Exploring the Triangular Relationship of Risk, Capital, and Efficiency Under ESG Practices

Abstract

This study investigates the dynamic relationship between risk-taking, capital adequacy, and operational efficiency in the MENA banking sector, with a particular emphasis on the role of ESG performance in shaping sustainable financial behavior. Using an unbalanced panel of 167 commercial banks from 2015 to 2024, we develop a three-equation framework and estimate it using the two-step System-GMM method to address endogeneity, simultaneity, and dynamic effects. The empirical results reveal significant interdependencies among risk, capital, and efficiency, confirming the existence of a sustainable risk–capital–efficiency nexus. The results reveal that bank risk is strongly persistent; however, ESG performance significantly mitigates credit risk, particularly through its social and governance dimensions, which enhance transparency, borrower discipline, and stakeholder trust. Efficiency also acts as a stabilizing force by reducing overall risk. Capital adequacy is positively influenced by ESG performance and efficiency, indicating that sustainable and well-managed banks maintain stronger capital buffers and more resilient balance sheets. Furthermore, bank efficiency improves with profitability, capitalization, favorable macroeconomic conditions, and socially oriented ESG engagement. These findings demonstrate that ESG adoption is a strategic driver of financial soundness, simultaneously lowering risk, reinforcing capital strength, and enhancing operational performance. The paper offers important implications for regulators and bank managers, highlighting the need to embed ESG metrics into supervisory frameworks, risk-management systems, and long-term strategic planning.

1. Introduction

In recent years, the global financial system has undergone a structural transformation driven by the growing emphasis on sustainability. The traditional conception of banking stability, centered mainly on capital adequacy and risk containment, has evolved to integrate Environmental, Social, and Governance (ESG) factors that influence transparency, stakeholder discipline, and long-term value creation. This shift reflects growing recognition that sustainability practices can shape banks’ risk profiles, funding conditions, and operational resilience. For the Middle East and North Africa (MENA) region, this transformation is particularly significant. The region’s banking sector plays a pivotal role in financing economic diversification and development strategies, including national visions. However, MENA banks also operate in environments characterized by governance heterogeneity, concentrated ownership structures, weak insolvency regimes, and macroeconomic volatility, making the evaluation of sustainability driven prudential outcomes particularly relevant. In such settings, ESG adoption may simultaneously discipline risk-taking, alter capital incentives, and affect operational efficiency in non-linear or offsetting ways, leaving the overall prudential impact empirically unresolved. Understanding how sustainability practices influence prudential outcomes has thus become essential for ensuring long-term resilience. From a theoretical perspective, ESG performance affects bank soundness through multiple, interconnected channels. Strong governance and social responsibility enhance stakeholder discipline and internal monitoring, reducing moral hazard and credit risk. At the same time, ESG disclosure mitigates information asymmetry, improving funding conditions and supporting capital accumulation, while prudential authorities increasingly view ESG as a signal of risk governance quality. However, these channels may also generate trade-offs, as compliance costs and institutional constraints can affect efficiency and capital allocation. As a result, the net effect of ESG on risk, capital, and efficiency is theoretically ambiguous and likely context-dependent, particularly in emerging banking systems.

Classical empirical evidence from [1,2,3] shows that regulatory capital frameworks generally succeed in strengthening buffers without necessarily inducing portfolio risk-shifting. Other studies, such as [4,5] find that higher capital levels are associated with lower credit risk, supporting the moral hazard argument that well-capitalized banks tend to take fewer risks. In contrast, ref. [6] identify a U-shaped relationship between capital and credit risk, suggesting that risk declines initially as capital increases but rises again beyond a certain threshold. These competing findings illustrate the theoretical ambiguity surrounding the capital–risk relationship, especially in emerging markets with diverse institutional features.

Research linking capital regulation, risk-taking, and bank efficiency originates with [7], who show that efficiency and capitalization shape banks’ risk profiles. Similarly, ref. [8] argue that efficient banks are better able to manage risk and adjust capital, whereas inefficient banks may assume additional risk to compensate for weak performance. Together, this literature implies that risk, capital, and efficiency are jointly determined and dynamically interdependent rather than separable dimensions of bank behavior.

In this context, ESG performance has emerged as a potential moderating force. Studies [9,10,11] show that ESG engagement enhances monitoring quality, reduces information asymmetry, and lowers default risk, while [12,13,14] document improvements in loan quality, efficiency, and liquidity creation. However, ESG adoption may also involve short-term compliance and restructuring costs [15], suggesting that its effects on prudential outcomes may be heterogeneous or non-linear.

Efficiency represents another core determinant of banking soundness. According to the efficient structure hypothesis, well-managed and cost-efficient banks achieve superior profitability and risk control through better resource allocation [16]. Conversely, inefficiency weakens risk management capacity and erodes capital buffers [17]. Moreover, ESG practices can enhance operational discipline and organizational processes, though their impact may vary across institutional environments. Recent evidence, including [15], suggests that ESG practices may reinforce or reshape this risk–capital–efficiency nexus, but existing findings remain fragmented.

Despite extensive literature on risk–capital and efficiency relationships, most studies particularly in the MENA region, examine these dimensions in isolation, rely on static or single-equation models, and implicitly assume exogeneity. As a result, the joint, dynamic ESG–risk–capital–efficiency nexus remains insufficiently understood, especially in banking systems where ESG adoption is recent and institutional constraints may alter transmission mechanisms. This study addresses this gap by empirically examining the dynamic and jointly endogenous interactions between ESG performance, credit risk, capital adequacy, and efficiency in MENA banks. Using an unbalanced panel of 167 commercial banks from 13 MENA countries over 2015–2024, the analysis explicitly models feedback effects across all three prudential dimensions, overcoming limitations of prior static and partial approaches.

The contributions of this paper are threefold. First, it fills a gap in the MENA literature by integrating ESG performance directly into a dynamic risk–capital–efficiency framework. Second, it advances the efficiency literature by treating efficiency as both an outcome and a driver of prudential behavior rather than an exogenous control. Third, it provides new regional evidence using a dynamic System-GMM approach that addresses persistence, simultaneity, and endogeneity shortcomings that characterize much of the existing MENA banking literature.

The remainder of this paper is organized as follows. Section 2 presents the literature review and develops the hypotheses. Section 3 describes the data and methodology. Section 4 discusses the empirical results, and Section 5 concludes with policy implications and directions for future research.

2. Literature Review and Hypothesis Development

The relationship between bank risk, capital, and efficiency has long been analyzed through the lenses of regulation, moral hazard, market discipline, and managerial efficiency. However, empirical findings remain heterogeneous and context-dependent, particularly in emerging markets and ESG-driven environments, highlighting the need for a system-based approach to jointly analyze these dimensions.

From a regulatory perspective, early contributions by [18,19,20] show that higher capital requirements may, under some conditions, induce banks to increase portfolio risk in order to preserve target returns, creating a potential regulatory paradox. Ref. [1] formalize the joint adjustment of capital and risk in a partial adjustment framework, while the buffer capital theory and franchise value hypothesis [3,21] argue that well-capitalized banks hold capital above the regulatory minimum to absorb shocks and protect charter value, thereby restraining risk-taking.

From a moral hazard angle, undercapitalized or poorly supervised banks may take excessive risk when deposit insurance and implicit guarantees weaken market discipline [22]. Taken together, these strands indicate that capital regulation does not exert a uniform disciplining effect; instead, its impact on risk depends on institutional quality, supervisory pressure, and market incentives. Consequently, the sign and magnitude of the capital–risk relationship may be non-linear [6] or asymmetric across bank types, justifying a joint modeling framework rather than isolated single-equation approaches.

Efficiency adds a third dimension. The efficient structure hypothesis suggests that more efficient banks generate higher profits, accumulate capital internally, and can manage risk exposures more effectively, while the bad management hypothesis posits that inefficiency in screening and monitoring leads to higher non-performing loans and risk [8,23,24]. This duality implies that efficiency simultaneously influences both risk and capital, reinforcing the argument that these three dimensions are endogenously determined by managerial quality, regulatory constraints, and market discipline. Prior studies document both positive and negative associations between efficiency and risk or capital, suggesting that outcomes vary with ownership structure, competition, and institutional settings rather than reflecting a universal mechanism.

The integration of ESG into the capital–risk–efficiency framework introduces additional theoretical foundations that clarify how ESG performance can influence bank behavior. Stakeholder theory [25] argues that addressing the interests of multiple stakeholders reduces conflicts, strengthens long-term relational stability, and ultimately lowers risk-taking incentives. From the perspective of the resource-based view [26,27], ESG engagement builds valuable intangible assets such as trust, legitimacy, and reputational capital that enhance operational efficiency and improve capital adequacy through better profitability and access to stable funding. Risk management theory views ESG practices as mechanisms for identifying, mitigating, and controlling environmental, social, and governance risks, thereby directly reducing both credit and operational risk [28,29]. However, ESG effects may not be uniformly positive. ESG initiatives may involve short-term compliance costs, non-financial investments, or heterogeneous impacts across environmental, social, and governance pillars [15], particularly in emerging markets where disclosure standards and supervisory enforcement differ.

According to slack resource theory, financially strong banks with surplus resources are more capable of investing in ESG, which in turn reinforces stability and strengthens their capital buffers [30,31], while reputation theory [32,33] highlights ESG’s role in reducing information asymmetry and disciplining risk-taking. Collectively, these theories identify multiple ESG transmission channels, disciplinary, informational, reputational, and prudential, operating simultaneously across risk, capital, and efficiency. This theoretical multiplicity directly motivates treating ESG and the three prudential dimensions as jointly endogenous.

2.1. Modeling Bank Risk: The Roles of Capital, Efficiency, and ESG

A large empirical literature examines risk as the outcome of capital decisions, bank efficiency, and more recently, ESG performance.

On the capital side, many studies find that higher capital reduces bank risk, consistent with the buffer capital and franchise value views. Ref. [34] shows that the introduction of risk-based capital standards in U.S. banks increased capital ratios and reduced portfolio risk, while [5] reports that declines in capital precede higher risk in Spanish banks. Cross-country evidence further supports the stabilizing role of capital [35,36].

However, other works find non-linear or ambiguous patterns; ref. [6] identifies a U-shaped relationship between capital and risk, and [37,38] report weak or insignificant links. Rather than implying contradictory evidence, these fundings indicate that capital affects risk conditionally depending on regulation, market discipline, and bank-specific incentives.

Regulatory pressure constitutes an additional channel shaping banks’ risk-taking behavior. Ref. [1] show that banks facing regulatory capital pressure adjust capital and risk simultaneously to avoid supervisory intervention while [3,21] emphasize precautionary buffer-building. In this sense, regulatory pressure conditions how capital, efficiency, and ESG translate into observed risk outcomes, rather than acting as an independent determinant.

Efficiency also drives risk outcomes. Inefficient banks are more prone to credit problems due to weak underwriting or compensatory risk-taking [23,39]. Ref. [17] finds that certain risk measures are positively associated with efficiency in Chinese banks, while others show opposite patterns. This ambiguity supports modeling efficiency without imposing a priori sign restrictions.

Recent ESG studies explicitly link sustainability practices to risk reduction. Refs. [9,40,41] show that ESG engagement lowers default and idiosyncratic risk, while [10] and [42] document reduced insolvency and liquidity risk in emerging markets. Ref. [43] distinguishes proactive from passive risk-taking, and [44] highlights ESG’s role in improving credit screening and regulatory oversight. Importantly, ESG effects differ across pillars, suggesting heterogeneous transmission mechanisms.

Overall, the literature suggests that ESG performance is expected to discipline risk-taking on average, but with heterogeneous effects across institutional contexts and ESG components.

Risk emerges as the outcome of interacting capital decisions, efficiency conditions, regulatory pressure, and ESG engagement. These interactions are dynamic and mutually reinforcing, motivating a risk equation embedded within a jointly endogenous system.

Building on the theoretical foundations and empirical findings, we propose the following hypothesis.

H1. 

Higher ESG performance and bank efficiency are generally associated with lower credit risk, while the effect of capital adequacy on risk depends on regulatory pressure and institutional conditions.

2.2. Modeling Capital Adequacy: The Roles of Risk, Efficiency, and ESG

The reverse relationship capital as an outcome of risk and efficiency decisions is also well-documented.

Ref. [1] shows that banks jointly adjust capital and risk, while [3,21] confirm buffer-building behavior in response to risk and regulatory pressures. Capital adequacy therefore reflects both precautionary motives and responses to changing risk profiles.

Regulatory pressure plays a central role in shaping banks’ capital adequacy decisions. Banks hold buffers above minima to reduce supervisory intervention costs [3,21] and ref. [1] shows that tighter constraints induce simultaneous capital increases and risk adjustment. Thus, regulatory pressure governs how risk and ESG performance translate into capital accumulation rather than acting independently. Thus, capital adequacy reflects endogenous responses to risk exposure, efficiency-driven internal capital generation, and supervisory constraints.

Efficiency influences capital through internal capital generation. Efficient banks generate higher retained earnings and sustain stronger capital ratios [7,24], with similar evidence from Malaysia, Korea, and the MENA region [45,46,47].

However, higher capital may also entail costs. Refs. [16,48] show that excessive capitalization can reduce efficiency, particularly in smaller banks.

ESG may affect capital adequacy indirectly through improved risk profiles and profitability. Refs. [49,50,51] show that ESG performance enhances stability and resilience, while [42] link ESG disclosure to higher profitability and market valuation.

These studies justify specifying a capital adequacy equation in which capital depends on risk, efficiency, and ESG performance, with the expectation that higher efficiency, and stronger ESG engagement are associated with stronger and more stable capital ratios. Banks experiencing higher credit risk often respond by strengthening their capital positions to absorb potential losses and satisfy regulatory expectations, in line with the capital buffer theory [3]. At the same time, ESG-oriented institutions benefit from enhanced transparency, reputational capital, and lower funding costs, which facilitate the accumulation of stronger capital ratios [52,53]. Moreover, operationally efficient banks generate higher retained earnings and enjoy better internal capital formation, allowing them to maintain higher levels of regulatory capital compared with less efficient peers [16,54]. Given these channels, capital adequacy depends on risk, efficiency, and ESG performance, but the direction of effects may vary across contexts.

Capital buffers emerge from a combination of internal efficiency, external regulatory pressure, risk exposure, and ESG-driven credibility. This reinforces the treatment of capital as an endogenous adjustment variable within a multi-equation framework.

Considering prior evidence, the second hypothesis is formulated as follows:

H2. 

Capital adequacy increases with ESG performance and efficiency, while its response to higher credit risk depends on regulatory pressure and buffer-building dynamics.

2.3. Modeling Efficiency: The Roles of Risk, Capital, and ESG

Efficiency itself can be viewed as an outcome of risk-taking, capitalization, and ESG policies.

Well-capitalized banks tend to be more efficient due to prudent management and lower funding costs [7,37], while refs. [47,55] confirm these patterns internationally and in MENA. However, refs. [16,48] document trade-offs where high capital reduces cost efficiency. Risk also feeds into efficiency. High risk may reflect poor management or underinvestment in monitoring [16,23,56], while conservative risk strategies may constrain efficiency. Empirical evidence remains mixed [17], reinforcing the need for dynamic modeling. ESG increasingly emerges as a driver of efficiency. Socially responsible banks often exhibit better performance [57,58,59,60,61], and ESG compliance improves efficiency in emerging regions [62]. However, ref. [15] highlight short-term cost pressures associated with ESG adoption. These findings suggest that ESG’s efficiency effects are conditional and pillar-specific rather than uniformly positive.

Efficiency reflects the net outcome of prudential strength, risk discipline, and ESG-related organizational practices, supporting its endogenous role within the triadic system. Drawing on the literature, we advance the following hypothesis:

H3. 

Bank efficiency is positively influenced by ESG and by capital adequacy, while higher risk weakens efficiency.

Although the literature provides mixed and sometimes non-linear evidence regarding the relationships between risk, capital, and efficiency, the present study adopts a linear dynamic specification for two reasons. First, System-GMM is designed to accommodate endogenous feedback effects, thereby capturing the bidirectional nature of prudential adjustments without requiring explicit non-linear functional forms. Second, our hypotheses reflect the most prevalent patterns identified in prior research and are contextualized to the institutional and regulatory structures of MENA banking.

3. Research Design

3.1. Data Sources and Sample Selection

This study analyzes the dynamic relationship between credit risk, capital adequacy, efficiency, and ESG performance in a panel of 167 commercial banks operating in 13 MENA countries (

Table 1. Presentation of the sample.

N°	Countries	Bank in the Country	Bank in Sample	Percentage in Sample
1	Bahrain	83	13	7.8%
2	Egypt	37	14	8.3%
3	Jordan	20	15	9%
4	Kuwait	11	10	6%
5	Iraq	54	33	19.7%
6	Lebanon	57	6	3.6%
7	Morocco	19	8	4.8%
8	Oman	19	7	4.2%
9	Qatar	18	8	4.8%
10	Saudi Arabia	30	10	6%
11	Syria	21	13	7.8%
12	Tunisia	23	12	7.2%
13	United Arab Emirates	40	18	10.8%
	Total	432	167	100%

), namely Saudi Arabia, the United Arab Emirates, Qatar, Kuwait, Bahrain, Oman, Iraq, Jordan, Syria, Lebanon, Egypt, Morocco, and Tunisia, over the period 2015–2024.

The dataset is an unbalanced panel because observations for some banks and years are missing particularly ESG scores, financial statements, or macro-level gaps. Instead of imputing missing values, all available observations are retained, as the System-GMM estimator explicitly accommodates unbalanced panels and provides consistent estimation without requiring balanced data structures. After accounting for missing observations, lagged variables, and DEA score construction, the final estimation sample comprises 1026 bank-year observations.

The study period captures the post-Basel III regulatory phase and the increasing diffusion of ESG reporting practices across the region.

Bank-level financial data were collected from Refinitiv Eikon, complemented by annual reports when needed. Macroeconomic indicators, including GDP growth, inflation, and domestic credit to the private sector in percentage of GDP (DCPS), were collected from the World Bank’s World Development Indicators (WDIs).

ESG variables (overall ESG score and E, S, G sub-components) were sourced from the Refinitiv ESG database, which provides standardized metrics ensuring comparability across institutions and years. The sample selection is motivated by both conceptual and empirical considerations. First, the MENA banking sector represents a unique context in which financial institutions play a dominant role in financing national development strategies, diversification plans, and sustainability transitions. Banks in this region operate within regulatory frameworks that are evolving toward Basel III and ESG integration, making them an ideal setting for studying the interaction between risk, capital, efficiency, and sustainability performance. Second, the inclusion of 167 commercial banks across 13 MENA countries ensures broad coverage of the region’s heterogeneous banking systems, allowing the analysis to capture cross-country regulatory, institutional, and macroeconomic differences that influence prudential behavior. Third, the study period (2015–2024) is chosen to reflect the post-Basel III implementation phase and the period during which ESG reporting and sustainability initiatives expanded significantly across MENA markets. This timeframe also allows the model to incorporate recent economic shocks, structural reforms, and sustainability-driven financial policies. Finally, the choice of commercial banks is driven by data availability and comparability, as these institutions consistently disclose financial statements and ESG indicators, enabling a reliable analysis of the risk–capital–efficiency nexus.

3.2. Variable Selection and Definition

To empirically examine the dynamic relationship between risk, capital strength, efficiency, and ESG performance in MENA banks, this study constructs a set of dependent, explanatory, and control variables grounded in established banking and sustainability literature (

Table 2. Presentation of variables.

Variable	Definition	Measure
Dependent variable
Risk	Credit risk/deterioration in asset quality	${\displaystyle \frac{Non Performing Loans}{Gross Loans}}$
EQTA	A bank’s capital adequacy	${\displaystyle \frac{Equity}{Total Assets}}$
Eff	Technical banking efficiency	Two-stage DEA: output-oriented model using two inputs (interest expenses, labor expenses) and two outputs (loans, investment securities) computed in R
Independent variable
ESG	Overall Environmental, Social and Governance performance	ESG composite score from Refinitiv Eikon
ENVIR	Environmental pillar of ESG	Environmental score from Refinitiv Eikon
SOC	Social pillar of ESG	Social score from Refinitiv Eikon
GOV	Governance pillar of ESG	Governance score from Refinitiv Eikon
Control variables
REG	Regulatory pressure/distance from minimum capital requirement	Regulatory pressure (REG): Regulatory stringency is computed according to the work of Shrieves and Dahl (1992) [1] REG = 0, if CARi,t > min reg + σCAR REG = min reg + σCAR − CAR, if CARi,t < min reg + σCAR
LLRNL	Loan loss reserves ratio (forward-looking credit risk buffer)	${\displaystyle \frac{Loan loss reserves}{Gross loans}}$
CTI	Cost-to-income ratio	${\displaystyle \frac{operating expenses}{operating income}}$
ROA	Return on assets	${\displaystyle \frac{Net Income}{Total Assets}}$
LDR	Loan-to-deposit ratio (liquidity and lending intensity)	${\displaystyle \frac{Net Loans}{Total Deposit}}$
GDPG	GDP growth	Annual GDP growth rate from World Development Indicators
INF	Inflation rate	Consumer Price Index growth from World Development Indicators
DCPS	domestic credit to the private sector (% of GDP)	From World Development Indicators

).

3.2.1. Dependent Variable

Risk (RISK): Credit risk is measured using the ratio of non-performing loans to total gross loans. This is the standard proxy for asset quality deterioration and has been widely used in prior research [9,15,17,48,63,64,65].

Capital Adequacy (EQTA): Capital strength is captured by the equity to total assets ratio, a standard solvency indicator reflecting banks’ capacity to absorb losses [15,17,64]. Higher EQTA generally indicates stronger capitalization and aligns with Basel regulatory principles.

Banking Efficiency (EFF): Efficiency is measured using output-oriented technical efficiency scores derived from a two-stage Data Envelopment Analysis (DEA) implemented in R. DEA is a non-parametric frontier approach based on the Pareto Koopmans notion of efficiency, which defines a bank as efficient only if no input or output can be improved without worsening another [66].

Unlike parametric methods, DEA does not impose a functional form on the production process and can handle multiple heterogeneous inputs and outputs [67]. Consistent with the intermediation approach in the banking literature, the model treats banks as institutions that transform financial and human resources into earning assets. The DEA was estimated under variable returns to scale (VRS) and an output orientation, which is appropriate when banks seek to maximize the production of earning assets given available inputs.

This approach evaluates how effectively banks transform inputs (interest expenses and labor expenses) into outputs (loans and investment securities). DEA-based measures are consistent with recent applications in sustainable banking research [15,17,68,69,70].

Efficiency scores range from 0 to 1, where 1 indicates full efficiency on the frontier, and scores below 1 reflect relative inefficiency requiring input reduction and/or output enhancement.

3.2.2. Independent Variable

ESG Performance (ESG): Banks’ sustainability engagement is proxied by a composite ESG score (0–100), encompassing environmental, social, and governance indicators. This variable reflects the extent of responsible practices and disclosure, consistent with recent work in MENA and global contexts [15,68,71,72].

3.2.3. Control Variable

To isolate the core effects, several financial and macroeconomic controls are included.

Regulatory pressure (REG): Regulatory stringency is computed as described by Shrieves and Dahl (1992) [1]. The measure captures the deviation of a bank’s capital adequacy ratio (CAR) from the minimum required level:

$$REG=\left\{\begin{array}{c}0 if {CAR}_{it}>MinREG+{\sigma }_{CAR}\\ MinREG+{\sigma }_{CAR} if {CAR}_{it}<MinREG+{\sigma }_{CAR}\end{array}\right.$$

where $MinREG$ is the minimum required capital ratio and ${\sigma }_{CAR}$ is its standard deviation.

Bank Size (SIZE): measured by the log of total assets, capturing scale effects and diversification benefits [17,63,64,73].

Profitability (ROA): the return on assets, reflecting management quality and operational performance [17,47,63,74,75] measured by the ratio of net income to total assets.

Loan-to-deposit ratio (LDR): measuring liquidity transformation and funding structure [16,47].

Loan loss reserves to net loans (LLRNLs): this captures provisioning behavior and expected credit losses [65].

Cost-to-income ratio (CTI): a representation of operational efficiency and cost structure measured by the operating expenses to operating income ratio [75].

Macroeconomic conditions: GDP growth (GDPCG), inflation (INF), and domestic credit to the private sector (% of GDP, DCPS), controlling for systemic and financial development effects [17,47,63,76].

3.3. Model Specification and Estimation Method

To examine the simultaneous and dynamic interactions among risk, capital, efficiency, and ESG performance, this study estimates the following system of three equations, adapted from [15,16,48]:

$${RISK}_{i,t}={\alpha }_0+{\alpha }_1{RISK}_{i,t-1}+{\alpha }_{2}EQT{A}_{i,t}+{\alpha }_{3}ES{G}_{i,t}+{\alpha }_4{EFF}_{i,t}+{\alpha }_{5}RE{G}_{i,t}+{\alpha }_6{X}_{i,t}+{\mu }_i+{\epsilon }_{i,t}$$

(1)

$$EQT{A}_{i,t}={\beta }_0+{\beta }_{1}EQT{A}_{i,t-1}+{\beta }_{2}NP{L}_{i,t}+{\beta }_3{EFF}_{i,t}+{\beta }_{4}ES{G}_{i,t}+{\beta }_{5}RE{G}_{i,t}+{\beta }_6{Z}_{i,t}+{\eta }_i+{\nu }_{i,t}$$

(2)

$$E{FF}_{i,t}={\gamma }_0+{\gamma }_{1}E{FF}_{i,t-1}+{\gamma }_{2}EQT{A}_{i,t}+{\gamma }_{3}NP{L}_{i,t}+{\gamma }_{4}ES{G}_{i,t}+{\gamma }_5{W}_{i,t}+{\theta }_i+{\xi }_{i,t}$$

(3)

where $X_{i,t}$, $Z_{i,t}$, and $W_{i,t}$ are vectors of control and macroeconomic variables (SIZE, ROA, LTDR, GDP growth, inflation, and DCPS).

To empirically assess the dynamic interactions among risk, capital, and efficiency, while accounting for the growing importance of sustainability practices, this study employs the System Generalized Method of Moments (System-GMMs) estimator. This approach is particularly suited for addressing endogeneity, simultaneity, and dynamic persistence in panel data where the dependent variable is influenced by its past realizations. The mutual interdependence between bank risk-taking, capitalization, and operational efficiency makes traditional static estimators (fixed or random effects) inappropriate, as they would produce biased and inconsistent estimates due to correlation between lagged variables and unobserved effects.

Following the procedures developed by [77,78], the System-GMM estimator combines equations in first differences and levels, using lagged levels as instruments for the differenced equation and lagged differences as instruments for the level equation. This dual instrumentation improves efficiency and mitigates the weak-instrument bias commonly associated with the Difference-GMM estimator of [79]. As emphasized by [80], the System-GMM estimator is particularly robust when the number of cross-sectional units (N) is larger than the number of time periods (T), as in our case where N = 167 MENA banks and T = 10 years (2015–2024). This method also effectively controls for unobserved heterogeneity, simultaneity bias, and the potential endogeneity of explanatory variables, ensuring consistent and efficient parameter estimation.

To capture the multidimensional nature of bank soundness, we specify a system of three dynamic equations estimated simultaneously under the System-GMM framework. Each equation represents a key dimension of banking performance:

The first equation seeks to explain bank credit risk (RISK) as a function of its own lag, capital adequacy, efficiency, ESG performance, and control variables (bank size, profitability, liquidity, and macroeconomic indicators).

The second equation models capital adequacy (EQTA) as a dynamic outcome influenced by lagged value, risk exposure, efficiency, ESG performance, and the same set of controls.

The third equation explains bank efficiency (EFF) measured by DEA scores as a function of lagged efficiency, risk, capital adequacy, ESG performance, and relevant controls.

In all three equations, the ESG performance score is introduced as a key explanatory variable, capturing how sustainability engagement influences banks’ prudential behavior and operational effectiveness. This tri-equation setup allows us to explore not only the direct effects of ESG on risk, capital, and efficiency, but also the feedback mechanisms among these three dimensions of financial soundness. The structure is consistent with the integrated frameworks developed by [16,17], and extended by [15,48] to include sustainability and governance factors in the risk–capital–efficiency nexus.

The validity of the System-GMM estimations is verified through the Hansen and Sargan tests of over-identifying restrictions and the Arellano–Bond AR(1) and AR(2) serial correlation tests. The models are estimated using collapsed instruments to avoid instrument proliferation, and robust standard errors are applied as described in [81], correction to ensure finite-sample reliability. Overall, this estimation strategy provides a rigorous and dynamic framework for analyzing the sustainable banking and financial soundness nexus in MENA countries.

Model validity is assessed through the Hansen test (and Sargan test) for over-identifying restrictions and the Arellano–Bond AR(1) and AR(2) tests for serial correlation.

Based on prior theoretical and empirical evidence, capital adequacy is expected to negatively affect credit risk, consistent with the buffer-capital and franchise-value hypotheses.

Higher efficiency and stronger ESG performance are also anticipated to reduce credit risk through enhanced monitoring and governance quality.

Efficiency and ESG are expected to positively influence capital, reflecting that efficient and sustainable banks accumulate capital internally and attract long-term investors.

Finally, credit risk is expected to negatively influence efficiency, as rising non-performing loans increase provisioning costs and constrain productive intermediation.

This methodology enables a comprehensive assessment of how prudential soundness and sustainability interact within the MENA banking system, offering dynamic evidence on the ESG-driven risk–capital–efficiency nexus.

Risk, capital adequacy, and efficiency are treated as jointly endogenous because they are determined simultaneously through banks’ managerial decisions, regulatory constraints, and past performance. Higher risk affects capital buffers and efficiency, while capital and efficiency in turn influence risk-taking incentives, generating reverse causality and dynamic feedback effects. The System-GMM estimator addresses these identification challenges by using internal instruments based on lagged values of the endogenous variables, thereby controlling for simultaneity, unobserved bank-specific heterogeneity, and dynamic persistence. Compared with static fixed-effects or three-stage least squares estimators, System-GMM provides consistent and efficient estimates of the causal relationships within the risk–capital–efficiency–ESG nexus.

4. Empirical Results and Analysis

4.1. Descriptive Statistics

Table 3. Descriptive statistics.

Stats	Obs	Mean	Median	SD	p25	p75
SIZE	1615	24.719	24.934	2.613	22.558	26.692
EQTA	1615	0.216	0.133	0.198	0.104	0.201
EFF	1670	0.338	0.251	0.275	0.136	0.437
ROA	1670	0.013	0.011	0.116	0.005	0.017
NPL	1670	0.082	0.026	0.505	0	0.062
LLR	1670	0.084	0.029	0.511	0	0.054
LTDP	1670	6.225	0.842	79.322	0.627	1.043
REG	1615	0.036	0.0025	0.063	0	0.042
ESG	1090	39.12	38.657	17.594	26.492	52.246
CTI	1615	0.019	0.021	0.111	0.015	0.0287
GDP	1651	−0.0049	−0.00025	0.048	−0.026	0.020
INF	1567	0.071	0.021	0.215989	0.0057	0.0504
DCPS	1233	3.667	4.112	0.873765	2.635	4.383

reports the descriptive statistics for the main variables used in the analysis for 167 MENA banks over the period 2015–2024. The average bank size, measured as the natural logarithm of total assets, is 24.72, with moderate dispersion (SD = 2.61), indicating substantial variation across banks but a generally balanced representation of both large and mid-sized institutions. The mean equity to total assets ratio (EQTA) is 0.22, reflecting moderate capitalization levels typical of MENA banks under Basel III frameworks. The average technical efficiency (EFF) derived from DEA equals 0.34, with values ranging from near 0 to 1, suggesting significant room for improving operational efficiency. The mean return on assets (ROAs) of 1.3% and the low median indicate that profitability is modest, while the high standard deviation reveals heterogeneous performance across banks. The RISK measured by the non-performing loan ratio has a mean of 8.3% and a median of 2.6% suggest that, although most banks maintain healthy asset quality, several institutions suffer from episodes of credit deterioration, a pattern which is common in emerging financial systems.

The loan loss reserves ratio (LLRNL) displays low means but high variability, confirming cyclical provisioning behavior in response to changing credit risk. The mean capital adequacy ratio of 30.7% and median of 18.1% indicate that, while most banks are well-capitalized, some hold exceptionally high capital buffers, leading to a large dispersion (SD = 50.8). The average ESG score is 39.1 (out of 100), implying that MENA banks are still in an early stage of integrating sustainability practices, though top quartile performers achieve scores above 52. Macroeconomic indicators show modest real GDP growth (mean ≈ −0.5%) and low inflation (7.1%), consistent with a relatively stable macro-financial environment during the sample period. The domestic credit to private sector ratio (DCPS) averages 3.67, indicating the banking sector’s central role in regional financing. Overall, the descriptive statistics reveal substantial heterogeneity among MENA banks in terms of size, capitalization, risk exposure, and sustainability performance, providing a suitable basis for investigating the dynamic risk–capital–efficiency–ESG nexus through the System-GMM framework.

Table 4. Correlation matrix.

	EFF	SIZE	EQTA	ROA	LDR	NPL	LLR	ESG	GDP	CTI	INF	DCPS	REG
EFF	1
SIZE	0.2867 ***	1
EQTA	0.0574 ***	0.2696 ***	1
ROA	−0.0545 **	0.1291 ***	−0.0113 ***	1
LDR	0.1224 ***	0.0205	0.0944 **	−0.0243	1
NPL	−0.0152 *	0.0247	0.0084 **	−0.0517 **	−0.0068	1
LLR	−0.0022	0.0745 ***	0.0714	−0.004	−0.0115	0.6538 ***	1
ESG	0.0305 *	0.3336 ***	−0.0017 ***	0.1229 ***	0.0751 **	−0.1607 ***	−0.0847 **	1
GDP	0.0147	−0.0089	−0.0503 **	−0.0336 *	−0.019	0.0211	−0.0211	0.0101	1
CTI	−0.0795 ***	0.0679 ***	−0.034 ***	0.9242 ***	−0.0216	−0.0429 *	0.0074	−0.0357	−0.0271	1
INF	0.18 ***	0.3032 ***	−0.102 *	0.0097	−0.0176	−0.0036	−0.0044	−0.0528	−0.1501 ***	0.0184	1
DCPS	−0.1292 ***	−0.5398 ***	−0.6165 **	0.047 *	−0.0786 ***	−0.0642 **	−0.0995 ***	0.2209 ***	−0.1031 ***	0.0956 ***	−0.0661 ***	1
REG	−0.0698 ***	−0.3342 ***	−0.3125 ***	−0.0229 *	−0.0338	−0.0221	−0.0373	−0.0448	−0.019	0.025	−0.0453 *	0.2828 ***	1
VIF	1.6	2.23	1.16	2.97	3.37	13.75	13.88	1.51	1.19	3.6	2.12	2.82	1.76

Notes: (***), (**) and (*) significative, respectively, at 1%, 5%, and 10%.

presents the Pearson correlation coefficients among the main variables. The results show relatively low to moderate correlations, suggesting the absence of serious multicollinearity problems and supporting the suitability of the variables for multivariate estimation. Efficiency (EFF) exhibits a positive correlation with bank size (r = 0.29) and equity ratio (r = 0.06), implying that larger and better-capitalized banks tend to operate more efficiently. This relationship reflects the scale and resource advantages that enable larger institutions to invest in technology and managerial capabilities. Efficiency is also weakly and negatively related to NPLs (r = −0.02) and regulatory pressure (r = −0.07), suggesting that banks with higher inefficiency tend to face tighter supervisory constraints and elevated credit risk consistent with the “bad management” hypothesis [23].

Capital adequacy (EQTA) is moderately correlated with size (r = 0.27) and efficiency (r = 0.06), confirming that capitalization is higher among larger and more efficient banks. ESG performance (ESG) shows a positive correlation with size (r = 0.33) and ROA (r = 0.12), indicating that more profitable and larger banks are typically more engaged in sustainability initiatives. The negative correlations between ESG and RISK (r = −0.16) and between ESG and REG (r = −0.04) further suggest that banks with better ESG profiles face lower risk and less regulatory pressure. Among macroeconomic indicators, GDP growth and inflation rate (INF) display negligible correlations with bank-specific variables, while DCPS is negatively correlated with size (r = −0.54) and capital (r = −0.62), reflecting structural financial depth differences across MENA economies. Overall, the correlation structure confirms that the variables are economically meaningful and statistically distinct, justifying their joint inclusion in the dynamic System-GMM estimation of the risk–capital–efficiency–ESG nexus.

Before estimating the System-GMM models, multicollinearity was assessed using both the Pearson correlation matrix and the Variance Inflation Factor (VIF). The analysis revealed that the loan loss reserves ratio (LLR) and the non-performing loans ratio (RISK) are strongly correlated (r = 0.65), confirming that they measure closely related dimensions of credit risk. The VIF values for these variables exceeded the acceptable threshold of 10 when included simultaneously, indicating a potential multicollinearity problem.

To address this issue and ensure model stability, the two variables were not entered together as explanatory variables within the same equation. Instead, non-performing loan was retained as an indicator of credit risk, consistent with prior studies such as [15,16,48,64,65], while LLR was alternatively used in robustness tests as a measure of forward-looking provisioning behavior. This specification strategy prevents redundancy and ensures that estimated coefficients capture distinct effects of credit risk and capital adequacy on efficiency and ESG performance.

4.2. Multivariate Analysis

Before analyzing the coefficients, several post-estimation diagnostics confirm the robustness of the System-GMM specification.

Across the three estimated equations for risk, capital, and efficiency, the Arellano–Bond AR(1) test is statistically significant, as expected for the first differences, while the AR(2) test is insignificant, indicating the absence of second-order serial correlation.

The Hansen and Sargan tests of over-identifying restrictions consistently show high p-values (above 0.20), confirming the validity of the instrumental variables and the absence of instrument proliferation bias.

Overall, these results validate the dynamic panel structure and the appropriateness of the GMM estimator in addressing endogeneity and simultaneity among the main variables.

4.2.1. Determinants of Banking Risk

The first equation focuses on explaining credit risk (RISK), which is captured by the ratio of non-performing loans to total gross loans.

In the MENA context, this finding reflects how banks with higher ESG commitments tend to adopt prudent lending strategies aligned with national sustainability frameworks, thereby reducing non-performing loans.

The results from the four dynamic System-GMM models (

Table 5. System-GMM estimation of the risk equation.

	RISK	RISK	RISK	RISK
L.RISK	0.699 ***	0.6658 ***	0.743 ***	0.652 ***
	(0.0380)	(0.0312)	(0.051)	(0.028)
EQTA	0.209 *	0.2407 **	0.112	0.247 **
	(0.1094)	(0.1168)	(0.1148)	(0.123)
ESG	−0.1057 ***
	(0.0342)
ENVIR		−0.0315
		(0.0218)
SOC			−0.13 ***
			(0.041)
GOV				−0.066 **
				(0.025)
EFF	−0.0491 ***	−0.0465 **	−0.0299 *	−0.057 ***
	(0.0172)	(0.0211)	(0.0175)	(0.0202)
SIZE	0.0183 **	0.0027	0.0262 ***	0.007
	(0.0074)	(0.005)	(0.0093)	(0.0051)
ROA	−0.0710	−0.0867	−0.0757	−0.069
	(0.1133)	(0.0884)	(0.120)	(0.1292)
LTDP	−0.0633	−0.0546	−0.0677	−0.0689 *
	(0.0443)	(0.042)	(0.048)	(0.037)
LLR	0.5441 ***	0.5957 ***	0.4988 ***	0.584 ***
	(0.0323)	(0.0299)	(0.0455)	(0.0248)
REG	0.5293 *	−0.0921	0.8578 **	0.1343
	(0.2733)	(0.210)	(0.3896)	(0.2065)
GDP	−0.1062 ***	−0.1065 **	−0.116 ***	−0.093 ***
	(0.0331)	(0.0421)	(0.0329)	(0.0329)
INF	−0.0249	−0.0105	0.009	−0.031
	(0.0344)	(0.0306)	(0.0525)	(0.0271)
DCPS	0.0515 ***	0.0287 ***	0.065 ***	0.0414 ***
	(0.0138)	(0.009)	(0.0189)	(0.0126)
_cons	−0.6127 ***	−0.1491	−0.866 ***	−0.2780 *
	(0.2095)	(0.1133)	(0.2848)	(0.1529)
Observation	1026	1026	1026	1026
Fisher	3445.80 ***	4893.16 ***	3029.89 ***	4310 ***
Arellano–Bond test for AR(1)	−2.92 *** (p = 0.003)	−2.79 *** (p = 0.005)	−2.98 *** (p = 0.003)	−2.93 *** (p = 0.003)
Arellano–Bond test for AR(2)	−0.09 (p = 0.929)	−0.09 (p = 0.932)	0.24 (p = 0.809)	−0.03 (p = 0.975)
Sargan test	18.63 (p = 0.481)	21.26 (p = 0.322)	12.92 (p = 0.881)	23.11 (p = 0.233)
Hansen test	19.19 (p = 0.445)	19.29 (p = 0.438)	14.82 (p = 0.787)	19.31 (p = 0.437)

Notes: (***), (**) and (*) are significative at 1%, 5%, and 10%. Values between parentheses indicate standard deviation error.

) show that credit risk in MENA banks exhibits a clear and significant persistence across time, as indicated by the positive and highly significant lagged NPL coefficient in all specifications. This finding is fully consistent with earlier studies such as [9,15] for US banks and ref. [48] for GCC banks, which document the slow-moving nature of NPLs. This reflects structural weaknesses in loan quality management due to weak recovery processes, judicial inefficiencies, and structural rigidities in emerging markets. There is risk persistence due to long-term credit exposures and cyclical effects.

In addition, the positive relationship between capital adequacy and risk (significant in three out of four models) supports the risk-shifting hypothesis, whereby better capitalized banks may engage in riskier lending activities. This pattern echoes the evidence presented by [17,64,82], who observe that higher capital buffers do not always translate into safer credit portfolios in developing financial systems. In the MENA context, the positive coefficient can be interpreted as indicating that higher capital allows banks more freedom to expand lending aggressively, consistent with pro-risk incentive structures.

A central insight of this study concerns the role of ESG performance and its components in shaping bank credit risk. Across the four models, the overall ESG score significantly reduces NPLs, reinforcing the global evidence that sustainability engagement promotes prudent risk-taking and strengthens asset quality.

This supports the stakeholder and reputational capital hypotheses, according to which sustainable banks benefit from enhanced transparency, stakeholder confidence, and better screening of borrowers [9,43]. Our results are consistent with the findings of [44], who demonstrate that stronger ESG engagement contributes to lower credit risk by enhancing credit risk management processes. Similarly, our findings show that higher ESG performance significantly reduces non-performing loans, confirming the stabilizing effect of ESG integration on banks’ risk profiles.

Comparable findings are observed in [9,43] for Chinese banks, and [71] for European banks, all showing that ESG-oriented institutions exhibit better credit risk profiles.

When decomposing ESG into its environmental, social, and governance pillars, we find that the social and governance components significantly reduce credit risk, while the environmental pillar remains insignificant. This asymmetry is consistent with [68,72] for emerging countries; these studies show that social responsibility and governance strengthen monitoring mechanisms, reduce information asymmetry, and foster responsible lending practices. Governance, in particular, is frequently reported as the strongest pillar affecting bank risk [68,83], because board quality and transparency directly influence lending standards.

This pattern reflects the structure of MENA banking systems, where credit risk is more immediately shaped by borrower screening, transparency, governance quality, and stakeholder relationships than by environmental exposure. The social pillar captures practices related to customer protection, employee stability, and community engagement, which improve information flows and borrower discipline, thereby lowering default risk. The governance pillar directly strengthens board oversight, risk committees, and internal controls, reducing agency problems and excessive risk-taking. By contrast, environmental risks in MENA banks are often indirect, long-term, or weakly priced in loan portfolios, which explains their limited short-run effect on non-performing loans.

Efficiency, measured through DEA scores, consistently shows a negative and significant association with risk, reinforcing the “bad management” hypothesis originally proposed by [23]. Similar results are found in [69] in Pakistan’s banking sector, and [84], all concluding that inefficient banks are more likely to accumulate problem loans due to weaker monitoring, lower screening capacity, and higher operational frictions. The control variables further align with international evidence: larger banks display higher risk in some specifications, consistent with the too-big-to-fail literature [64,73]; GDP growth reduces NPLs, as shown in [65,85]; and excessive credit expansion increases risk, echoing findings by [86] on credit cycles in emerging economies. Loan loss reserves remain strongly and positively related to NPLs, confirming that provisioning behavior is reactive, in line with [52,65]. Finally, the significance of ESG, efficiency, and capital across models confirms the multidimensional nature of credit-risk formation and situates the MENA region within global patterns identified in both developed and emerging markets.

Across the four System-GMM risk specifications, the Arellano–Bond AR(1) test is statistically significant (p-values ≈ 0.003–0.005), which is expected in first-differenced equations because differencing mechanically induces first-order serial correlation in the residuals. More importantly, the AR(2) test is consistently insignificant (p = 0.929, 0.932, 0.809, and 0.975), indicating no evidence of second-order serial correlation in the differenced residuals. This is a key requirement for System-GMM validity because it supports the assumption that lagged levels are valid instruments for the differenced equation, strengthening the credibility of the dynamic specification.

Instrument validity is further supported by the Hansen J-test results. For all four models, Hansen p-values are comfortably above conventional significance levels (p = 0.445, 0.438, 0.787, and 0.437), so we fail to reject the null hypothesis that the instrument set is jointly valid (i.e., orthogonal to the error term). In practical terms, these results suggest that the internal instruments used in the System-GMM framework are not overfitting the endogenous components and that the estimated negative effects of ESG (overall and pillars) and efficiency on credit risk are not driven by invalid instrumentation. In addition, we limit instrument proliferation by using collapsed instruments, and we ensure that the number of instruments remains reasonable relative to the number of banks (a standard rule of thumb is keeping instruments well below N) to avoid weakening the Hansen test and inflating finite sample bias.

4.2.2. Determinants of Capital Adequacy

The results of the System-GMM estimations for the capital adequacy equation (

Table 6. System-GMM estimation of the capital adequacy equation.

	EQTA	EQTA	EQTA	EQTA
L.EQTA	0.484 ***	0.527 ***	0.509 ***	0.485 ***
	(0.051)	(0.049)	(0.051)	(0.0527)
ESG	0.044 ***
	(0.015)
ENVIR		0.024 **
		(0.0095)
SOC			0.0416 ***
			(0.0134)
GOV				0.037 **
				(0.0144)
EFF	0.009 **	0.0107 **	0.0084 *	0.0109 **
	(0.004)	(0.0042)	(0.0047)	(0.0044)
SIZE	−0.011 ***	−0.008 ***	−0.011 ***	−0.008 ***
	(0.002)	(0.0019)	(0.0021)	(0.0019)
ROA	0.201 ***	0.1808 ***	0.203 ***	0.184 ***
	(0.021)	(0.0199)	(0.0226)	(0.2007)
LTDP	−0.034 ***	−0.036 ***	−0.041 ***	−0.026 ***
	(0.009)	(0.0087)	(0.0103)	(0.0086)
NPL	0.002	−0.0096	−0.0044	0.0081
	(0.007)	(0.0061)	(0.006)	(0.0081)
REG	−0.537 ***	−0.432 ***	−0.527 ***	−0.453 ***
	(0.096)	(0.0929)	(0.1005)	(0.093)
GDP	0.013 *	0.0118	0.0212 **	0.0055
	(0.007)	(0.0087)	(0.0088)	(0.0077)
INF	0.039 **	0.0465 ***	0.0361 **	0.049 ***
	(0.016)	(0.0168)	(0.017)	(0.0156)
DCPS	−0.004	0.0036	−0.0026	−0.0042
	(0.004)	(0.0037)	(0.0039)	(0.00428)
_cons	0.379 ***	0.283 ***	0.388 ***	0.303 ***
	(0.059)	(0.054)	(0.061)	(0.0525)
Observations	1026	1026	1026	1026
Fisher	607.63 ***	666.11 ***	685.76	582.61
Arellano–Bond test for AR(1)	−2.89 (p = 0.004)	−3.23 *** (p = 0.001)	−2.93 *** (p = 0.003)	−3.09 (p = 0.002)
Arellano–Bond test for AR(2)	0.46	0.28 (p = 0.780)	0.30 (p = 0.764)	−0.61 (p = 0.544)
Sargan test	(p = 0.646)	8.45 (p = 0.993)	8.04 (p = 0.995)	7.13 (p = 0.998)
Hansen test	7.23 (0.998)	23.13 (p = 0.337)	24.81 (p = 0.256)	24.17 (p = 0.285)

Notes: (***), (**), and (*) are significative at 1%, 5%, and 10%. Values between parentheses indicate standard deviation errors.

) reveal a strong degree of persistence, with the lagged equity to total assets ratio remaining positive and highly significant in all four models, with coefficients ranging from 0.48 to 0.53. This indicates that capitalization practices in MENA banks are structurally stable and adjust slowly over time, consistent with previous findings by [48] for GCC banks, which show that capital ratios exhibit strong inertia due to regulatory constraints and internal capital-planning frameworks. The positive and significant coefficients for ESG and its individual pillars demonstrate that banks with stronger sustainability profiles tend to hold higher capital buffers. This result is in line with [87], who report that socially responsible banks adopt more conservative financial structures and accumulate larger solvency cushions. In our results, the social and governance pillars exert stronger effects on capitalization than the environmental pillar, confirming that ESG mechanisms related to responsible behavior, transparency, and governance strengthen market discipline and reinforce capital-building incentives.

The positive association between ESG performance and capital adequacy reflects several reinforcing channels. First, under Basel III and evolving supervisory expectations, ESG-oriented banks tend to adopt more conservative balance-sheet structures to signal resilience and regulatory compliance. Second, stronger ESG performance enhances transparency and reputation, reducing information asymmetry and enabling banks to access cheaper and more stable funding, which facilitates capital accumulation. Third, ESG engagement is associated with lower risk volatility and earnings instability, encouraging banks to retain profits and build precautionary capital buffers. These mechanisms jointly explain why ESG-oriented banks in the MENA region hold systematically higher equity ratios.

Efficiency has a positive and significant impact on capital adequacy in all specifications, indicating that more efficient banks maintain higher equity ratios. This finding supports the efficient-structure hypothesis, originally articulated by [88], which argues that efficient banks generate higher income, face lower operating costs, and therefore accumulate capital more easily. Similar patterns are reported by [84] who show that efficiency improvements enhance profitability and internal capital generation. In contrast, bank size consistently shows a negative and significant coefficient, suggesting that larger banks tend to operate with thinner capital buffers. This behavior is consistent with the too-big-to-discipline or managerial discretion hypotheses [17], which argue that large banks may rely more heavily on implicit government support or market power, allowing them to operate with lower solvency margins.

The results for the control variables further reinforce established empirical patterns. Profitability (ROA) is strongly and positively associated with capital adequacy, confirming the pecking order theory: profitable banks rely more on retained earnings to strengthen their capital position [15,63]. The loan-to-deposit ratio is significantly negative, implying that banks with aggressive lending relative to their deposit base accumulate lower equity buffers, reflecting higher liquidity pressure and funding fragility in emerging markets [17]. Interestingly, credit risk (RISK) is insignificant in all models, suggesting that MENA banks adjust their capital ratios primarily through regulatory requirements rather than through internal responses to loan deterioration. Regulatory pressure (REG) is consistently negative and significant, meaning that banks facing tighter regulatory constraints reduce their capital. Among macroeconomic variables, GDP growth and inflation show weak but positive effects on capital, indicating that favorable economic conditions allow banks to improve capitalization, in line with [76,85,89]. These results highlight that the capital structure of MENA banks is shaped jointly by sustainability performance, efficiency conditions, macroeconomic factors, and internal profitability dynamics. They also underscore that ESG, particularly its social and governance dimensions, contributes to stronger capital buffers and enhances the prudential soundness of banks in the region.

As in the risk equation, the System-GMM diagnostics for the capital adequacy models confirm the validity of the dynamic specification. The AR(1) test is significant, while the AR(2) test is consistently insignificant across all specifications (p-values > 0.54), indicating no second-order serial correlation in the differenced residuals. The Hansen test p-values remain above conventional significance levels, supporting the joint validity of the instrument set and suggesting that the endogenous regressors are appropriately instrumented. Overall, these results confirm that the estimated effects of ESG, efficiency, risk, and regulatory pressure on capital adequacy are not driven by misspecification or instrument proliferation, reinforcing the credibility of the dynamic capital adjustment process identified for MENA banks.

4.2.3. Determinants of Banking Efficiency

The results of the efficiency equation (

Table 7. System-GMM estimation of the efficiency equation.

	EFF	EFF	EFF	EFF
L.EFF	0.483 ***	0.485 ***	0.489 ***	0.463 ***
	(0.057)	(0.057)	(0.058)	(0.056)
ESG	0.069
	(0.046)
ENVIR		0.0916
		(0.064)
SOC			0.102 **
			(0.041)
GOV				0.034
				(0.048)
EQTA	0.91	0.740 **	1.062 ***	0.765 **
	(0.328)	(0.287)	(0.330)	(0.338)
SIZE	−0.003	−0.0026	−0.0053	−0.001
	(0.0054)	(0.0062)	(0.0054)	(0.0057)
ROA	2.175 ***	2.17 ***	2.182 ***	2.250 ***
	(0.217)	(0.2107)	(0.216)	(0.236)
CTI	2.182	2.103 ***	2.024 ***	2.402 ***
	(0.451)	(0.425)	(0.437)	(0.465)
LTDP	−0.115	−0.0661	−0.108	−0.136 *
	(0.081)	(0.073)	(0.0804)	(0.081)
NPL	0.23	0.2045 **	0.255 ***	0.199 **
	(0.084)	(0.092)	(0.082)	(0.086)
GDP	0.856	0.813 ***	0.833 ***	0.913 ***
	(0.097)	(0.0916)	(0.095)	(0.105)
INF	−0.652	−0.6525 ***	−0.677 ***	−0.645 ***
	(0.122)	(0.1164)	(0.127)	(0.123)
DCPS	−0.311	−0.268 ***	−0.307 ***	−0.322 ***
	(0.037)	(0.0376)	(0.037)	(0.040)
_cons	1.463	1.248 ***	1.467 ***	1.508 ***
	(0.218)	(0.207)	(0.2124)	(0.248)
Observations	1026	1026	1026	1026
Fisher	104.49 ***	169.35 ***	114.63 ***	123.27 ***
Arellano–Bond test for AR(1)	z = −3.40 p = 0.001	−3.50 *** (p = 0.000)	−3.35 *** (p = 0.003)	−3.42 (p = 0.001)
Arellano–Bond test for AR(2)	−0.59 (p = 0.557)	−0.68 (p = 0.932)	0.54 (p = 0.591)	−0.62 (p = 0.538)
Sargan test	z = 11.70 p = 0.926	13.73 (p = 0.844)	11.44 (p = 0.934)	11.15 (p = 0.942)
Hansen test	24.59 (p = 0.218)	23.28 (p = 0.275)	24.90 (p = 0.205)	24.03 (p = 0.241)

Notes: (***), (**) and (*) are significative at 1%, 5%, and 10%. Values between parentheses indicate standard deviation errors.

) indicate a strong and significant persistence in banking efficiency across all four specifications. The coefficient of lagged efficiency (0.46–0.49) shows that efficiency levels are path-dependent and adjust gradually over time. This confirms the arguments of [15,84,90], who document that bank efficiency reflects structural managerial practices and technological capabilities that do not change abruptly from year to year. Across all models, equity to total assets (EQTAs) positively and significantly affects efficiency, suggesting that well-capitalized banks in the MENA region operate more efficiently. This result is consistent with [47,48,74,84,91], who argue that stronger capital buffers reduce funding costs, mitigate risk-shifting incentives, and enhance operational stability, allowing banks to allocate resources more effectively.

The positive and significant effect of efficiency on capital adequacy indicates that operationally efficient banks generate higher and more stable retained earnings, which constitute the primary source of internal capital formation in bank-dominated financial systems. Lower operating costs, better asset utilization, and improved risk control allow efficient banks to accumulate equity without relying excessively on external capital issuance. In the MENA context, where capital markets are less developed and dividend smoothing is common, this retained-earnings channel is particularly important, explaining why efficiency improvements translate directly into stronger capital buffers.

Regarding ESG variables, the results show that ESG performance generally improves bank efficiency, with the social pillar being the most influential. The social score (SOC) exhibits a positive and significant effect, while environmental and governance dimensions show non-significant effects. This finding contrasts with that of [15] in the context of U.S. banking. Ref. [92] find a U-shaped link between corporate social performance and bank efficiency, indicating that both highly socially responsible banks and those with low social performance tend to be the most efficient. The result demonstrates that banks that engage more actively in socially responsible practices such as employee well-being, fair treatment, community development, and customer protection tend to experience better internal processes and stronger stakeholder relationships. These mechanisms ultimately translate into operational efficiencies. The relatively weaker impact of environmental and governance scores suggests that in emerging markets such as the MENA region, socially oriented ESG dimensions may be more immediately relevant for internal cost structures and operational performance.

Bank profitability (ROA) also exerts a strong positive effect on efficiency, confirming the “efficient-structure hypothesis” whereby more profitable banks tend to operate with superior managerial quality, better cost control, and more productive asset portfolios. This supports the findings of [15,47,74]. A positive coefficient of the cost-to-income ratio (CTI) on efficiency indicates that operating expenses in MENA banks may represent productivity-enhancing investments such as digitalization, risk-management systems, and ESG-related processes rather than managerial inefficiency. Although these costs initially increase the cost-to-income ratio, they ultimately contribute to improved technical efficiency, consistent with findings in the recent literature. Similar evidence is reported by [75].

Loan-to-deposit ratios exhibit a negative or weakly significant impact on efficiency, indicating that banks with aggressive lending strategies face higher operating and risk-management costs, which reduces efficiency. This is in line with the studies of [15,23], which presented the “bad management” and “moral hazard” hypotheses, where poorly controlled credit expansion increases non-performing assets and operational burdens. Indeed, RISK shows a positive significant effect: banks with higher non-performing loans ratios appear to adjust their efficiency upward, possibly reflecting corrective restructuring efforts, a pattern which was also observed by [47] in the MENA region.

Macroeconomic controls behave as expected. GDP growth (gdpg) positively and significantly influences efficiency, indicating that banks operate more efficiently during expansionary periods when credit demand is stable and asset quality improves. This is in line with [93] for Jordanian banks. Inflation has a negative and significant effect, consistent with [47,94], who argue that inflation creates uncertainty, increases administrative costs, and amplifies inefficiencies. Finally, higher domestic credit to the private sector (DCPS) negatively affects efficiency, reflecting greater competition and pressure on bank margins, a mechanism aligned with [95], who show that competitive markets can initially lower cost efficiency when banks adjust business models.

The results confirm that sustainability performance, capital strength, and stable macroeconomic environments act as key contributors to efficiency, while aggressive lending and inflationary pressures reduce it.

The System-GMM diagnostics for the efficiency equation confirm the validity of the estimation strategy. The Arellano–Bond AR(1) test is significant, while the AR(2) test is consistently insignificant across specifications, indicating no second-order serial correlation in the differenced residuals. The Hansen and Sargan test p-values remain well above conventional significance levels, supporting the validity of the instrument set and rejecting concerns of over-identification. Moreover, the number of instruments remains well below the number of banks, limiting instrument proliferation. Overall, these diagnostics confirm the robustness of the dynamic efficiency estimates and lend credibility to the identified ESG–capital–risk–efficiency interactions.

5. Conclusions

This study investigates the interconnected relationship between banking risk, capital strength, and efficiency in the MENA region, while incorporating the growing role of ESG performance as a driver of sustainable financial behavior. Using an unbalanced panel of 167 banks from 13 MENA countries over 2015–2024 and applying a three-equation System-GMM framework, the findings highlight several key insights. First, banking risk measured through NPLs exhibits strong persistence, but ESG performance significantly reduces credit risk, particularly through the governance and social dimensions. These results confirm that sustainable practices enhance transparency, stakeholder trust, and borrower discipline, thereby contributing to lower asset deterioration. Second, capital ratios are influenced positively by ESG engagement and efficiency, suggesting that sustainable banks experience stronger capital buffers, improved financial resilience, and more stable funding structures. Third, efficiency is shaped by well-capitalized balance sheets, profitability, and favorable macroeconomic conditions, while the social pillar of ESG yields the strongest improvements in operational performance. Together, these findings underscore that ESG adoption is not merely a reputational investment but a strategic lever that enhances bank stability, capital strength, and efficiency.

What the three-equation system teaches us is that ESG operates through distinct but connected channels in MENA banking: (i) the Social and Governance pillars are the most consistent drivers of improved prudential outcomes (lower risk and stronger capital), and (ii) the Social pillar is the most robust channel for operational efficiency gains, while the Environmental pillar shows weaker short-run effects. This triadic structure indicates that sustainability affects bank soundness not in isolation, but through mutually reinforcing feedback pertaining to risk, capitalization, and efficiency.

The study offers several important implications. Theoretically, the results support both the risk-mitigation hypothesis and the stakeholder value creation theory, demonstrating that ESG practices reduce moral hazard behavior, strengthen governance mechanisms, and enhance productive efficiency. The positive effect of capital on efficiency and the mitigating effect of ESG on risk are in line with recent evidence from emerging markets and developed regions alike. Managerially, bank executives should recognize that incorporating sustainability principles into core operations—particularly social responsibility initiatives—can improve internal processes, reduce credit losses, and boost long-term competitiveness. Operational risk managers may also benefit from integrating ESG indicators into risk assessment frameworks, credit scoring, and early warning systems.

From a policy perspective, the findings emphasize the need for MENA regulators and central banks to promote ESG disclosure standards and integrate sustainability metrics into prudential supervision. Supervisors could operationalize ESG information through (a) a Pillar 2-style supervisory review that incorporates ESG signals into bank-specific risk assessments and supervisory capital expectations, (b) risk-based monitoring and early warning systems where weak Governance/Social scores trigger closer credit-risk scrutiny, and (c) guidance on capital buffer planning for banks with weaker ESG risk management. Strengthening ESG reporting frameworks aligned with IFRS-S1/S2 and national sustainability agendas would enhance transparency, reduce information asymmetry, and improve the reliability of ESG metrics used by banks and investors. Given the significant role of the social dimension in improving efficiency, policies that encourage financial inclusion, consumer protection, and human capital development within the banking sector may yield broader economic benefits.

Despite its contributions, this study is subject to several limitations that also open up important avenues for future research. First, although the analysis covers the period 2015–2024, it does not explicitly isolate the effects of major exogenous shocks such as the COVID 19 pandemic. While the dynamic System-GMM framework partly absorbs macroeconomic disturbances, future research could incorporate crisis-specific dummy variables or sub-sample analyses to better disentangle pandemic-related effects on bank risk, capital adequacy, and efficiency. Second, ESG data availability remains uneven across banks and years in the MENA region. Although the use of an unbalanced panel is methodologically appropriate and reflects market realities, future studies could benefit from alternative ESG databases, refined disclosure quality indicators, or textual ESG measures to reduce potential measurement bias and better capture ESG heterogeneity across institutions. Third, banking efficiency is measured using a two-stage DEA approach, which relies on specific assumptions regarding returns to scale, input–output selection, and frontier construction. While DEA is well established in the banking literature, future research could complement this approach with stochastic frontier analysis, network DEA, or machine learning-based efficiency models to test the robustness of efficiency estimates.

Future research can extend this framework by exploring alternative risk measures (e.g., Z-score, expected credit loss proxies), richer ESG quality metrics, and heterogeneity by bank type and ownership (including Islamic banks), as well as non-linear or regime-dependent dynamics under differing regulatory environments.

Overall, the findings demonstrate that sustainable banking is not only compatible with financial soundness but actively reinforces it. By embedding ESG considerations into their strategic and operational frameworks, banks in the MENA region can enhance their resilience, improve efficiency, and contribute more effectively to the region’s transition toward a sustainable and stable financial system.