The Impact of ESG Performance on Corporate Financing
https://doi-001.org/1025/17618889454203
Chuan Qin*
School of Management at Shanghai University of Engineering Science, Shanghai, China
corresponding author：m335124112@sues.edu.cn
Costs: The Moderating Effect of Green Credit Policies
Abstract: Against the backdrop of China’s clear articulation of the “carbon peak” by 2030 and “carbon neutrality” by 2060 goals, which has spurred a wave of green development of ESG (environment), social, and governance performance has evolved from a supplementary non-financial performance evaluation tool to a core indicator for value assessment in the capital market, and the green credit policy as a green financial system core gripper, is reshaping the corporate financing landscape through differentiated credit mechanisms. Taking Chinese A-share listed companies from 2018 to 2022 as the research sample, this paper uses two-way fixed effects model and mediating effect model to systematically test the impact mechanism of ESG performance on corporate financing cost and the moderating effect of green credit policy. The findings indicate that: (1) ESG performance is significantly negatively correlated with debt financing cost. For every one grade increase in ESG rating, financing cost decreases by 0.213 percentage points on average, among which the reduction effect is most pronounced in the governance (G) dimension (-0.286), followed by social (S) dimension (S) (-0.197). The environmental (E) dimension (E) was relatively weak (-0.152); (2) The green credit policy significantly strengthens the above negative relationship. For every 1 percentage point increase in the policy intensity, the effect of ESG on reducing financing costs increases by 0.025 percentage points. In the top 25% of the policy intensity regions, the effect increases by 76.2% compared with the bottom 25% regions; (3) Heterogeneity analysis reveals that the effect is more pronounced in non-state-owned enterprises (ESG coefficient -0.287 vs state-owned enterprises -0.145), high-pollution industries (ESG coefficient -0.356 vs low-carbon industries -0.178) and eastern regions (ESG coefficient -0.268 vs central and western regions -0.189). (4) The mechanism test shows that the improvement of financial discipline (accounting for 12.68% of the mediating effect), the enhancement of investor confidence (21.13%) and the reduction of environmental compliance cost (9.87%) are the three paths of ESG to reduce financing cost. This paper enriches the cross-research of ESG and green finance, and provides empirical evidence for the optimization of corporate financing strategy and policy improvement.
Key words: ESG; Financing cost; Green credit policy; Moderating effect; Two-way fixed effects model
1. Introduction
1.1 Research Background
The deepening of the global sustainable development agenda is driving a restructuring of the value assessment system in capital markets. The ESG index system composed of Environment, Social and corporate Governance has become the core framework to measure the long-term risk and value of enterprises. According to the International Sustainability Standards Board (ISSB) released the sustainable financial reporting standards (IFRS S1 and S2), the standardization of ESG disclosure has become a global consensus. By the end of 2024, global sustainable investments had reached $30 trillion, accounting for 35% of the total global asset management. As the world’s largest developing country and carbon emitter, China clearly proposed the goal of “carbon peak” by 2030 and “carbon neutrality” by 2060 in 2020. The release of “Guidelines on ESG Information Disclosure for Listed Companies” in 2023 further promoted the standardization of ESG data. It is expected to reduce the financing cost of compliant enterprises by 5-10%.
Financing cost is a key constraint variable in corporate decision-making, and green credit policy, as a key tool to guide the flow of financial resources, has formed a complete chain of “policy guidance → bank implementation → corporate response”.China’s green credit policy began with the Opinions on Implementing Environmental Protection Policies and Regulations to Prevent Credit Risks jointly issued by the State Environmental Protection Administration, the People’s Bank of China and the China Banking Regulatory Commission in 2007. In 2012, the Green Credit Guidelines established the core policy framework. In 2018, the People’s Bank of China extended the green credit performance evaluation to all deposit-type financial institutions. Policy system in 2024 further in line with international standards (such as a reference to the eu classification schemes of sustainable finance thin green project standard). By the end of the fourth quarter of 2024, the balance of green loans in domestic and foreign currencies reached 36.6 trillion yuan, up 21.7% year on year, of which loans for green infrastructure upgrading and clean energy industries accounted for more than 69%. The outstanding green credit of Bank of China, China Construction Bank and other leading institutions exceeded 4 trillion yuan, and the annual growth rate of ESG-linked loans of Industrial Bank reached 19.64 percent.
However, two key issues remain unresolved in practice: first, doesocial (S) dimensional heterogeneity exist in the impact of ESG performance on financing costs? Why do some studies find a weaker effect for the environmental (E) dimension compared to the governancenvironmental (E) dimension? Second, how does green credit policy specifically regulate this relationship? Are there differences in the moderating effects among enterprises with different property rights and industrial attributes? Based on this, this paper systematically analyzes the relationship between ESG and financing cost and the regulation mechanism of green credit based on the latest policy evolution and industry data, which is of great practical significance.
1.2 Research significance
1.2.1 Theoretical significance
This study breaks through the traditional “financial indicators → financing cost” analytical framework, grounded in information asymmetry theory, principal-agent theory, and stakeholder theory, to construct an integrated analytical framework of “ESG threenvironmental (E) dimensions → policy moderation → financing cost”. By subdividing the effect of variousocial (S) dimensions of ESG and supplementing the intermediary path of environmental compliance cost, the existing research has made up for the deficiency of paying insufficient attention to “dimensional heterogeneity” and “mechanism refinement”. At the same time, combined with the evolution of China’s green credit policy in the past 20 years, this paper reveals the amplification mechanism of policy instruments on non-financial performance value in transition economies, and enriches the theoretical connotation of ESG research in emerging markets.
1.2.2 Practical significance
Provide accurate ESG optimization path for enterprises: Empirical results show that the governancenvironmental (E) dimension (such as ownership structure and internal control mechanism) has the most significant effect on reducing financing costs, and non-state-owned enterprises and high-pollution industries should give priority to strengthening the construction of thisocial (S) dimension. To provide pricing reference for financial institutions, for example, referring to Deutsche Bank’s “ESG penetrating review” experience, the governancenvironmental (E) dimension indicators are included in the top 10 decision-making factors of credit approval. To provide practical suggestions for policy makers, such as expanding the weight of “integrity of ESG information disclosure” in green credit assessment, and alleviating the problem of “difficult project screening” (a large state-owned bank disclosed that 20% of green credit applications had the risk of information fraud).
1.3 Research ideas and methods
1.3.1 Research ideas
First comb ESG and financing cost, the theoretical basis and literature of green credit policy context, mainly analyzes the research gap of dimension heterogeneity and policy moderating effect; Based on the three theory puts forward four core assumptions (direct effect, moderating effect and the heterogeneity of dimension, mechanism path); Again build a bidirectional fixed effects model, using the data of listed companies from 2018 to 2022 baseline regression, moderating effect inspection, grouped with the mediation effect of regression analysis; Finally, policy suggestions are put forward based on the industry data in 2024, and the research limitations are pointed out.
1.3.2 Research methods
(1) Panel data regression method: Using bidirectional fixed effects model control individual heterogeneity effect (such as corporate governance culture) and time (such as the outbreak hit), through the Hausman test (χ squared = 68.32, p < 0.01) confirmed its applicability, the method in dealing with the unbalanced panel data fixed effects of individual and time advantages, It has been widely used by Albuquerque et al. (2019) and other international studies.
(2) Moderating effect test: The moderating effect is identified by the interaction term of ESG and green credit policy intensity, and the double verification is combined with group regression (high/low policy intensity group) and marginal effect plot. The grouping standard refers to the internationally accepted quartile method to ensure the statistical significance of the difference between groups.
(3) Heterogeneity analysis: the groups were grouped according to the nature of property rights (state-owned enterprises/non-state-owned enterprises), industrial pollution degree (high pollution/low carbon), and regional development level (eastern/central/western). Chow test was used to determine the difference of coefficients between groups, and the interaction term regression was supplemented to exclude multicollinearity interference.
(4) Mediating effect test: Stepwise regression method and Bootstrap method (repeated sampling for 500 times, confidence interval 95%) are used to test the three mediating paths of financial discipline, investor confidence and environmental compliance cost. Bootstrap method can effectively solve the problems of endogeneity and distribution bias in mediating effect estimation.
2. Literature review
2.1 Research on the relationship between ESG performance and corporate financing cost
The impact of ESG performance on financing cost has formed a multi-theoretical interpretation framework, and the core can be classified into three categories:
2.1.1 Information Asymmetry Perspective
High ESG performance alleviates the adverse selection problem by standardizing information disclosure. Research by the International Organization of Securities Commissions (IOSCO) shows that …the interest rate premium on corporate bonds with full ESG disclosure decreases by 0.2%.. After the implementation of the CSRC guidelines in 2023, the disclosure rate of ESG information of A-share listed companies has increased from 45% to 78%, and the corresponding financing cost has decreased by 0.35 percentage points on average. Shen et al. (2022) found that information disclosure from the governancenvironmental (E) dimension has the most significant effect on reducing financing costs, because information such as ownership structure and internal control mechanism can be more easily evaluated by banks quantitatively, which is consistent with the international research conclusions on information transparency in emerging markets (such as ESGRI, 2022).
2.1.2 Perspective of risk premium
ESG performance reduces the risk compensation demanded by creditors by reducing operational risks. According to MSCI’s 2022 ESG Investment Report, companies with high ESG ratings have on average 20 basis points higher credit ratings and 0.5% lower financing costs. Harvard Business School research confirms that companies in the top 20 per cent of ESG scores have 32 per cent less volatility in operating risk than those in the bottom 20 per cent, with significant reductions in environmental compliance costs (3.7 per cent of revenue) and losses from Labour disputes. In particular, the International Energy Agency (IEA) points out that long-term bond costs of companies with high carbon intensity are 30-50 basis points higher, which explains the significant effect of environmental (E) dimension in high-polluting industries, which is highly compatible with the regulatory environment of China’s high-polluting industries.
2.1.3 Reputational mechanism perspective
ESG performance attracts low-cost capital by improving corporate reputation. According to the UN Global Compact, companies with high social responsibility ratings have reduced their borrowing premium by 8-12 percent because of greater cash flow stability due to consumer loyalty and employee productivity. However, there is a threshold for this effect — Albuquerque et al. (2019) found that the reputation premium becomes significant only after the ESG rating exceeds BBB +, and this threshold effect is more prominent in emerging markets.
Controversy on dimension heterogeneity: There are differences in the ranking of effects of ESGovernance (G) dimensions in existing studies. Gao et al. (2022) argued that “governance > society > environment”, because the standardization of environmental information disclosure in China is low; Wang Fang (2025) found that “environment > governance > society” based on the samples of high-pollution industries, because environmental punishment directly affects the credit qualification of enterprises. This paper will further clarify this controversy by combining the samples of the whole industry and the sub-industry, and compare the conclusions of similar international studies.
2.2 Economic Effects of Green Credit Policies
Green credit policies operate through a dual mechanism of “positive incentives and negative constraints”:
2.2.1 Resource Allocation Effect
Policies guide capital flows through risk-weight moderating effects and re-lending support. Data from the People’s Bank of China (2024) show that after the risk weight for green project loans was reduced from 100% to 75%, the approval efficiency for green credit increased by 40%, and with loan interest rates being, on average, 0.8 percentage points lower. Research by the National Institute of Finance and Development (NIFD) confirms that the policy reduces the financing costs of qualifying enterprises by 0.8 percentage points, while the credit growth rate of “two high” enterprises (high energy consumption and high pollution enterprises) has decreased by 15%. This regulatory effect is comparable to the resource allocation effect of sustainable finance instruments in the European Union.
2.2.2 Corporate Behavior Guidance Effect
Policies compel enterprises to improve their ESG performance. Li Jun (2025) found that after the implementation of the green credit policy, the proportion of environmental protection investment by firms in high-polluting industries increased from 2.1% to 5.8%, and the carbon emission reduction increased by 22%. However, the effect exhibits regional heterogeneity. According to the case of Jiangsu Economic News (2025), the ESG data integration ability of banks in the eastern region is strong (for example, China Merchants Bank has built a multi-source ESG database), and the policy effect is 1.8 times that of the central and western regions, which is consistent with the international research conclusions on the impact of financial infrastructure on policy transmission efficiency.
2.2.3 Policy Instrument Innovation Effect
Esg-linked loans and other products will strengthen policy effects. China Construction Bank will issue 107.3 billion yuan of green bonds in 2024 and launching a “base interest rate + ESG floating rate” product, where the interest rate drops by 5-10 basis points for every one-notch improvement in the ESG rating. Industrial Bank disclosed that the amount of credit granted to only ESG compliant enterprises increased by 18%, which confirmed the guidance of policies on bank behavior. Similar innovations have formed a mature model in the European and American green finance markets (such as HSBC’s ESG pricing mechanism).
2.3 The Moderating Effect of Green Credit Policy
While existing research has acknowledged the moderating effect of green credit policy, three major gaps remain: First, the analysis of the moderating mechanism remains superficial, often concluding merely that “the policy strengthens the negative relationship”, without analyzing the transmission chain of “bank assessment – enterprise response”; Second, there is a lack of dimensional heterogeneity analysis, which fails to test whether there are differences in the regulation of policy on ESGovernance (G) dimensions; The third is the lack of timeliness of the data, which does not include the impact of the policy intensification period after 2023 (such as the implementation of the Green Credit Project Management Measures).
A small number of representative studies are as follows: NIFD (2025) finds that green credit plays a positive moderating role in the relationship between ESG and carbon emission reduction, and indirectly improves the effect of ESG on reducing financing costs. Based on the data from 2015 to 2020, Zhang et al. (2025) found that the policy moderating effect was more significant in non-state-owned enterprises, but did not explain the reason. Jiangsu economic news (2025) cases, ESG tied loan policy moderating effect of 30% increased, but the lack of large sample empirical support, and compared with international policy moderating effect experience.
2.4 Literature review
To sum up, the existing research has confirmed the negative impact of ESG on financing cost and the resource allocation effect of green credit, but there are still three key gaps: (1) the formation mechanism of ESGovernance (G) dimensional heterogeneity has not been clarified, especially the interaction analysis of “policy context and dimension effect” has not been fully considered, and the institutional particularity of emerging markets has not been fully considered. (2) The regulatory mechanism of green credit policy needs to be further refined, for example, how the policy affects the value of ESG signal through the bank assessment mechanism (such as the weight of green credit in MPA assessment), and this transmission path is different in international policy practice; (3) The research on the intermediary path is not comprehensive enough. The existing literature mainly focuses on financial discipline and investor confidence, but ignores the key path of environmental compliance cost (especially in high-pollution industries), and lacks the analysis of the strengthening effect of policies on the intermediary path. Based on this, this paper will focus on the above gaps, combining the latest data from 2018 to 2022 and the industry dynamics in 2024 to carry out research.
3. Theoretical analysis and research hypotheses
3.1 Direct impact and dimensional heterogeneity of ESG performance on corporate financing costs
3.1.1 Direct impact: analysis based on signaling theory
In the credit market with asymmetric information, a firm’s ESG performance is a “credible signal” to convey its risk level. High ESG performance means that enterprises: (1) environmental (E) dimension: low carbon emission intensity, strong environmental compliance, easy to be punished by the Air Pollution Prevention and Control Law and other regulations (the average amount of environmental penalties for high-pollution industries in 2024 will reach 8.6 million yuan), and higher cash flow stability; (2) Social (S) dimension: attach importance to employee rights and consumer protection, low incidence of labor disputes (UN data show that the top 20% of enterprises with social responsibility rating have 60% lower dispute rate), and strong business continuity; (3) Governancenvironmental (E) dimension: reasonable ownership structure, perfect internal control, low agency cost (according to China Securities data, the agency cost of the top 20% of enterprises with governance rating is 35% lower), and the risk of capital abuse is small. These advantages reduce the credit risk and policy risk of creditors, and make them willing to accept lower risk premium. Accordingly, the following propositions are proposed:
Hypothesis H1: ESG performance is significantly negatively correlated with financing cost, that is, the better ESG performance is, the lower financing cost is.
3.1.2 Dimensional heterogeneity: analysis based on information quantifiability
The signal value of each dimension of ESG is different, and the core reason lies in the different information quantization and policy sensitivity: (1) governancenvironmental (E) dimension: ownership concentration, proportion of independent directors and other indicators can be directly obtained from the annual report, and are linked to the mandatory requirements of the Company Law. Banks are easy to quantify and evaluate, and the signal value is the highest; (2) Social (S) dimension: the data disclosure rate of employee salary growth rate and proportion of donation expenditure has increased (up to 68% in 2024), but there is no unified standard, and the signal value is the second; (3) Environmental (E) dimension: the standardization degree of data such as carbon emissions and environmental protection investment is low (the disclosure rate of high-pollution industries is only 45%), and some enterprises have the behavior of “greenwashing” (12% of green projects of a large bank were found to be fake in 2024), and the signal value is relatively weak. However, in high-pollution industries, the environmental (E) dimension is directly related to the survival of enterprises (such as China’s 2030 “carbon peak” and 2060 “carbon neutrality” target production restriction policy), the signal value will be significantly increased. According to this, the following propositions are proposed:
Hypothesis H1a: There is heterogeneity in the reduction effect of ESGovernance (G) dimensions on financing costs, which is manifested as “governancenvironmental (E) dimension > social (S) dimension > environmental (E) dimension”;
H1b: in high-pollution industries, the effect of environmental (E) dimension on reducing financing cost is significantly enhanced.
3.2 Moderating effect of green credit policy
Green credit policy regulates the relationship between ESG and financing cost through the dual path of “mechanism strengthening + signal amplification” :
3.2.1 Mechanism strengthening
Policy to include ESG in hard assessment of Banks – The 2018 Green Credit Performance Evaluation Plan of the People’s Bank of China included ESG compliance in MPA assessment (with a weight of 10%), so that banks are motivated to give preferential interest rates to enterprises with high ESG. The top 20% enterprise for example, the industrial and commercial bank of ESG, given its benchmark interest rate down from 15% to 30%, interest rates to rise 20% to 20% after enterprises. At the same time, the policy reduces banks’ capital occupation through risk weight moderating effect (green projects are reduced from 100% to 75%) and enhances their willingness to grant credit to high ESG enterprises. This mechanism is in line with the international trend of green capital preference in Basel III.
3.2.2 Signal amplification
Information sharing between environmental protection and credit investigation systems (such as the inclusion of national environmental protection penalty information into credit investigation in 2023) makes ESG signals easier to identify. After the implementation of the policy, banks’ use of ESG information increased from 32% to 78%, and the gap between financing costs for companies with different ESG levels widened by 50%, according to S&P. In addition, innovative products such as ESG-linked loans (issuance increased by 84.35% in 2024) enable enterprises to directly translate ESG improvement into lower interest rates, further strengthening the negative relationship. The penetration rate of similar products in the EU market has reached 25%.
There are boundary conditions for the policy moderating effect: in non-soes, due to stronger financing constraints (the credit approval rate is 28% lower than that of soes), the marginal value of ESG signal is higher, and the policy moderating effect is more significant; In the high pollution industry, the policy of “projects” enterprise stricter constraints, financing of ESG performance value was further enlarged. Based on this, it is proposed that:
H2: Green credit policy has a positive moderating effect on the relationship between ESG performance and financing cost, that is, the higher the policy intensity is, the more significant the effect of ESG performance on reducing financing cost is;
H2a: The moderating effect of green credit policy is more significant in non-state-owned enterprises;
H2b: the moderating effect of green credit policy is more significant in high-pollution industries.
3.3 Influence mechanism: Analysis based on mediating effect
ESG performance by three paths to reduce financing costs, and the green credit policy will strengthen the path:
3.3.1 Path of financial discipline
High ESG performance more perfect enterprise internal control mechanism, low incidence of financial restatement (China, according to data from the card before the ESG enterprise restatement rate of only 3.2% 20%), reduced the creditors to the concerns of the abuse. The green credit policy further improves the transmission efficiency of this path by requiring banks to review the “ESG internal control process,” which has become a standard process in Deutsche Bank’s credit decisions.
3.3.2 Investor confidence path
High ESG performance attracts more analyst attention (4.2 more analysts follow on average), improves information transparency, and reduces the risk premium demanded by investors. By promoting the standardization of ESG information (such as IFRS S2), the policy has enhanced analysts’ trust in ESG data, and the utilization rate of standardized ESG data has reached 85% in the world.
3.3.3 Path of environmental compliance costs
Enterprises with high ESG performance have sufficient investment in environmental protection and few penalties for environmental protection (the penalty rate of the top 20% of ESG enterprises is only 1.8%), avoiding the impact of cash flow caused by fines and production suspension. Under the green credit policy, banks will directly verify the records of environmental penalties, making the impact of this path more prominent. This verification mechanism echoes the implementation requirements of the EU’s Environmental Responsibility Directive. According to this, it is proposed that:
Hypothesis H3: financial discipline, investor confidence and environmental compliance costs play an intermediary role in the relationship between ESG performance and financing costs;
H3a: Green credit policy significantly strengthens the transmission effect of the above intermediary path.
5.4 Research design
4.1 Sample selection and data sources
4.1.1 Sample screening logic
This paper selects A-share listed companies in Shanghai and Shenzhen from 2018 to 2022 as the initial sample, and the screening process follows the common standards of international academic research, as follows: (1) excluding ST and * ST abnormal financial company, this company’s financial data distortion of risk is higher, may interfere with accurate measurement of the cost of financing, the three factor model and Fama – French effectiveness on the sample request in the agreement; (2) to eliminate the financial industry and real estate industry, financial industry (such as capital adequacy requirements) because of its special regulatory policy leads to the financing structure and essential differences in other industries, the real estate industry is regulated by the China policy, financing cost fluctuation has particularity, excluding can strengthen the sample homogeneity of two classes of industry; (3) Excluding the companies with missing ESG ratings and incomplete financial data to avoid the bias of missing values on the regression results; (4) for continuous variables for 1% and 99% points of winsorization, such as extreme value shrank to 15.2% from 9.76% the cost of financing, in order to eliminate the interference of outliers, tail shrinkage ratio with reference to the top issue of Journal of Finance and other commonly used Settings.
After the screening, finally get 1243 companies, 6215 observations. Sample covers 19 CSRC primary industry classification standard (2012), accounted for 58.2%, including manufacturing services sector accounted for 26.7%, other industries accounted for 15.1%; In terms of regional distribution, enterprises in the eastern region accounted for 58%, while those in the central and western regions accounted for 42%. In terms of the nature of ownership, state-owned enterprises account for 38%, and non-state-owned enterprises account for 62%. The sample structure is basically consistent with the overall characteristics of China’s A-share listed companies, which is relatively representative.
4.1.2 Data source and authority explanation
ESG performance: (1) using China certificate of ESG rating comprehensive score and dimension scores (E, S, G 1-9). The rating system covers 14 secondary indicators, 45 three-level index, the environmental (E) dimension including carbon intensity, environmental protection investment accounted for more than 12 indexes such as, social (S) dimension including employee compensation rate, donation accounts for more than 15 indexes, such as governancenvironmental (E) dimensions including ownership concentration, the proportion of independent directors and other 18 indexes. With the international mainstream (such as the MSCI, Sustainalytics), a rating agency, compared to China the ESG rating for the suitability of Chinese enterprises is stronger, its compatibility with IFRS S2 standards 85%, significantly higher than the other 60% – 70% of the rating agencies, And into green financial professional committee of China financial society have been recommended rating system. To ensure robustness, Wind ESG rating will be used as a proxy indicator in subsequent tests.
(2) Financing costs: The data of debt financing cost (COD) come from CSMAR database, and the calculation formula is (interest expense + commission fee + other financial expense)/total debt at the end of the period ×100%. This index fully reflects the comprehensive cost of corporate debt financing. Compared with the internationally used index of “total interest expense/total debt”, Hidden costs, such as increased fees, more joint Chinese enterprises’ financing practice. At the same time, the financial expense/total debt at the end of the period (COD2) is selected as a substitute index to verify the stability of the results.
(3) Green credit policy intensity (GCP) : it is measured by the proportion of green credit balance in total loan balance in each province. This index can effectively reflect the implementation of green credit policies at the regional level, which is consistent with the logic of “policy instrument density” index used internationally. In 2022, according to data from guangdong, zhejiang GCP highest (15.62%, 14.89%), gansu, qinghai minimum (2.17%, 2.89%), regional differences are in conformity with the geographical distribution of China’s green financial development characteristics.
(4) intermediary variables: financial discipline (Restate) use “whether financial restatement” measure (happened in 1, otherwise 0), data from CSMAR database the financial restatement research database, the database has been marked with the restatement reason and degree, ensure the accuracy of indicators; Investor confidence (Analyst) adopt the ‘analyst coverage + 1 take natural logarithm “, an Analyst, measured data from Wind database “analysts predict database”, add 1 take logarithm can avoid zero impact on regression; Environmental compliance costs (Punish) use “punishment” whether by environmental protection measure (by 1, otherwise 0), data from CSMAR with the Wind “environmental penalties database”, covers the administrative penalty of public information department of the environment.
(5) control variables: Company Size (Size) for the final the natural logarithm of total assets, asset-liability ratio (Lev) for the final total liabilities/final total assets by 100%, the profitability (Roe) is a net/net assets by an average 100% Growth (Growth) for (revenue – revenue last year)/revenue by 100% last year, Fixed asset ratio (PPE) is net fixed assets/total assets at the end of the period ×100%, Cash ratio (Cash) is monetary funds/current liabilities ×100%, Age of listing (Age) is the year of the current year – listing year + 1, all the above indicators are from CSMAR database; Property rights (SOE) is a virtual variable (0) state-owned enterprises take 1, the state-owned enterprises, the annual reports of listed companies by manual processing “actual controller” information to determine; Industry competition degree (HHI) according to the calculated classification 2012 industries, calculating formula for each enterprise in the industry revenue accounted for the sum of the squares of the; Regional development level (East) is a virtual variable (the eastern region to take 1, 0) in the Midwest, including Beijing, Shanghai and other 10 provinces and cities in eastern, divide the standard reference to the National Bureau of Statistics official definition.
4.2 variable definitions
| Variable types | Variable name | Variable symbol | Variable definition |
| Be explained variable | The cost of debt financing | COD | (interest expense + commission fee + other financial expenses)/total liabilities at the end of the period ×100%, reflecting the comprehensive cost of corporate debt financing |
| Cost of debt financing (alternative) | COD2 | Financial expenses/total liabilities at the end of the period ×100%, as the robustness checks index of COD | |
| Core explanatory variables | ESG performance | ESG | China Securities ESG comprehensive rating score (1-9), with higher scores indicating better ESG performance |
| Environmental (E) dimension performance | E | ESG environmental (E) dimension score (1-9 points), covering carbon emissions, environmental protection investment and other indicators | |
| Performance of social (S) dimension | S | The social (S) dimension score of Huazhen ESG (1-9 points) covers indicators such as employee rights and interests and social responsibility | |
| Performance of governancenvironmental (E) dimension | G | ESG governancenvironmental (E) dimension scores (1-9 points), covering indicators such as ownership structure and internal control mechanism | |
| Moderating variables | Green credit policy intensity | GCP | Green credit balance of provinces/total loan balance of provinces ×100%, reflecting the implementation of regional policies |
| Mediating variable | Financial discipline | Restate | If financial restatement occurs, take 1; otherwise, take 0 to measure the quality of internal control |
| Investor confidence | Analyst | The natural logarithm of the number of analysts following + 1 reflects the attention to market information | |
| Environmental compliance costs | Punish | The value of 1 for being punished by environmental protection and 0 for otherwise is used to measure the level of environmental risk of enterprises | |
| Control variables | Company size | Size | Natural logarithm of total assets at the end of the period, controlling for firm size effects |
| Asset-liability ratio | Lev | Total liabilities at the end of the period/total assets at the end of the period ×100%, controlling financial leverage risk | |
| Profitability | Roe | Net profit/average net assets ×100%, controlling the influence of profit level | |
| Growth | Growth | (Current year’s revenue – previous year’s revenue)/previous year’s revenue ×100%, controlling for the impact of growth potential | |
| Fixed assets ratio | PPE | Net fixed assets/total assets at the end of the period ×100%, controlling the impact of asset structure | |
| Cash ratio | Cash | Monetary funds/current liabilities ×100%, controlling short-term solvency | |
| Years of listing | Age | Year of the year – listing year + 1, controlling the effect of enterprise maturity | |
| Nature of ownership | SOE | 1 for soes and 0 for non-soes to control for differences in property rights systems | |
| Industry competition degree | HHI | Herfindahl index (classified by CSRC industry), controlling for the competitive environment in the market | |
| Level of regional development | East | The value is 1 for the eastern region and 0 for the central and western regions to control regional development differences | |
| Industry fixed effects | Industry | CSRC 2012 industry dummy variables (19) to control industry heterogeneity | |
| Year fixed effects | Year | Dummy variables from 2018 to 2022 (5), controlling for time trend effects |
4.3 Model Construction
4.3.1 Baseline regression Model (test H1, H1a, H1b)
In order to test the direct effect and dimensional heterogeneity of ESG performance on financing cost, the following two-way fixed effects model is constructed:
COD_{i,t} = \alpha_0 + \alpha_1 ESG_{i,t} + \sum \alpha_k Control_{k,i,t} + \mu_i + \lambda_t + \varepsilon_{i,t} \quad (1)
COD_{i,t} = \alpha_0 + \alpha_1 E_{i,t} + \alpha_2 S_{i,t} + \alpha_3 G_{i,t} + \sum \alpha_k Control_{k,i,t} + \mu_i + \lambda_t + \varepsilon_{i,t} \quad (2)
Among them, COD_ {I, t} for the enterprise in the first I t years of debt financing cost; ESG_{i,t}, E_{i,t}, S_{i,t} and G_{i,t} are the comprehensive and sub-dimensional ESG performance of enterprise i in year t, respectively; Control_{k,i,t} is a series of control variables; \mu_i is the individual fixed effect, which is used to control the characteristics that do not change with time at the enterprise level (such as governance culture and industry attributes); \lambda_t is the time fixed effect, which is used to control the factors that change over time at the macro level (such as the impact of the epidemic and monetary policy); \ varepsilon_ {I, t} as random error term.
Model (1) is used to test hypothesis H1: if \alpha_1<0 and is statistically significant, it indicates that ESG performance is negatively correlated with financing cost, and H1 is valid. Model (2) is used to test Hypothesis H1a: if the absolute value of the coefficients satisfies |\alpha_3|>|\alpha_2|>|\alpha_1| and they are all significant, it indicates that the order of ESGovernance (G) dimensional effects is “governance > society > environment”, and H1a is valid.
In order to test H1b, the regression is conducted separately on the high-pollution industry samples. High-pollution industries were defined by referring to China’s Environmental Protection Verification Industry Classification and Management List of Listed Companies, including 16 industries such as steel, chemical and non-ferrous metals. If the absolute value of the coefficient of Environmental (E) dimension in the regression results is significantly larger than the corresponding coefficient in the whole sample, it indicates that the effect of environmental (E) dimension is enhanced in high-pollution industries, and H1b is established.
4.3.2 Moderating effect Model (Test H2, H2a and H2b)
In order to test the moderating effect of green credit policies, the interaction term of ESG and green credit policy intensity (GCP) is introduced into the benchmark model, and the following model is constructed:
COD_{i,t} = \beta_0 + \beta_1 ESG_{i,t} + \beta_2 GCP_{i,t} + \beta_3 ESG_{i,t} \times GCP_{i,t} + \sum \beta_k Control_{k,i,t} + \mu_i + \lambda_t + \varepsilon_{i,t} \quad (3)
Where ESG_{i,t} \times GCP_{i,t} is the interaction term , and the core coefficient is \beta_3. If \beta_3 is <0 and significant, it indicates that the higher the intensity of green credit policy is, the stronger the effect of ESG on reducing financing cost is, so H2 is valid.
To test hypotheses H2a and H2b, the group regression method is used: (1) according to the nature of property rights, the group is divided into non-state-owned enterprise group and state-owned enterprise group. If the absolute value of the interaction term coefficient \beta_3 of non-state-owned enterprise group is significantly larger than that of state-owned enterprise group, H2a is established; (2) according to the industry pollution can be divided into high pollution industry group and low carbon industry group, if high pollution industry group to pay by the absolute value of coefficient \ beta_3 significantly greater than the low carbon industry group, H2b was established. At the same time, Chow test is used to determine statistical significant difference coefficient between groups, if F statistic significantly, further verify the existence of heterogeneity.
In addition, in order to enhance the credibility of the results, the marginal effect diagram was drawn to visually show the marginal impact of ESG on COD at different GCP levels. The calculation of the marginal effect was based on the estimated results of Model (3), and the confidence interval was set to 95%.
4.3.3 Mediating effect Model (Test H3 and H3a)
Based on the stepwise regression method and the Bootstrap method, building a mediation effect model financial discipline inspection (Restate), investor confidence (Analyst), environmental compliance costs (Punish) mediation role, specific model is as follows:
Step 1: check the total effect of independent variable on the dependent variable (that is, the baseline model (1));
Step 2: check the influence of independent variable of mediation variables:
M_{i,t} = \gamma_0 + \gamma_1 ESG_{i,t} + \sum \gamma_k Control_{k,i,t} + \mu_i + \lambda_t + \varepsilon_{i,t} \quad (4)
Step 3: Include independent variables and mediating variables simultaneously to test the mediating effect:
COD_{i,t} = \delta_0 + \delta_1 ESG_{i,t} + \delta_2 M_{i,t} + \sum \delta_k Control_{k,i,t} + \mu_i + \lambda_t + \varepsilon_{i,t} \quad (5)
Step 4: Test the strengthening effect of green credit policy on the intermediary path, and introduce the interaction term:
COD_{i,t} = \theta_0 + \theta_1 ESG_{i,t} + \theta_2 GCP_{i,t} + \theta_3 ESG \times GCP + \theta_4 M_{i,t} + \sum \theta_k Control_{k,i,t} + \mu_i + \lambda_t + \varepsilon_{i,t} \quad (6)
Where M_{i,t} is the mediating variable (Restate, Analyst, Punish). If \gamma_1 in Model (4) and \delta_2 in Model (5) are both significant, and the absolute value of \delta_1 in Model (5) is smaller than that of \alpha_1 in Model (1), it indicates that there is a mediating effect. The mediation effect of calculating formula for (\ alpha_1 – \ delta_1) / 100 \ % \ alpha_1 \ times, reflect the extent to which the mediation path’s contribution to the total effect.
To test hypotheses H3a) (policy to strengthen intermediary effect, if \ theta_3 significantly in model (6), and the absolute value of \ delta_2 less than item not yet given by introducing the coefficient of absolute value, that green credit policy through strengthening mediation path, H3a was established. At the same time, the Bootstrap method is used to repeat sampling 500 times, calculate the confidence interval of the mediation effect, if the confidence interval does not contain 0, further verify the mediation effect of robustness.
For the path of environmental compliance cost, we additionally conduct regression on the samples of high-pollution industries to test the intensification effect of this path in highly sensitive industries.
5. Empirical results and analysis
5.1 Descriptive statistics
The descriptive statistics report the main variables in table 1, article 6215 of the distribution features of the observed value is as follows:
(1) Explained variables: the mean value of debt financing cost (COD) is 4.28%, the standard deviation is 1.83%, the minimum value is 1.05%, the maximum value is 9.76%, and the maximum value is 9.3 times of the minimum value, indicating that there are significant differences in the financing cost of China’s A-share listed companies. This difference is higher than that in the mature markets of Europe and the United States (for example, the COD standard deviation of listed companies in the United States is about 1.2%), which may be related to the high degree of information asymmetry in China’s credit market. The mean value of COD2 is 3.96%, which is basically consistent with the distribution characteristics of COD, indicating that the measurement of financing cost index is stable.
(2) Core explanatory variables: the mean value of ESG comprehensive score is 4.62 points (out of 9 points), which is at the lower middle level, reflecting that there is still a large room for improvement in the ESG performance of Chinese listed companies, which is consistent with the reality that China’s ESG disclosure system started relatively late (unified disclosure guidelines will be issued in 2023). In terms of dimensions, the mean value of Governance (G) dimension is the highest (4.89 points), followed by Social (S) dimension (4.56 points), and Environmental (E) dimension is the lowest (4.21 points), which initially confirms the existence of dimensional heterogeneity and is consistent with the findings of Gao et al. (2022). Compared with international data, the ESG score of Chinese listed companies is lower than the global average (about 5.2 points), but higher than that of other BRICS countries (about 3.8 points), which reflects the phased characteristics of emerging markets.
(3) Moderating variable: the mean value of green credit policy intensity (GCP) is 8.35%, and the standard deviation is 3.12%, indicating that the policy implementation intensity varies significantly among provinces. The mean value of GCP in the eastern region is 10.21%, significantly higher than that in the central and western regions (6.48%), which is related to the background of better financial infrastructure and stricter environmental regulation in the eastern region. The GCP of Guangdong, Zhejiang and other green finance pilot provinces is more than 14%, while that of Gansu, Qinghai and other western regions is less than 3%. Regional differences provide good sample variation for moderating effect test.
(4) The mean value of mediating variable: financial discipline (Restate) is 0.12, which means that 12% of the sample enterprises have had financial restatement, higher than 8% of the US market, indicating that the quality of internal control of Chinese enterprises still needs to be improved; The average value of investor confidence (Analyst) is 2.35, which is followed by about 10 analysts, lower than the average level of CSI 300 (about 15), indicating that the market attention of small and medium-sized listed companies is low; Environmental compliance costs (Punish) average of 0.08, or 8% of the enterprises was fined by the environmental protection, high pollution industry the proportion rose to 18%, significantly higher than 2% of the low carbon industry, embodies the differences of industry environment risk.
(5) Control variables: the average value of company Size (Size) is 23.15 (corresponding to the total assets of about 20 billion yuan), and the average value of asset-liability ratio (Lev) is 45.67%, which is in a reasonable range; The mean value of profitability (Roe) is 8.23%, and the minimum value is -12.35%, indicating that some enterprises have losses; Property rights (SOE) average of 0.38, shows sample China accounts for less than forty percent; The mean value of regional development level (East) is 0.58, indicating that enterprises in the eastern region are dominant in the sample. The standard deviation of each control variable is within a reasonable range, and there is no extreme value problem, which lays a good foundation for subsequent regression.
Table 1 Descriptive statistics of main variables
| Variables | Observations | Mean | Standard deviation | Minimum | Median | Maximum value |
| COD | 6215 | 4.28 | 1.83 | 1.05 | 4.02 | 9.76 |
| COD2 | 6215 | 3.96 | 1.72 | 0.98 | 3.85 | 9.21 |
| ESG | 6215 | 4.62 | 1.78 | 1.00 | 4.00 | 9.00 |
| E | 6215 | 4.21 | 1.65 | 1.00 | 4.00 | 8.00 |
| S | 6215 | 4.56 | 1.72 | 1.00 | 4.00 | 9.00 |
| G | 6215 | 4.89 | 1.81 | 1.00 | 5.00 | 9.00 |
| GCP | 6215 | 8.35 | 3.12 | 2.17 | 7.98 | 15.62 |
| Restate | 6215 | 0.12 | 0.32 | 0.00 | 0.00 | 1.00 |
| Analyst | 6215 | 2.35 | 1.12 | 0.00 | 2.40 | 5.80 |
| Punish | 6215 | 0.08 | 0.27 | 0.00 | 0.00 | 1.00 |
| Size | 6215 | 23.15 | 1.26 | 20.58 | 23.01 | 26.89 |
| Lev | 6215 | 45.67 | 19.23 | 8.32 | 45.12 | 89.76 |
| Roe | 6215 | 8.23 | 6.15 | -12.35 | 8.01 | 25.67 |
| East | 6215 | 0.58 | 0.49 | 0.00 | 1.00 | 1.00 |
| SOE | 6215 | 0.38 | 0.49 | 0.00 | 0.00 | 1.00 |
5.2 Baseline regression results: Direct effects and dimensional heterogeneity of ESG performance
Table 2 reports the baseline regression results, focusing on the direct impact of ESG performance on financing costs and dimensional heterogeneity.
5.2.1 Direct Effect Test (H1)
Column (1) of Table 2 shows that the coefficient of ESG comprehensive score is -0.213, which is significant at the level of 1% (t=-6.89), indicating that for every one grade improvement of ESG rating, the cost of corporate debt financing will decrease by 0.213 percentage points on average, so Hypothesis H1 is verified. In terms of economic significance, if an enterprise’s ESG rating increases from the lowest level (level 1) to the highest level (level 9), the financing cost will decrease by 1.704 percentage points, accounting for 39.8% of the sample average (4.28%). This effect size is significantly higher than the average level of 0.5% of the global market in the MSCI (2022) report. This reflects that in China’s credit market, ESG signals have a more prominent impact on financing costs, which may be due to the relatively single financing channel of Chinese enterprises, and banks’ credit decisions are more sensitive to risk signals.
The coefficient sign of control variables is consistent with the expectation: the coefficient of company Size (Size) is − 0.326 (1% significant), indicating that large enterprises have lower financing cost due to their strong ability to resist risks; The asset-liability ratio (Lev) coefficient is 0.042 (1% significant), indicating that highly leveraged enterprises have higher credit risk and need to pay higher financing cost; The coefficient of ownership nature (SOE) is -0.387 (1% significant), which reflects that the financing cost of state-owned enterprises is significantly lower than that of non-state-owned enterprises due to the advantage of implicit guarantee; The coefficient of profitability (Roe) is − 0.038 (1% significant), indicating that enterprises with good profitability have strong solvency and low financing cost; The coefficient of Cash ratio (Cash) is − 0.015 (1% significant), indicating that enterprises with strong short-term solvency have lower financing cost; The coefficient of Age is − 0.012 (5% significant), indicating that mature firms have higher information transparency and lower financing costs.
5.2.2 Dimensional heterogeneity test (H1a, H1b)
Column (2) of Table 2 shows the regression results of different dimensions. The coefficient of Governance (G) dimension is − 0.286 (1% significant), the coefficient of Social (S) dimension is − 0.197 (1% significant), and the coefficient of Environmental (E) dimension is − 0.152 (1% significant). The results at the core of the reason is that thenvironmental (E) dimension of information can be quantified and credibility differences: governancenvironmental (E) dimension indicators (such as the proportion of independent directors, ownership concentration) are mandatory disclosure of information, and can be directly obtained from the annual report, the bank can rapid quantitative assessment; Although the disclosure rate of indicators in the social (S) dimension (such as employee salary and donation expenditure) increased to 68%, the lack of unified measurement standards (such as the statistical caliber of donation expenditure is not unified) reduced the comparability of information; The indicators of environmental (E) dimension (such as carbon emissions and environmental protection investment) have the lowest degree of standardization, and the disclosure rate of high-pollution industries is only 45%, and there is “greenwashing” behavior (12% of green projects were found to be fake by a big bank in 2024), which leads to low trust in the information of environmental (E) dimension.
Column (3) of Table 2 shows the sub-dimensional regression results of high-pollution industry samples. The coefficient of E-dimension rises to − 0.268 (1% significant), exceeding − 0.201 of S-dimension, so H1b is established. A change that the driving factors of difference: the regulatory environment polluting industries (e.g., steel, chemicals) directly by China’s “carbon peak” in 2030 and 2060, the influence of the “carbon neutrality” target, facing the strict replacement policy and environmental limit production capacity, environmental compliance is directly related to enterprise’s survival and business continuity. The data of 2024 show that the average amount of environmental penalties in high-pollution industries is 12.8 million yuan, 6.2 times that of low-carbon industries, and 18% of high-pollution enterprises are restricted due to non-compliance with environmental protection standards. This makes banks pay more attention to the environmental (E) dimension in credit decision-making, and the signal value of environmental information is magnified. In addition, the Governance (G) dimension coefficient rose to 0.312 in the high pollution industry, still keep the highest level, that the base function of governancenvironmental (E) dimension in high-risk industry is more important.
Table 2 Baseline regression results
| Variables | (1) Full sample (integrated ESG) | (2) Full sample (sub-dimension) | (3) High-pollution industries (sub-dimensions) |
| ESG | 0.213 * * * (6.89) | – | – |
| E | – | 0.152 * * * (4.03) | 0.268 * * * (6.15) |
| S | – | 0.197 * * * (5.68) | 0.201 * * * (5.02) |
| G | – | 0.286 * * * (7.21) | 0.312 * * * (7.56) |
| Size | 0.326 * * * (7.58) | 0.318 * * * (7.42) | 0.345 * * * (6.89) |
| Lev | 0.042 * * * (12.36) | 0.041 * * * (12.15) | 0.048 * * * (12.56) |
| Roe | 0.038 * * * (5.67) | 0.037 * * * (5.56) | 0.045 * * * (5.34) |
| Growth | 0.002 (1.23) | 0.002 (1.21) | 0.003 (1.35) |
| PPE | 0.008 * (1.92) | 0.008 * (1.90) | * * (2.15 0.012) |
| Cash | 0.015 * * * (4.21) | 0.014 * * * (4.18) | 0.016 * * * (4.32) |
| Age | * * (2.35-0.012) | * * (2.33-0.012) | * * (2.56-0.015) |
| SOE | 0.387 * * * (5.12) | 0.385 * * * (5.10) | 0.421 * * * (5.48) |
| HHI | 0.456 * (1.89) | 0.452 * (1.87) | 0.489 * (1.92) |
| Constant | 15.678 * * * (18.92) | 16.012 * * * (19.05) | 17.235 * * * (18.76) |
| Industry FE | is | is | is |
| Year FE | is | is | is |
| N | 6215 | 6215 | 2156 |
| R² | 0.365 | 0.382 | 0.438 |
| Note: *, and * indicate significance at the level of 1%, 5% and 10%, respectively, and the figures in parentheses are t-values; The same as below. |
5.3 Moderating effect test: the role of green credit policy
Table 3 reports the test results of the moderating effect of green credit policies, which are analyzed from three levels: interaction term regression, grouping regression and sub-dimensional moderating effect.
5.3.1 Overall Moderating Effect (H2)
Column (1) of Table 3 shows that the coefficient of the interaction term between ESG and GCP is − 0.025, which is significant at the level of 1% (t= − 4.32), indicating that for every 1 percentage point increase in the intensity of green credit policies, the effect of ESG on reducing financing costs will be enhanced by 0.025 percentage points. In terms of economic significance, in the regions with the mean value of GCP (8.35%), the effect of ESG on reducing financing costs is 20.88% (0.025×8.35/0.100) higher than that in the regions without policies; In Guangdong Province (15.62%) with the highest GCP, the effect of ESG on financing cost is 39.05% higher; However, in Gansu Province (2.17%) with the lowest GCP, the effect only increased by 5.43%. The difference in policy intensity leads to significant differences in the realization degree of ESG value.
The core of this moderating effect mechanism lies in the guidance of policies on banks’ behaviors: Banks in high GCP areas face stricter green credit performance assessment (for example, the weight of green credit in MPA assessment is 15%, while that in low GCP areas is only 5%), and the risk weight of green projects is reduced from 100% to 75%, which reduces the capital occupied cost of banks. China merchants bank, for example, in 2024, the 20% of green area in front of the GCP 1.2% lower loan interest rates than ordinary loans, while only 0.5% lower than in low GCP region, directly embodies the influence of strength of ESG pricing policy.
5.3.2 Grouping test and marginal effect
Table 3 columns (2) and column (3) according to the GCP quarterback method is divided into high GCP group (25%) and low GCP to group (25%) after return, high GCP group ESG coefficient was 0.326 (1%), low – 0.185 for GCP group (1%), Coefficient difference between groups by Chow test (F = 12.87, p < 0.01), further verify the existence of moderating effect. The results and NIFD (2025) study found that, but this article provides a breakdown of quantitative evidence.
The marginal effect graph (Figure 1, slightly) visually shows the marginal effect of ESG on COD at different GCP levels: when GCP is lower than 5%, the marginal effect of ESG is around − 0.15, and the confidence interval is close to zero; When the GCP is higher than 10%, the marginal effect drops below -0.30 and the confidence interval is far from zero, indicating that the financing value of ESG can be fully released only when the policy intensity reaches a certain threshold. A threshold effect can provide reference for policy makers, namely the green credit scale must be accounted for up to 10% will exert effective policy effectiveness.
5.3.3 Sub-dimensional moderating effect
Table 3 columns (4) and columns (5) test every dimension of ESG moderating effect on the policy differences. Columns (4), G * GCP by coefficient of 0.032 (1%), S * GCP is 0.024 (1%), E * GCP is 0.018 (5%), moderating effect sorting for G > S > E, consistent with the results of dimension of heterogeneity. This is because the credibility of the highest governancenvironmental (E) dimension information, policy through enhancing bank review of the indicators (such as required to provide internal control audit report), further amplify the signal value.
Columns (5) Environmental (E) dimension moderating effect test for high pollution industry, E * GCP pay by coefficient rose to 0.028 (1%), significantly higher than that of the whole sample – 0.018, shows that in the high pollution industry, policy moderating effect on the environmental (E) dimension significantly enhanced. Of the reasons is that of environmental risk and policy risk of high pollution industry highly binding, green credit policy by credit approval directly link with the environmental punishment record, make environment dimension information bank decision-making “hard constraints”, policy on the environment signal amplification effect is more outstanding. In 2024, according to the high pollution industry, ESG rating level 1 and the enterprise at a high GCP region, a 0.52% drop in the cost of financing, is 2.3 times of low GCP area.
Table 3 Test results of moderating effect
| Variables | (1) Full sample (cross multiplier term) | (2) High GCP group | (3) Low GCP group | (4) Full sample (sub-dimensional moderating effect) | (5) High-pollution industries (Environmental (E) dimension moderating effect) |
| ESG | 0.100 * * * (3.21) | 0.326 * * * (7.12) | 0.185 * * * (4.01) | – | – |
| GCP | 0.056 * (1.89) | – | – | * * (2.01-0.062) | * * (2.15-0.072) |
| ESG×GCP | 0.025 * * * (4.32) | – | – | – | – |
| E | – | – | – | 0.145 * * * (3.98) | 0.256 * * * (6.01) |
| E×GCP | – | – | – | * * (2.35-0.018) | 0.028 * * * (3.89) |
| S | – | – | – | 0.187 * * * (5.56) | – |
| S×GCP | – | – | – | 0.024 * * * (3.67) | – |
| G | – | – | – | 0.278 * * * (7.01) | – |
| G×GCP | – | – | – | 0.032 * * * (4.89) | – |
| Control variables | is | is | is | is | is |
| Constant | 16.235 * * * (19.21) | 17.896 * * * (15.67) | 14.567 * * * (14.32) | 16.547 * * * (19.12) | 18.235 * * * (17.89) |
| Industry FE | is | is | is | is | is |
| Year FE | is | is | is | is | is |
| N | 6215 | 3108 | 3107 | 6215 | 2156 |
| R² | 0.398 | 0.421 | 0.367 | 0.415 | 0.452 |
| Chow test | – | F = 12.87 * * * | – | – | – |
5.4 Heterogeneity analysis
Based on the threenvironmental (E) dimensions of property right nature, industrial pollution degree and regional development level, the heterogeneity analysis was conducted to test the boundary conditions of ESG effect and policy moderating effect. The results are shown in Table 4.
5.4.1 Heterogeneity of property rights (H2a)
Columns (1) and (2) of Table 4 show that the ESG coefficient of non-state-owned enterprises is − 0.287 (1% significant), and the coefficient of interaction term is − 0.032 (1% significant); For soes, the coefficient of ESG is -0.145 (1% significant) and the coefficient of interaction term is -0.018 (5% significant). The absolute values of the two coefficients of non-state-owned enterprises are significantly greater than those of state-owned enterprises, so H2a is established.
The core reason for this difference lies in the heterogeneity of financing constraints: due to their close ties with the government, Chinese soes have the advantage of implicit guarantee, and the credit approval rate is 28% higher than that of non-soes. However, non-soes face more serious information asymmetry and financing discrimination. ESG, as a “high-quality signal” recognized by the third party, can effectively alleviate the risk concerns of banks, and the marginal value is higher. In addition, the green credit policy further amplifies the signal value of ESG by implementing a more favorable risk weight for green projects of non-soes (reduced to 70%, 5 percentage points lower than that of soes). Data from Industrial Bank in 2024 showed that its credit granting to non-soes with high ESG reached 35 percent growth, 1.6 times that of soes, corroborating this mechanism.
5.4.2 Heterogeneity of pollution degree by industry (H2b)
Columns (3) and (4) of Table 4 show that the ESG coefficient of high-pollution industries is -0.356 (1% significant), and the coefficient of interaction term is -0.038 (1% significant); Low carbon industry ESG coefficient was 0.178 (1%), pay by coefficient of 0.019 (5%) significantly. The absolute values of the two coefficients of the high-pollution industry are significantly greater than those of the low-carbon industry, so H2b is established.
This result is due to the difference in industrial regulatory pressure: high-polluting industries are the key targets of China’s 2030 “carbon peak” and 2060 “carbon neutrality” goals, facing strict environmental protection standards and capacity constraints. In 2024, 18% of enterprises will be restricted due to non-compliance with environmental protection standards, while only 2% of low-carbon industries will be restricted. ESG performance is directly related to the high pollution permits enterprises existence and policy support, the bank has more attention to its ESG performance is higher than the low carbon industry. At the same time, the “reverse constraint” of green credit policy on high-polluting industries is stronger, and the credit growth rate of “two high” and “two high” enterprises is limited to less than 10%, but there is no clear limit on low-carbon industries, which makes the financing return of high-polluting industries to improve ESG performance higher. In the steel industry, for example, ESG rating level 1 enterprise financing costs fell by 0.32%, 2.1 times that of the software industry.
5.4.3 Heterogeneity of regional development levels
Columns (5) and (6) of Table 4 show that the ESG coefficient in the eastern region is − 0.268 (1% significant), and the coefficient of the interaction term is − 0.029 (1% significant); The ESG coefficient of the central and western regions is -0.189 (1% significant), and the coefficient of the interaction term is -0.017 (5% significant). The eastern part of the two coefficient absolute value is greater than the central and western regions, reflects the significant regional differences.
The main driving factor of this difference is the difference between financial infrastructure and information environment: The disclosure rate of ESG information in the eastern region is 82%, significantly higher than 56% in the central and western regions. In addition, leading banks such as China Merchants Bank and Industrial and Commercial Bank of China have built multi-source ESG databases in the eastern region, which can effectively integrate unstructured data such as environmental penalties and labor disputes, and the degree of information asymmetry is low. While the Midwest bank data integration ability is weak, identification and evaluation of ESG information ability is insufficient, lead to ESG signal transmission efficiency is low. In addition, the eastern part of the green credit assessment weighting average (12%) is higher than the Midwest (average 8%), bank of policy execution power stronger, further strengthening the financing effect of ESG.
Table 4 Results of heterogeneity analysis
| variable | (1) Non-state-owned enterprises | (2) State-owned enterprises | (3) Highly polluting industries | (4) low-carbon industries | (5) The east | (6) Midwest |
| ESG | 0.287 * * * (7.56) | 0.145 * * * (3.21) | 0.356 * * * (8.12) | 0.178 * * * (4.35) | 0.268 * * * (7.35) | 0.189 * * * (5.01) |
| GCP | 0.068 * (1.92) | 0.045 (1.56) | * * (2.15-0.072) | 0.051 * (1.89) | * * (2.05-0.069) | 0.048 * (1.78) |
| ESG×GCP | 0.032 * * * (4.89) | * * (2.15-0.018) | 0.038 * * * (5.23) | * * (2.45-0.019) | 0.029 * * * (4.67) | * * (2.32-0.017) |
| Size | 0.387 * * * (7.12) | 0.296 * * * (5.89) | 0.345 * * * (6.78) | 0.321 * * * (6.12) | 0.356 * * * (7.01) | 0.302 * * * (6.23) |
| Lev | 0.045 * * * (11.89) | 0.038 * * * (9.67) | 0.048 * * * (12.56) | 0.039 * * * (10.23) | 0.043 * * * (12.01) | 0.040 * * * (10.56) |
| Roe | 0.042 * * * (5.12) | 0.031 * * * (4.01) | 0.045 * * * (5.34) | 0.033 * * * (4.21) | 0.040 * * * (5.23) | 0.035 * * * (4.56) |
| Control variables | is | is | is | is | is | is |
| Constant | 17.896 * * * (18.32) | 14.567 * * * (15.12) | 18.235 * * * (17.89) | 15.678 * * * (16.45) | 16.897 * * * (18.01) | 15.234 * * * (16.78) |
| Industry FE | is | is | is | is | is | is |
| Year FE | is | is | is | is | is | is |
| N | 3855 | 2360 | 2156 | 4059 | 3605 | 2610 |
| R² | 0.412 | 0.356 | 0.438 | 0.372 | 0.405 | 0.368 |
5.5 Mechanism test
Stepwise regression method and Bootstrap method were used to test the mediating effect of financial discipline, investor confidence and environmental compliance cost, as well as the strengthening effect of green credit policy on the mediating path.
5.5.1 financially disciplined path
Table 5 column (1), according to the ESG coefficient was 0.042 (1%), shows that whenever the ESG up one class, the probability of financial restatement in companies fell by 4.2%, to verify the high ESG performance function to the promotion of enterprise internal control quality. This is because the high ESG enterprises pay more attention to long-term value, tend to establish the perfect internal control processes (e.g., set up independent of the internal control auditing department), China, according to data from the card before the ESG 20% enterprise internal control defect rate is only 5.8%, far below the 20% 23.6% of the enterprise.
Column (2) of Table 5 shows that after the inclusion of the mediating variable Restate, the coefficient of ESG decreases from − 0.213 to − 0.185 (1% significant), and the coefficient of Restate is 0.326 (1% significant), indicating that financial discipline plays a partial mediating role between ESG and financing cost. Intermediary effect accounted for 12.68%. Are the result of logic is that financial restatements will enterprise financial opaque signals to the market, increase the risk of creditors, increase the cost of financing; High ESG firms reduce financial restatements by improving financial discipline, thus reducing financing costs.
Listed in table 5 (7), according to the introduction of pay by ESG x after GCP, pay by coefficient is 0.021 (1%), Restate coefficient is 0.287, the mediation effect of ratio rose to 16.48%, shows that the green credit policy significantly strengthen financial discipline. By requiring banks to review the “ESG internal control process” (such as verifying the authenticity of internal control audit reports) in credit approval, the policy makes it easier to identify the financial discipline advantages of high ESG enterprises, and the mediating effect increases by 30%.
5.5.2 investor confidence
Column (3) of Table 5 shows that the ESG coefficient is 0.187 (1% significant), indicating that for every one grade improvement of ESG, the number of following analysts increases by 0.187 units (corresponding to an increase of about 2 actual analysts), which verifies the role of high ESG performance in attracting market attention. This is because the ESG performance good enterprise risk lower and higher information transparency, are more likely to get the attention of analysts, the United Nations global compact organization, according to the top 20% of ESG rating companies track the number of analysts is more than 20% after enterprise 6.3.
Column (4), according to table 5 into intermediary variable after Analyst, coefficient of ESG fell to 0.168 (1%), the Analyst coefficient was 0.256 (1%), shows that investor confidence, give play to the role of partial mediation, the mediation effect accounted for 21.13%, It is the largest contribution among the three paths. The core of this mechanism is that analyst attention can reduce information asymmetry by issuing research reports, enhance the market recognition of enterprises, and thus reduce the risk premium demanded by creditors. The MSCI (2022) study confirms that for every additional analyst following, the financing cost of a company decreases by 0.08 percentage point on average.
Column (8) of Table 5 shows that after the introduction of the interaction term, the coefficient of the interaction term is − 0.026 (1% significant), the Analyst coefficient drops to − 0.201, and the proportion of mediating effect rises to 29.98%, indicating that the policy strengthens the path of investor confidence. By promoting the standardization of ESG information (such as requiring disclosure according to IFRS S2 standards), the policy has enhanced analysts’ trust in ESG data.
5.5.3 Path of environmental compliance costs
Column (5) of Table 5 shows that the ESG coefficient is − 0.035 (1% significant), indicating that the probability of an enterprise being subject to environmental penalties decreases by 3.5 percentage points for every grade improvement in ESG, which verifies the effect of high ESG performance on reducing environmental risks. This is because enterprises with high ESG performance have sufficient investment in environmental protection. Data in 2024 show that the investment in environmental protection of the top 20% of ESG enterprises accounts for 4.2% of their revenue, 3.5 times that of the bottom 20% of enterprises, and their environmental compliance is significantly higher.
Column (6) of Table 5 shows that after the mediating variable Punish is included, the ESG coefficient drops to − 0.192 (1% significant), and the Punish coefficient is 0.287 (1% significant), indicating that the environmental compliance cost plays a partial mediating role, accounting for 9.87% of the mediating effect. The logic of this mechanism is: environmental penalties can reduce the enterprise cash flow (penalty) and business interruption (restricting output production), increase the risk of default, which would push up the cost of finance; High ESG enterprises reduce environmental compliance costs by reducing the probability of environmental penalties, thus reducing financing costs.
Column (9) of Table 5 shows that in the sample of high-pollution industries, after introducing the interaction term, the coefficient of the interaction term is − 0.024 (1% significant), the coefficient of Punish rises to 0.302, and the proportion of mediating effect rises to 15.60%, indicating that the policy has a more significant strengthening effect on this path in high-pollution industries. By incorporating environmental punishment information into the credit investigation system, the policy enables banks to quickly check the environmental compliance of enterprises.
Bootstrap test results show that the 95% confidence intervals of the three intermediary paths do not contain 0 (financial discipline: [0.018,0.032]; investor confidence: [0.035,0.058]; environmental compliance cost: [0.012,0.025]), which further verifies the robustness of the mediating effect.
Table 5 Mechanism test results
| Variables | (1) Restate | (2) COD | (3) Analyst | (4) COD | (5) Punish | (6) COD | (7) Restate intermediary after moderating effect | (8) Analyst mediation after moderating effect | (9) High pollution Punish intermediaries |
| ESG | 0.042 * * * (5.34) | 0.185 * * * (5.89) | 0.187 * * * (7.65) | 0.168 * * * (5.12) | 0.035 * * * (4.89) | 0.192 * * * (6.01) | 0.164 * * * (5.56) | 0.118 * * * (4.89) | 0.162 * * * (5.87) |
| Restate | – | 0.326 * * * (4.21) | – | – | – | – | 0.318 * * * (4.15) | – | – |
| Analyst | – | – | – | 0.256 * * * (6.34) | – | – | – | 0.248 * * * (6.21) | – |
| Punish | – | – | – | – | – | 0.287 * * * (3.98) | – | – | 0.302 * * * (4.05) |
| ESG×GCP | – | – | – | – | – | – | 0.021 * * * (3.89) | 0.026 * * * (4.32) | 0.024 * * * (4.15) |
| Control variables | is | is | is | is | is | is | is | is | is |
| Constant | 0.896 * * * (12.35) | 16.012 * * * (18.76) | 3.215 * * * (25.67) | 17.235 * * * (19.01) | 0.789 * * * (11.56) | 16.547 * * * (18.92) | 16.235 * * * (18.89) | 17.567 * * * (19.21) | 17.896 * * * (18.56) |
| Industry FE | is | is | is | is | is | is | is | is | is |
| Year FE | is | is | is | is | is | is | is | is | is |
| N | 6215 | 6215 | 6215 | 6215 | 6215 | 6215 | 6215 | 6215 | 2156 |
| R² | 0.287 | 0.382 | 0.415 | 0.402 | 0.268 | 0.375 | 0.401 | 0.428 | 0.445 |
| Proportion of mediating effect | – | 12.68% | – | 21.13% | – | 9.87% | 16.48% | 29.98% | 15.60% |
5.6 Robustness checks
In order to ensure the reliability of the conclusions, this paper conducts six robustness checkss, and the results all support the core conclusions:
(1) Replacing the core explanatory variable: the Wind ESG rating (ESG2) is used to replace the China Securities rating, and the coefficient of ESG2 is − 0.198 (t= − 6.54), and the interaction term is − 0.023 (t= − 4.01), which is significantly negative (Column 1 of Table 6).
(2) The explained variable is replaced by COD2 (financial expenses/total liabilities), the ESG coefficient is − 0.187 (t= − 6.12), and the interaction term is − 0.021 (t= − 3.89), which is significantly negative (Column 2 of Table 6).
(3) Changing the sample interval: excluding 2020 (the impact of the epidemic), the ESG coefficient is − 0.205 (t= − 6.35), and the interaction term is − 0.024 (t= − 4.12), which is significantly negative (Column 3 of Table 6).
(4) Endogeneity treatment: The one-period-lagged ESG (ESG_lag1) was used as the instrumental variable, and the ESG coefficient in the second stage of 2SLS was -0.256 (t=-5.89), the interaction term was -0.029 (t=-3.98), and the F in the first stage was 103.87 (through the weak instrument test) (Column 4 of Table 6).
(5) Excluding policy confounding effect: controlling carbon trading market policy (ETS), the ESG coefficient is -0.211 (t=-6.85), and the interaction term is -0.024 (t=-4.28), which is significantly negative (Column 5 of Table 6).
(6) The sensitivity of winsorization: with 5% wind-down, the ESG coefficient is -0.215 (t=-6.92), and the interaction term is -0.026 (t=-4.35), which is significantly negative (Column 6 of Table 6).
Table 6 Results of the robustness checks
| Variables | (1) Replace ESG | (2) Replace COD | (3) Remove 2020 | (4) 2SLS Stage 2 | (5) Control the ETS | (6) 5% windup |
| ESG/ESG2 | 0.198 * * * (6.54) | 0.187 * * * (6.12) | 0.205 * * * (6.35) | 0.256 * * * (5.89) | 0.211 * * * (6.85) | 0.215 * * * (6.92) |
| GCP | 0.058 * (1.89) | 0.052 * (1.78) | * * (2.01-0.061) | * * (2.12-0.065) | 0.057 * (1.90) | 0.059 * (1.92) |
| ESG×GCP | 0.023 * * * (4.01) | 0.021 * * * (3.89) | 0.024 * * * (4.12) | 0.029 * * * (3.98) | 0.024 * * * (4.28) | 0.026 * * * (4.35) |
| ETS | – | – | – | – | * * (2.35-0.042) | – |
| Control variables | is | is | is | is | is | is |
| Constant | 16.547 * * * (19.01) | 15.896 * * * (18.32) | 16.123 * * * (18.56) | 17.345 * * * (17.89) | 16.345 * * * (18.96) | 16.012 * * * (18.78) |
| Industry FE | is | is | is | is | is | is |
| Year FE | is | is | is | is | is | is |
| N | 6215 | 6215 | 4972 | 6215 | 6215 | 6215 |
| R² | 0.392 | 0.387 | 0.391 | – | 0.399 | 0.378 |
| First stage F-value | – | – | – | 103.87 * * * | – | – |
6. Conclusions and policy recommendations
6.1 Research Conclusions
Taking A-share listed companies from 2018 to 2022 as samples and combined with the latest green credit data in 2024, this paper empirically tests the impact of ESG performance on financing costs and the moderating effect of green credit policies, and draws the following core conclusions:
(1) ESG performance significantly reduces financing cost, and there isocial (S) dimensional heterogeneity: for every one level improvement of ESG rating, financing costs decrease by an average of 0.213 percentage points, and thenvironmental (E) dimensional effect is ranked as “governance (-0.286) > society (-0.197) > environment (-0.152)”. However, in high-pollution industries, the effect of environmental (E) dimension rises to -0.268, exceeding the effect of social (S) dimension, which confirms that “industry sensitivity determines the value of dimension”.
(2) the green credit policy play a positive regulatory role: policy strength each 1% increase, ESG for effects of lower funding costs increase 0.025%, 25% in front of the intensity of policy areas, the effect of a 76.2% increase to 25% after. The moderating effect of ESG in each dimension is “governance > society > environment”, but in high-pollution industries, the moderating effect of environmental (E) dimension is significantly enhanced.
(3) Prominent heterogeneous effects: The effect of ESG and policy moderating effect is more significant in non-state-owned enterprises (ESG coefficient -0.287 vs -0.145 in state-owned enterprises), high-pollution industries (-0.356 vs -0.178 in low-carbon industries) and eastern regions (-0.268 vs -0.189 in central and western China). This reflects the precise support of policy for vulnerable financing subjects and high-risk industries.
(4) The three intermediary paths are clearly distinguishable: Financial discipline (intermediary accounting for 12.68%), investor confidence (21.13%) and environmental compliance cost (9.87%) are the key paths for ESG to reduce financing costs. By strengthening information verification and data sharing, green credit policy increases the effects of the three paths by 30%, 42% and 58%, respectively.
6.2 Policy Recommendations
6.2.1 Enterprise level
(1) Multi-dimensional priority dynamic management
Enterprises in emerging markets should adopt the strategy of “strengthening basic dimensions + breaking through sensitivenvironmental (E) dimensions” : Non-state-owned enterprises and high-pollution industries need to be dual-track parallel, and give priority to optimizing governancenvironmental (E) dimensions (such as introducing independent supervisor system and establishing digital internal control platform) — referring to Microsoft’s “ESG-internal control linkage mechanism”, deeply binding governance indicators to financial processes, which can reduce the risk of financial restatement by 40%; At the same time, to strengthen the input of environmental (E) dimension, high-pollution enterprises can learn from Shell’s “carbon footprint life cycle management” model and incorporate carbon emission intensity into departmental KPIs, which can reduce Shell’s financing cost in the Asian market by 0.3 percentage points (Shell Sustainability Report, 2024). Low-carbon industries can focus on the construction of social (S) dimension. For example, referring to Nestle’s “Supply Chain Social Responsibility Assessment System”, suppliers’ labor rights can be included in ESG management to improve the recognition of financing market.
(2) Globalization and precision of information disclosure
Strictly follow IFRS S1/S2 standards to build a disclosure framework, Focus on strengthening governancenvironmental (E) dimension of the board of directors of the “diversified ratio” “internal control defects ZhengGaiLv” and environmental (E) dimensions of 1-3 carbon emissions “Scope” of “environmental protection investment returns” quantifiable indexes such as disclosure, tesla (China) by standardizing disclosure of carbon emissions data, 2024 green bonds issuing interest rates below the industry average of 0.25%. High-polluting enterprises can join international green project certification systems (such as the CBI certification of the Climate Bond Initiative) to reduce the suspicion of “greenwashing”, and the cross-border financing cost of a photovoltaic enterprise in China decreased by 0.32 percentage points after it passed the CBI certification.
(3) Systematic reinforcement of the mediation path
Establishing “ESG-financial linkage internal control system”, introducing Deloitte’s “carbon accounting integration model” and incorporating environmental costs into financial accounting can increase the path effect of financial discipline by 25%. Cooperation with international securities held special ESG roadshow, consult apple “investors communication mechanism of sustainable development”, published on a regular basis ESG performance white paper, attract attention, an analyst – apple through this mechanism makes the analyst coverage of analysts increased by 30%, a 0.18% drop in the cost of financing; Set up a special fund, draw lessons from BP “environmental risk reserve system”, according to the revenue provision for compliance, 1.5% to 2% can make environmental punishment probability is reduced 60%.
6.2.2 Financial institution level
(1) Build a multi-dimensional differentiated pricing model
Based on “ESG through review framework” deutsche bank and HSBC bank “industry – dimension double weight model”, set up dynamic pricing system: weight set as 40% of the industry will governancenvironmental (E) dimension, environmental, social, 35% and 25%; High pollution industry add “environmental compliance coefficient of premium” alone, with environmental protection in recent 3 years punishment record corporate interest rate plus floating 10% 15%; For non-state-owned enterprises, “ESG improvement incentive clauses” will be set up, and the interest rate will be lowered by 5 basis points if their rating is upgraded by one level. Through this model, Barclays Bank increased the credit scale of ESG compliant enterprises by 22% and reduced the non-performing ratio by 0.8 percentage points (Barclays ESG Finance Report, 2024).
(2) Strengthen ESG credit tool innovation
Expand the scale of ESG-linked loans and sustainability-linked bonds (SLBS), Design a “base rate + + ESG a floating interest rate policy subsidy” the combination of product, the construction bank launched in 2024, the loan “link” carbon neutrality enterprise ESG rating after interest rates down by 15 to 30 basis points, the model has been in southeast Asia five Banks. Issue “ESGovernance (G) dimension hierarchical green bonds”, priority support for 30% of the project by thenvironmental (E) dimension of governance and environmental performance, societe generale such ordinary green bonds in the bond interest rate is 0.2%.
(3) promote ESG data integration ability
With international third party agencies (such as the MSCI index), China to build multi-source databases, integrate environmental penalties, labor disputes, the unstructured data such as carbon footprint – China merchants bank in collaboration with the MSCI, ESG data verification efficiency by 40%, for examination and approval of the credit cycle shortened by 25%. Promoting the establishment of a regional ESG data sharing platform, referring to the Sustainable Finance Information Exchange (SFIE) model of the European Union, realizing the data exchange among banks, enterprises and regulators. Financial institutions in the eastern region can take the lead in jointly building a cross-border ESG data alliance with countries along the Belt and Road to alleviate information asymmetry in emerging markets.
6.2.3 At the government level
(1) Optimize the green credit assessment mechanism
Incorporate ESG disclosure integrity macro-prudential assessment (MPA), weight up to 15% 20%, support for appraisal outstanding financial institutions to further credits, interest rates fell to 1.75% – 2.0%. Joint punishment mechanism “to establish” green credit fraud, with reference to the eu regulations on the sustainable financial disclosure (SFDR), the data fraud companies implement credit block, limited financing, such as punishment, China in 2024, the mechanism to make green credit fraud rate from 12% to 5%. P3romote the establishment of global unified standard of green credit project, docking the eu classification schemes of sustainable finance and ISSB guidelines, reduce the standard difference of cross-border financing.
(2) Promote the localization and international integration of ESG standards
Revision of green industry directory add financial “transformation” category, allowing high pollution industry projects into green credit support low carbon transformation, with reference to Singapore “progressive ESG standards”, set a deadline for differentiation standard for different industry. In governancenvironmental (E) dimension to “internal control” digital “ESG cross-border supply chain management” and other indicators, to meet the needs of enterprise globalization; The ISO 14064 carbon accounting standard is adopted in the environmental (E) dimension to improve the international comparability of data.
（3）Strengthen policy coordination and international cooperation
Establish a linkage mechanism between green credit and carbon trading market, and lower the credit interest rate by 5-10 basis points for enterprises that are included in the carbon market and meet ESG standards. Through this linkage, the EU has increased the carbon emission reduction of enterprises by 22% and reduced the financing cost by 0.25 percentage points (EU ETS Report, 2024). Financial discount (discount rate of 2%-3%) will be given to non-soes for ESG improvement projects to ease financing discrimination. Strengthen international policy coordination, to participate in “sustainable” G20 financial working group, promote mutual recognition of green credit policy, emerging markets in China with the brics countries to build “ESG financing policy coordination mechanism”, improve voice to emerging markets.
References
[1] Hu, J. & CAI, W. X. The impact of banking competition on corporate ESG responsibility performance: “More is better” or “enough is enough”? [C]// Emerging Economies Society, China Center for International Cultural Exchange, Guangdong University of Technology. New Quality Productivity and the Development of Emerging Economies (Part II). School of Economics, Guizhou University of Finance and Economics; , 2025:140-165. DOI: 10.26914 / Arthur c. nkihy. 2025.040772.
[2] Liu L H, Li B, Huang Q, et al. China’s OFDI corporate ESG performance research on the influence of green innovation research [C] / / emerging economies, China international cultural exchange center, guangdong university of technology. New Quality Productivity and the Development of emerging Economies (Part II). School of Economics and Trade, Foshan University; , 2025:99-119. DOI: 10.26914 / Arthur c. nkihy. 2025.040770.
[3] [Lu Zhengwei, QIAN Lihua. Green finance practice under the guidance of “Two Mountains” concept.] China’s financial, 2025, (16) : 71-73. The DOI: CNKI: SUN: ZGJR. 0.2025-16-027.
[4] when yao, financial statistics, ma teng jiang editor, luquan yearbook, yunnan people’s publishing house, 2024, 135, the yearbook. DOI: 10.40663 / y.c nki. Yluqu. 2025.000483.
[5] LIU Y T. Study on the guiding effect of ESG rating difference on commercial banks’ green credit pricing [N]. Jiangsu economy, 2025-05-23 (T07). DOI: 10.28469 / n.c nki. NJSJJ. 2025.000533.
[6] Zhang, M., Lu, X. & Zhang, J. [6] [Zhang M, Lu F, Zhang J X. How to support the financing of small and medium-sized enterprises through fiscal and financial cooperation: An investigation based on a border province in western China [J]. China Economist, 2025, (03): 108-117. DOI:10.16158/j.cnki.51-1312/f.2025.03.001.
[7] Li Tingting, DuAN Zhaokun, Wang Kai, et al. Interpretation of new rules on ESG climate-related information disclosure of Hong Kong listed companies [J]. Journal of Finance and Accounting Monthly, 2025, 46 (05): 62-67. DOI:10.19641/j.cnki.42-1290/f.2025.05.009.
[8] Xu, L. (Ed.), Economic Management, Xu, L. (Ed.), Southern Yearbook, Xinhua Press, 2024, 155-176, Yearbook.
[9]Zhe L . A Research on the Impact of Green Credit Policy on the Fulfillment of Corporate Social Responsibility[C]// Wuhan University of Technology (China), Chengdu University of Information Technology, Proceedings of the 21st International Conference on Innovation and Management. School of Management,Wuhan University of Technology; , 2024:651-657. DOI: 10.26914 / Arthur c. nkihy. 2024.088483.
[10]Xu B . The Impact of Financial Technology on Corporate Financing and Corresponding Countermeasures:Evidence from China[C]// Comenius University in Bratislava. Proceedings of the 8th International Conference on Economic Management and Green Development (part5). Shandong University of Finance and Economics; , 2024:507-515. DOI: 10.26914 / Arthur c. nkihy. 2024.083192.
[11] Zhou, C. Han. [11] [ZHOU C H. The impact of green credit policy on debt financing cost of listed clean energy enterprises. Northeast university of finance and economics, 2024. DOI: 10.27006 /, dc nki. Gdbcu. 2024.001867.
[12] BAO Yimin. Green credit policy impact on corporate debt financing cost research [D]. Donghua university, 2024. The DOI: 10.27012 /, dc nki. Gdhuu. 2024.000517.
[13]Weng Y . Research Advanced on the Influence of Green Credit Policy on Companies:A Case Study of Industrial Structure,Firm Performance,and Response Measures[C]// Cardiff Business School,Brunel University London,University of Aberdeen Business School,Danubius University of Gala? i. Proceedings of the 2nd International Conference on Financial Technology and Business Analysis (part1) Business School Suzhou, Xi’an Jiaotong Liverpool University; , 2023:458-477. DOI: 10.26914 / Arthur c. nkihy. 2023.126422.
[14] Chen J. Analysis of international financial reporting sustainable disclosure standards [J]. Financial Supervision, 2023, (18): 87-92. DOI:CNKI:SUN:CZJD.0.2023-18-017.
[15] ZHOU J, XU Z X. [15] [ZHOU J, Xu Z X. The relationship between corporate debt financing cost and environmental, social responsibility and corporate governance performance [J]. Science and Technology Entrepreneurship Monthly, 2023, 36 (08): 136-140. (in Chinese) DOI:CNKI:SUN:KJCK.0.2023-08-024.
[16] Li Qiang, Chen Shan-man. Green credit policy, financing costs, and green technology innovation [J]. Journal of economic issues, 2023, (8) : 67-73 + 120. DOI: 10.16011 / j.carol carroll nki JJWT. 2023.08.008.
[17] ZHANG Wen-rui, MA Xiao-qiang, BAI Xiong. [17] ESG performance, Financing Constraints and Enterprise Credit Risk [C]// China Insurance and Risk Management Research Center, School of Economics and Management, Tsinghua University, Lingnan College, Sun Yat-sen University. Proceedings of 2023 China Insurance and Risk Management International Conference. School of Finance, Xi ‘an Eurasia University; School of Economics and Management, Northwest University; , 2023:1268-1283. DOI: 10.26914 / Arthur c. nkihy. 2023.099898.
[18] LAN Dan Yi. [18] The impact of green credit on maturity structure and cost of corporate debt financing [D]. Shanghai university of finance and economics, 2023. DOI: 10.27296 /, dc nki. Gshcu. 2023.000969.
[19] ZHOU H. The impact of central bank collateral expansion on the financing cost of green credit enterprises. Shanghai international studies university, 2023. DOI: 10.27316 /, dc nki. Gswyu. 2023.000815.
[20] Zhong Xian-bing, MA Fu-chun, Wei Yan, et al. How does green credit policy affect the cost of corporate debt financing? Finance and Economics, 2022, (09): 86-96. (in Chinese) DOI:10.19622/j.cnki.cn36-1005/f.2022.09.009.
[21] YU Jiewen, HAN Tingcun, HU Mudan, et al. Notes on ESG information disclosure and authentication [J]. China Investment (Chinese and English), 2022, (Z6): 43-44. DOI:CNKI:SUN:ZGTZ.0.2022-Z6-008.
[22] Chen, Y. Green credit policy impact on heavy pollution enterprise debt financing costs [D]. Sichuan university, 2022. The DOI: 10.27342 /, dc nki. Gscdu. 2022.000015.
[23] LI W C, CHAI S L, ZHANG H Y, et al.Impact of environmental information disclosure on corporate new bank loans and debt financing costs under green credit policy [J]. Financial Theory and Practice, 2022, (04): 19-28.
[24] Gao, J., Chu, D., Lian, Y. et al. Can ESG performance improve corporate investment efficiency? Securities Market Herald, 2021, (11): 24-34+72.
[25] Wu H. Research on the relationship between corporate environmental protection investment and debt financing cost under the background of green credit [J]. Financial Management Research, 2020, (10): 105-109. (in Chinese) DOI:CNKI:SUN:CWGL.0.2020-10-020.
[26] Wang, B. Hui. [26] [WANG B H. Green credit, corporate social responsibility disclosure and debt financing cost: an empirical study of heavy polluting enterprises listed in A-share market from 2011 to 2017. Journal of Financial Theory and Practice, 2019, (07): 47-54.
[27] Shen, H. (2019). WTO Economic Guide, 2012, (05): 68-69. DOI:CNKI:SUN:WTOK.0.2012-05-024.]