The Moderating Role of Board Characteristics in the Relationship Between CSR and Bank Stability: Evidence from MENA Banks

Abstract

This study uses a dataset of conventional banks from 2010 to 2022 to investigate the moderating effect of board characteristics (BC) on the relationship between corporate social responsibility (CSR) and bank stability in the MENA region. Bank stability is measured using the Z-ROA index, which captures a bank’s ability to withstand financial shocks. The study addresses endogeneity and heterogeneity concerns using the system generalized method of moments (SGMM), with diagnostic tests confirming the validity of instruments and the absence of second-order autocorrelation. Three main conclusions are presented. First, CSR has a major detrimental impact on bank stability, indicating that when poorly managed or misaligned with strategic objectives, CSR initiatives may weaken financial resilience. Second, board attributes such as independence, diversity, and experience have a positive impact on bank stability, highlighting the importance of sound governance in ensuring prudent financial management. Third, the interaction between CSR and board characteristics exerts a positive and significant influence on bank stability, suggesting that well-structured boards can enhance the strategic value of CSR initiatives. As a robustness check, the study re-estimates the model using non-performing loans (NPLs) as an alternative measure of bank stability. The results remain consistent with the baseline findings, confirming the robustness and credibility of the conclusions. CSR continues to show a positive association with NPLs, while board characteristics and their interaction with CSR maintain negative and significant effects. These findings reinforce that effective board governance can transform CSR practices into stability-enhancing strategies. For policymakers and banking executives seeking to integrate sustainability into governance frameworks, the results underscore the crucial role of corporate governance in translating CSR efforts into tangible stability outcomes. The study calls for greater regulatory focus on board structures to maximize the stability benefits of CSR in the banking sector, contributing to the growing body of research on CSR and financial stability in developing economies.

1. Introduction

The link between bank stability and CSR has attracted growing academic attention in recent years, particularly in the aftermath of financial crises and rising stakeholder expectations for responsible and sustainable banking behavior (Cornett et al., 2016). CSR is argued to support bank stability through increased reputation, risk minimization, and improved stakeholder relationships, ultimately enhancing long-term financial performance (Beck et al., 2013). However, the extent to which these benefits materialize often depends on a bank’s internal governance structure, particularly the design and effectiveness of its board of directors (García-Sánchez et al., 2018).

Banks differ from non-financial firms as systemically important institutions subject to strong regulatory oversight and stakeholder scrutiny. For this reason, the CSR and bank stability relationship is especially relevant in the banking sector. The issue becomes even more important in the MENA region, where banks act as primary financiers and play a central role in maintaining economic resilience. Many MENA economies are also undergoing institutional reform and financial sector modernization, increasingly emphasizing corporate governance and CSR as mechanisms to promote transparency, risk management, and sustainable growth (Makhlouf et al., 2017; Luo & Bhattacharya, 2009). In such an environment, CSR functions not only as reputational insurance but also as a strategic alignment mechanism between banks and broader social and environmental objectives.

The MENA region provides a compelling setting to examine the CSR and stability relationship due to its economic, institutional, and political diversity. The region contains both oil-rich and emerging economies, leading to substantial variation in institutional quality, regulatory frameworks, and governance practices. Many MENA countries have also experienced political change and social unrest, which heightens the need for financial stability and corporate accountability. At the same time, growing pressure from international investors and global institutions is pushing banks to adopt CSR and ESG practices aligned with international standards. However, CSR in the region remains at an evolutionary stage, often driven by external pressures or image concerns rather than deeply embedded corporate values. These dynamics create a suitable context for assessing how governance mechanisms, specifically board attributes, influence the effectiveness of CSR in enhancing financial resilience (Eriqat et al., 2024; Hakimi et al., 2023a; Jamali et al., 2017).

Although theory generally supports a positive CSR and stability relationship, empirical evidence remains mixed. Some studies argue that CSR can reduce profitability or divert attention away from core banking activities, particularly when governance mechanisms are symbolic or weak (Khan et al., 2013). Conversely, strong governance through independent, diverse, and effective boards ensures that CSR is incorporated strategically into decision-making and allows banks to capture its long-term benefits (Adams & Ferreira, 2009; Pathan & Faff, 2013). Yet, few studies directly test how board characteristics moderate the CSR and bank stability relationship, especially within the distinctive institutional environment of the MENA region.

Several scholars suggest that CSR enhances bank stability by promoting openness, trust, and long-term stakeholder relationships (Cornett et al., 2016; Goss & Roberts, 2011). CSR activities can also improve risk management, reduce informational asymmetries, and increase resilience during periods of financial distress (Ogachi & Zoltan, 2020). According to stakeholder theory, banks that consider the interests of a wider range of actors, including customers, employees, regulators, and communities, are better able to withstand systemic shocks (Freeman, 1984). However, CSR may also generate agency problems if poorly monitored, leading managers to invest in symbolic rather than strategic activities (Prior et al., 2008; Khan et al., 2013). These contrasting arguments highlight the need to identify the conditions under which CSR strengthens bank stability.

Board characteristics are a key moderating factor in this relationship. Governance theory suggests that effective boards align managerial behavior with shareholder and stakeholder interests (Jensen & Meckling, 2019; Adams & Ferreira, 2007). Empirical studies demonstrate that board size, independence, gender diversity, and CEO duality influence CSR implementation and risk-taking behavior in financial institutions (Hasan et al., 2018; García-Sánchez et al., 2018). For example, diverse and independent boards can enforce responsible and risk-aware strategies, thereby enhancing the stability-building effect of CSR (Liao et al., 2018). Conversely, poorly structured boards may fail to direct CSR toward long-term financial objectives, limiting its effectiveness. Despite its theoretical importance, limited empirical research examines how CSR and board composition jointly affect bank stability in developing economies such as those in the MENA region. Prior studies often analyze CSR and governance separately or focus on high-income countries (Beltratti & Stulz, 2012; Nkundabanyanga et al., 2014). This creates an important gap in understanding how internal governance mechanisms shape the risk–return profile of socially responsible banking under different institutional conditions.

To address this gap, the present study investigates the moderating effect of board characteristics on the CSR and stability relationship using data from MENA banks. The study makes three key contributions. First, it extends the CSR and bank-stability literature to the MENA region, which is under-studied but institutionally diverse, featuring a mix of Islamic and conventional banking systems. Second, it applies multidimensional measures to capture both CSR and board characteristics, providing a more nuanced understanding of their interaction. Third, it offers practical implications for regulators, investors, and policymakers by identifying the conditions under which CSR strengthens or fails to strengthen bank stability.

The remainder of this paper is structured as follows. Section 2 reviews the theoretical and empirical literature. Section 3 explains data, variables, and methodology. Section 4 presents the empirical results. Section 5 reports the robustness checks considering theory and previous studies. Section 6 concludes with policy implications and directions for future research.

2. Literature Review and Hypotheses Development

2.1. CSR and Bank Stability

The linkage between bank stability and CSR has become more prominent in the recent literature. CSR is no longer viewed only as an ethical responsibility but also as a strategic asset that strengthens organizational performance. Several theoretical perspectives support this argument. Stakeholder theory (Freeman, 1984) suggests that banks which address the interests of customers, employees, regulators, and society build long-term legitimacy and trust, both of which are essential for financial stability. Legitimacy theory extends this idea and argues that CSR practices help institutions preserve societal approval and avoid sanctions based on regulatory breaches or reputational damage. From the resource-based perspective, CSR can function as a unique internal capability that provides sustainable competitive advantage, particularly in highly regulated and trust-sensitive industries such as banking (Barney, 1991).

Empirical findings reinforce these theoretical claims. Octrina and Hanif (2024) show that Indonesian banks with higher CSR ratings were more efficient, indicating that responsible practices can enhance operational performance. In Islamic banking, the relationship can be more complex. Khémiri and Alsulami (2023) document a reversed U-shaped relationship between stability and financial inclusion in GCC countries, where excessive inclusion increased financial risk. However, CSR moderated this effect, underscoring its stabilizing role. In Europe, Bouattour et al. (2024) demonstrate that banks with strong ESG records were less likely to face financial distress, particularly in digital banking environments where reputational risks are more visible. By contrast, Li et al. (2024) report evidence from the United States indicating that deregulation and heightened competition reduced CSR engagement and increased financial vulnerability.

These contrasting findings emphasize the importance of institutional settings in shaping the CSR and stability relationship. CSR contributes to stability in several ways. First, CSR enhances risk management by promoting transparency, corporate ethics, and sound governance practices, which reduces the likelihood of misconduct (Cornett et al., 2016). Second, CSR strengthens stakeholder trust, which becomes crucial during periods of market stress. Third, CSR attracts long-term investors who prefer sustainable business models, thereby ensuring a more stable capital base (C. S. Fernando et al., 2017).

In summary, theoretical and empirical literature generally supports CSR as a stabilizing force in banking. However, negative outcomes can occur in the absence of effective governance. Agency theory explains that managers may pursue CSR for personal image or social expectations rather than shareholder value (Barnea & Rubin, 2010; Krüger, 2015). Such behavior can divert resources away from core activities and weaken financial strength. Conversely, when governance systems are robust, CSR initiatives are better aligned with long-term value creation (Ferrell et al., 2016; Liang & Renneboog, 2017).

Based on the development above, we can formulate the following hypothesis:

H1. 

More responsible banks are more stable.

2.2. The Moderating Role of Board Characteristics in the CSR–Bank Stability Relationship

Recent research identifies board characteristics as important moderators of the CSR and stability relationship. Elements such as board size, independence, gender diversity, and specialized committees influence how effectively CSR practices translate into financial strength. Agency theory suggests that effective boards limit managerial opportunism and ensure that CSR serves long-term objectives rather than reputational motives. Resource dependence theory asserts that diverse and well-connected boards improve access to knowledge and resources, allowing CSR to be embedded strategically into organizational processes.

Empirical studies subscribe to these theoretical frameworks. Endrikat et al. (2021) offered a meta-analysis that established that board size, independence, and gender composition positively impact CSR engagement directly and indirectly through mediation by CSR committees. Their study affirms the importance of board composition as a CSR outcome determinant. In addition, the presence of specialized committees such as CSR or sustainability committees has been correlated with the adoption of improved CSR strategies. Such committees facilitate focused monitoring and integration of CSR into corporate governance structures, thus improving the contribution of CSR as a source of financial stability. In the context of emerging economies, Djalilov and Hartwell (2024) indicated that those banks with sound social and environmental capabilities, together with good board governance, have enhanced stability. This finding supports the function of board compositions in managing institutional changes as well as aligning CSR with broader societal aims. Overall, the literature predicts that board features significantly mediate the relationship of CSR with bank stability. Favorable board designs, diversity, independence, and specialized committees permit more intense strategy implementation of CSR, raising the resilience of banks. These relationships need to continue to be considered in future work given varying institutional conditions and shifting patterns of governance.

The interplay between corporate social responsibility (CSR), stakeholder management, and firm performance has been under increasing scholarly attention over the past several years, particularly in the banking and financial services sector. Rising recognition of CSR as a strategic tool rather than a peripheral practice has encouraged researchers to examine its multifaceted implications on financial performance and organizational behavior. García-Sánchez et al. (2018) emphasize stakeholder engagement as a key driver of CSR effectiveness in banking. Through empirical research drawing on a sample of European banks, they conclude that the degree of stakeholder engagement in the CSR decision-making processes of banks positively affects banks’ CSR performance and legitimacy. They argue that stakeholder-oriented strategies foster responsiveness and transparency that enhance the perceived trustworthiness and credibility of financial institutions.

The observation underscores the strategic role of stakeholder engagement in shaping both CSR communication and impact. Similarly, the work of S. Fernando and Lawrence (2014) explores the role of CSR in building firm performance but introduces the mediating role of organizational culture. They emphasize that firms with an ethical corporate culture are better positioned to leverage CSR initiatives for financial gain. Their Australian sample illustrates a positive relationship between CSR practices and financial performance, especially when supported by internal cultural alignment. This emphasizes that the impact of CSR is not merely external; it is also an outcome of internal governance and values alignment.

Expanding the debate to emerging economies, Shakil (2024) focuses on the relationship between environmental, social, and governance (ESG) performance and firm value in the context of sustainability. The study finds a positive and significant relationship between ESG disclosure and market-based performance metrics such as Tobin’s Q. Specifically interesting is the finding that transparency in ESG reporting reduces information asymmetry and thereby enhances investor confidence. This confirms earlier evidence emphasizing the signaling function of CSR and ESG practices in financial markets. Cek and Eyupoglu (2020) examine the influence of CSR performance on the financial performance of European banks listed on a stock exchange. Their analysis demonstrates a positive yet nonlinear relationship, showing diminishing marginal returns to CSR investments beyond a certain level. They also discover sectoral heterogeneity, in the sense that the CSR–financial performance link is contingent and moderated by firm-level factors such as size, geographical presence, and business model.

Boateng et al. (2025) used a panel ARDL model from 2010 to 2023, confirming that CSR expenditure negatively impacts return on assets. Bank size, however, serves as a moderator, and larger banks are better placed to benefit from CSR. The findings confirm that the financial impact of CSR is not homogeneous and is highly dependent on institutional variables like size. That more nuanced perspective resolves previous conflicting findings in the CSR–performance literature.

Collectively, these studies emphasize that CSR and ESG programs, implemented strategically and in parallel with robust stakeholder engagement, can exert a material influence on a firm’s financial and reputational capital. However, they also indicate that the impact is moderated by internal and external contingencies, including organizational culture, market expectations, and the degree of transparency. The emerging evidence base indicates the necessity of adopting a multidimensional perspective on CSR, one that recognizes its contribution not only to social legitimacy but also to creating sustainable financial performance.

Based on the development above, we can formulate the following hypothesis:

H2. 

Board characteristics can moderate the CSR–bank stability relationship.

Recent studies have increasingly applied dynamic panel and system-based estimators to better address endogeneity, unobserved heterogeneity, and simultaneity in the CSR performance stability nexus. Do et al. (2024) examine the impact of ESG performance on bank stability in ASEAN countries from 2015 to 2022 using the System Generalized Method of Moments (SGMM) to address endogeneity and dynamic effects. Their results reveal a negative and significant relationship between ESG engagement and bank stability. Khémiri et al. (2024) investigate the relationship between financial inclusion and Islamic bank stability using System Generalized Method of Moments (SGMM) to address endogeneity and dynamic effects. Their results reveal a nonlinear trade-off, where greater financial inclusion initially reduces bank stability. Ali et al. (2025) examine the relationship between climate risk, credit risk, and corporate social responsibility (CSR) in MENA banks using the System Generalized Method of Moments (SGMM) to address endogeneity and dynamic effects. Their findings reveal that CSR plays a mitigating role, reducing the adverse impact of climate risk on banks’ credit risk.

3. Empirical Design

3.1. The Sample

Using a sample of conventional banks spread across ten MENA countries between 2010 and 2022, we investigated whether board characteristics can moderate the link between corporate social responsibility and bank stability in the MENA region. We used an initial sample made up of 76 banks. However, several banks have been excluded because of the consistency and accessibility of bank information. As a result, 40 traditional banks made up the final sample (see Table 1). To ensure consistency and comparability, we retained only conventional banks with complete financial data for the study period. Islamic banks were excluded because their operational and financial frameworks differ from conventional banking models, which could introduce heterogeneity into the analysis. This approach also minimizes potential survivorship bias, as only banks with continuous data availability were included. The years 2010–2022 were chosen to highlight the post-global financial crisis recovery, significant governance and financial reforms, and growing MENA region CSR and ESG practice emphasis. This chronology comprises pertinent institutional changes and legislative developments that provide a significant window for analyzing how corporate behavior and board qualities affect bank stability.

Table 1. Distribution of the sample by country.

Middle East North Africa Countries
Countries	Number of Banks	%
Jordan	4	10%
Kuwait	4	10%
Oman	2	5%
Lebanon	1	2.5%
Qatar	4	10%
Saudi Arabia	6	17.5%
United Arab Emirates	6	15%
Egypt	1	2.5%
Morocco	2	5%
Tunisia	10	25%
Number of banks	40	100%

3.2. Variable Selection and Theoretical Justification

3.2.1. Dependent Variable: Bank Stability

In this paper, we extend the literature by investigating the effect of corporate social responsibility and board characteristics on banking stability. To capture this relationship, the dependent variable is the bank stability measured by $Z-\mathrm{S}\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{E}\left(\mathrm{R}\mathrm{O}\mathrm{A}\right)$. Referring to Zaghdoudi (2019), and Hakimi et al. (2022), we employ these metrics that represent different dimensions of bank stability. In Equation (1), the $Z-\mathrm{S}\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{E}\left(\mathrm{R}\mathrm{O}\mathrm{A}\right)$ may be calculated by dividing the standard deviation of return on assets by the means of return on assets plus the capital adequacy ratio. A bank’s efforts to mitigate risks and absorb losses were reflected in the Z-score. The bank was steady when the Z-score number was high, and vice versa.

$$Z-\mathrm{S}\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{E}\left(\mathrm{R}\mathrm{O}\mathrm{A}\right)={\displaystyle \frac{\mathrm{R}\mathrm{O}\mathrm{A}+\mathrm{E}\mathrm{Q}\mathrm{T}\mathrm{A}}{\partial \left(\mathrm{R}\mathrm{O}\mathrm{A}\right)}}$$

(1)

3.2.2. Main Explanatory Variable: Corporate Social Responsibility

According to Boussaada et al. (2023), the study applies a pillar score of Environmental, Social, and Governance (ESG). This pillar score provides a general and balanced indication of the performance of a firm based on environmental, social, and governance dimensions. The values of this score range between 0 and 100.

3.2.3. Moderating Variable: Board Characteristics

The board characteristics index (BC_index) includes the following characteristics: board size (number of directors on the board), board independence (proportion of independent directors), duality (a dummy variable equal to 1 if the CEO also serves as Chair of the Board, and 0 otherwise), board gender diversity (percentage of women on the board), board tenure (average length of board members’ terms), and board compensation (total remuneration of directors in U.S. dollars relative to total assets).

In this study, the board characteristics index (BC_index) was constructed through two main steps. Following Hakimi et al. (2023a), we applied empirical normalization and weighting using equal coefficients (1/N). Once the indices were calculated, the resulting values ranged from 0 to 1.

3.2.4. Control Variables

As outlined, our econometric model incorporates several control variables. The first category pertains to bank-specific factors, including bank size (BS), which is used to explain variations in bank stability (Ghenimi et al., 2017), and the capital adequacy ratio (CAR), a key determinant of bank stability (Anginer et al., 2014). We also included the bank diversification measured by the non-interest income ratio (NII) (Mercieca et al., 2007). The second category relates to industry-specific variables, such as bank concentration (CONC), recognized as significant drivers of bank stability (Mercieca et al., 2007). The third category encompasses macroeconomic conditions and the financial environment represented by the GDP growth rate (GDPG) and the inflation rate (INF) (Djebali & Zaghdoudi, 2017). Finally, we included the quality of governance institutions in the MENA region is assessed using the World Governance Indicators (WGIs). Six policy variables, sourced from the World Bank’s global governance dataset and analyzed by Kaufmann et al. (2011), are included. These indicators are control of corruption, government effectiveness, political stability and absence of violence/terrorism, regulatory quality, rule of law, and voice and accountability. As outlined by Kaufmann et al. (2011), institutional quality is measured as the average of these six indicators, with values ranging from −2.5 (indicating poor governance) to 2.5 (indicating strong governance).

Bank-level data, including financial and accounting variables, were sourced from the RefinitivEikon database. The data on macroeconomic variables, bank concentration, and institutional quality were obtained from the World Bank database, specifically the World Bank Indicators (WDIs), Global Financial Development (GFD), and Worldwide Governance Indicators (WGIs).

The choice of variables stems from accepted theoretical frameworks. Widely used in financial stability research as a consistent assessment of a bank’s risk-adjusted return profile is the Z-score (ROA). Consistent with stakeholder and legitimacy theories, the ESG score offers a complete and balanced proxy for CSR performance. Agency and resource dependence theories guided the selection of board characteristics; these theories define board size, independence, diversity, and structure as fundamental factors affecting strategic decision-making and risk control. Following earlier studies in banking and governance research, control variables were included to account for bank-specific, industry-specific, and macroeconomic factors known to influence stability.

3.3. Empirical Approach and Model Specification

To investigate the moderating effect of board characteristics in the relationship between corporate social responsibility and bank stability, we performed the SGMM as an empirical technique. The SGMM technique may be used to solve endogeneity, one of the crucial issues in corporate and banking finance. Furthermore, OLS and fixed- and random-effect (FE and RE) models frequently encounter two issues: measurement errors and omitted variables bias. To do this, we employed the SGMM technique, which Blundell and Bond (1998) suggested in their investigation. According to Zhou (2014), Teixeira and Queiros (2016), Danisman and Tarazi (2020), and Hakimi et al. (2023b), the SGMM approach yields more dependable and practical results.

Since endogeneity problems are common in studies involving financial performance and governance, the system generalized method of moments (SGMM) approach was selected because it is robust when managing dynamic panel data. To provide more accurate and consistent parameter estimates, this approach also accounts for unobserved heterogeneity and possible reverse causality. The methodology adheres to accepted practices in financial econometrics, specifically in research examining how governance and corporate social responsibility affect business results.

We followed the moderator effects analysis approach of Sharma et al. (1981) and Ping (1995), which is based on two main steps. The first step consists of analyzing the main effects of the explanatory variables on the dependent variable. The second step aims to explore the interaction effects.

The empirical technique outlined in this research consists of three phases. First, we examined the relationship between corporate social responsibility and bank stability. The econometric model to be validated in this step is presented in Equation (2):

$${Z-ROA}_{i,t}={\beta }_0+{\beta }_1{Z-ROA}_{i,t-1}+{\beta }_2{\mathrm{E}\mathrm{S}\mathrm{G}}_{i,t}+{\beta }_3{\mathrm{B}\mathrm{S}}_{i,t}+{\beta }_4{CAR}_{i,t}+{\beta }_5{\mathrm{N}\mathrm{I}\mathrm{I}}_{i,t}+{\beta }_6{\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{C}}_{i,t}+{\mathsf{\beta }}_7{\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{G}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_8{\mathrm{I}\mathrm{N}\mathrm{F}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_9{\mathrm{I}\mathrm{Q}}_{\mathrm{i},\mathrm{t}}+{\epsilon }_{i,t}$$

(2)

In the second stage, we investigated how board characteristics affected bank stability. The following Equation (3) displays the econometric model:

$${Z-ROA}_{i,t}={\beta }_0+{\beta }_1{Z-ROA}_{i,t-1}+{\beta }_2{\mathrm{B}\mathrm{C}}_{i,t}+{\beta }_3{\mathrm{B}\mathrm{S}}_{i,t}+{\beta }_4{CAR}_{i,t}+{\beta }_5{\mathrm{N}\mathrm{I}\mathrm{I}}_{i,t}+{\beta }_6{\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{C}}_{i,t}+{\mathsf{\beta }}_7{\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{G}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_8{\mathrm{I}\mathrm{N}\mathrm{F}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_9{\mathrm{I}\mathrm{Q}}_{\mathrm{i},\mathrm{t}}+{\epsilon }_{i,t}$$

(3)

In the third step, we investigated whether board characteristics act as a moderator in the relationship between corporate social responsibility and bank stability. To do this, we incorporated an interaction variable into the econometric model that captures the interplay between corporate social responsibility and board characteristics. The econometric model to be verified is outlined in Equation (4):

$${Z-ROA}_{i,t}={\beta }_0+{\beta }_1{Z-ROA}_{i,t-1}+{\beta }_2{\mathrm{E}\mathrm{S}\mathrm{G}\mathrm{x}\mathrm{B}\mathrm{C}}_{i,t}+{\beta }_3{\mathrm{B}\mathrm{S}}_{i,t}+{\beta }_4{CAR}_{i,t}+{\beta }_5{\mathrm{N}\mathrm{I}\mathrm{I}}_{i,t}+{\beta }_6{\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{C}}_{i,t}+{\mathsf{\beta }}_7{\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{G}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_8{\mathrm{I}\mathrm{N}\mathrm{F}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_9{\mathrm{I}\mathrm{Q}}_{\mathrm{i},\mathrm{t}}+{\epsilon }_{i,t}$$

(4)

All variables’ definitions are given in Table 2.

Table 2. Definition and measurement of variables.

Variables	Definitions	Measures	Source
Dependent variables (Z-ROA)
Z-ROA	Bank stability	The ratio of the sum of the averaged ROA and the CAP to the standard deviations of ROA. We applied the natural logarithm on Z-ROA.	Refinitiv Eikon database
Corporate Social Responsibility and Board Characteristics
ESG	ESG score	Composite environmental/social/governance score	Refinitiv Eikon database
BC_index	Board characteristics index	Composite Index of Board Characteristics.	Authors’ calculation
ESG*BC_index	Interactional variable	The interaction between ESG and BCI	Authors’ calculation
Bank specifics
BS	Bank size	Natural logarithm of total assets	Refinitiv Eikon database
CAR	Capital adequacy ratio	Bank capital to total assets (%)	Refinitiv Eikon database
NII	Bank diversification	Non-interest income in % of total income	Refinitiv Eikon database
Industry specifics
CONC	Bank Concentration	Bank concentration (%)	World Bank database
Financial environment and macroeconomic conditions
GDPG	The growth rate of GDP	Annual growth rate of GDP (%)	World Bank database
INF	The inflation rate	Consumer price index (%)	World Bank database
QI_index	Institutional quality	An index of IQ	World Bank database

4. Empirical Results

4.1. Summary Statistics and Correlation Matrix

As shown in Table 3, the average level of bank stability, measured by Z-ROA, is 2.811, with a maximum value of 4.333 and a minimum of −0.5525. The average ESG score (ESG) is 40.358, with a maximum of 80.794 and a minimum of 12.843. In the MENA region, the board characteristics index (BC_index) ranges from a high value of 0.623 to a low level of 0.06.

Table 3. Descriptive Statistics.

Variable	Mean	Std. Dev.	Min	Max
ZROA	2.811	2.512	−0.525	4.333
ESG	40.358	12.703	12.843	80.794
BS	23.5	1.313	20.942	26.512
CAR	15.839	10.987	1.256	40.350
CONC	80.886	13.979	56.035	100
NII	38.121	17.5	9.552	96
INF	3.816	10.791	−3.749	171.205
GDPG	2.547	4.035	−21.4	19.592
Qi_index	0.662	0.167	0.171	0.971
BC_index	0.327	0.104	0.06	0.623

Regarding bank-specific factors, the average bank size (BS) is 23.5, with values ranging from a minimum of 20.942 to a maximum of 26.512. CAR stands for Capital Adequacy Ratio, which averages 15.839%, ranging from 1.25% to 40.350%. Bank diversification, measured by NII, records a mean value of 38.121, with a minimum of 9.552 and a maximum of 96. Concerning industry specifics, Table 3 reports that the mean level of bank concentration (CONC) is 80.888%, with a maximum of 100% and a minimum of 56.035%. For macroeconomic conditions, the MENA region experienced the highest GDP growth rate of 19.592% and the lowest level of −21.4%. The inflation rate ranges from a low value of −3.749% to a maximum value of 171.205%, with an average of 3.878%. The highest level of inflation was registered by Lebanon in 2022.

The correlation matrix offers insights into the level and nature of relationships between variables by calculating the coefficients of their linear connections. Table 4 below displays the correlation matrix for each variable used in this study. From this table, we conclude that there is no significant issue with collinearity problems.

Table 4. Pairwise correlations.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
(1) ZROA	1.000
(2) ESG	0.000	1.000
	(0.996)
(3) BS	0.256 *	0.253 *	1.000
	(0.000)	(0.000)
(4) CAR	0.163 *	0.009	0.239 *	1.000
	(0.000)	(0.836)	(0.000)
(5) CONC	0.233 *	−0.152 *	0.416 *	0.208 *	1.000
	(0.000)	(0.001)	(0.000)	(0.000)
(6) NII	−0.073	−0.150 *	0.342 *	0.151 *	0.225 *	1.000
	(0.097)	(0.001)	(0.000)	(0.001)	(0.000)
(7) INF	0.005	−0.018	−0.047	−0.063	−0.149 *	−0.153 *	1.000
	(0.907)	(0.682)	(0.332)	(0.154)	(0.001)	(0.001)
(8) GDPG	0.027	0.011	0.085	0.039	0.139 *	0.111 *	−0.181 *	1.000
	(0.544)	(0.814)	(0.080)	(0.381)	(0.002)	(0.012)	(0.000)
(9) Qi_index	0.013	0.058	0.413 *	0.105 *	0.525 *	0.085	−0.323 *	0.214 *	1.000
	(0.762)	(0.198)	(0.000)	(0.017)	(0.000)	(0.056)	(0.000)	(0.000)
(10) Bc_index	−0.166 *	0.146 *	−0.212 *	−0.068	−0.412 *	−0.184 *	−0.080	−0.032	−0.059	1.000
	(0.000)	(0.002)	(0.000)	(0.141)	(0.000)	(0.000)	(0.083)	(0.482)	(0.195)

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

To further ensure the robustness of the regression results reported in Table 4, the Variance Inflation Factor (VIF) test was conducted to detect potential multicollinearity among the explanatory variables. The VIF reflects the extent to which the variance of an estimated coefficient is inflated due to linear dependence among independent variables. A value of 1 indicates no multicollinearity, values between 1 and 5 indicate moderate correlation, and values above 5 are commonly interpreted as evidence of severe multicollinearity. Table 5 reports the VIF results for all three models. The first model, which examines the impact of corporate social responsibility on bank stability, has an average VIF of 1.38, indicating the absence of multicollinearity and only moderate correlation among regressors. The second model, focusing on the effect of board characteristics on stability, yields a mean VIF of 1.37, again showing no multicollinearity concerns. The third model, which incorporates the interaction between CSR and board characteristics, produces a mean VIF of 1.39. Together, these values confirm that multicollinearity does not threaten the validity of the coefficient estimates and that the regression models are both stable and reliable.

Table 5. Variance Inflation Factor (VIF).

Model 1			Model 2			Model 3
Variable	VIF	1/VIF	Variable	VIF	1/VIF	Variable	VIF	1/VIF
BS	1.75	0.572	CONC	1.81	0.551	CONC	1.85	0.539
QI	1.68	0.594	QI	1.73	0.578	QI	1.68	0.593
CONC	1.65	0.606	BS	1.54	0.648	BS	1.61	0.622
NII	1.32	0.759	BCI	1.28	0.782	ESG*BCI	1.36	0.737
ESG	1.29	0.774	NII	1.22	0.818	NII	1.31	0.765
INF	1.18	0.848	INF	1.19	0.837	INF	1.19	0.843
CAR	1.09	0.915	CAR	1.10	0.912	CAR	1.09	0.914
GDPG	1.07	0.938	GDPG	1.06	0.945	GDPG	1.06	0.940
Mean VIF	1.38		Mean VIF	1.37		Mean VIF	1.39

Table 6 reports the CD test results, which confirm the presence of cross-sectional dependence across all variables (p < 0.05). This indicates that banks in the MENA region experience common macroeconomic shocks, financial linkages, and regional spillover effects. The high correlation coefficients for variables such as GDP growth (GDPG) and bank size (BS) further reflect the shared exposure of financial institutions to systemic regional dynamics. Given these findings, estimators that are robust to cross-sectional dependence are required. The use of the two-step System GMM estimator is therefore justified, as it accommodates both endogeneity and cross-sectional correlation.

Table 6. Pesaran (2004) Cross-Sectional Dependence (CD) Test Results.

	CD-Statistic	p-Value
ZROA	7.48	0.000
ESG	10.72	0.000
BS	56.66	0.000
CAR	2.15	0.032
CONC	17.51	0.000
NII	14.61	0.000
INF	23.38	0.000
GDPG	59.09	0.000
Qi_index	12.10	0.000
BC_index	11.56	0.000

Table 7 presents the results of the Levin et al. (2002) and Im et al. (2003) and Im–Pesaran–Shin (IPS) panel unit root tests, which were employed to examine the stationarity of variables prior to estimating the dynamic model. For both tests, the null hypothesis states that the panels contain a unit root, while the alternative hypothesis indicates stationarity in at least part of the panel. A p-value below 0.05 leads to rejection of the null hypothesis. The results reveal that both LLC and IPS reject the unit root hypothesis, confirming that all variables are stationary in levels. This provides statistical support for proceeding with the dynamic estimation.

Table 7. Panel unit root test.

	LLC Adjusted t *	IPS W-t-Bar	Z[t-Bar]/p-Value
ZROA	−78.7730 *** (0.0000)	−12.5717 *** (0.0000)	3.735 (1.000)
ESG	−9.0147 *** (0.0000)	66.7149 * (0.0762)	−1.304 * (0.096)
BS	−12.3291 *** (0.0000)	−4.3731 *** (0.0000)	−2.397 ** (0.008)
CAR	−38.7789 *** (0.0000)	−8.1042 *** (0.0000)	1.342 (0.910)
CONC	−2.3712 * (0.0991)	−32.3088 *** (0.0000)	−3.457 *** (0.000)
NII	−9.9645 *** (0.0000)	−6.9332 *** (0.0000)	−3.868 *** (0.000)
INF	−8.0879 *** (0.0000)	−4.3847 *** (0.0000)	−0.421 (0.337)
GDPG	−8.5778 *** (0.0000)	−6.4609 *** (0.0000)	−4.842 *** (0.000)
Qi_index	−2.6428 *** (0.0041)	5.9536 * (0.0829)	2.908 (0.998)
BC_index	−5.1433 *** (0.0000)	−3.8460 *** (0.0001)	1.615 (0.947)

*, **, and *** indicate statistical significance at the 1%, 5%, and 10% level.

To account for the cross-sectional dependence identified by the Pesaran (2004) CD test, this study applied the second-generation panel unit root test developed by Pesaran et al. (2008). The outcomes, summarized in Table 7, indicate that ZROA, CAR, INF, QI-INDEX, and BC-INDEX are non-stationary in their original form (at level) but become stationary after first differencing, suggesting they are integrated of order one (I(1)). On the other hand, BS, CONC, NII, and GDPG are stationary at levels (I(0)), while ESG shows weak stationarity at the 10% significance level. The presence of both I(0) and I(1) variables justifies the use of dynamic panel estimators such as System GMM, which can handle mixed integration orders without requiring all variables to be different.

Table 8 reports the results of the Pesaran and Yamagata (2008) slope heterogeneity test, which evaluates whether the estimated slope coefficients are consistent across cross-sectional units. The null hypothesis assumes slope homogeneity, while the alternative hypothesis indicates heterogeneity. A p-value below 0.05 leads to rejection of the null hypothesis.

Table 8. Pesaran and Yamagata (2008) Slope Heterogeneity Test Results.

Statistic	Value	p-Value
Delta (Δ)	1.190	0.234
Adjusted Delta (Δ_adj)	3.068	0.002

The adjusted Delta statistic (Δ_adj = 3.068, p = 0.002) is statistically significant at the 1 percent level, confirming rejection of slope homogeneity. This suggests that the relationships between bank stability (ZROA) and its determinants, including ESG, BS, CAR, CONC, NII, INF, GDPG, QIINDEX, and CAINDEX, differ significantly across banks. In other words, individual bank characteristics and regional conditions produce variations in how explanatory variables influence stability. The confirmation of slope heterogeneity underscores the need for estimation techniques that are robust to parameter variation across cross-sections. The use of the two-step System GMM estimator is therefore appropriate, as it accommodates both endogeneity and heterogeneity within a dynamic panel framework.

4.2. Discussion of the Empirical Findings

4.2.1. Findings on the Effect of Corporate Social Responsibility on Bank Stability

The first step in the empirical strategy of this paper is to examine the effect of CSR on bank stability in the MENA region, using the Z-ROA as the measure of stability. The empirical results are outlined in Table 5. Results in Table 9 indicate that Sargan test of over-identification and serial correlation tests cannot reject the over-identification null hypothesis and no autocorrelation. The p-values of the Arellano and Bond and Sargan tests are greater than 5%. The Sargan test indicates a statistic of 34.629 with the probability of 1.000, which confirms the instruments employed are exogenous and valid and the model is over-identified. The AR (1) test shows the p-value 0.015, which shows weak first-order autocorrelation of the residuals, even though this is not significant for the SGMM model and will never affect proper model specification. The AR (2) test gives the statistic 0.759 with probability 0.448, showing the absence of second-order serial autocorrelation of first-differenced errors, which establishes the validity of lagged instruments.

Table 9. Regression results: the effect of ESG on ZROA.

ZROA	Coef.	St. Err.	t-Value	p-Value
ZROA (−1)	0.873	0.01	89.83	0.000 ***
ESG	−0.019	0.008	−2.34	0.02 **
BS	0.952	0.232	4.10	0.000 ***
CAR	0.006	0.004	1.74	0.082 *
CONC	0.017	0.012	1.42	0.155
NII	0.003	0.007	0.49	0.625
INF	−0.036	0.006	−5.80	0.000 ***
GDPG	0.067	0.012	5.46	0.000 ***
QI_index	9.868	1.867	5.29	0.000 ***
Constant	−27.232	4.814	−5.66	0.000 ***
obs	416
AR (1)	−2.444
Prob	0.015
AR (2)	−0.759
Prob	0.448
Sargan test	3.46292
Prob	1.0000

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

The findings displayed in Table 6 indicate that the lagged dependent variable has a positive and significant effect, meaning that bank stability in the previous year positively influences the current year’s bank stability.

Additionally, Table 6 shows that the coefficient of ESG is negatively and significantly associated with the dependent variable (Z-ROA). This suggests that corporate social responsibility significantly decreases bank stability. Although CSR has the potential to enhance long-term resilience; its implementation has the potential to damage bank stability when weakly regulated, strategically misguided, or excessively costly. Over-expenditure on CSR, symbolic actions, or abuse by managers can compromise profitability, drain resources from the core business, and create reputational and operational risks, thereby weakening banking institution stability. This result is divergent from the findings of Djalilov and Hartwell (2024). Therefore, we accept H1.

Concerning the effect of bank specifics, findings indicate that the bank size is positively significant for banking stability at a 1% significance level. Bank size enhances stability by promoting diversification, providing easier access to capital markets, and permitting the utilization of sophisticated risk management structures. Larger banks also have greater regulatory oversight and reputational advantages, which increase depositor confidence and reduces the risk of running during financial distress. This result aligns with the findings of Adusei (2015). The capital adequacy ratio positively and significantly influences banking stability. Capital adequacy is important in enhancing the stability of banks because it serves as a shock absorber for losses, sound risk management, and boosts confidence in the market. Higher CAR values decrease the risk of bank failure and result in a more robust financial system. This finding is consistent with the works of Bourke (1989); Molyneux and Thornton (1992).

Regarding macroeconomic effects, the results show that GDP growth has a positive and significant impact on bank stability. GDP growth plays a significant role in ensuring the stability of banks by increasing the repayment capacity of borrowers, loan demand, and credit risk mitigation. A favorable macroeconomic environment enhances bank profitability as well as asset quality, thereby stabilizing financial distress probability and enhancing the resilience of the banking system. This result corroborates the findings of Espinoza and Prasad (2010) and Klein (2013).

We also found that inflation is inversely related to bank stability. Inflation has a negative impact on the stability of banks by lowering the real value of assets, increasing credit risk, and increasing interest rate and exchange rate volatility. Excessive inflation undermines borrowers’ ability to repay and confidence among depositors, thereby lowering banks’ overall resilience. This result supports the findings of Khemiri (2025).

The institutional quality index is found to be positively and significantly associated with bank stability. Findings indicate that the index of institutional quality significantly increases the level of bank stability in the MENA region. Institutional quality contributes to bank stability by enabling efficient regulation, enforcement of contracts, protection of property rights, and reducing governance risks. Strong institutions make banks more resilient by promoting prudent risk management, reducing non-performing loans, and strengthening stakeholders’ confidence. This finding is in line with the works of Laeven and Levine (2009) and Kim and Choi (2020) and Alnabulsi et al. (2022, 2023a).

4.2.2. Findings of the Effect of Board Characteristics on Bank Stability

The results presented in Table 10 pertain to the impact of board characteristics on bank stability, as measured by Z-ROA. The findings indicate that the board characteristics index significantly enhances bank stability in the MENA region. This suggests that strong governance practices are associated with increased stability among banks in the MENA region. Board characteristics such as independence, expertise, heterogeneity, and how often they meet positively influence bank stability by enhancing the quality of governance, risk management, and decreasing managerial risk-taking. Efficient boards are a guarantee of sound strategic choices and early warnings of weaknesses, hence increasing the resilience of the banking system.

Table 10. Regression results: the effect of BC_index on ZROA.

ZROA	Coef.	St. Err.	t-Value	p-Value
ZROA (−1)	0.864	0.007	123.67	0.000 ***
BC_index	6.424	1.301	4.94	0.000 ***
BS	1.393	0.204	6.82	0.000 ***
CAR	0.009	0.003	2.50	0.012 **
CONC	0.028	0.017	1.66	0.097 *
NII	0.01	0.007	1.46	0.144
INF	−0.031	0.003	−9.82	0.000 ***
GDPG	0.077	0.008	10.16	0.000 ***
QI_index	6.912	1.844	3.75	0.000 ***
Constant	−39.61	4.138	−9.57	0.000 ***
obs	405
AR (1)	−2.463
Prob	0.014
AR (2)	−0.701
Prob	0.483
Sargan test	32.12072
Prob	1.0000

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

Regarding the effects of bank-specific factors and macroeconomic conditions, there are no significant changes compared to the results presented in Table 10. For industry specifics, bank concentration becomes positive and significant at a 10% significance level. Concentration in the banking sector has a positive impact on financial stability by way of higher profitability, efficiency of operations, and customer confidence. Large banks in concentrated markets have economies of scale, fewer incentives to assume excessive risk, and are more likely to attract regulatory attention, which makes them more resilient and the banking system more stable overall. These results are consistent with the findings of Berti et al. (2017).

4.2.3. Findings of the Interactional Effect of Corporate Social Responsibility and Board Characteristics on Bank Stability

The empirical results presented in Table 11 are relative to the moderating effect of board characteristics on the relationship between corporate social responsibility and bank stability. The findings indicate that the coefficient of the interaction between corporate social responsibility and board characteristics (ESG*BC_index) is positive and significant at the level of 10% of significance.

Table 11. Regression results: The moderating effect of Esg*BC_index on ZROA.

ZROA	Coef.	St. Err.	t-Value	p-Value
ZROA (−1)	0.883	0.008	115.18	0.000 ***
ESG*BC_index	0.029	0.017	1.69	0.092 *
BS	0.908	0.18	5.05	0.000 ***
CAR	0.008	0.003	2.70	0.007 ***
CONC	0.038	0.017	2.23	0.026 **
NII	0.008	0.006	1.37	0.17
INF	−0.036	0.008	−4.48	0.000 ***
GDPG	0.068	0.01	6.76	0.000 ***
QI_index	6.39	2.34	2.73	0.006 ***
Constant	−27.256	3.322	−8.20	0.000 ***
obs	419
AR (1)	−2.471
Prob	0.014
AR (2)	−0.834
Prob	0.405
Sargan test	32.90142
Prob	1.0000

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

The findings indicate that the interplay between corporate social responsibility and board characteristics positively impacts the stability of the bank. Effective boards enhance the integration, monitoring, and strategic alignment of CSR efforts so that they contribute to risk mitigation, stakeholder trust, and long-term resilience. This synergy enhances both governance and external reputation, thus aiding in the stability of the bank. This result is in line with the work of Djalilov and Hartwell (2024). Hence, we accept H2.

Regarding the effects of bank-specific factors, industry characteristics, and macroeconomic conditions, there are no significant changes compared to the results presented in Table 9 and Table 10.

5. Robustness Check: Non-Performing Loans as Measure of Bank Stability

As a robustness check, we test and report in this section the moderating role of board characteristics in the relationship among corporate social responsibility and bank stability measured by non-performing loans. To ensure the robustness of our findings, we re-estimated the baseline model using an alternative proxy for bank stability, namely the ratio of non-performing loans (NPLs). This indicator is widely recognized in the banking literature as a measure of asset quality and credit risk exposure (e.g., Alnabulsi et al., 2023b; Ghosh, 2015; Beck et al., 2013; Hakimi et al., 2023c). A lower NPL ratio reflects higher financial stability and better risk management practices.

Table 12 and Table 13 present the results of the System GMM estimations when NPLs are used as the dependent variable. The findings remain consistent with our baseline results, confirming the reliability of our main conclusions. First, CSR (ESG) exhibits a positive and significant relationship with NPLs, suggesting that excessive or poorly aligned CSR initiatives may deteriorate asset quality and increase credit risk. This result reinforces our earlier conclusion that CSR, when not strategically integrated, can undermine financial resilience. Second, board characteristics (BC) maintain a negative and significant coefficient, implying that effective governance mechanisms, particularly in terms of board independence, diversity, and expertise, reduce the share of non-performing loans and enhance overall stability.

Table 12. Robustness check.

NPLs	Coef.	St.Err.	t-Value	p-Value
NPLs (−1)	0.693 ***	0.005	138.69	0.000 ***
ESG	0.022 ***	0.004	5.53	0.000 ***
BS	−2.663 ***	0.176	−15.05	0.000 ***
CAR	−0.013 **	0.005	−2.34	0.019 **
CONC	−0.025 ***	0.009	−2.81	0.005 ***
NII	−0.029 ***	0.004	−6.43	0.000 ***
INF	0.017 ***	0.005	3.39	0.001 ***
GDPG	−0.171 ***	0.009	−17.49	0.000 ***
QI_index	12.210 ***	2.012	6.07	0.000 ***
Constant	60.200 ***	3.791	15.88	0.000 ***
obs	399
AR (1)	−2.9374
Prob	0.0033
AR (2)	−0.21233
Prob	0.8318
Sargan test	33.47399
Prob	1.0000

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

Table 13. Robustness check.

NPLs	Coef.	St.Err.	t-Value	p-Value
NPLs (−1)	0.673 ***	0.010	62.86	0.000 ***
BC_index	−8.800 ***	2.807	−3.13	0.002 ***
BS	−2.471 ***	0.186	−13.23	0.000 ***
CAR	−0.006	0.003	−1.60	0.110
CONC	−0.034 ***	0.010	−3.37	0.001 ***
NII	−0.042 ***	0.004	−9.91	0.000 ***
INF	0.022 ***	0.004	4.67	0.000 ***
GDPG	−0.161 ***	0.009	−17.51	0.000 ***
QI_index	11.809 ***	2.278	5.18	0.000 ***
Constant	54.930 ***	4.549	12.07	0.000 ***
obs	390
AR (1)	−2.8653
Prob	0.0042
AR (2)	−0.306
Prob	0.759
Sargan test	33.50375
Prob	1.0000

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

Third, and most importantly, the interaction term between CSR and board characteristics (ESG × BC) remains negative and significant, confirming the moderating role of the board. This indicates that well-structured boards can transform CSR practices into sound, stability-enhancing strategies by ensuring alignment with long-term financial goals. Overall, these robustness estimations confirm that our key findings are not sensitive to the choice of the bank stability measure. The moderating role of board characteristics on the CSR–stability nexus remains significant whether bank stability is captured by ZROA or by NPLs, thereby reinforcing the robustness and credibility of our results.

6. Concluding Remarks and Policy Recommendation

The objective of this paper was to explore the relationship between corporate social responsibility (CSR) and bank stability in the MENA region and analyze whether board characteristics (BC) can moderate this relationship. We used a sample of MENA banks over the 2010–2022 period and the system generalized method of moments (SGMM) as our econometric approach. CSR was measured using a composite CSR index, while board attributes were proxied using board independence, expertise, and diversity measures. Bank stability was proxied using the Z-score.

The empirical findings reveal a negative and significant relationship between CSR and bank stability in the MENA region, suggesting that CSR can be harmful to financial resilience, particularly if not strategically aligned with governance structures. Yet, board characteristics were found to affect bank stability positively significantly. Moreover, our results show that the interaction between CSR and board characteristics significantly increases bank stability, emphasizing the point that good governance structures can manage to turn CSR into a stabilizer rather than a risk.

The findings of this study have important implications for policymakers, regulators, and managers in the banking sector across different MENA sub-regions. Strengthening board governance through greater independence, diversity, and financial expertise can enhance the positive impact of CSR on bank performance and stability. However, policy priorities may differ depending on regional characteristics. In oil-dependent economies, where state ownership and resource rents often shape governance structures, regulatory efforts should focus on improving board independence and transparency to ensure that CSR initiatives contribute to diversification and sustainable development. In contrast, non-oil economies should emphasize aligning CSR practices with financial inclusion, innovation, and SME support to foster broader economic resilience. Similarly, in countries with greater political stability, long-term CSR strategies can be institutionalized more effectively, while in less stable environments, short-term risk management and stakeholder engagement policies may be more relevant. Overall, promoting regulatory frameworks that encourage sound corporate governance and strategic CSR integration can help banks strengthen stakeholder trust, reduce reputational risk, and achieve sustainable growth across diverse MENA contexts.

Despite the significance of these findings, several limitations should be acknowledged. First, although the use of the system generalized method of moments (SGMM) mitigates potential endogeneity concerns arising from unobserved heterogeneity and dynamic relationships, it may not fully eliminate all sources of endogeneity. Issues such as measurement errors, omitted variable bias, and potential reverse causality between CSR and bank stability could still affect the robustness of the estimates. Future studies could adopt complementary identification strategies such as difference-in-differences approaches, instrumental variable techniques with more exogenous instruments, or natural experiments to further validate causal inferences. Second, the results remain constrained by the availability and reliability of governance and CSR data in the MENA region. Data inconsistency and limited disclosure practices may have influenced the accuracy of the findings. Third, the study’s regional focus on MENA banks provides valuable contextual insights but restricts the generalizability of the results to other regions with differing institutional, cultural, and regulatory frameworks. Future research could benefit from cross-regional analyses to test whether similar relationships hold under alternative governance and market structures. Finally, both board characteristics and CSR practices are dynamic and may evolve over time in response to shifting socio-economic, political, and technological conditions. These temporal changes were not fully captured in the current study, which may affect the stability of the observed relationships.

Future studies could extend this research by incorporating longitudinal designs to examine the evolving moderating role of board structures on CSR–stability dynamics. Additionally, exploring emerging themes such as ESG integration, digital transformation in governance, and AI-driven risk management could provide deeper insights into the modern CSR–governance–stability nexus. Addressing these issues would enhance the validity, causal interpretation, and policy relevance of future research on sustainable bank governance.