What Drives Banks to Provide Green Loans? Corporate Governance and Ownership Structure Perspectives of Vietnamese Listed Banks

Abstract

This study delves into the influence of banks’ governance and ownership structures on green lending. To examine this, we utilized the two-step system GMM and PCSE methods on the panel data of Vietnamese commercial banks spanning from 2010 to 2023. The findings suggest that board characteristics, precisely board size, board independence, and gender diversity, play a significant role in encouraging banks to provide green credit. The study highlights the importance of ownership structure in green lending. Banks with a high percentage of government ownership tend to fund more green projects, while foreign counterparts are reluctant to fund green finance. A mechanism test is also conducted to point out that banks’ disclosure of their green loan commitments is an influential channel whereby corporate governance and ownership structure impact green loans. Additionally, this research finds that the issuance of the Green Loan Principles in 2018 can facilitate banks’ governance of sustainable lending.

1. Introduction

In 2023, developing countries faced a shortfall of USD 71 billion in investment in climate adaptation (United Nations Environment Programme 2022). To address this problem, financial institutions require deeper participation in providing funds for long-lived low-carbon projects through various green finance instruments and services (Farhad and Naoyuki 2019). The green finance products developed by banks, such as green loans or green credit, are considered to be key drivers in fighting against climate change (Mirovic et al. 2023; Zhou et al. 2022) and a win–win solution for financial providers and society as a whole (Xi et al. 2022). However, the high risk and low returns of green projects make them less attractive, and banks are reluctant to finance such initiatives (Farhad and Naoyuki 2019, 2020). Thus, it has become more crucial than ever to have a comprehensive understanding of the factors that impact banks’ green lending.

In development banking, corporate governance is the core means by which banks can improve green credit (Felipe 2020; Yao et al. 2023). With regard to sustainable growth, several studies have recognized the role of corporate governance practices (Zheng and Kouwenberg 2019) and boards of directors in banks’ strategic decision making related to environmental, social, and governance performance (Menicucci and Paolucci 2022). However, minimal effort has been devoted to investigating corporate governance’s effect on green loans. Felipe (2020), Yao et al. (2023) were pioneers in examining the impacts of the governance and ownership structures on the green lending behaviours of banks; nevertheless, their research studies were only exploratory works. Therefore, our paper aims to examine the empirical impact of corporate governance and ownership structures on banks’ green lending decision making in emerging markets in the Vietnamese context.

There are several worthy reasons to include Vietnamese banks in this research stream. First, the Vietnamese banking sector, overseen by the State Bank of Vietnam (SBV), has undergone substantial changes since the economic reform (Doi Moi) of 1986. The SBV has consistently focused on enhancing the banking industry by implementing programs and regulations aimed at restructuring banks and the Vietnamese economy. The “Restructuring the System of Credit Institutions during 2011–2015” project, approved by the Vietnamese Prime Minister, outlined three key stages for bank restructuring in Vietnam. Firstly, efforts were made to enhance the liquidity of weaker banks through mergers and acquisitions. Secondly, banks were encouraged to improve financial transparency by reducing non-performing loans. Lastly, overall restructuring was carried out to address operational, strategic, and risk management aspects. Consequently, from a banking system that previously operated within a centrally planned economy, Vietnamese banks now play vital roles in Vietnam’s financial market, aligning with global trends and serving diverse client needs across all sectors. Vietnam is one of the most vulnerable countries to climate change (Kouwenberg and Zheng 2023). As the banking sector is a primary part of Vietnam’s financial system (Anwar and Nguyen 2011), most of Vietnam’s climate protection and mitigation programs largely depend on banking channels to access financial support funds. Secondly, with the ambition of achieving a net-zero emissions target by 2050, Vietnam is a pioneer of promoting a green economy. The Vietnamese government has been actively advancing the development of a green economy by implementing the National Green Growth Strategy. The latest strategy spans from 2021 to 2030, with a vision extending to 2050, underscoring the significance of green finance in achieving sustainable growth. In pursuing this goal, the government has emphasized the integration of blended financial resources to support green projects. Additionally, it aims to establish a market-based funding mechanism that complies with global standards, increasing its access to international funding opportunities such as the Green Climate Fund and other Official Development Assistance programs. Furthermore, the government is exploring adaptive fiscal policies, including eco-taxes and a carbon tax, to align with its green objectives. It proactively provides policies for its financial institutions and cooperates with international organizations such as the International Finance Corporation to attract green finance for the economy. Consequently, outstanding green loans steadily increased from 2015 to 2022, at an annual average of 25% (State Bank of Vietnam 2023), focusing on the renewable and clean power sectors. However, green loans account for only 4.2% of the entire economy’s outstanding loans, and less attention has been paid to essential fields in environmental protection, including sustainable transportation, construction, and waste management (State Bank of Vietnam 2023).

Given the fact that the literature on green finance in emerging nations is fragmented, this study aims to make a contribution to the emerging literature on green finance and the policy implications for regulators and banks. Thus, by using data obtained from Vietnam’s economy, our research provides a somewhat unique arena to identify the extent to which corporate governance and ownership structures impact the green loans of banks.

Using the two-step system generalized method of moments (GMMD) and panel-corrected standard errors (PCSE) models for the panel data of Vietnamese listed banks from 2010 to 2023, the main empirical findings of this study are fourfold. Firstly, banks’ board characteristics, including board size, independence, and gender diversity, are positively and significantly associated with green loans. Secondly, ownership structure, including government and local nongovernment ownership, yields a significantly positive nexus with green credit, while foreign ownership has a significantly negative relationship with green loans. Third, the disclosure of green loan commitments by banks has a partial mechanism effect on the roles of governance and ownership structures in banks’ green lending behaviours. Finally, after the turning point of the Green Loan Guideline initiated in 2018, banks have tended to facilitate sustainable lending.

This study makes several important contributions. Firstly, to the best of our knowledge, it is the first instance of empirical research to examine the impact of corporate governance and ownership structure on banks’ decision making with regard to green lending and their effect mechanism in the context of an emerging market. Secondly, the findings provide significant recommendations for bank managers, policymakers, and practitioners to enhance green loans.

The remainder of this paper is structured as follows. Section 2 presents the literature review and hypothesis development. Section 3 explains the design of the empirical study. Section 4 summarizes the research results and initiates a discussion. Section 5 proposes mechanism test. Section 6 provides additional robustness checks. Section 7 concludes the research.

2. Literature Review and Hypothesis Development

2.1. Literature Review

Current studies on green finance focus on banks’ green loans and their substantial impact on economic development, firm production, and bank performance (Luo et al. 2021). Green loans benefit the economy by minimizing credit risk and improving financial stability (Goodhart 2005; Oguntuase 2020). As one of a bank’s green finance channels, green credit helps better manage risk and allocate capital to environmentally friendly projects, supporting a sustainable economy (Yin et al. 2021). Green loans promote a country’s development through an environmental and economic system and resource-saving intensive industries (Xi et al. 2022). Moreover, green loans serve to improve economic performance and reduce pollution (Koval et al. 2022). Thus, as a crucial part of mitigating climate change, green loans can help a country pursue its net-zero carbon strategy (Zhou et al. 2022). In addition to economic development, green loans play an essential role in enterprises’ development. For small firms, green loans enhance their capital structure and help them avoid financial hardship (Isaac et al. 2022). Large firms fund their research and development activities in innovative products (Díaz-García et al. 2015; Islam et al. 2014) and foster green innovations (Huang et al. 2022).

Green loans benefit not only the economy and firms but also banks; therefore, it is said that they are a win–win mechanism for banks, enterprises, and society at large (Xi et al. 2022). From banks’ perspective, green loans help them improve their operating performance (Cilliers et al. 2010), financial indicators (Belinda et al. 2022; Luo et al. 2021; Xi et al. 2022), and sustainable development (Ding et al. 2022), whilst also minimizing their credit risk (Zhou et al. 2022). Additionally, green credit significantly strengthens listed banks’ green reputation, giving them a competitive advantage in the market (Xi et al. 2022). It also moderates banks’ profitability characteristics, such as size, liquidity, and capital adequacy. It is a tool for banks to reach their social and environmental targets on the international stage (Mirovic et al. 2023).

Minimal research has investigated the direct nexus between green loans and banks’ corporate governance and ownership structure, except for qualitative studies by the authors of (Felipe 2020; Yao et al. 2023). Using Chinese bank data, Yin et al. (2021) found that government-owned banks tended to lend more green credit. In contrast, Zhou et al. (2022) implied that state-owned banks did not influence the ratio of green lending to total loans. Other branches of the literature on corporate governance and ownership structure in relation to green growth mostly combined corporate governance and ownership structure with green banking disclosures (Bose et al. 2018), sustainable development (Francesco et al. 2019), sustainability reporting (Amidjaya and Widagdo 2020), green performance (Sudipta et al. 2021), and ESG performance (Menicucci and Paolucci 2022, 2023), rather than with green loans, green credit, or green finance. For instance, Bose et al. (2018) investigated the impact of corporate governance and ownership structure on green banking disclosure in the banking industry in Bangladesh. The results positively associated board size and government ownership with banks’ green disclosure, while board independence and foreign ownership did not affect the practice. Francesco et al. (2019) researched factors affecting the environmental responsibility of banks and showed that corporate governance positively affected banks’ environmental engagement. Specifically, among the governance mechanisms introduced in that paper, board gender diversity, counting a higher number of female directors on boards, was positively associated with banks’ environmental activities. Amidjaya and Widagdo (2020) examined the effect of ownership structure and corporate governance on Indonesian banks’ sustainability reporting. The results indicated that general corporate governance, boards with a higher percentage of foreign shareholders, and banks with a higher proportion of government ownership appeared to generate better sustainability reports. Sudipta et al. (2021) constructed an index for the green banking performance of banks in Bangladesh and examined the relationship between this indicator and banks’ financial activities. They also considered ownership structure, finding that a governmental share in a bank’s ownership structure lowered the bank’s green performance, while that of foreign investors had no significant impact. Menicucci and Paolucci (2022, 2023) explored the role of board characteristics and diversity in the environmental, social, and governance (ESG) performance of Italian banks, documenting that more extensive, more independent, and diversified boards were related to higher ESG scores, while board age did not impact the banks’ ESG performances.

Two points should be noted from the research on green loans and corporate governance and the previously described ownership structure. First, the direct relationship between corporate governance and the ownership structure of banks and their green loans is rarely considered by scholars. If any, this relationship is qualitative or indirectly reflected in the context of investigating the association between green loans and other factors, as in the studies of Yin et al. (2021) and Zhou et al. (2022). In the global climate finance literature, the knowledge domain is characterized by the four following major themes: renewable energy financing, risks associated with climate change for the financial sector, the impact of green investment on corporations and investor preferences, and risk pricing and hedging in the financial industry related to climate change (Kouwenberg and Zheng 2023). Among these main topics, there is an absence of research on the determinants of banks’ green finance behaviours, including green loans (Isaac et al. 2022) in general and the role of corporate governance and ownership structure in green loans in particular. A review of global research on corporate governance and board attributes by Zheng and Kouwenberg (2019) suggested that the influence of corporate governance on corporate social responsibility and sustainability is an emerging issue and called for more research on it. Second, given the large corpus of research on green finance as a whole, most studies are based on the context of developed economies, which may not apply to emerging markets such as Vietnam, whose financial system is bank-centric and where banks play an essential role in providing loans to the economy (Sarath and Pham 2015). Consequently, this study aims to fill this gap in the literature on corporate governance and green loans in developing countries.

2.2. Hypothesis Development

Based on the studies by the authors in (Bose et al. 2018; Felipe 2020; Yao et al. 2023; Menicucci and Paolucci 2022, 2023), who investigated green banking disclosure and ESG performance from the perspective of corporate governance, this study introduces a modified model with the independent variables of banks’ corporate governance and ownership structure.

2.2.1. Corporate Governance and Green Loans

Given the vital importance of corporate governance in banks’ green practices, as previously discussed, existing research on the corporate governance of banks has emphasized the crucial role of boards of directors in ESG dimensions, socially responsible practices, and sustainability (Menicucci and Paolucci 2022; Zheng and Kouwenberg 2019). In the literature, some commonly used factors capturing board characteristics are the size, independence, and gender diversity of boards (Menicucci and Paolucci 2022). Therefore, this study utilizes these three independent variables to investigate banks’ governance.

Research on emerging markets suggests that a larger board size can lead to a broader range of perspectives being considered during decision-making processes, which can help stabilize corporate performance (Bryan and José 2019). Furthermore, a board with more members brings a more diverse set of skills and managerial viewpoints to the table of discussion, which may result in a more substantial commitment to ESG practices and sustainable development (Menicucci and Paolucci 2022). In addition to this, a larger board size may reflect more significant experience in and more effective communication of environmental issues, potentially leading to increased support for environmental initiatives (Bose et al. 2018). Studies exploring the relationship between board characteristics and sustainability have positively linked board size with green performance, as demonstrated by researchers such as Spitzeck (2009); Kiliç et al. (2015); and Menicucci and Paolucci (2023). Based on this evidence, we hypothesize the following:

H1. 

There is a positive relationship between board size and green loans.

Regarding board independence, non-executive directors are typically concerned with their company’s reputation (Fama and Jensen 1983). As environmentally friendly practices become more popular and reflect positively on firms, a board with more independent directors may prioritize environmental issues (Francesco et al. 2019). Moreover, independent directors play a significant role in governing the board and advocating for stakeholders (Patelli and Prencipe 2007). As a result, they aim to enhance a firm’s long-term performance by implementing environmental policies and managing risks. Studies have also highlighted the role of board independence in CSR and ESG, as shown by Bose et al. (2018), Francesco et al. (2019), and Menicucci and Paolucci (2022, 2023). In light of this analysis, we propose the following hypothesis:

H2. 

There is a positive relationship between board independence and green loans.

According to research by Francesco et al. (2019), women on boards exhibit a strongly risk-averse attitude, prioritize reductions in environmental risks, and are committed to environmental engagement. This aligns with studies suggesting that women directors on boards are closely linked to a firm’s sustainability performance. Valeria (2019) indicated that female directors tend to prioritize social reasoning and are more inclined towards social issues than their male counterparts, which positively impacts how environmental problems are addressed. Post et al. (2011) highlighted the critical role of female board directors in addressing environmental challenges, while Paul et al. (2017) suggested that they play a crucial role in improving CSR performance. Moreover, Menicucci and Paolucci (2022) found that women board directors are also more likely to prioritize CSR disclosure. Based on these findings, we hypothesize the following:

H3. 

There is a positive relationship between board gender diversity and green loans.

2.2.2. Ownership Structure and Green Loans

The ownership structure of firms has been shown to significantly influence the essence of the agency problem (Kose et al. 2016). As a vital corporate governance mechanism, it has the potential to impact the dynamics of corporate social responsibility (Huang 2010). Governmental ownership was proven to have positive impacts on green growth in the research conducted by Amidjaya and Widagdo (2020); Bose et al. (2018); Yin et al. (2021); and Zhou et al. (2022). Additionally, as governments strongly support green finance, state-owned banks are mainly selected to implement green policies (Yin et al. 2021). Since they also have advantages stemming from reliable networks and government agencies (Zhou et al. 2022), they may have a higher chance of accessing funds (Sudipta et al. 2021). Thus, we hypothesize the following:

H4. 

There is a positive relationship between government ownership and green loans.

In the context of exposure to foreign markets, foreign banks derive significant benefits from accessing substantial funds and global markets (Gillan and Starks 2003), and as a result, they have access to international green funds to lend to green projects. While foreign ownership is a relatively new area in green finance research, scholars have incorporated it into sustainable reporting literature. According to Haniffa and Cooke (2005), companies with a higher proportion of foreign investors tend to disclose more on corporate social responsibilities. Foreign investors also bring distinct values and knowledge to their businesses compared to local shareholders, placing significant emphasis on social and environmental issues (Khan et al. 2013; Bose et al. 2018) and viewing sustainability reporting as a long-term benefit for sustainable development (Amidjaya and Widagdo 2020). A recent qualitative study conducted by (Anh et al. 2023) documents that foreign banks show their direct engagement with green project lending and active contributions to the growth of green investment regulations. Therefore, it can be inferred that banks with a higher proportion of foreign ownership fund green projects more to please ethical foreign shareholders. This leads to the fifth hypothesis, as follows:

H5. 

There is a positive relationship between foreign ownership and green loans.

3. Research Design

3.1. Data

This paper investigates the impact of corporate governance and ownership structure on banks’ green loans in Vietnam. Based on data availability, the empirical analysis utilized the panel data of 29 Vietnamese banks from 2010 to 2023, yielding a research sample of 405 observations. The research data on dependent and independent variables were gathered from banks’ audited financial and annual reports, which were publicly available on their websites and the Vietnamese economic and financial information platform (Cafef). The data on control variables were obtained from the FiinPro database. Specifically, information on green loans was manually extracted from the footnotes of the audited financial reports of each bank. Similarly, data on each bank’s governance and ownership information were collected from their public annual reports. To ensure consistency and reliability, one author was responsible for data extraction, followed by a cross-check carried out by the others.

3.2. Variables

3.2.1. Dependent Variable

The dependent variable for this research is green loans. Vietnamese banks identify green credit based on the Green Loan Principles (GLPs), which were originally published in 2018 by the Loan Market Association and the Asia Pacific Loan Market Association and the Law on Environmental Protection (LEP) No. 72/2020/QH14 dated November 2020 legislated by the Vietnamese National Assembly. Accordingly, “green loans are any type of loan instrument made available exclusively to finance or re-finance, in whole or in part, new and/or existing eligible green projects” (GLP 2018). Nevertheless, the practice of green loans in Vietnam started late compared to other developed banking systems, so the amounts of green loan have been inconsistently reported across its entire banking sector. We examined the footnotes in Vietnamese banks’ financial statement to clarify the green loan amounts, particularly loan classification by sector. We found that, among the indicative categories of eligibility for green projects listed on the GLP and LEP, lending for sustainable water and wastewater management projects is reported by Vietnamese banks. Hence, this research employs loans for these projects as proxies for green loans in Vietnam. Several researchers studying green loans also have measured this variable by environmental loans (e.g., Huang et al. 2022; Yin et al. 2021; Zhou et al. 2022).

3.2.2. Independent Variables

As previously described, this paper identifies the effects of corporate governance, particularly the characteristics of the boards of directors and ownership structure, on decisions by Vietnamese commercial banks to fund green projects. Thus, following Amidjaya and Widagdo (2020); Bose et al. (2018); Francesco et al. (2019); Menicucci and Paolucci (2022), and to be in line with our hypotheses, we include board size, board independence, board gender diversity, foreign ownership, government ownership, and private and domestic ownership as independent variables in our research model.

3.2.3. Control Variables

Yin et al. (2021) showed that the internal determinants of banks, including profitability, size, liquidity, and capital adequacy ratio, positively impact green credit. Bose et al. (2018) and Menicucci and Paolucci (2022) documented the similar effects of bank age, leverage, and profitability on the green banking disclosure and ESG performance. Considering these essential elements and eluding model misspecification, we added six control variables: bank size, return on asset, capital adequacy, age, leverage, and liquidity. The descriptions of all variables in our research model are presented in

Table 1. Variable definitions.

Variables	Notation	Measurement
Dependent variables
Green loan	GL	The ratio of environmental loans to total loans of bank.
Independent variables
Board size	BRSZ	Natural logarithm of the total number of directors on the bank’s board
Board independence	BRIND	Percentage of independent directors on the bank’s board
Board gender diversity	GEN	Percentage of female directors of the bank’s board
Government ownership	GSHA	Percentage of total shares held by domestic governmental shareholders
Foreign ownership	FSHA	Percentage of total shares held by foreign shareholders
Control variables
Return on asset	ROA	Return of bank’s total assets
Bank size	BKSZ	Natural logarithm of bank’s total assets
Capital adequacy	CAR	Bank’s capital adequacy ratio
Leverage	LEV	The ratio of total debt to total assets of bank
Liquidity	LOQ	The ratio of total loans to total deposits of bank
Age	AGE	Natural logarithm of the number of established years

.

3.3. Empirical Model

To check the research hypotheses, we employed a two-step system GMM method which was developed by the authors in (Arellano and Bover 1995; Blundell and Bond 1998) as an efficient approach to the dynamic panel data model. They started with an autoregressive panel data model:

$$y_{it}=\alpha y_{i,t-1}+{{\beta }^{\prime }}_1{x}_{it}+{{\beta }^{\prime }}_2{x}_{it-1}+{\eta }_i+{\nu }_{it}$$

(1)

where i = 1, …, N (large), t = 2, …, T (fixed), and $u_{it}={\eta }_{it}+{\nu }_{it}$ is the error term’s decomposed fixed effects. Then, three assumptions are proposed for the model: (i) individual fixed effects are uncorrelated across i: E (η_i) = 0, E (ν_it) = 0, E (ν_it, η_i) = 0; (ii) an idiosyncratic component, ν_it, is also uncorrelated over time: E (ν_it, ν_is) = 0 with i ≠ s; (iii) the initial observation, y_it, and the composite fixed effects, u_it, are constantly correlated in subsequent time periods.

Based on these conditions, to assess the validity of the general GMM, diagnostic procedures include the testing of overidentifying restrictions, verified by the Hansen or Sargan statistics, and the autocorrelation structure of the first differenced residuals, developed by (Arellano and Bond 1991). Accordingly, the GMM may be valid because of the following:

• The null hypothesis of overidentifying moment conditions is not rejected.
• The null hypothesis of there being no serial correlation of the order of one in the first difference residuals is rejected.
• The null hypothesis of there being no serial correlation of the order of two in the first difference residuals is not rejected.

In comparison to the standard GMM estimator developed by Arellano and Bond (1991), the system GMM approach demonstrates enhanced precision, particularly in scenarios characterized by high autoregressive parameter values and a limited number of observations in the time series (Blundell and Bond 1998). The two-step system GMM is appropriate for our research in terms of three aspects. First, it performs well in unbalanced datasets (Khattak et al. 2023) and enables bank lending to be modelled dynamically (Cheng and Qu 2020). Second, it minimizes to the potential endogeneity issue in panel data research by eliminating the assumption of strict exogeneity and bank-specific effects being unobserved in regression (Khattak et al. 2023). Third, the GMM estimator allows the research to control the dependent and explanatory variables from unobservable factors.

Following the authors in (Felipe 2020; Yin et al. 2021; Yao et al. 2023), the research estimates different models to address the main objectives of the study by constructing a baseline regression model as follows:

$${GL}_{it}={\alpha }_0+{{\alpha }_{1}GL}_{it-1}+{\alpha }_{2}corgo{v}_{it}+{\alpha }_3{control}_{it}{+\epsilon }_{it}$$

(2)

$${GL}_{it}={\beta }_0+{{\beta }_{1}GL}_{it-1}+{\beta }_2{ownership}_{it}+{\beta }_3{control}_{it}{+\mu }_{it}$$

(3)

where i refers to the bank and t refers to the year. Corporate governance (corgov) and ownership structure (ownership) as well as control factors (control) are expressed by the above-defined alternative variables in Table 1. Model (2) serves to identify the impact of banks’ corporate governance on green loans. Model (3) was developed to investigate the relationship between banks’ ownership structure and green loans.

4. Empirical Analysis

This section examines the research hypotheses. First, we present the descriptive statistics and the correlations between the research variables, and then analyse the main estimation results.

4.1. Descriptive Statistics and Correlations

The descriptive statistics of the variables used for the entire sample are reported in

Table 2. Descriptive statistics of the variables.

Variables	Obs.	Mean	Std. Dev.	Min	Max
Dependent variables
GL	405	0.0011	0.0039	0.0000	0.0405
Independent variables
BRSZ	405	1.9546	0.2455	1.3862	2.7080
BRIND	405	0.1234	0.0887	0.0000	0.4000
GEN	405	0.1780	0.1622	0.000	0.8000
GSHA	404	0.1748	0.2990	0.000	1.0000
FSHA	404	0.1176	0.1140	0.000	0.4479
DSHA	404	0.7099	0.3164	0.000	1.0000
Control variables
ROA	405	0.0126	0.0479	−0.0068	0.9310
BKSZ	405	18.721	1.2802	13.425	21.556
CAR	403	0.1347	0.0532	0.0000	0.5492
LEV	405	0.9047	0.0706	0.0944	1.0300
LIQ	404	0.8918	0.1830	0.3632	1.7893
AGE	406	3.0776	0.5202	0.6931	4.1896

Notes: All the variables are defined in Table 1.

. The results from the descriptive analysis show that, on average, the independence and gender diversity of banks’ boards were relatively low, with mean values of 12.34% and 17.80%, respectively. In particular, the highest percentage of independent directors on the banks’ boards was only 40%, and 80% for that of female directors on boards, even though there were several banks without independent and female directors on their boards (the minimal value was 0). Regarding ownership structure, governmental and foreign shareholders accounted for the low percentages in Vietnamese banks, with mean estimations of 17.48% and 11.76%, respectively. At the same time, private and domestic investors dominated banks’ ownership structures. There was a wholly government-owned bank (maximum value was 1), which was the Vietnamese Bank of Agriculture and Rural Development (Agribank), while many banks did not have state or foreign investors. Our data on ownership structure also revealed that most foreign investors in banks were institutional.

We calculated the Pearson correlations among the variables in the regression model and reported their coefficients with significance levels in

Table 3. Pearson correlations.

Variable	GL	BRSZ	BRIND	GEN	GSHA	FSHA	DSHA	ROA	BKSZ	CAR	LEV	LIQ	AGE
GL	1.00
BRSZ	−0.0534	1.00
BRIND	0.0816	−0.3105 ***	1.00
GEN	0.1649 ***	−0.0730	0.1630 ***	1.00
GSHA	0.0724	0.4317 ***	−0.4021 ***	−0.1181 **	1.00
FSHA	−0.0891 *	0.2709 ***	0.1032 **	−0.1347 ***	−0.0179	1.00
DSHA	−0.0309	−0.4360 ***	0.3318 ***	0.1619 ***	0.4167	−0.3516	1.00
ROA	−0.0054	0.0117	0.0772	−0.0199	−0.0268	0.0893	0.0060	1.00
BKSZ	0.1388	0.4213 ***	0.0520	0.0146	0.4736 ***	0.4089 ***	−0.4779	0.0054	1.00
CAR	−0.1615 **	−0.0600	−0.0515	−0.1731 ***	−0.1224 **	−0.0715	0.1349 **	−0.0152	−0.4703	1.00
LEV	0.1140 **	0.1128 **	0.0757	0.0923 *	0.0720	0.1417	−0.1133 *	−0.1310 *	0.3889 **	−0.3947	1.00
LIQ	0.0498	0.0093	−0.0386	−0.1203 **	0.0820 *	0.0526	−0.0989 **	0.0912 *	0.1173 **	0.0312	−0.2137 ***	1.00
AGE	0.1040 **	0.1425 ***	0.1800 ***	0.1671 ***	0.3137 ***	0.1890 ***	−0.3572 ***	0.0535	0.4535 ***	−0.4300 **	0.0824 **	0.1566 ***	1.00

Notes: This table shows the Pearson pair-wise correlation matrix of variables. All the variables are defined in Table 1. *** p < 0.01, ** p < 0.05, * p < 0.1.

. The matrix displays the fact that there was no strong correlation between the variables. Additionally, the variance-inflation factors (VIF test) indicated that the mean VIF of the variables in our research model was 1.8. Therefore, the model employed was satisfactory, and multicollinearity was not severe. The correlation results also showed that most variables had a significant relationship with the explained factor.

4.2. Results and Discussion

4.2.1. Corporate Governance and Green Loans

The empirical research results for the nexus between corporate governance and green loans are reported in

Table 4. Impact of corporate governance and green loans.

Variables	(GLs)
L.GL	0.7150 *** (0.0013)
BRSZ	0.001 *** (0.000)
BRIND	0.002 *** (0.000)
GEN	0.004 *** (0.000)
ROA	−0.002 (0.002)
BKSZ	−0.001 ** (0.000)
CAR	−0.010 ** (0.001)
LEV	−0.001 * (0.000)
LIQ	−0.002 * (0.000)
AGE	0.001 ** (0.000)
Cons	0.013 *** (0.001)
Obs.	374
AR (1) (z value)	0.053 (−1.628)
AR (2) (z value)	0.234 (1.191)
Sagan (chi2)	1.000 (25.585)

Notes: This table shows the coefficients of estimates from the two-step SYS-GMM method on banks’ green loans (GLs). All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, ** p < 0.05, * p < 0.1.

. Models (1), (2), and (3) show the estimations of the relationship between board size, board independence, board gender diversity, and green loans, respectively. The lagged dependent variable (L.GL) is highly significant in three models, indicating the dynamic nature of bank green loan measures and endorsing our choice for the GMM estimator. Following the above assumptions, the GMM estimator is valid in our research. Specifically, the null hypothesis of no serial correlation of the order of one in the first-difference residuals is rejected, and the p values AR (1) are less than 0.1. The insignificant values of AR (2) show that there is no issue of second-order autocorrelation. The Sargan test probability values are insignificant, validating the overidentifying condition of the GMM.

The regression results show that the size of the boards had a positive association with banks’ green loans with a high significance level. Specifically, a 1% increase in board size is correlated with a 0.001% increase in green lending. Although the magnitude is humble, this reveals that the number of directors on board is meaningful in the discussion table of green lending decision. This result is also consistent with studies that document the positive and statistically significant nexus between board size and banks’ green practices as well as CRS activities (e.g., Birindelli et al. 2018; Bose et al. 2018; Jizi et al. 2014; Kiliç et al. 2015; Menicucci and Paolucci 2022, 2023; Spitzeck 2009).

A regression estimation indicated that banks’ board independence played a positive and statistically significant role in their green loan behaviour. The coefficient for this is associated with a 1% increase in independent directors on banks’ boards and a 0.002% increase in Vietnamese banks’ loans for green projects. This positive relationship also aligns with studies on related fields in the banking sector (e.g., Bose et al. 2018; Menicucci and Paolucci 2022). The crucial effect of outside directors on Vietnamese banks’ boards induces such boards to deliver a more objective consultancy to boards of management than inside directors (Lin and Nguyen 2022). Additionally, the more significant presence of independent directors on boards helps reduce banks’ conflicts of interest and agency problems (Liang et al. 2013). Menicucci and Paolucci (2022) also appreciated the influence of independent directors on banks’ boards regarding their expertise, experience, and reputation, thereby promoting banks’ sustainable development.

Besides outside directors, female directors on banks’ boards also displayed a significant role in green loan decision making. The regression results reported that a 1% increase in female directors on a bank’s board led to a 0.004% increase in the bank’s green loans, meaning that female directors have an influential voice in banks’ governance. This finding is supported by previous studies conducted by Bear et al. (2010); Francesco et al. (2019); and Menicucci and Paolucci (2022) related to gender diversity on banks’ boards. Research has documented that women raise greater awareness of environmental risk as an integrated part of operational risk (Francesco et al. 2019) and concern about people’s welfare (Sabina and Morten 2010). Women board members are believed to be more sensitive toward sustainability than men (Birindelli et al. 2018; Samara et al. 2019) and express more expertise than men on ESG issues (Arayssi et al. 2016; Velte 2016; Williams 2003). Moreover, banks should use women board directors’ intellectual and relational abilities to achieve valuable ESG performance (Menicucci and Paolucci 2022). Vietnamese banks, have only enjoyed the joint presence of female board directors in recent years (since 2016). Nevertheless, from our research data, it is exciting and noteworthy that some women now hold potent positions on committees, such as chair or vice-chair. Thus, it is reasonable that the broader appearance of female directors on banks’ boards fosters banks’ green lending in Vietnam.

4.2.2. Ownership Structure and Green Loans

The regression analysis between ownership structure and green loans was conducted, and the results are reported in

Table 5. Impact of ownership structure and green loans.

Variables	GL
L.GL	0.721 *** (0.005)
GSHA	0.0003 * (0.000)
FSHA	−0.002 *** (0.000)
ROA	−0.005 *** (0.001)
BKSZ	−0.001 *** (0.000)
CAR	−0.011 (0.001)
LEV	−0.000 (0.000)
LIQ	−0.002 (0.000)
AGE	0.002 *** (0.000)
Cons	0.017 *** (0.001)
Obs.	374
AR (1) (z value)	0.054 (−1.724)
AR (2) (z value)	0.176 (1.353)
Sagan (chi2)	1.000 (28.301)

Notes: This table shows the coefficients of estimates from the two-step system GMM method on banks’ green loans (GLs). All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, * p < 0.1.

. Models (4), (5), and (6) show the impacts of governmental shareholders, foreign shareholders, and nongovernmental domestic shareholders on green loans, respectively. All the specifications of the two-step system GMM pass the tests for second-order autocorrelation and the Sargan test.

From a structural standpoint, it is intriguing that the coefficients for governmental ownership were significantly positive while those for foreign ownership were notably negative. Specifically, a 1% increase in the percentage of state ownership leads to a 0.0003% increase in green loans, respectively, which is in line with the research conducted by the authors in (Amidjaya and Widagdo 2020; Bose et al. 2018; Yin et al. 2021; Zhou et al. 2022) on the impact of ownership structure on banks’ green practices. Conversely, a 1% increase in the share of foreign ownership in banks resulted in a decrease of 0.002% in green lending.

Studies have found that state-owned banks have an advantage in green finance due to their government ownership. In the Vietnamese context, governmental shareholders in banks are mostly institutional investors who have a high demand not only in terms of stable financial returns but also in terms of the environmental activities of banks (Bose et al. 2018). The significant positive relationship between government ownership and green loans is also attributed to the intention of the Vietnamese government to leverage state-owned banks as a mirror in green finance. Figure 1 shows the percentage of government ownership in our research data.

The significantly negative impact of foreign ownership on banks’ green loans in Vietnam followed a similar pattern to that observed in China by Bose et al. (2018) and Liang et al. (2013). While foreign investors in banks are expected to enhance corporate governance and improve business practices (Amidjaya and Widagdo 2020), this is not the case for Vietnamese banks. Our research data indicate that foreign investors are present in a limited number of Vietnamese banks with a relatively low proportion of such ownership, as shown in Figure 2. As a result, the voice of foreign shareholders may not be as influential as anticipated in green lending decision making.

Additionally, foreign shareholders may have a limited understanding of the local business culture (Liang et al. 2013) and prioritize green practices and growth, emphasizing sustainability reporting (Amidjaya and Widagdo 2020) and adhering to international standards when making loan decisions. This can pose challenges for Vietnamese enterprises, especially small and medium-sized ones, since these struggle to secure access to general loans and specialized funds for the eco-friendly initiatives that foreign banks provide.

In the realm of control variables, it is worth noting that, except for bank age, the other factors including bank profitability, bank size, bank capital adequacy ratio, bank leverage, and bank liquidity are negatively related to green loans. Banks with higher profitability and larger total assets may indicate maturity and limited future growth opportunities (Roll et al. 2009). Consequently, they are unwilling to invest in high-risk and low-return green projects (Farhad and Naoyuki 2019). Banks are reluctant to fund green projects due to Basel capital requirements (Farhad and Naoyuki 2020). However, in Vietnam, green loans are still relatively small in scale. Thus, they may not be under pressure from capital requirements. From a bank leverage perspective, as previously stated, outstanding green loans in the Vietnamese banking industry in 2022 only made up 4.23% of outstanding total loans, and the source of funds was mainly from international organizations such as the International Finance Corporation and the Green Credit Trust Fund. Therefore, banks may not depend on leverage to provide green loans. In relation to the age of banks, it has been observed that the adoption of green finance in Vietnam is a relatively recent phenomenon.

5. Mechanism Test

The results of the two-step system GMM regression evidence the significant impact of bank governance and ownership structure on banks’ green lending behaviours. Nevertheless, the channel through which the significant relationships are confirmed is questionable. In this section, we conduct a mechanism test to find out how bank governance and ownership structure can be significantly associated with green credit in Vietnamese banks. The authors in (Felipe 2020; Yao et al. 2023) revealed the mechanism between these factors; however, these were qualitatively analysed. Specifically, Felipe (2020) documented that the bank governance and ownership structure impact green lending through reporting processes which facilitate banks’ disclosure and transparency in engaging sustainable lending. Accordingly, disclosure and transparency are categorized into three claim types: promise-based claims (banks’ commitment in supporting environmental and social loans), negative claims (banks’ unwillingness to disburse green credit), and evidence-based claims (banks’ claim that the funding is provided for green projects). Based on their analysis, we manually collected the data on banks’ disclosure from their annual and ESG reports to empirically test the mechanism. We quantified the disclosure information in a binary variable and employed it as the mechanism in our test. Consequently, the variable takes a value of 1 if an annual or ESG reports reveals promise- and evidence- based disclosure; otherwise, it is 0. As such, a stepwise regression was employed to examine the mediation effect of reporting processes on the impact of bank governance and the ownership structure on banks’ green lending behaviours. The stepwise equations are given below:

$${REP}_{it}={\alpha }_0+{\alpha }_{1}corgo{v}_{it}+{\alpha }_2{control}_{it}{+\epsilon }_{it}$$

(4)

$${GL}_{it}={\gamma }_0+{{\gamma }_1{REP}_{it}+\gamma }_{2}corgo{v}_{it}+{\gamma }_3{control}_{it}{+\omega }_{it}$$

(5)

$${REP}_{it}={\beta }_0+{\beta }_1{ownership}_{it}+{\beta }_2{control}_{it}{+\mu }_{it}$$

(6)

$${GL}_{it}={\delta }_0+{\delta }_1{REP}_{it}+{\delta }_2{ownership}_{it}+{\delta }_3{control}_{it}{+\theta }_{it}$$

(7)

REP is a reporting process serving as a mechanism effect in our research. The paper also estimates different models for the research objectives. Accordingly, equation sets (4) and (5) test the mediation effect of reporting processes in the impact of bank governance on green lending while equation sets (6) and (7) test this effect in the role of ownership structure in green credit. Given the significant results of Equations (2) and (3), the stepwise regression is conducted in two steps. First, we test examine the significant levels of coefficients ${\alpha }_1$, ${\gamma }_1$, ${\beta }_1$, and ${\delta }_1$; if they are all significant, the mediation effect of the reporting processes is confirmed, indicating the significant channel of banks’ disclosure and transparency through which bank governance and ownership structure impact banks’ green lending behaviours. Second, we test the significant levels of coefficients ${\gamma }_2$ and ${\delta }_2$. If they are all significant, the mediation effect is partial; otherwise, it is complete. The results of the mediation effect tests are reported in

Table 6. Results of mediation effect regression on bank governance.

Variables	REP	GL
REP		0.003 *** (0.000)
BRSZ	0.172 * (0.101)	−0.002 ** (0.001)
BRIND	0.476 * (0.254)	−0.000 (0.002)
GEN	0.347 ** (0.132)	0.003 ** (0.001)
Obs.	403	403
Adj R-squared	0.323	0.119
Controls	YES	YES

Notes: This table shows the coefficients of estimates for the mediation effect of reporting processes in the impact of bank governance on bank green loans (GLs). All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, ** p < 0.05, * p < 0.1.

and

Table 7. Results of mediation effect regression on ownership structure.

Variables	REP	GL
REP		0.003 *** (0.000)
GSHA	−0.231 *** (0.083)	0.000 (0.001)
FSHA	−0.862 *** (0.209)	−0.004 * (0.002)
Obs.	403	403
Adj R-squared	0.342	0.128
Controls	YES	YES

Notes: This table shows the coefficients of estimates for the mediation effect of reporting processes on the impact of bank ownership structure on bank green loans (GLs). All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, * p < 0.1.

.

The estimated coefficients of corporate governance and ownership structure in Equations (3) and (5) and reporting processes in Equations (4) and (6) are highly significant in all regressions, confirming the mediation effect of the information disclosure for the impact of corporate governance and ownership structure on the green lending behaviours of banks. The reported coefficients of board size, board gender, and foreign ownership in Equations (4) and (6) are significant at high levels, indicating that the mechanism of reporting processes in supporting green lending through bank governance and ownership structure has a partial effect. The partial mediation effect is understandable because sustainable lending impacts Vietnamese banks’ green reputation and other financial aspects, and may be affected by other factors besides bank governance and ownership. The results demonstrate that the coefficients associated with board independence and state ownership are not statistically significant. This indicates the complete mediation of these factors in the decision-making process regarding green loans in Vietnamese banks.

6. Additional Analysis

6.1. Robustness Check

To add robustness to the findings of this research, we employed the panel-corrected standard errors (PCSE) method to re-estimate the baseline models. There are two primary reasons why the PCSE method fits our additional analysis. First, PCSE considers sequential correlation and cross-sectional dependence (Hasanul et al. 2021). Secondly, this method performs well even under the error conditions’ high heteroscedasticity and contemporaneous correlation (Beck and Katz 1995).

Table 8. Robustness with the PCSE method: impact of corporate governance and green loans.

Variables	GL
BRSZ	0.000 * (0.001)
BRIND	0.005 ** (0.002)
GEN	0.004 ** (0.002)
Controls	YES
Obs.	403
R-squared	0.064
chi2	27.83 ***

Notes: This table shows the coefficients of estimates from the PCSE method on banks’ green loans (GLs). All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, ** p < 0.05, * p < 0.1.

and

Table 9. Robustness with the PCSE method: impact of ownership structure and green loans.

Variables	GL
GSHA	0.001 ** (0.000)
FSHA	−0.004 *** (0.002)
Controls	YES
Obs.	403
R-squared	0.136
chi2	145.39 ***

Notes: This table shows the coefficients of estimates from the PCSE method on banks’ green loans (GLs). All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, ** p < 0.05.

report the robustness checks with the PCSE method.

The coefficients of the independent variables from the robustness test are significant and consistent with the results of the benchmark regression, which further strengthens our research findings.

6.2. The Role of Green Loan Principles Issuance

This study is a response to the regulatory guidance which is the Green Loan Principles (GLPs) published in 2018 by the Loan Market Association and the Asia Pacific Loan Market Association. We are concerned that there may be a difference in the impact of corporate governance and ownership structure on bank green credit before and after this guideline. Hence, in this section, we divide the research sample into two subgroups: banks before and after the issuance of the 2018 guidelines to test for the difference. Then, following Bose et al. (2018), we re-estimate the baseline models to compare the differences in the two subsamples. The regression results are reported in

Table 10. Sub-sample regressions for corporate governance.

Variables	2010–2017	2018–2023
L.GL	0.900 *** (0.027)	0.668 *** (0.007)
BRSZ	−0.000 * (0.000)	0.001 *** (0.000)
BRIND	−0.000 (0.000)	0.005 *** (0.001)
GEN	−0.000 (0.000)	0.002 *** (0.001)
Controls	YES	YES
Obs.	203	142
AR (1) (z value)	0.097 (−1.659)	0.087 (−1.619)
AR (2) (z value)	0.177 (−1.466)	0.493 (0.684)
Sagan (chi2)	0.987 (8.559)	0.135 (18.614)

Notes: This table shows the coefficients of the estimates from the two-step system GMM method on governance and bank green loans (GLs) for the different periods. All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, * p < 0.1.

and

Table 11. Sub-sample regressions for ownership structure.

Variables	2010–2017	2018–2023
L.GL	0.900 *** (0.027)	0.663 *** (0.004)
GSHA	0.001 *** (0.000)	0.001 *** (0.001)
FSHA	0.002 *** (0.001)	−0.003 *** (0.001)
Controls	YES	YES
Obs.	203	142
AR (1) (z value)	0.087 (−1.708)	0.080 (−1.614)
AR (2) (z value)	0.111 (−1.592)	0.494 (0.683)
Sagan (chi2)	0.944 (11.069)	0.152 (22.168)

Notes: This table shows the coefficients of estimates from the two-step system GMM method on ownership structure and bank green loans (GLs) for the different periods. All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01.

.

The regression results between the two subsamples before and after the proposal of the Green Loan Principles in 2018 confirm the differences in the impact of corporate governance and ownership structure on the green lending behaviours of Vietnamese banks. Before 2018, the criteria for receiving sustainable loans were not officially and clearly assessed by banks; therefore, they treated loans for sustainable water and wastewater management projects as normal loans. As a result, bank governance may not have a strong influence on these credits. Nevertheless, after the issuance of the Green Loan Principles, these loans are classified as green loans; hence, banks’ boards of directors and shareholders may convene with them discuss how they may support them as a means of improving their level of green credit. The results of the adaptive behaviour of banks’ boards and shareholders are also consistent with the conclusion of the authors in (Felipe 2020; Yao et al. 2023) that bank governance and ownership structure find ways to engage in sustainable lending.

6.3. The Role of Bank Size

We also examine whether the size of a bank influences its decision to finance green projects in Vietnamese banks. We establish a threshold using the average total assets of the banks to divide our sample into two categories: small banks with total assets below the mean, and large banks with total assets exceeding the mean. The results are shown in

Table 12. Bank size regressions for corporate governance.

Variables	Small Banks	Large Banks
L.GL	0.623 *** (0.004)	0.143 *** (0.008)
BRSZ	−0.002 (0.001)	0.001 ** (0.001)
BRIND	0.004 (0.003)	0.002 *** (0.001)
GEN	0.006 (0.001)	−0.001 ** (0.001)
Controls	YES	YES
Obs.	177	180
AR (1) (z value)	0.08 (−1.618)	0.031 (−1.512)
AR (2) (z value)	0.551 (0.596)	0.363 (−0.908)
Sagan (chi2)	0.999 (10.815)	1.000 (12.983)

Notes: This table shows the coefficients of estimates from the two-step system GMM method on governance and bank green loans (GLs) in terms of bank size. All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, ** p < 0.05.

and

Table 13. Bank size regressions for ownership structure.

Variables	Small Banks	Large Banks
L.GL	0.891 *** (0.016)	0.058 *** (0.009)
GSHA	0.017 ** (0.009)	0.001 ** (0.000)
FSHA	0.006 ** (0.002)	−0.004 ** (0.001)
Controls	YES	YES
Obs.	177	180
AR (1) (z value)	0.015 (−1.705)	0.018 (−1.997)
AR (2) (z value)	0.367 (0.902)	0.518 (−0.646)
Sagan (chi2)	0.997 (7.919)	1.000 (13.081)

Notes: This table shows the coefficients of estimates from the two-step system GMM method on ownership structure and bank green loans (GLs) in terms of bank size. All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, ** p < 0.05.

.

The findings suggest that board characteristics, particularly board size and independence, have a more significant influence on large banks as compared to small banks. In terms of ownership structure, state shareholders contribute to the promotion of green loans across all bank sizes, whereas foreign ownership only appears to benefit small banks.

6.4. The Role of Interaction between Banks’ Governance and Ownership Structure

We also test whether the interaction between governance and ownership structure casts any effects on the green lending behaviour of Vietnamese banks. As such, we perform the regressions for six interactions for these factors based on the two-step system GMM. The estimated findings are reported in

Table 14. Regression of interaction between banks’ governance and ownership structure.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
L.GL	0.745 *** (0.005)	0.712 *** (0.006)	0.805 *** (0.006)	0.713 *** (0.003)	0.809 *** (0.008)	0.858 *** (0.004)
BRSZ × GSHA	0.001 *** (0.000)
BRSZ × FSHA		0.002 * (0.001)
BRIND × GSHA			0.000 (0.001)
BRIND × FSHA				0.006 *** (0.002)
GEN × GSHA					0.001 (0.001)
GEN × FSHA						0.009 * (0.005)
BRSZ	0.001 *** (0.000)	0.000 (0.000)	0.001 *** (0.000)	0.000 * (0.000)	0.001 *** (0.000)	0.001 *** (0.000)
BRIND	0.001 ** (0.001)	0.002 *** (0.001)	0.001 * (0.001)	0.003 *** (0.001)	0.001 (0.001)	0.001 *** (0.000)
GEN	0.003 *** (0.001)	0.003 *** (0.001)	0.002 *** (0.001)	0.004 *** (0.000)	0.002 ** (0.001)	0.002 *** (0.001)
GSHA	0.000 (0.000)	0.001 *** (0.000)	0.001 ** (0.000)	0.001 *** (0.000)	0.001 * (0.000)	0.002 *** (0.000)
FSHA	−0.001 * (0.000)	−0.005 ** (0.002)	−0.001 (0.001)	−0.001 ** (0.001)	−0.001 (0.001)	−0.004 ** (0.002)
Controls	YES	YES	YES	YES	YES	YES
Obs.	374	374	374	374	374	374
AR (1) (z value)	0.074 (−1.627)	0.073 (−1.626)	0.075 (−1.622)	0.073 (−1.629)	0.072 (−1.629)	0.072 (−1.623)
AR (2) (z value)	0.222 (1.221)	0.220 (1.225)	0.204 (1.270)	0.223 (1.217)	0.203 (1.274)	0.212 (1.248)
Sagan (chi2)	0.898 (22.321)	1.000 (19.141)	0.508 (21.201)	1.000 (22.919)	0.351 (23.933)	0.496 (21.406)

Notes: This table shows the coefficients of estimates from the two-step system GMM method on the interaction between governance and ownership structure on bank green loans (GLs) in terms of bank size. All the variables are defined in Table 1. The robust standard errors of the estimated coefficients reported in parentheses are clustered at the bank level. *** p < 0.01, ** p < 0.05, * p < 0.1.

.

The results indicate that the interplay between board characteristics and ownership structure, such as the correlation between board size and ownership structure, foreign ownership in conjunction with board independence, and board gender, is positively associated with the promotion of eco-friendly lending by Vietnamese banks. The results imply that banks with larger boards are more inclined to engage in green financing, particularly when they have both domestic and foreign shareholders. Furthermore, the influence of board independence and gender diversity on green lending is more evident in banks with the significant participation of foreign investors.

7. Conclusions and Policy Implications

In this study, we discovered how banks’ governance and ownership structures can affect their green loan portfolio. Our study focuses on Vietnamese commercial banks from 2010 to 2023, utilizing the two-step system GMM and PCSE methods. Additionally, we conducted a mechanism test to find the mediation effect of these association. Subsample research is also reported to examine the effect of the Green Loan Principles to classify green loans from Vietnamese banks.

The research yielded four key findings. Firstly, banks’ green lending behaviour is positively and significantly linked to board characteristics including board size, board independence, and boar gender diversity. This result highlights the influential role that bank board characteristics play in banks’ loan decision making and their commitment to social responsibility. Secondly, the presence of shareholders has varying impacts on Vietnamese banks’ green loan behaviours. Specifically, banks with a higher percentage of state and private local shareholders tend to fund more green projects, while those with a greater proportion of foreign investors seem to limit the level of green credit. Thirdly, the role of bank governance and ownership structure in banks’ green lending behaviours is mediated by banks’ disclosure of their commitment to finance green projects. Fourthly, the issuance of the Green Loan Principles by Loan Market Association and the Asia Pacific Loan Market Association in 2018 facilitates the governance of sustainable lending by the awareness of banks’ board of directors and shareholders to support more green loans as documented in our research.

This article aims to support sustainable development by offering recommendations to policymakers and practitioners in Vietnam and other developing nations. First, the research findings evidence the critical role of the bank board in promoting green credit growth. Therefore, to leverage green finance in specific and sustainable development in general, governments should focus on banks who are key players in financial system. Secondly, among three types of shareholders, foreign ones may not improve the green credit level, which may result from the fact that foreign investors place a high demand loan assessment because of their international standards. Hence, it is imperative for the government to promote the establishment of a flexible and efficient loan assessment mechanism by foreign banks that considers local culture and practices. This approach will facilitate access to green finance for environmentally conscious borrowers. Thirdly, banks’ outsiders realize their commitment to finance green projects via their disclosure on annual and ESG reports; thus, to attract green loan borrowers, banks should share their green message through this channel to increase their green lending level. Fourthly, green loan guidelines can serve as a catalysis for sustainable governance, and then policymakers and banks in Vietnam and other developing countries can develop official green loan policies to improve green credit situation.

While the findings above and contributions are noteworthy, it is essential to acknowledge the limitations of this research. One significant limitation is the lack of an alternative measurement for green loans for Vietnamese banks. Additionally, our research results show that the coefficients of banks’ governance and ownership structures are low in economic magnitude, yielding minor effects on green lending in Vietnam. Therefore, future research is encouraged to incorporate more comprehensive data on green loans and different emerging market contexts to examine the effects of bank governance and ownership structure on green lending.