0.007	0.012	0.009	0.035	0.018	−0.027	0.050	−0.058	−0.043	0.018	−0.163 ***	1.000

*** p < 0.01, ** p < 0.05, * p < 0.1.

presents the Pearson correlation matrix for all variables. Several statistically significant correlations (p-value < 0.01) are observed among the independent variables. However, all correlation coefficients remain below the widely accepted threshold of 0.8 (Gujarati, 2003), suggesting that multicollinearity is unlikely to pose a serious issue.

This conclusion is reinforced by the variance inflation factor (VIF) and tolerance statistics in

Table 4. VIF test.

Variables	VIF	1/VIF
BS	1.501	0.666
DUA	1.131	0.884
IND	1.543	0.648
DGND	1.129	0.886
DNAT	1.26	0.794
DAGE	1.083	0.923
FS	1.135	0.881
AGE	1.109	0.902
FAM	1.107	0.903
TNG	1.069	0.936
lnLIQ	1.109	0.902
GO	1.035	0.966
Mean VIF	1.184	.

. The highest VIF value is 1.543, well below the conventional threshold of 10. Tolerance values range from 0.648 to 0.966, confirming that multicollinearity does not compromise the reliability of the regression estimates.

4.3. Multivariate Analysis

4.3.1. Linear Model

Table 5. Linear model results.

Variables	Model 1(a) LEV				Model 1(a’) LEV	Model 1(b) DM				Model 1(b’) DM
	Static	Dynamic				Static	Dynamic
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
	OLS	OLS	FE	GMM-Sys	GMM-Sys	OLS	OLS	FE	GMM-Sys	GMM-Sys
L.LEV		0.850 ***	0.242 **	0.683 ***	0.647 ***
		(0.024)	(0.105)	(0.115)	(0.129)
L.DM							0.794 ***	0.092	0.692 ***	0.718 ***
							(0.033)	(0.065)	(0.114)	(0.095)
BS	−0.098 ***	−0.042 ***	0.028	−0.136 *	−0.194 *	0.099 ***	0.019	0.024	−0.221 *	−0.180 *
	(0.032)	(0.015)	(0.036)	(0.069)	(0.106)	(0.035)	(0.022)	(0.063)	(0.120)	(0.097)
DUA	0.054 ***	0.007	0.066 *	0.030	0.127	0.004	0.005	0.128 ***	0.111	0.043
	(0.017)	(0.009)	(0.037)	(0.060)	(0.105)	(0.019)	(0.012)	(0.033)	(0.109)	(0.096)
IND	0.158 ***	0.055 **	0.039	−0.021	0.194	0.024	0.016	0.132	0.156	−0.361
	(0.047)	(0.022)	(0.057)	(0.121)	(0.170)	(0.053)	(0.038)	(0.092)	(0.230)	(0.349)
DGND	−0.040	−0.011	−0.034	0.073	0.188	−0.047	−0.056	−0.092 *	−0.008	0.140
	(0.040)	(0.021)	(0.045)	(0.067)	(0.133)	(0.046)	(0.035)	(0.053)	(0.101)	(0.201)
DNAT	−0.014	−0.041 **	−0.150	−0.109	−0.021	−0.181 ***	−0.075 ***	0.111	−0.163 *	−0.339 **
	(0.036)	(0.019)	(0.093)	(0.077)	(0.158)	(0.036)	(0.026)	(0.091)	(0.082)	(0.163)
DAGE	−0.580 ***	−0.074	0.324	−0.109	0.868	−0.078	−0.080	−0.135	0.087	−0.466
	(0.175)	(0.075)	(0.213)	(0.559)	(0.868)	(0.178)	(0.145)	(0.354)	(0.681)	(1.076)
FS	0.010 **	0.002	0.088 ***	0.070 **	0.060 *	0.024 ***	0.004	0.111 ***	0.029	0.017
	(0.005)	(0.002)	(0.031)	(0.033)	(0.032)	(0.005)	(0.004)	(0.038)	(0.029)	(0.016)
AGE	0.021	0.012	−0.043	−0.018	−0.001	−0.011	−0.007	−0.102	−0.008	−0.010
	(0.019)	(0.008)	(0.049)	(0.027)	(0.035)	(0.019)	(0.015)	(0.074)	(0.032)	(0.035)
FAM	−0.025	−0.004		0.088	0.197	−0.078 ***	−0.014		0.095	−0.303
	(0.021)	(0.009)		(0.151)	(0.189)	(0.021)	(0.016)		(0.197)	(0.317)
TNG	−0.183 ***	−0.007	−0.010	−0.116	0.022	0.457 ***	0.094 **	0.155 *	0.177	0.139
	(0.043)	(0.019)	(0.045)	(0.140)	(0.127)	(0.059)	(0.041)	(0.087)	(0.122)	(0.138)
lnLIQ	−0.188 ***	−0.040 ***	−0.064 ***	−0.055 **	−0.058 *	0.123 ***	0.036 ***	0.138 ***	0.072 *	0.052 *
	(0.011)	(0.007)	(0.017)	(0.023)	(0.029)	(0.014)	(0.009)	(0.022)	(0.039)	(0.026)
GO	0.086 ***	0.050 ***	0.025 **	0.007	0.044	0.082 ***	0.005	0.009	0.105	0.073
	(0.025)	(0.015)	(0.011)	(0.045)	(0.031)	(0.024)	(0.019)	(0.018)	(0.079)	(0.059)
FAM*BS					0.092					0.060
					(0.150)					(0.117)
FAM*DUA					−0.116					−0.034
					(0.126)					(0.134)
FAM*IND					−0.239					0.324
					(0.252)					(0.339)
FAM*DGND					−0.096					−0.208
					(0.152)					(0.208)
FAM*DNAT					−0.081					0.199
					(0.200)					(0.172)
FAM*DAGE					−1.340					0.833
					(0.868)					(1.368)
Constant	0.650 ***	0.108 *	−1.430 **	−1.080 *	−1.043 *	−0.672 ***	−0.067	−2.101 ***	−0.390	0.149
	(0.121)	(0.056)	(0.553)	(0.610)	(0.599)	(0.120)	(0.102)	(0.755)	(0.526)	(0.449)
Industry	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Year	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Observations	600	480	480	480	480	600	480	480	480	480
R-squared	0.464	0.904	0.390			0.430	0.790	0.353
F-statistic (p-value)	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
No. Groups			120	120	120			120	120	120
No. Instruments				50	62				52	100
AR (1) (p-value)				0.001	0.000				0.002	0.002
AR (2) (p-value)				0.518	0.486				0.530	0.402
Hansen (p-value)				0.638	0.791				0.685	0.287
Difference-in-Hansen for GMM instruments (p-value)				0.421	0.417				0.119	0.097

Note: The table shows the regression coefficients and corrected standard errors (in parentheses). *, **, *** represent the 1%, 5% and 10% significance level, respectively. The results are obtained using various estimation methods, including static OLS, dynamic OLS, fixed effects, and two-step system GMM. Among these, the system GMM estimator is considered the most robust, as it effectively addresses potential endogeneity, unobserved heterogeneity, and autocorrelation.

reports the results of the linear regressions. Columns 1–4 present estimates for the leverage model (Model 1(a)), while columns 6–9 report estimates for the debt maturity model (Model 1(b)). The lagged dependent variables in both models are statistically significant at the 1% level, with coefficients of 0.683 for lagged leverage (L.LEV) and 0.692 for lagged debt maturity (L.DM), confirming the presence of dynamic endogeneity.

In Model 1(a), board size (BS) exhibits a negative and statistically significant effect on leverage (coefficient = −0.136, p < 0.1). This finding is consistent with prior evidence from European firms (Dimitropoulos, 2014) and South African firms (Sewpersadh, 2019). Other board characteristics, CEO duality (DUA), director independence (IND), gender diversity (DGND), nationality diversity (DNAT), and age diversity (DAGE), do not have statistically significant effects on leverage. Among the control variables, the log of liquidity (lnLIQ) is negatively associated with leverage (coefficient = −0.055, p < 0.05), while firm size (FS) has a positive but marginally significant effect (coefficient = 0.070, p < 0.1).

For Model 1(b), the results show that lnLIQ positively affects debt maturity, with a coefficient of 0.072 (p < 0.1). This suggests that more liquid firms tend to rely more on long-term debt.

In this study, the moderating effect of ownership structure on the relationship between board characteristics and capital structure was examined by introducing interaction terms into Models 1(a) and 1(b), resulting in Models 1(a’) and 1(b’), respectively. The interaction terms, FAM*BS, FAM*DUA, FAM*IND, FAM*DGND, FAM*DNAT, FAM*DAGE, account for potential differences between family-owned and non-family-owned firms. Results presented in columns 5 and 10 of Table 5 indicate that none of these interaction terms are statistically significant. This implies that the ownership structure does not significantly influence the relationship between board characteristics and either leverage or debt maturity. Therefore, the findings from Models 1(a) and 1(b) are robust across both ownership types.

4.3.2. Nonlinear Model

Table 6. Nonlinear model results.

Variables	Model 2(a) LEV				Model 2(b) DM
	Static	Dynamic			Static	Dynamic
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	OLS	OLS	FE	GMM-Sys	OLS	OLS	FE	GMM-Sys
L.LEV		0.849 ***	0.241 **	0.600 **
		(0.024)	(0.105)	(0.239)		0.785 ***	0.093	0.530 ***
L.DM						(0.033)	(0.065)	(0.142)
BS	0.786 ***	0.030	0.330	−2.169 *	1.331 ***	0.475 **	0.237	3.271 **
	(0.291)	(0.143)	(0.308)	(1.283)	(0.356)	(0.236)	(0.435)	(1.271)
BS2	−0.300 ***	−0.026	−0.105	0.731 *	−0.423 ***	−0.157 **	−0.074	−1.156 **
	(0.098)	(0.048)	(0.097)	(0.423)	(0.120)	(0.079)	(0.134)	(0.443)
DUA	0.064 ***	0.006	0.073 **	0.024	0.016	0.008	0.139 ***	−0.007
	(0.018)	(0.009)	(0.036)	(0.149)	(0.020)	(0.012)	(0.032)	(0.116)
IND	0.043	−0.073	−0.081	−0.064	−0.394 **	−0.191	−0.124	−0.604
	(0.150)	(0.070)	(0.108)	(0.335)	(0.164)	(0.121)	(0.167)	(0.823)
IND2	0.136	0.153 *	0.150	−0.004	0.505 ***	0.246 *	0.316	0.716
	(0.190)	(0.087)	(0.137)	(0.397)	(0.190)	(0.139)	(0.194)	(0.835)
DGND	−0.050	−0.004	−0.029	0.115	−0.046	−0.052	−0.090	−0.052
	(0.040)	(0.022)	(0.049)	(0.115)	(0.046)	(0.036)	(0.055)	(0.106)
DNAT	−0.031	−0.044 **	−0.161	−0.067	−0.211 ***	−0.087 ***	0.113	−0.294 **
	(0.037)	(0.019)	(0.099)	(0.182)	(0.036)	(0.027)	(0.090)	(0.126)
DAGE	−0.553 ***	−0.072	0.343	−0.745	−0.044	−0.066	−0.045	0.820
	(0.174)	(0.075)	(0.235)	(0.752)	(0.178)	(0.144)	(0.372)	(0.899)
FS	0.012 ***	0.002	0.087 ***	0.047	0.028 ***	0.006	0.108 ***	0.032
	(0.005)	(0.002)	(0.031)	(0.075)	(0.005)	(0.004)	(0.041)	(0.023)
AGE	0.023	0.012	−0.017	−0.023	−0.010	−0.006	−0.095	0.001
	(0.019)	(0.008)	(0.056)	(0.042)	(0.019)	(0.015)	(0.091)	(0.048)
FAM	−0.032	−0.007		0.297	−0.095 ***	−0.023		−0.117
	(0.021)	(0.009)		(0.400)	(0.021)	(0.016)		(0.155)
TNG	−0.194 ***	−0.007	−0.010	0.053	0.444 ***	0.094 **	0.154 *	0.123
	(0.043)	(0.020)	(0.045)	(0.183)	(0.058)	(0.041)	(0.087)	(0.224)
lnLIQ	−0.188 ***	−0.040 ***	−0.065 ***	−0.047	0.125 ***	0.038 ***	0.138 ***	0.024
	(0.011)	(0.007)	(0.017)	(0.029)	(0.013)	(0.009)	(0.022)	(0.040)
GO	0.084 ***	0.049 ***	0.025 **	0.082 *	0.080 ***	0.004	0.007	0.068
	(0.025)	(0.015)	(0.011)	(0.049)	(0.023)	(0.019)	(0.019)	(0.057)
Constant	−0.001	0.069	−1.690 ***	0.984	−1.585 ***	−0.392 **	−2.189 **	−2.775 **
	(0.235)	(0.117)	(0.583)	(1.256)	(0.293)	(0.199)	(0.974)	(1.110)
Industry	YES	YES	YES	YES	YES	YES	YES	YES
Year	YES	YES	YES	YES	YES	YES	YES	YES
Observations	600	480	480	480	600	480	480	480
R-squared	0.471	0.904	0.399		0.445	0.792	0.358
F-statistic (p-value)	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
No. Groups			120	120			120	120
No. Instruments				49				69
AR (1) (p-value)				0.011				0.009
AR (2) (p-value)				0.548				0.407
Hansen (p-value)				0.661				0.501
Difference-in-Hansen for GMM instruments (p-value)				0.387				0.348

Note: The table shows the regression coefficients and corrected standard errors (in parentheses). *, **, *** represent the 1%, 5%, and 10% significance level, respectively. The results are obtained using various estimation methods, including static OLS, dynamic OLS, fixed effects, and two-step system GMM. Among these, the system GMM estimator is considered the most robust, as it effectively addresses potential endogeneity, unobserved heterogeneity, and autocorrelation.

presents the results from the nonlinear models of capital structure, Model 2(a) for total debt and Model 2(b) for debt maturity, estimated using the system GMM method.

Model 2(a) reveals a U-shaped relationship between board size (BS) and leverage. The coefficient for BS (β₁) is negative and significant at the 10% level (β₁ = −2.169), while the squared term BS² (β₂) is positive and significant at the same level (β₂ = 0.731). This suggests that leverage initially decreases by increasing board size but begins to rise beyond a certain point. The turning point occurs at approximately 4.39 board members (calculated as −β₁/(2*β₂) = 1.48 in logarithmic terms).

In Model 2(b), the relationship between board size and debt maturity is found to be inverted U-shaped. The coefficient for BS is positive and significant at the 5% level (β₁ = 3.271), while BS² is negative and significant (β₂ = −1.156), indicating that debt maturity increases with board size up to an optimal level, after which it declines. The turning point is approximately 4.1 members.

Additionally, some findings from the linear model are replicated in the nonlinear analysis. Notably, nationality diversity (DNAT) continues to exhibit a negative and statistically significant effect on debt maturity (coefficient = −0.294, p < 0.05). However, the positive relationship between lnLIQ and debt maturity is no longer statistically significant in the nonlinear specification.

Figure 1 represents the fitted values from the nonlinear specification (quadratic form) estimated using System GMM.

4.4. Robustness Tests

A key concern when using the GMM estimator is the proliferation of instruments, which can bias coefficient estimates and weaken the Hansen test’s ability to assess instrument validity (Roodman, 2009a). To address this, we follow the recommendations of Roodman (2009a) by

• Restricting the number of lags used to instrument endogenous variables (one or two lags).
• Applying the ‘collapse’ option, as recommended by Roodman (2009b), to further reduce the instrument count.

We also pay particular attention to the validity of instruments through the Hansen test of overidentifying restrictions. Across all model specifications, the Hansen test results indicate that the null hypothesis of instrument validity cannot be rejected, suggesting that our instruments are appropriate and do not overfit the endogenous variables. Furthermore, we check for second-order serial correlation using the Arellano–Bond test, and the results show no evidence of autocorrelation in the differenced residuals, confirming the consistency of the GMM estimator.

Table 7. Robustness tests.

Linear Model					Nonlinear Model
Variables	(1) Model 1(a) LEV	(2) Model 1(b) DM	(3) Model 1(a’) LEV	(4) Model 1(b’) DM	Variables	(5) Model 2(a) LEV	(6) Model 2(b) DM
L.LEV	0.644 ***		0.652 ***		L.LEV	0.650 **
	(0.113)		(0.129)			(0.249)
L.DM		0.663 ***		0.693 ***	L.DM		0.496 ***
		(0.138)		(0.088)			(0.147)
BS	−0.147 **	−0.284 **	−0.197 *	−0.190 *	BS	−2.231 *	4.172 **
	(0.071)	(0.128)	(0.103)	(0.108)		(1.322)	(1.663)
DUA	0.050	0.128	0.122	0.020	BS2	0.758 *	−1.457 **
	(0.075)	(0.116)	(0.102)	(0.107)		(0.438)	(0.580)
IND	−0.011	0.300	0.201	−0.358	DUA	0.048	0.041
	(0.115)	(0.278)	(0.167)	(0.361)		(0.134)	(0.127)
DGND	0.083	0.049	0.180	0.091	IND	−0.190	−0.450
	(0.059)	(0.107)	(0.128)	(0.210)		(0.371)	(0.759)
DNAT	−0.102	−0.165 *	−0.011	−0.341 **	IND2	0.085	0.475
	(0.084)	(0.083)	(0.157)	(0.168)		(0.417)	(0.943)
DAGE	−0.256	0.150	0.869	−0.582	DGND	0.123	−0.003
	(0.552)	(1.004)	(0.829)	(1.284)		(0.105)	(0.108)
FS	0.071 *	0.029	0.060 *	0.015	DNAT	−0.004	−0.244 *
	(0.040)	(0.025)	(0.032)	(0.016)		(0.173)	(0.139)
AGE	−0.012	−0.006	0.000	−0.014	DAGE	−0.688	0.829
	(0.033)	(0.035)	(0.036)	(0.036)		(0.774)	(0.911)
FAM	0.106	0.096	0.189	−0.342	FS	0.036	0.037
	(0.194)	(0.199)	(0.188)	(0.307)		(0.067)	(0.026)
TNG	−0.123	−0.017	0.018	0.165	AGE	−0.022	−0.016
	(0.161)	(0.187)	(0.130)	(0.166)		(0.040)	(0.057)
lnLIQ	−0.064 **	0.065*	−0.057 *	0.056 *	FAM	0.220	−0.236
	(0.032)	(0.039)	(0.029)	(0.030)		(0.371)	(0.155)
GO	0.004	0.115	0.042	0.074	TNG	0.065	0.047
	(0.041)	(0.075)	(0.030)	(0.057)		(0.191)	(0.196)
					lnLIQ	−0.030	0.088
						(0.031)	(0.059)
					GO	0.103*	0.068
						(0.053)	(0.077)
Constant	−1.086	−0.305	−1.054 *	0.242	Constant	1.196	−3.540 **
	(0.772)	(0.457)	(0.598)	(0.449)		(1.194)	(1.357)
Observations	480	480	480	480	Observations	480	480
No. Groups	120	120	120	120	No. Groups	120	120
Instruments	46	41	61	95	Instruments	46	65
AR (1)	0.001	0.001	0.000	0.002	AR (1)	0.009	0.012
AR (2)	0.529	0.591	0.494	0.394	AR (2)	0.596	0.304
Hansen	0.698	0.573	0.748	0.265	Hansen	0.575	0.565

Note: The table shows the regression coefficients and corrected standard errors (in parentheses). *, **, *** represent the 1%, 5% and 10% significance level, respectively.

presents the robustness test results, which confirm that the primary findings are stable. Statistically significant relationships observed in both linear and nonlinear models generally remain significant, with only minor changes in significance levels. The signs and magnitudes of the estimated coefficients are largely consistent with those reported earlier, reinforcing the robustness and reliability of our results to instrument reduction.

4.5. Discussion

This study investigates how board characteristics influence capital structure decisions in Algerian firms. The findings provide several important insights.

First, the linear model reveals a significant negative association between board size and both leverage and debt maturity. These results align with evidence from other emerging or developing economies such as Ghana (Ahmed, 2019), Nigeria (Sani et al., 2020), and Saudi Arabia (Bazhair, 2023), where smaller boards are associated with more conservative financial policies. In contrast, the negative relationship between board size and debt maturity contrasts with findings in developed markets, such as France and Europe (Maurice et al., 2015; García & Herrero, 2021), where no significant link has been observed.

The nonlinear models offer additional insights. A U-shaped relationship between board size and leverage, and an inverted U-shaped relationship between board size and debt maturity, support hypotheses H1a and H1b. These patterns suggest that both very small and very large boards may be less effective in managing capital structure decisions. Smaller boards may reinforce control over managers, as proposed by agency theory, whereas larger boards may suffer from coordination challenges that reduce monitoring efficiency, prompting firms to rely on debt as a substitute for weak internal governance. Moderately sized boards thus appear to strike a better balance between control and decision-making.

Regarding debt maturity, the inverted U-shaped relationship suggests that firms with smaller boards tend to favor long-term debt, possibly because effective oversight reduces reliance on short-term debt as a disciplinary mechanism. Conversely, larger boards may suffer from a lack of cohesion, weakening their monitoring capacity and prompting firms to shift toward short-term debt to enforce financial discipline.

Among the other board attributes examined, only nationality diversity shows a statistically significant effect, specifically, a negative association with debt maturity. A possible explanation is that foreign directors, by enhancing board control, may favor short-term debt to increase discipline in contexts with weak external governance.

Other board characteristics, CEO duality, board independence, gender diversity, and age diversity, were not significantly related to either leverage or debt maturity. Therefore, hypotheses H2a, H2b, H3a, H3b, H4a, H4b, H6a, and H6b are not supported. Regarding nationality diversity, H5a is rejected due to the lack of significance for leverage, while H5b is also rejected as the effect on debt maturity, although significant, is in the opposite direction from what was expected.

These results call into question the universal applicability of conventional governance models in emerging markets. Weak institutional enforcement, symbolic board independence, and limited director authority may reduce the effectiveness of formal governance mechanisms in such settings.

From a theoretical standpoint, the findings offer partial support for agency theory, particularly in illustrating how debt may serve as a substitute for internal governance mechanisms when board oversight is weak. In underdeveloped financial systems like Algeria’s, where external governance mechanisms are limited, firms may strategically use capital structure, especially short-term debt, as a disciplinary tool. However, the mixed nature of the findings highlights the limits of applying universal governance frameworks to transitional economies without adaptation to institutional realities.

From a practical perspective, these results underscore the need for context sensitive governance reforms. Policymakers should avoid directly replicating governance practices from developed economies and instead prioritize locally relevant strategies that strengthen internal control and enhance board effectiveness. Future research should also explore informal governance mechanisms, such as family ownership, political connections, and relational ties, that may exert significant influence on financial decisions in emerging markets.

5. Conclusions

This study examined how board characteristics influence capital structure decisions in Algerian firms over the 2015–2019 period. Using the Generalized Method of Moments (GMM) to address endogeneity concerns, the analysis revealed significant nonlinear effects of board size on both leverage and debt maturity. It also found that nationality diversity on the board plays an important role in shaping firms’ financing behavior.

To the best of our knowledge, this is the first study of its kind in the Algerian context, employing a relatively large panel of firms to investigate these relationships in detail. By focusing on a transitional economy with underdeveloped financial markets and limited external governance mechanisms, the study provides new insights into how internal governance structures operate in environments where external oversight is weak or ineffective.

A central contribution of this research is the identification of an optimal board size threshold, beyond which board effectiveness declines. This nonlinear effect remains largely unexplored in the governance-capital structure literature, particularly within emerging and transitional economies. While studies in developed countries often emphasize board independence or gender diversity (e.g., García & Herrero, 2021; Elmoursy et al., 2025), our findings suggest that in Algeria, factors such as board size and nationality are more relevant to financial decision-making.

Although Algeria shares characteristics with other emerging markets, such as concentrated ownership and underdeveloped financial systems, its governance dynamics are strongly influenced by specific legal, cultural, and institutional factors. For example, unlike Detthamrong et al. (2017), who report no significant relationship between governance and leverage in Thailand, or Kijkasiwat et al. (2022), who finds that larger boards are associated with lower leverage in emerging markets, our results reveal a distinct nonlinear pattern. Furthermore, compared to MENA countries such as Egypt (Abdel-Wanis & Rashed, 2023), and Bahrain (Alabdullah & Mohamed, 2023), where board characteristics often show a positive association with leverage, Algerian firms exhibit different dynamics. These differences underscore the need for a context-specific understanding of governance mechanisms.

By empirically demonstrating how traditional governance mechanisms operate differently in the Algerian setting, this study not only fills a gap in the regional literature but also contributes to the broader debate on the transferability of governance models across institutional environments.

From a practical perspective, the findings offer useful implications for policymakers and financial regulators. For instance, the Algerian securities regulator (COSOB) could use these insights to revise governance guidelines, placing greater emphasis on board effectiveness and diversity, rather than simply replicating international norms. Such reforms could help enhance investor confidence and support the development of Algeria’s underdeveloped stock market. Similarly, the Algerian Institute of Corporate Governance “Hawkama El Djazaïr”, which updates the national governance code, could consider reforms that reflect local realities, especially since traditional mechanisms like CEO duality separation and board independence seem to have little effect in the Algerian context. Financial institutions may also benefit from including board governance indicators in their credit assessments to better evaluate borrower risk and encourage stronger governance practices.

However, this study has limitations. The final sample was reduced to 120 firms due to missing data, which may introduce sample selection bias and limit the generalizability of the results. This limitation reflects broader issues of data availability and corporate transparency in many emerging economies, including Algeria.

While this study focused on demographic board characteristics, future research could explore other dimensions, such as cognitive traits, managerial expertise, board processes, and the presence of audit or strategy committees. These aspects could offer a more complete understanding of how boards influence financial decision-making in emerging and transitional economies.

In conclusion, this study highlights the critical role of board composition, especially board size and diversity, in influencing capital structure decisions in contexts with weak institutional environments. Improving internal governance, adapted to local conditions, could be an effective way to support better financial decisions and promote more sustainable corporate financing strategies.