The development of the new energy industry is vital for achieving China’s “dual carbon” goals; however, financing constraints resulting from the significant disparity in bargaining power between banks and new energy enterprises (NEEs) remain a key obstacle. While existing research acknowledges that financial support policies can reduce financing costs, it has yet to adequately examine the complex interactions among different policy instruments, thereby limiting the development of effective policy portfolios. This study aims to address the issue of high financing costs faced by NEEs due to unequal bargaining power by quantitatively assessing the power imbalance between banks and NEEs and systematically evaluating both the individual and synergistic effects of green credit policies (GCPs) and government subsidies to inform optimized policy design. Using A-share listed NEEs in China as the sample, a two-tier stochastic frontier analysis (SFA) model is employed to measure the bargaining power and transaction surplus between NEEs and banks. Statistical and regression analyses are then conducted to investigate the individual and synergistic effects of GCPs and government subsidies, distinguishing between fiscal subsidies and research and development (R&D) subsidies to assess their respective impacts. The results reveal that NEEs possess substantially weaker bargaining power relative to banks; GCPs effectively weaken banks’ bargaining power with a stronger impact on banks holding greater initial power, thereby indirectly enhancing NEEs’ bargaining power. Fiscal subsidies do not significantly enhance NEEs’ bargaining power and show no synergy with GCPs. In contrast, R&D subsidies improve the bargaining power of NEEs, particularly for those with initially weaker positions, and exhibit a synergistic effect with GCPs. These findings provide practical insights for constructing a more effective policy framework to support the development of NEEs.

1. Introduction

Achieving the “peak carbon” and “carbon neutrality” goals is central to China’s strategy for transforming its energy structure, with the development of the new energy industry representing a critical pathway toward a clean and low-carbon economy [1–3]. Currently, new energy enterprises (NEEs) face multiple challenges including strong positive externalities, output uncertainties, extended investment return cycles, and elevated risk levels. These factors contribute to persistently high financing costs, which impede progress in low-carbon economic development [4]. Although China has made some progress in reducing financing costs for the new energy sector—for instance, since 2015, the average financing costs of photovoltaic, onshore wind, and offshore wind projects have decreased by 20%, 15%, and 33%, respectively, the current CoC (cost of capital) for renewable energy in China (3.9%, 3.0%, and 2.8%) remains higher than Germany’s lowest CoC benchmarks (1.4%, 1.1%, and 2.4%). This disparity indicates that policy support for China’s new energy sector remains insufficient, underscoring the urgency of optimizing policy portfolios to reduce financing costs amid an accelerated global energy transition and post-pandemic economic recovery [3, 5].

As an emerging economy, China’s capital market remains underdeveloped, with banks serving as the primary source of financing. Bank loans valued for their speed, flexibility, and interest deductibility are the preferred financing channel for NEEs [6, 7]. Within China’s bank-dominated financial system, corporate financing costs largely depend on firms’ bargaining power with banks [8]. However, research indicates that banks hold a dominant position in the credit market [9], a result of their franchise base monopoly over credit resources [10] and the inherently high-risk nature of NEEs. Enhancing the credit bargaining power of NEEs is therefore essential for reducing financing costs and promoting the green economic transition.

To address the financing difficulties and high costs faced by NEEs, the Chinese government has implemented a series of support policies from both financial and fiscal perspectives [11], primarily comprising green credit policies (GCPs) and government subsidies. Although prior studies have preliminarily examined the effects of these policies, several research gaps persist. First, limited attention has been given to analyzing how such policies influence the financing costs of NEEs through the lens of bargaining power, constraining the ability to identify underlying mechanisms and policy targets. Second, there has been no systematic comparison of the individual and synergistic effects of GCPs and government subsidies, which overlooks policy interactions and hindering the design of optimized policy combinations. Third, the heterogeneous effects of different types of government subsidies have not been comparatively analyzed, limiting the formulation of targeted strategies for enhancing subsidy mechanism.

This study uses data from Chinese A-share listed NEEs from 2008 to 2021 to examine the impact of GCPs and heterogeneous government subsidies on the credit bargaining power of NEEs. The contributions of this research are threefold. First, by investing the credit bargaining process, it reveals the mechanisms through which GCPs and government subsidies influence the bargaining power of NEEs and banks, addressing a gap in prior studies that focus solely on the final impact of policies on financing costs. This provides a more comprehensive analytical framework for understanding policy impacts. Second, by jointly analyzing the individual and synergistic effects of GCPs and government subsidies, the study uncovers the interactive relationships between policies and offers empirical support for their combined application and optimization. Third, by distinguishing between fiscal subsidies and research and developmeny (R&D) subsidies, it compares their respective synergistic effects with GCPs, offering valuable insights for refining the subsidy framework within the new energy sector. Accordingly, this research provides both constructive theoretical insights and practical guidance for addressing the challenge of insufficient credit bargaining power of NEEs, as identified in previous literature [9].

2. Literature Review and Theoretical Analysis

2.1. Literature Review

Existing studies have identified high financing costs as a key constraint on the high-quality development of NEEs in China. One primary reason for this is the disparity in bargaining power between NEEs and banks. Given that bank loans constitute the primary source of external financing [7], banks often dominate credit negotiations. Leveraging their monopoly over credit resources, banks can compel enterprises to accept higher loan rates, thereby generating excess profits and increasing financing costs for borrowers [12]. This imbalance is particularly acute in the case of NEEs, which are typically knowledge and technology intensive. As the relevant technologies remain in early stages, significant risks and uncertainties persist in their development [13]. Additionally, the initial fixed asset investments required for new energy projects are substantial, and their payback periods are relatively long [4]. Compared with traditional energy sources, China’s new energy sector still suffers cost disadvantages and lower market competitiveness. These characteristics result in higher risk coefficients for NEEs, often exposing them to pronounced credit discrimination during the financing process [14, 15]. As a result, bargaining power of NEEs is markedly weaker than that of banks, and new energy projects frequently incur elevated loan costs and risk premiums, with financing expenses exceeding fair market levels [16–18].

Previous research has also examined the role of government policy in reducing NEEs’ credit financing costs. The effectiveness of GCPs, in particular, has been preliminarily validated. Scholtens and Dam [19] found that GCPs in Australia encouraged companies to disclose green business information. Disclosure related to corporate social responsibility was found to enhance corporate performance and reduce financing costs. Lian [20] compared green enterprises with “two-high” enterprises and found that while GCPs raised financing costs for polluting enterprises, they did not reduce those of green enterprises. Li et al. [13] investigated the influence of GCPs on corporate green innovation and found that these policies improved financing conditions, thereby reducing financing costs and significantly enhancing innovation. Xu and Li [21] revealed that GCPs had asymmetric effects—raising financing costs for “two-high” enterprises but lowing them for green enterprises. However, Niu et al. [22] using panel data from 2001 to 2015, concluded that GCPs did not significantly reduce financing costs for green enterprises. Guo and Fang [23] found that including green credit assets as medium-term lending facility collateral increased credit availability and lowered financing costs for green credit enterprises. Zhou et al. [24] confirmed that GCPs enhanced credit access and reduced financing costs for green enterprises, leading to a Pareto improvement in resource allocation. Zhang and Zhong [25] demonstrated that GCPs promoted technological innovation, with mediation effects indicating that green credit substantially reduced interest expenses. Wang et al. [26] also found that green finance pilot policies eased financing constraints for NEEs.

The effectiveness of government subsidies has also been validated, with a growing body of literature exploring the varying effects of heterogeneous subsidies. Wu et al. [27] found that, in inefficient credit markets, government intervention could reduce information asymmetry, lower financing costs, and improve credit resource allocation. However, rent-seeking behaviors might lead to over lending risks. Shen [28] argued that fiscal subsidies act as implicit guarantees, lowering banks’ risk assessments and reducing debt financing expenses for enterprises—although only when subsidies are large enough to shift firm profitability from negative to positive. Wang [29] studied the impact of R&D subsidies, finding that R&D subsidies facilitate technological and regulatory certification, which enhances investor confidence and lowers external financing costs. Zeng et al. [30] pointed out that subsidies offer implicit guarantees that reduce debt financing costs burdens. Liu and Du [31] identified a signaling effect, whereby subsidies alleviate information asymmetry between firms and external investors. Yao et al. [32] further showed that innovation related subsidies had stronger effects than noninnovation subsidies in stimulation investment in R&D.

More recently, scholars have begun to analyze the joint impacts of GCPs and government subsidies. Yang [33] compared direct subsidies and indirect tools (such as GCPs) on renewable energy generation enterprises, finding that mixed subsidies enhanced both renewable energy output and social welfare. Yang et al. [34] examined Chinese renewable energy enterprises and found that GCPs positively influenced performance under financing constraints, and that the combined use of GCPs and production subsidies generated stronger effects. Hong et al. [35] investigated how coordinated fiscal and financial policies improve corporate environmental governance and found that local green credit interest subsidy policies incentivized banks to expand provide green credit, thereby strengthening environmental governance.

In summary, the literature has extensively investigated the financing challenges faced by NEEs and largely attributes them to the substantial bargaining power imbalance between NEEs and banks. Prior research has established that GCPs can effectively reduce the debt financing costs by improving credit conditions and enhancing credit access. Government subsidies, through mechanisms such as implicit guarantees and certification effects, also help to mitigate information asymmetry and lower financing costs. Nonetheless, important gaps remain. First, most existing studies treat financing costs as the final outcome, lacking analysis from the perspective of bargaining power—thereby failing to illuminate how GCPs and subsidies affect the negotiation dynamics between NEEs and banks. This limits our understanding of the mechanisms underlying policy impacts on different stakeholders. Second, the synergistic effects of GCPs and government subsidies have received limited attention, and the interaction between these policy tools remain underexplored. Third, although some research has begun to investigate the differing effects of heterogeneous government subsidies, few studies have systematically compared how distinct subsidy types (e.g., fiscal versus R&D) interact with GCPs. This restricts the development of precision and effective policy framework. This study addresses these gaps by developing a systematic analytical framework grounded in bargaining power theory. Using empirical methods, it evaluates both the individual and combined effects of GCPs and government subsidies. Importantly, by distinguishing between fiscal and R&D subsidies, this research provides a detailed comparative analysis of their synergistic impacts when paired with GCPs. The findings contribute meaningful theoretical insights and offer practical guidance for optimizing the financial support policy system for NEEs.

2.2. Theoretical Analysis and Hypothesis

A comprehensive review of the literature reveal that GCPs did not initially enhance the bargaining power of green enterprises, and therefore failed to reduce their financing costs [20, 22]. However, after a period of policy development and refinement, GCPs began to improve bargaining power and ultimately reduced financing costs [23, 25]. This suggests that GCPs may exhibit a threshold effect in influencing the bargaining power of NEEs, indicating a nonlinear relationship. As GCPs mature, they can reduce the bargaining power of banks and, in turn, enhance that of NEEs. This effect operates through three primary mechanisms: risk compensation, administrative regulation, and signaling effects. Specifically, in terms of risk compensation, the GCPs provide interest subsidies and refinancing support to banks, helping to offset the additional risks associated with lending to NEEs. These mechanisms internalize externalities into profits, incentivizing banks to reduce the risk premiums on credit offered to NEEs. In terms of administrative regulation, the implementation of GCPs involves institutional mechanisms such as the “special statistical system for green loans”, which mandates the disclosure of green loan issuance and incorporates this data into performance assessments. These measures put pressure on banks and encourage their adherence to green credit standards. In terms of signaling effects, GCPs clarify long term expectations for the new energy sector, enhancing the confidence of financial institutions, including banks, in supporting NEEs. This, in turn, strengthens the bargaining power of NEEs. Based on this analysis, the following hypothesis is proposed.

H1: GCPs can indirectly enhance the bargaining power of NEEs by reducing the bargaining power of banks, and this effect is nonlinear.

Both fiscal subsidies and R&D subsidies can also improve the bargaining power of NEEs. These subsidies operate through three shared channels: signaling effects, implicit government guarantees, and improvements in business operations. In terms of signaling attributes, when enterprises receive government subsidies, it signals official endorsement of their growth potential and operational quality. This reduces information asymmetry, enabling banks to make more accurate risk assessments and consequently lowering financing costs [36]. Regarding implicit government guarantees, subsidies often imply a form of governmental backing, which alleviates creditor concerns about default risk and lowers the cost of debt [37]. In terms of operational improvements, subsidies provide enterprises with additional liquidity, enhancing cash flow, profitability, and financial stability. This reduces reliance on external financing and strengthens NEEs’ position in credit negotiations [38, 39].

However, R&D subsidies differ significantly from general fiscal subsidies in two respects. First, R&D subsidies involve a more selective allocation process that includes application, government review and final distribution. The review phase is particularly profitable enterprises typically qualify [40], resulting in stronger signaling effects. Second, R&D subsidies sever as a form of “certification” of enterprise quality from the perspective of external investors [41], thereby exerting a more pronounced influence on bargaining power compared to general fiscal subsidies. Accordingly, the second hypothesis is formulated.

H2: Both fiscal subsidies and R&D subsidies can directly enhance the bargaining power of NEEs, with R&D subsidies being more effective.

Although both the GCPs and government subsidies contribute to improving the bargaining power of NEEs, they target different entities. GCPs primarily affect banks, compelling them to make credit concessions, thereby indirectly benefiting NEEs. In contrast, government subsidies act directly on NEEs, improving their negotiation position. When implemented together, these policies jointly address risk compensation and information asymmetry, improve the credit negotiation environment, and facilitate redistribution of credit surplus in favor of NEEs. Given that R&D subsidies exhibit stronger selectivity and certification effects than fiscal subsidies, their synergy with GCPs is expected to be greater. Thus, the third hypothesis is proposed.

H3: GCPs and government subsidies can jointly enhance the bargaining power of NEEs. And R&D subsidies demonstrating a stronger synergistic effect with GCPs than fiscal subsidies.

The specific impact mechanisms underlying these policy impacts are illustrated in Figure 1.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

The influence mechanism of GCPs and subsidies in the credit bargaining process.

3. Research Design

3.1. Measurement of Bargaining Power and Analysis of Influencing Factors

3.1.1. Measurement of Bargaining Power

To measure the bargaining power of NEEs, this study adopts the methodology proposed by Xu et al. [9] which conceptualizes the bargaining process through the lens of pricing surplus allocation. Specifically, the two-tier stochastic frontier analysis (SFA) model is employed to quantify bargaining power. In a standard credit market, banks and NEEs negotiate loan terms, ultimately arriving at an agreed upon lending rate. In this process, the party with greater bargaining power secures more favorable pricing outcomes. Therefore, the credit financing costs of NEEs can be set as follow:

()

In this context, represents the lowest acceptable price for the banks, is the highest acceptable price for the NEEs, and η represents the banks’ bargaining power, where 0 ≤ η ≤ 1. represents the surplus that the banks can gain in reaching the equilibrium price.

To simultaneously characterize the bargaining power of NEEs and banks, introducing a “fair price”, also known as the loan rate under perfect competition conditions, would be necessary. Some studies also refer to this as a fair loan price, benchmark price, or market equilibrium price [42]. This fair price is not directly observable, but it objectively exists and is primarily influenced by the individual characteristics, represented by x. Therefore, it is defined as μ(x), satisfying

. Thus,

and

denote the anticipated surplus for NEEs and banks, respectively. The ultimate surplus acquired by the banks or NEEs hinges on their individual bargaining power. This concept can be restated by employing the surplus definition to reiterate Equation (2):

()

Equation (2) indicates that banks can raise the cost and obtain a final surplus scale of

. Likewise, NEEs can decrease the debt expense by extracting banks’ surplus, which is

. The surplus differential between banks and NEEs quantifies the discrepancy between the actual credit financing price and the “fair price”. The banks’ bargaining power is positively correlated with the credit price, whereas the NEEs’ bargaining power is negatively correlated with the credit price. Hence, Equation (2) can be abbreviated as:

()

here, μ(x_i) = βx_i, where β represents the coefficients to be estimated and x_i represents the NEEs’ feature variables, which are the primary factors determining credit pricing.

represents the banks’ surplus, and

implies that the enterprises can lower credit expenditure by plundering a portion of the banks’ surplus. v_i is the conventional disturbance term.

Due to the one-sided nature of the effects of w_i and u_i in the above equation, it is assumed that both w_i and u_i follow independent and xidentical exponential distributions, that is, and . Furthermore, v_i is assumed to follow a standard normal distribution, that is, . Simultaneously, it is assumed that w_i, u_i, and v_i are mutually independent and unrelated to the characteristic variables x_i influencing the credit financing price of the enterprise mentioned above. Under these assumptions, w_i and u_i both follow one-sided distributions, and in this case, ordinary least squares (OLS) estimates may lead to a biased composite disturbance term ε_i. Hence, to concurrently estimate the parameters and the transaction surplus, maximum likelihood estimation (MLE) should be employed for estimating Equation (3).

Following the assumptions above, the probability density function (PDF) of ε_i can be obtained by Equation (4):

()

In Equation (4),

, and

. Here, Φ(⋅) represents the cumulative distribution function (CDF) of the standard normal distribution. For a sample with size n, the log−likelihood function can be written as:

()

In Equation (5), ζ = {δ, σ_w, σ_u, σ_v} are the main parameters of concern. MLE is applied, that is, the corresponding parameters are obtained by maximizing the function value of the Equation (5). It is important to note that since σ_w and σ_u appear separately in the likelihood equation, σ_w appears in a_i and b_i, while σ_u appears in α_i and β_i. Thus, the identification of these three standard deviations is achieved.

This study primarily investigates the credit financing transaction surplus. Hence, estimating the conditional expectations of w_i and u_i becomes essential. The conditional PDF of these two variables are as follows:

()

Here, ,χ_1i = Φ(b_i) + exp(α_i − a_i) · Φ(β_i), and χ_2i = Φ(β_i) + exp(a_i − α_i) · Φ(b_i) = exp(a_i − α_i) · χ_1i.

Based on Equations (6) and (7), the conditional expectations of u_i and w_i can be further derived, and their estimation formulas are as follows:

()

Based on Equations (8) and (9),

and

can be used to calculate the surplus of NEEs and banks, which more intuitively represent the magnitude of bargaining power between NEEs and banks. The result represents the average tendency of NEEs (banks) to pay less (charge more) in actual loan costs as a percentage on the basis of the “fair price” for loans. The difference between these two values is the net surplus (NS), which can be calculated using Equation (10). If NS > 0, NEEs exhibit stronger bargaining power; otherwise, banks possess greater bargaining power.

()

3.1.2. Analysis of Factors Influencing Bargaining Power

By estimating Equation (3) above, the bargaining power variables of NEEs and banks (u_i and w_i) can be obtained after variance decomposition. A preliminary statistical analysis of u_i and w_i can reveal the current bargaining power of NEEs and banks in the credit market. Furthermore, this paper focuses on the impact of GCPs and heterogeneous government subsidies on NEEs, as well as the synergistic effects of policies. Thus, this work draws on the research by Yan et al. [36], GCPs (GC), fiscal subsidies (Sub), and R&D subsidies (RD) are selected as the policy variables of interest, and according to the theoretical analysis above, Equations (11)–(14) are constructed to investigate whether these policy factors have an impact on the bargaining power of NEEs.

()

3.2. Sample, Variables, and Data

3.2.1. Sample Selection

Due to data availability constraints, this study uses China’s A-share listed NEEs as the research sample. The selection criteria are based on the methodology proposed by Xu et al. [43]. Annual reports were manually reviewed to identify the presence of new energy products and business activities for each firm in each year. Based on this process, a total of 2460 firm-year observations were selected, covering the period from 2008 to 2021.

3.2.2. Variable Selection

3.2.2.1. Variables for Measuring Bargaining Power

The selection of variables in this study is informed by a dual theoretical framework that integrates resource dependance theory and credit rationing theory. The financing costs variable follows the definitions used in Zhou et al. [24] and Guo and Fang [23]. Drawing on Guo and Fang [23] and Xu et al. [9], the following firm level characteristics are considered key determinants of financing costs: firm size, solvency, corporate governance structure, profitability, asset tangibility, asset liquidity, agency costs, and firm age. The names, symbols, and measurements of these variables are detailed in Table 1.

Table 1. Variables for measuring bargaining power.

Variable category	Variable name	Symbol	Measurement
Dependent variable	Financing costs	Cost	Financial expense/total liabilities

Characteristic variables	Enterprise size	Size	Natural logarithm of total assets
	Debt repayment ability	Lev	Debt-to-equity ratio
	Dual position integration	Dual	Assigned 1, if chairman and general manager are the same person, else 0
	Profitability	Roa	Asset profitability ratio
	Equity concentration	Top1	Number of shares held by the largest shareholder divided by total number of shares
	Free cash flow	Inca	Free cash flow/total assets
	Asset liquidity	Liquid	Current assets/current liabilities
	Asset tangibility	Tangible	Net fixed assets/total assets
	Agency costs	Turnover	Operating expenses/operating revenue
	Firm age	Age	Natural logarithm of the time since establishment of the firm

3.2.2.2. Variables for Analyzing Policy Factors Influencing Bargaining Power

The policy factors variables include GCPs, fiscal subsidies and R&D subsidies. The proxy for GCPs was selected based on the methodologies proposed by Song et al. [44] and Lin and Pan [45]. Specifically, the implementation intensity of GCPs in each province is calculated using the formula: (1) Interest expenses of six major energy-intensive industries in each province/interest expenses of all industrial enterprises above a certain scale. Thus, each enterprise is then assigned the GCPs intensity value corresponding to its province. The measurement of fiscal subsidies follows the approach in existing literature, whereby subsidies are extracted from the nonoperating income section of firms’ financial reports [46]. The R&D subsidies variable is identified in accordance with Xu et al. [43], by extracting information related to R&D specific subsidies disclosed in the annual reports of listed firms. The bargaining power of NEEs is calculated using the two-tier SFA model, derived from the estimation and variance decomposition of Equation (3).

To analyze the individual impacts of policy factors, GCPs, fiscal subsidies and R&D subsidies are used as grouping variables, and the full research sample is divided into corresponding subsamples. GC_D is a binary variable representing the GCPs implementation phase [47], assigned GC_D a value of 0 for the period before 2012 and 1 for 2012 onwards. Sub_D equals 1 if fiscal subsidies exceed the sample median, and 0 otherwise. RD_D equals 1 if a firm receives R&D subsidies, and 0 if not.

In summary of the variable’s names, symbols, and measurements methods is provided in Table 2.

Table 2. Variables for analyzing policy factors influencing NEEs’ bargaining power.

Variable category	Variable name	Symbol	Measurement
Dependent variable	NEEs’ bargaining power	σ_u	Obtained from estimation and variance decomposition for Equation (3)

Core explanatory variables	GCPs Intensity	GC	1–interest expenses of six major energy-intensive industries in each province/interest expenses of all industrial enterprises above a certain scale
	Fiscal subsidies	Sub	Natural logarithm of total subsidies received by the NEEs
	R&D subsidies	RD	Natural logarithm of R&D subsidies disclosed in nonoperating income

Grouping variables	GCPs	GC_D	Assigned 0, before 2012; otherwise 1
	Fiscal subsidies dummy	Sub_D	For values surpassing the median, set to 1; otherwise, set to 0.
	R&D subsidies dummy	RD_D	For R&D subsidies greater than 0, set to 1; otherwise, set to 0.

3.2.3. Data Sources

Data on credit financing costs and firm-level characteristics are obtained from the WIND and CSMAR databases, supplemented by manual collection from the annual reports of NEEs. Data on GCPs is sourced from the China Industrial Statistical Yearbook and Economic Census Data. Data on government subsidies are manually extracted from the annual reports of listed companies.

4. Empirical Analysis

4.1. Descriptive Statistics of Variables

Descriptive statistics and correlation results for the variables listed in Tables 1 and 2 are presented in Table 3, excluding the NEEs’ bargaining power metric and dummy variables, which will be analyzed separately in a subsequent section. The results show that all variable values fall within reasonable and expected ranges, thereby validating their suitability for further empirical investigation.

Table 3. Statistics and correlation coefficients.

Variables	Mean	S.D.	(1)	(2)	(3)	(4)	(5)	(6)
(1) Cost	0.018	0.023	1.00	—	—	—	—	—
(2) GC	0.583	0.123	0.11 ^∗∗∗	1.00	—	—	—	—
(3) Sub	6.816	2.237	− 0.03	− 0.01	1.00	—	—	—
(4) RD	4.045	2.946	− 0.07 ^∗∗∗	0.06 ^∗∗	0.17 ^∗∗∗	1.00	—	—
(5) Size	22.86	1.454	− 0.15 ^∗∗∗	− 0.10 ^∗∗∗	0.38 ^∗∗∗	0.02	1.00	—
(6) Lev	0.52	0.187	− 0.00	− 0.13 ^∗∗∗	0.19 ^∗∗∗	− 0.04 ^∗	0.51 ^∗∗∗	1.00
(7) Dual	0.238	0.426	− 0.02	0.17 ^∗∗∗	− 0.08 ^∗∗∗	0.05 ^∗∗	− 0.21 ^∗∗∗	− 0.20 ^∗∗∗
(8) Roa	0.047	0.06	0.07 ^∗∗∗	− 0.00	0.08 ^∗∗∗	0.00	0.08 ^∗∗∗	− 0.26 ^∗∗∗
(9) Top1	0.348	0.156	0.10 ^∗∗∗	− 0.05 ^∗∗	0.12 ^∗∗∗	− 0.04 ^∗	0.30 ^∗∗∗	0.06 ^∗∗
(10) Inca	0.043	0.061	0.12 ^∗∗∗	0.01	0.02	− 0.08 ^∗∗∗	0.18 ^∗∗∗	− 0.01
(11) Liquid	1.759	2.011	− 0.08 ^∗∗∗	0.02	−0.09 ^∗∗∗	0.02	− 0.33 ^∗∗∗	− 0.61 ^∗∗∗
(12) Tangible	0.306	0.201	− 0.04	− 0.18 ^∗∗∗	− 0.00	− 0.16 ^∗∗∗	0.28 ^∗∗∗	0.21 ^∗∗∗
(13) Turnover	0.470	0.311	0.04	0.10 ^∗∗∗	0.04 ^∗	0.06 ^∗∗	− 0.13 ^∗∗∗	− 0.03
(14) Age	2.872	0.395	0.15 ^∗∗∗	− 0.01	− 0.11 ^∗∗∗	− 0.10 ^∗∗∗	0.18 ^∗∗∗	0.11 ^∗∗∗

	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)

(7) Dual	1.00	—	—	—	—	—	—	—
(8) Roa	0.06 ^∗∗	1.00	—	—	—	—	—	—
(9) Top1	− 0.10 ^∗∗∗	0.15 ^∗∗∗	1.00	—	—	—	—	—
(10) Inca	− 0.05 ^∗	0.30 ^∗∗∗	0.09 ^∗∗∗	1.00	—	—	—	—
(11) Liquid	0.08 ^∗∗∗	0.14 ^∗∗∗	− 0.05 ^∗	− 0.06 ^∗∗	1.00	—	—	—
(12) Tangible	− 0.12 ^∗∗∗	0.01	0.09 ^∗∗∗	0.32 ^∗∗∗	− 0.26 ^∗∗∗	1.00	—	—
(13) Turnover	− 0.01	0.09 ^∗∗∗	− 0.01	0.08 ^∗∗∗	0.00	− 0.17 ^∗∗∗	1.00	—
(14) Age	− 0.13 ^∗∗∗	− 0.01	− 0.10 ^∗∗∗	0.19 ^∗∗∗	− 0.18 ^∗∗∗	0.17 ^∗∗∗	0.00	1.00

Note: N = 2460.
^∗p < 10%.
^∗∗p < 5%.
^∗∗∗p < 1%.

4.2. Measurement Estimation Results: Bargaining Power of NEEs and Banks

4.2.1. Estimation Results of the Credit Bargaining Model

Table 4 presents the results derived from the estimation of Equation (3), where Models 1 and 2 present the results from the multivariate linear regression approach, while Models 3 and 4 report the outcomes obtained from the two-tier SFA method. The results show that the fitting effects of all models are quite good, with strong explanatory power for the majority of variables. Among them, Size, Tangible, and Inca have a significantly negative impact on the financing costs. However, Lev, Top1, and Turnover significantly increase the financing costs. These results are consistent with expectations and existing research [42]. The random disturbance terms for Models 3 and 4 indicate that random disruptive factors, NEEs’ bargaining power, and banks’ bargaining power are all significant, causing the final loan transaction price to deviate from the fair price. The results estimated from Model 4 are used as the foundation for subsequent variance decomposition, bargaining power, and surplus measurement.

Table 4. Regression results of credit financing cost of NEEs.

Variables	Dependent variable: Cost
Variables	Model 1	Model 2	Model 3	Model 4
Size	− 0.045 ^∗∗∗ (− 2.70)	− 0.039 ^∗∗ (− 2.12)	− 0.050 ^∗∗∗ (− 6.17)	− 0.037 ^∗∗∗ (− 4.25)
Lev	0.602 ^∗∗∗ (5.12)	0.528 ^∗∗∗ (4.35)	1.156 ^∗∗∗ (15.78)	1.057 ^∗∗∗ (13.42)
Dual	− 0.022 (− 0.62)	0.012 (0.34)	−0.000 (−0.02)	0.023 (1.02)
Roa	− 1.279 ^∗∗∗ (− 5.73)	− 1.159 ^∗∗∗ (− 5.13)	− 0.011 ^∗ (− 0.86)	− 0.012 ^∗ (− 1.97)
Top1	0.034 (0.26)	−0.211 (−1.60)	− 0.049 (− 0.79)	− 0.138 ^∗∗ (− 2.19)
Inca	− 0.917 ^∗∗∗ (− 4.34)	− 0.986 ^∗∗∗ (− 4.60)	− 0.424 ^∗∗ (− 2.19)	− 0.658 ^∗∗∗ (− 3.38)
Liquid	0.138 ^∗∗∗ (16.31)	0.133 ^∗∗∗ (15.88)	0.091 ^∗∗∗ (12.39)	0.093 ^∗∗∗ (12.99)
Tangible	1.101 ^∗∗∗ (11.35)	0.794 ^∗∗∗ (7.30)	1.502 ^∗∗∗ (26.75)	1.148 ^∗∗∗ (15.35)
Turnover	0.064 (1.37)	0.146 ^∗∗∗ (2.84)	0.178 ^∗∗∗ (4.33)	0.222 ^∗∗∗ (5.36)
Age	− 0.285 ^∗∗∗ (− 6.06)	− 0.150 ^∗∗ (− 2.26)	−0.024 (−0.98)	−0.020 (−0.64)
Year and industry	—	Control	—	Control
Constant	2.010 ^∗∗∗ (5.85)	2.497 ^∗∗∗ (4.16)	0.647 ^∗∗∗ (3.51)	0.394 (1.64)
sigma_v	—	—	−1.459 ^∗∗∗ (− 13.52)	−1.960 ^∗∗∗ (− 13.62)
sigma_u	—	—	−1.711 ^∗∗∗ (− 12.85)	−1.576 ^∗∗∗ (− 25.55)
sigma_w	—	—	− 0.854 ^∗∗∗ (− 22.99)	− 0.832 ^∗∗∗ (− 25.00)
N	2460	2460	2460	2460
Adj. R²	0.239	0.329	—	—
ll	− 2304	− 2120	− 1985	− 1815

^∗p < 10%.
^∗∗p < 5%.
^∗∗∗p < 1%.

Table 5 presents the results of the variance decomposition based on Model 4, reflecting the impact of the bargaining power of various parties on loan prices in the credit market. The results reveal that the difference in bargaining power between banks and NEEs is positive, E(w − u) = σ_w − σ_u = 0.2281, indicating that banks hold a strong position in negotiations. Bargaining factors account for 92.12% of the total variance, with the proportion of banks’ bargaining power accounting for 81.56%. The proportion of NEEs’ bargaining power accounts for only 18.44% of the variance, further illustrating the absolute dominance of banks in the process of credit market price formation. This perspective supports the findings by Wei and Zeng [12] and Kochiyama and Nakamura [48].

Table 5. Measurement of bargaining power between NEEs and banks.

Variable category	Variable meaning	Symbol	Estimation result
Disturbance term	Random error term	σ_v	0.1408
	Bargaining power of NEEs	σ_u	0.2069
	Bargaining power of banks	σ_ω	0.4350

Variance decomposition	Total variance of the error term		0.2518
	Proportion of bargaining factors in the total variance		92.12%
	Proportion of NEEs’ bargaining power		18.44%
	Proportion of banks’ Bargaining Power		81.56%

Table 6 provides a quantile statistical analysis of surplus, highlighting the unilateral gains of each party. On average, the surplus obtained by NEEs reduces the loan price by 17.1%, whereas the surplus obtained by banks increases the price by 30.1%, leading to an average 13% loan premium borne by NEEs. At the 25^th percentile, loan prices exceed the fair price by 28.5%, while at the median, they exceed it by 10%. At the 75^th percentile, NEEs surpasses banks in bargaining power, achieving a positive surplus equivalent to 3.4% of the fair price though this remains a small portion of the sample. These results are consistent with those of Xu et al. [9] and Wei [10].

Table 6. Surplus quantile statistical results.

Variable	Mean %	P25 (%)	P50 (%)	P75 (%)
Firms	17.10	12.30	13	17.90
Banks	30.10	14.60	23	40.80
NS	− 13	− 28.50	− 10	3.40

4.2.2. Analysis of Temporal Trends in Bargaining Power and Surplus

Figure 2 illustrates the temporal evolution of bargaining power quantiles for both NEEs and banks. Between 2008 and 2021, banks at the 75^th percentile show the largest decrease in bargaining power from 70% in 2008 to 52% in 2021, reflecting a substantial 18% decline. In contrast, bargaining power for banks at the median and 25^th percentiles exhibits relatively minor changes. NEEs show no significant time trends across percentiles, suggesting that reductions in banks dominance have not been matched by uniform improvements in NEEs’ bargaining power. These findings indicate that the bargaining power of stronger banks has weakened more significantly over time.

Figure 3 presents the annual distribution of NS obtained by NEEs, with a 1% trimmed mean applied to exclude outliers. It shows that from 2008 to 2021, the trends in changes across quantile samples are not entirely consistent, exhibiting significant heterogeneity. At the 25^th percentile, the surplus shows a significant fluctuation during each year from 2008 to 2021, presenting an inverted U-shape and ultimately showing some growth. The lowest point, in 2008, is approximately −57%, while the highest point, in 2012, is around −30%, exhibiting an increase of 27%. And comparing 2021–2008, there is an increase of about 10%. This indicates that weak enterprises experience a substantial increase in the available surplus in the credit market. At the median, the NS value stays below 0 and exhibits no significant trends in any direction over time. Similarly, at the 75^th percentile, the NS value stays above 0 and exhibits no significant trends over time. The results indicate that, over time, the increase in NS for NEEs is primarily reflected in NEEs that are initially weaker, such as private enterprises and small enterprises, and this change is associated with the weakening bargaining power of dominant banks.

4.3. Analysis of the Impact of Policy Factors on Bargaining Power and Surplus

4.3.1. Individual Impact of Policies

Following the methodology of Xie [42], the full sample was divided into subgroups based on the policy related dummy variables defined in Table 2, and Equation (3) was reestimated for each subgroup.

4.3.1.1. Individual Impact of GCPs

Equation (3) was reestimated using subsamples grouped by GC_D, and the bargaining power, surplus for each party, and NS are calculated. Figure 4 presents the distribution of bargaining power for NEEs and banks before and after the implementation of the GCPs. Prior to implementation, NEEs’ bargaining power of NEEs was around 7%, indicating a clear disadvantage relative to banks. Following GCPs implementation, the bargaining power of NEEs increased markedly. Concurrently banks at the 75^th and 50^th percentiles experienced a substantial reduction in bargaining power, while the 25^th percentile showed little change.

Figure 5 displays the distribution of surplus before and after the implementation of GCPs. Before the policies, surplus of NEEs was below 5%, and banks at the 75^th percentile captured over 80% of the total surplus. Additionally, over 25% of firms had an NS below −80%, and more than 50% had NS below −40%. Post GCPs, the median surplus of NEEs exceeded 15%, reflecting a significant increase. Banks’ surplus decreased overall, especially among banks in the upper quantiles (i.e., the 75^th and above), suggesting that those with stronger initial bargaining power made greater concessions. Moreover, more than 25% of the sample achieved positive NS, and the 25^th percentile increases substantially, highlighting the greater benefits that accrued to NEEs with initially weaker bargaining position.

4.3.1.2. Individual Impact of Subsidies

Equation (3) was reestimated for subsamples grouped bySub_D and RD_D, respectively. The bargaining power, surplus, and NS were computed accordingly. Figure 6 illustrates the impact of fiscal and R&D subsidies on the bargaining power of NEEs and banks. Firms receiving higher fiscal subsidies (Sub_D = 1) experienced a decline in bargaining power. In contrast, R&D subsidies (RD_D = 1) significantly enhanced the bargaining power of NEEs. For banks, the opposite effects were observed: excessive fiscal subsidies increased banks’ bargaining power, while R&D subsidies reduced it. These findings suggest that while GCPs improve NEEs’ bargaining power and NS especially among weaker firms, fiscal subsidies may unintentionally strengthen banks’ positions, whereas R&D subsidies directly empower NEEs.

Figure 7 shows the impact of subsidy factors on NS. The results indicate that, when NEEs receive a higher fiscal subsidy (Sub_D = 1), it actually reduces the NS obtained by the NEEs. However, receiving R&D subsidies (RD_D = 1) increases the NS for NEEs, which is particularly evident in the improvement of NS for NEEs at the 25^th and 50^th percentiles. These findings reveal that excessive fiscal subsidies tend to reduce bargaining power of NEEs and NS, possibly due to issues such as budget constraints and rent-seeking encouraged by the fiscal subsidies, and R&D subsidies can enhance NEEs’ bargaining power and NS.

4.3.2. Synergistic Impact of Policies

First, to investigate the synergistic effects of GCPs and government subsidies, Equations (11)–(14) were estimated, and the results are reported in Models 5 to 8 of Table 7.

Table 7. Analysis results of the impact of policy factors on bargaining power of NEEs.

Variables	Dependent: σ_u
Variables	Model 5	Model 6	Model 7	Model 8
GC	− 0.360 (− 1.64)	− 5.859 ^∗∗∗ (− 3.56)	− 5.797 ^∗∗∗ (− 3.38)	− 6.579 ^∗∗∗ (− 3.87)
GC²	—	5.796 ^∗∗∗ (3.80)	5.962 ^∗∗∗ (3.88)	5.924 ^∗∗∗ (3.85)
Sub	—	—	− 0.043 (− 0.73)	—
GC × Sub	—	—	0.036 (0.36)	—
RD	—	—	—	0.076 ^∗ (1.70)
GC × RD	—	—	—	0.148 ^∗∗ (2.00)
Constant	− 4.366 ^∗∗∗ (− 33.32)	− 5.937 ^∗∗∗ (− 13.76)	− 5.692 ^∗∗∗ (− 10.29)	− 6.271 ^∗∗∗ (− 13.12)
N	2460	2460	2460	2460

^∗p < 10%.
^∗∗p < 5%.
^∗∗∗p < 1%.

The results in Table 7 show that, in Model 5, the effect of GC is not statistically significant. This might be attributable to that GCPs follow a nonlinear influence pattern on NEEs’ bargaining power according to the theoretical analysis. However, in Model 6, GC has a significantly negative impact, whereas GC² has a significantly positive impact, this suggests a U-shaped relationship between GCPs and NEEs’ bargaining power. In Model 7, the impact of GC and GC² remains consistent with that in Model 6, but Sub and the interaction term GC × Sub are not significant. This indicates that the impact of fiscal subsidies on the bargaining power of NEEs is not significant, and there is no substantial synergistic effect with GCPs. In Model 8, the impact of GC is significantly negative, while those of GC², RD, and GC × RD are all significantly positive. This implies that R&D subsidies can significantly enhance the bargaining power of NEEs and that there is a positive synergistic effect with GCPs. Based on the empirical results, hypothesis H1 is verified, while hypothesis H2 and hypothesis H3 are partially verified.

Second, to further validate the nonlinear impact mechanism of GCPs on the bargaining power of NEEs, this study employs fitted curve analysis to visually illustrate the relationship between key variables (Figure 8). The results confirmed a significant U-shaped relationship, supporting and reinforcing empirical findings. Notably, Figure 9 reveals a clear synergistic effect between heterogeneous government subsidies and GCPs, with R&D subsidies exhibiting particularly strong complementarity with GCPs in enhancing the bargaining power of NEEs.

4.4. Robustness Test

To ensure the robustness of the research findings, this study replaces explanatory variables and estimation method, respectively. First, the explanatory variable GC is replaced with the green finance development index (GFD) of each province [49], and Equations (11)–(14) were reestimated, as shown in Table 8. The results show that GFD², RD, and GFD × RD are significantly positively correlated with σ_u, meaning that the main conclusions shown in Table 7 are further verified.

Table 8. Robustness test regression results for replacing explanatory variables.

Variables	Dependent: σ_u
Variables	Model 9	Model 10	Model 11	Model 12
GFD	− 0.806 ^∗∗ (− 2.01)	− 0.732 ^∗ (− 1.82)	− 0.892 ^∗∗ (− 2.18)	0.055 (0.09)
GFD²	—	0.375 ^∗ (1.84)	0.360 ^∗ (1.75)	0.357 ^∗ (1.75)
Sub	—	—	− 0.033 ^∗∗ (− 2.52)	—
GFD × Sub	—	—	0.039 (0.40)	—
RD	—	—	—	0.148 ^∗ (1.72)
GFD × RD	—	—	—	0.211 ^∗ (1.90)
Constant	− 3.976 ^∗∗∗ (− 13.02)	− 3.901 ^∗∗∗ (− 12.65)	− 5.692 ^∗∗∗ (− 10.29)	− 4.463 ^∗∗∗ (− 13.12)
N	2460	2460	2460	2460

^∗p < 10%.
^∗∗p < 5%.
^∗∗∗p < 1%.

Second, directly examining the impact of policy variables on credit financing costs, as well as the synergistic relationships among these policies, indirectly illustrates the influence of policy variables on the bargaining power of NEEs. A two-way fixed effects model is constructed, as shown in Equation (15):

()

Here, Cost is the dependent variable; GC, GC², Sub, and RD, along with the corresponding interaction terms—GC × Sub and GC × RD are selected as explanatory variables denoted by Policy; Controls consists of Table 1, ∑Year denotes the fixed effects for years; and ∑Ind represents the industry fixed effects. The model is then estimated, and the results are presented in Table 9.

Table 9. Robustness test regression results for replacing estimation method.

Variables	Dependent Variable: Cost
Variables	Model 13	Model 14	Model 15	Model 16
GC	0.656 ^∗∗∗ (4.29)	5.355 ^∗∗∗ (5.58)	5.388 ^∗∗∗ (5.41)	5.530 ^∗∗∗ (5.75)
GC²	—	− 4.139 ^∗∗∗ (− 4.96)	− 4.016 ^∗∗∗ (− 4.79)	− 4.062 ^∗∗∗ (− 4.88)
Sub	—	—	0.021 (1.12)	—
GC × Sub	—	—	− 0.023 (− 0.74)	—
RD	—	—	—	0.053 ^∗∗∗ (3.34)
GC × RD	—	—	—	− 0.076 ^∗∗∗ (− 2.91)
Controls	Yes	Yes	Yes	Yes
Year and industry	Control	Control	Control	Control
Constant	1.643 ^∗∗∗ (3.88)	0.402 (0.82)	0.317 (0.62)	0.350 (0.71)
N	2460	2460	2460	2460
Adj. R²	0.382	0.379	0.380	0.395

^∗p < 10%.
^∗∗p < 5%.
^∗∗∗p < 1%.

The results in Table 9 indicate that, the impact of GCPs on Cost exhibits an inverted U-shaped relationship. The interaction term GC × Sub does not have a significant impact on Cost, but the interaction term GC × RD shows a significant negative effect on Cost. Therefore, R&D subsidies play a synergistic role with GCP in reducing the credit financing costs for NEEs. In summary, the main research conclusions remain robust.

4.5. Endogeneity Treatment

The Sub and RD variables may have endogeneity due to potential two-way causality. On one hand, NEEs may apply for subsidies based on their operational status and development prospects [50], on the other hand, governments assess these same enterprise characteristics during the subsidy review process—factors that banks also consider when setting loan terms. To address this issue, this study adopts a two-stage instrumental variable approach, following the methodology of Guo [50] and Xu et al. [51].

Instrumental variables must be strictly exogenous and highly correlated with the endogenous explanatory variables. First, reasonable instrumental variables are selected for government subsidies. As local governments in China depend on “land finance”, an increase in government fiscal revenue leads to a corresponding increase in the fiscal subsidies available to enterprises. Therefore, following Xu et al. [51], the natural logarithm of the land transfer fees in the enterprise’s city (lnland) is used as the first instrumental variable. In addition, referencing Feng and Zhu [52], the first-order lag term of fiscal subsidies (lSub) is used as the second instrumental variable. Second, reasonable instrumental variables are selected for R&D subsidies. Referring to Guo [50], the mean of industry R&D subsidies (mRD) and the lagged R&D subsidies (lRD) are chosen as the corresponding instrumental variables. Third, Models 7 and 8 in Table 7 are separately reestimated by two-stage regression. Finally, weak instrumental variables and overidentification tests are performed. The results are shown in Table 10.

Table 10. Endogenous treatment results.

Variables	Endogenous treatment of sub		Endogenous treatment of RD
	First-stage	Second-stage	First-stage	Second-stage
	Sub	σ_u	RD	σ_u
lnland	0.085 ^∗∗∗ (3.59)	—	—	—
LSub	0.496 ^∗∗∗ (26.96)	—	—	—
Sub	—	0.001 (0.15)	—	—
GC × Sub	—	− 0.518 (− 0.10)	—	—
GC	—	− 0.028 ^∗∗∗ (− 3.44)	—	− 0.045 ^∗ (− 1.70)
GC²	—	0.032 ^∗∗∗ (4.07)	—	0.049 ^∗∗ (1.99)
mRD	—	—	0.768 ^∗∗∗ (25.44)	—
lRD	—	—	0.424 ^∗∗∗ (27.18)	—
RD	—	—	—	0.001 ^∗ (1.92)
GC × RD	—	—	—	0.029 ^∗∗∗ (2.89)
Controls	Yes	Yes	Yes	Yes
Year and industry	Yes	Yes	Yes	Yes
Constant	− 6.616 ^∗∗∗ (−7.98)	0.467 (0.12)	− 6.584 ^∗∗∗ (− 6.83)	0.045 ^∗∗∗ (4.21)
N	2460	2460	2460	2460
F test of excluded instruments	—	28.84 ^∗∗∗ (0.000)	—	18.30 ^∗∗∗ (0.000)
K–P LM	—	107.645 ^∗∗∗ (0.000)	—	781.20 ^∗∗∗ (0.000)
K–P Wald F	—	80.037	—	109.8
Hansen J	—	0.922 (0.3369)	—	0.404 (0.525)

^∗p < 10%.
^∗∗p < 5%.
^∗∗∗p < 1%.

As reported in Table 10, standard diagnostic checks, including weak instrument tests and over identification tests, are conducted to validate instrument relevance and exogeneity, all instrumental variables meet statistical requirements, confirming their suitability. The reestimated results demonstrate that, after addressing endogeneity concerns, R&D subsidies significantly enhance the bargaining power of NEEs and show a synergistic interaction with GCPs. These findings further reinforce the robustness of the study’s benchmark regression outcomes.

5. Discussion

This study introduces a novel bilateral bargaining framework to analyze the dynamics of credit bargaining power between NEEs and banks, with a focus on the influence of GCPS and heterogeneous government subsidies. Moving beyond prior studies that emphasize only financing cost outcomes [20, 24], while overlooking policy influences on credit negotiation processes [9], or separately examined GCPs and subsidy mechanisms [13, 28]. This research directly examines the bargaining process, filling a critical gap in the literature on how policy instruments affect negotiation dynamics. The empirical results demonstrate that the measured bargaining power distribution substantiates established theories of institutional dominance [12, 48], while the observed surplus allocation patterns validate previous findings on asymmetric market power structures [9, 12].

Specifically, through theoretical and empirical analysis, several conclusions can be made. First, NEEs generally have weaker bargaining power, while banks exhibit stronger bargaining power. Most NEEs have a negative NS, indicating that NEEs often engage in credit financing at costs that are higher than the fair price. Second, the GCPs can influence banks, reducing their bargaining power and thereby improving the bargaining power of NEEs. This results in a decrease in the surplus obtained by banks, with more dominant banks experiencing a greater impact than less dominant ones. The impact of the GCPs on the bargaining power of NEEs follows a U-shaped distribution, indicating that the positive effects of the GCPs are realized only when the policy reaches a certain level of implementation; thus, hypothesis H1 is verified. Thirdly, the fiscal subsidies are ineffective in improving the bargaining power for NEEs, while R&D subsidies can enhance the bargaining power of NEEs. However, R&D subsidies also increase the surplus obtained by enterprises, with a more pronounced impact exerted on weaker enterprises. This suggests that different policies not only have distinct effects on different entities, but also that they vary in their degree of impact. Therefore, hypothesis H2 is partially verified. Finally, fiscal subsidies do not show a synergistic effect with the GCPs, though. R&D subsidies do positively influence the bargaining power of NEEs. Thus, hypothesis H3 is partially verified.

5.1. Theoretical Contributions

This study makes three significant theoretical contributions: First, it constructs a novel analytical framework for assessing the credit bargaining power of NEEs, thereby moving beyond the traditional one-dimensional focus on financing costs. By empirically verifying the asymmetric bargaining relationship between banks and NEEs, this study reveals the theoretical mechanism underpinning structural imbalances in the credit market and deepens the application of resource dependance theory in the context of green finance. Second, it proposes a theory of differential policy effects, demonstrating that fiscal and R&D subsidies have fundamentally distinct impacts on bargaining power. While fiscal subsidies fail to enhance NEEs’ bargaining positions, R&D subsidies significantly improve surplus acquisition, particularly for disadvantaged firms. This finding supports the targeting principle in policy tool theory and empirically validates the central role of knowledge capital accumulation in strengthening credit negotiation power, thus enriching the theoretical foundations of innovation subsidy theory. Third, the study identifies the conditions and boundaries under which policy synergies occur. It confirms that GCPs and R&D subsidies exhibit significant synergistic effects, while no such synergy exists with fiscal subsidies. This insight contributes critical empirical support to the development of policy mix theory in the context of credit markets.

5.2. Practical Implications

The findings of this study have several important implications for the refinement of GCPs and government subsidy policies. First, it is recommended to establish a policy evaluation mechanism that systematically monitors and assesses the long-term effectiveness of GCPs. Dynamic policy adjustments should be evidence based. A dedicated regulatory agency should be established to guide and supervise financial institutions to ensure effective and consistent policy implementation. Second, in light of the increasing fiscal burden and widening subsidy gap, it is essential to optimize subsidy structures. This includes setting reasonable subsidy levels to prevent resource misallocation due to indiscriminate disbursements. Given the evolving development stage of the new energy industry, production-oriented subsidies may be gradually phased out. A phased approach could involve reducing subsidies for production activities while temporarily retaining R&D subsidies to support innovation. Finally, the development of a new, integrated financing support system for NEEs is crucial. This system should prioritize green credit as the primary financing tool, supplemented by targeted government subsidies to identify and support high-quality NEEs. Such a combined approach would form a more efficient and effective policy framework, balancing fiscal sustainability with industrial development goals.

5.3. Limitations and Future Research Directions

Despite its theoretical and practical contributions, this study has several limitations: First, the research sample consists solely of A-share listed NEEs, excluding nonlisted firms. Due to their more standardized financial structures and higher levels of information disclosure, listed firms may systematically differ from nonlisted NEEs in terms of financings to the broader NEEs population. Second, data limitations constrain the precision of policy variable measurement. Specifically, due to the lack of microlevel data on GCPs, this study employs provincial-level GCPs intensity as a proxy variable. While this approach enables empirical estimation, it may not fully capture the actual level of green credit support experienced by individual firms. The considerable regional variation in policy enforcement introduces potential measurement error, which could lead to biased estimates. Third, the findings are context-dependent. China’s unique bank-dominated financial system, its centralized industrial policy framework, and the current stage of development of its new energy sector may jointly limit the external validity of the conclusions. The extent to which these findings apply to other countries or financial systems thus require further investigation.

These limitations also point to valuable directions for future research. First, the development of a comprehensive financing database for NEEs using digital technologies could significantly improve data quality and availability. In this context, blockchain can ensure data transparency and integrity, while artificial intelligence can facilitate dynamic data cleaning and analysis, thereby enhancing the robustness of empirical research [53, 54]. Second, adopting mixed-methods approaches that integrate both quantitative and qualitative analyses could help address current methodological constraints and provide deeper insights into policy mechanisms. Finally, future studies could incorporate policy text analysis and other advanced techniques to more accurately capture the implementation intensity of GCPs, allowing for the more precise measurement and evaluation of policy effects at the firm level.

6. Conclusions

Grounded in resource dependance theory, this study investigates how the effects of GCPs and heterogeneous government subsidies on the credit bargaining power of NEEs and explores the synergistic relationships between these policy instruments. The findings indicate that GCPs indirectly strengthen NEEs’ bargaining power by weakening banks’ bargaining power, with the effect exhibiting a U-shaped pattern. While fiscal subsidies do not significantly enhance NEEs’ bargaining power, R&D subsidies demonstrate a substantial positive impact. Furthermore, a synergistic effect exists between GCPs and R&D subsidies, whereas no significant interaction is observed between GCPs and fiscal subsidies. This study provides empirical support for policy designs aimed at alleviating financing constraints in the new energy sector and contribute to the broader objective of achieving carbon peaking and carbon neutrality goals.

Nomenclature

NEEs:: New energy enterprises
GCPs:: Green credit policies
CoC:: Cost of capital
R&D:: Research and development
Two-tier SFA:: Two-tier stochastic frontier
PDF:: Probability density function
CDF:: Cumulative distribution function
NS:: Net surplus
MLE:: Maximum likelihood estimation
GFD:: Green finance development index.

Conflicts of Interest

The authors declare no conflicts of interest.

Author Contributions

Lu Meng is responsible for data curation, methodology, software, visualization, writing – original draft, writing – review and editing. Chunyan Li is response for conceptualization, investigation, supervision, article review and editing. Zhe Xu is responsible for fund acquisition, article framework construction, article review and editing.

Funding

The authors are grateful for the financial support provided by the National Natural Science Foundation of China (Grant 72103142).

Acknowledgments

The authors are grateful for the financial support provided by the National Natural Science Foundation of China (Grant 72103142).

Open Research

Data Availability Statement

The data will be made available by the corresponding authors upon reasonable request.

References

1 Lin B. Q. and Li J. L., Transformation of China’s Energy Structure Under Environmental Governance Constraints: A Peak Value Analysis of Coal and Carbon Dioxide, Social Sciences in China. (2015) 205, no. 9, 84–107.
Google Scholar
2 Mehboob M. Y., Ma B. J., Sadiq M., and Mehboob M. B., How Do Nuclear Energy Consumption, Environmental Taxes, and Trade Globalization Impact Ecological Footprints? Novel Policy Insight From Nuclear Power Countries, Energy. (2024) 313, https://doi.org/10.1016/j.energy.2024.133661, 133661.
10.1016/j.energy.2024.133661
Google Scholar
3 Mehboob M. Y., Ma B. J., Sadiq M., and Zhang Y., Does Nuclear Energy Reduce Consumption-Based Carbon Emissions: The Role of Environmental Taxes and Trade Globalization in Highest Carbon Emitting Countries, Nuclear Engineering and Technology. (2024) 56, no. 1, 180–188, https://doi.org/10.1016/j.net.2023.09.022.
10.1016/j.net.2023.09.022
CAS Web of Science® Google Scholar
4 Wang K. S., Sun X. R., and Wang F. R., Development of Green Finance, Debt Maturity Structure and Investment of Green Enterprise, Financial Forum. (2019) 24, no. 7, 9–19.
Google Scholar
5 Qadeer A., Jiang X., Rehman R. U., and Deng B., Does State Participation Promote or Damage Green Innovation?, Creativity and Innovation Management. (2024) 33, no. 1, 61–76, https://doi.org/10.1111/caim.12581.
10.1111/caim.12581
Google Scholar
6 Brunnschweiler C. N., Finance for Renewable Energy: An Empirical Analysis of Developing and Transition Economies, Environment and Development Economics. (2010) 15, no. 3, 241–274, https://doi.org/10.1017/S1355770X1000001X, 2-s2.0-77955853314.
10.1017/S1355770X1000001X
Web of Science® Google Scholar
7 Best R., Switching Towards Coal or Renewable Energy? The Effects of Financial Capital on Energy Transitions, Energy Economics. (2017) 63, 75–83, https://doi.org/10.1016/j.eneco.2017.01.019, 2-s2.0-85013077975.
10.1016/j.eneco.2017.01.019
Web of Science® Google Scholar
8 Chang Z. Z., Li H. X., and Mao P., The Impact of Local Government Debt on Corporate Financing Structure: Evidence From A-Share Listed Companies, Reform. (2023) no. 6, 105–125.
Google Scholar
9 Xu Z., Su C., Meng L., Sheng L., and Feng Y., Bargaining Power of New-Energy Enterprises in China’s Credit-Financing Market, Environmental Science and Pollution Research. (2023) 30, no. 6, 16661–16672, https://doi.org/10.1007/s11356-022-23118-1.
10.1007/s11356-022-23118-1
PubMed Web of Science® Google Scholar
10 Wei Q., A Study on the Loan Bargaining Power of Listed Companies in China, Contemporary Finance & Economics. (2017) no. 12, 58–67.
Google Scholar
11 Mehboob M. Y., Ma B. J., Mehboob M. B., and Zhang Y. S., Does Green Finance Rreduce Environmental Degradation? The Role of Green Innovation, Environmental Tax, and Geopolitical Risk in China, Journal of Cleaner Production. (2024) 435, https://doi.org/10.1016/j.jclepro.2023.140353, 140353.
10.1016/j.jclepro.2023.140353
Google Scholar
12 Wei Q. and Zeng S. R., The Impact of Bank’s Franchise on Loan Price: With an Estimation About Bargaining Power of China’s Commercial Bank, Journal of Industrial Engineering and Engineering Management. (2017) 31, no. 4, 101–108.
CAS Google Scholar
13 Li Z., Liao G., Wang Z., and Huang Z., Green Loan and Subsidy for Promoting Clean Production Innovation, Journal of Cleaner Production. (2018) 187, 421–431, https://doi.org/10.1016/j.jclepro.2018.03.066, 2-s2.0-85044535363.
10.1016/j.jclepro.2018.03.066
Web of Science® Google Scholar
14 Hall B. H. and Lerner J., The Financing of R&D and Innovation, Handbook of the Economics of Innovation 1, 2010, Elsevier, 609–639.
10.1016/S0169-7218(10)01014-2
Google Scholar
15 Meng Z., Sun H., and Liu X., Impact of Green Fiscal Policy on the Investment Efficiency of Renewable Energy Enterprises in China, Environmental Science and Pollution Research. (2022) 29, no. 50, 76216–76234, https://doi.org/10.1007/s11356-022-20832-8.
10.1007/s11356-022-20832-8
PubMed Web of Science® Google Scholar
16 Myers S. C. and Majluf N. S., Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have, Journal of Financial Economics. (1977) 5, no. 2, 147–175, https://doi.org/10.1016/0304-405X(77)90015-0, 2-s2.0-49449125071.
10.1016/0304-405X(77)90015-0
Google Scholar
17 Freel M. S., Are Small Innovators Credit Rationed?, Small Business Economics. (2007) 28, no. 1, 23–35, https://doi.org/10.1007/s11187-005-6058-6, 2-s2.0-33846260823.
10.1007/s11187-005-6058-6
Web of Science® Google Scholar
18 Soundarrajan P. and Vivek N., Green Finance for Sustainable Green Economic Growth in India, Agricultural Economics. (2016) 62, no. 1, 35–44.
Web of Science® Google Scholar
19 Scholtens B. and Dam L., Banking on the Equator. Are Banks That Adopted the Equator Principles Different From Non-Adopters?, World Development. (2007) 35, no. 8, 1307–1328, https://doi.org/10.1016/j.worlddev.2006.10.013, 2-s2.0-34547101439.
10.1016/j.worlddev.2006.10.013
Web of Science® Google Scholar
20 Lian L. L., Does Green Credit Influence Debt Financing Cost of Business?—A Comparative Study of Green Businesses and ’Two High’ Businesses, Journal of Finance and Economics. (2015) 30, no. 5, 83–93.
Google Scholar
21 Xu X. and Li J., Asymmetric Impacts of the Policy and Development of Green Credit on the Debt Financing Cost and Maturity of Different Types of Enterprises in China, Journal of Cleaner Production. (2020) 264, https://doi.org/10.1016/j.jclepro.2020.121574, 121574.
10.1016/j.jclepro.2020.121574
Web of Science® Google Scholar
22 Niu H. P., Zhang X. Y., and Zhang P. D., Institutional Change and Effect Evaluation of China’s Green Finance Policy: Taking the Empirical Study of Green Credit as an Example, Management Review. (2020) 32, no. 8, 3–12.
Google Scholar
23 Guo Y. and Fang F., Green Effects of the New Monetary Policy Collateral Framework, Financial Research. (2021) 1, 91–110.
Google Scholar
24 Zhou G., Liu C., and Luo S., Resource Allocation Effect of Green Credit Policy: Based on DID Model, Mathematics. (2021) 9, no. 2, https://doi.org/10.3390/math9020159, 159.
10.3390/math9020159
Web of Science® Google Scholar
25 Zhang D. and Zhong S., Has the Green Credit Policy Promoted Enterprise Green Technology Innovation?, Journal of Dalian University of Technology (Social Sciences). (2024) 1, 1–14.
Google Scholar
26 Wang B. J., Cui L. Y., and Wang Z. L., et al.Research on the Impact of Green Finance Pilot Policies on Technological Innovation of New Energy Enterprises, Journal of Nanjing University of Finance and Economics. (2024) 2, 67–77.
Google Scholar
27 Wu Y., Wu F., and Lian J., et al.Analysis of Excessive Credit Risk, China Soft Science. (2000) 10, 43–47.
Google Scholar
28 Shen X. H. and Identification Bank Risk, Government Subsidies, and Corporate Debt Costs—An Empirical Study Based on Corporate Data of Listed Companies From 2007 to 2012, Finance and Trade Economics. (2014) 2012, no. 9, 62–71.
Google Scholar
29 Wang G. G., Research on R&D Subsidy Signal Financing Mechanism, 2017, Shanghai Jiao Tong University.
Google Scholar
30 Zeng Y., Guo P., and Li B., The Impact of Government Subsidies on Company Debt Financing: New Evidence From Chinese Listed Companies, Emerging Markets Finance and Trade. (2021) 57, no. 7, 1843–1858, https://doi.org/10.1080/1540496X.2019.1627519, 2-s2.0-85068569162.
10.1080/1540496X.2019.1627519
Web of Science® Google Scholar
31 Liu R. Z. and Du J. X., Signal Effect, Government Subsidies, and Financing Constraints, Journal of Harbin University of Commerce (Social Science Edition). (2022) 2, 58–70.
Google Scholar
32 Yao L. X., Peng R. J., and Xu J. B., Asymmetric Effects of Heterogeneous Government Subsidies on Corporate R&D Investment and R&D Output, Contemporary Finance & Economics. (2022) 10, 40–51.
Google Scholar
33 Yang D. X., Study on the Development of Renewable Energy Power Industry and Government Subsidy Mechanism, 2018, Jinan University.
Google Scholar
34 Yang D.-X., Chen Z.-Y., Yang Y.-C., and Nie P.-Y., Green Financial Policies and Capital Flows, Physica A: Statistical Mechanics and Its Applications. (2019) 522, 135–146, https://doi.org/10.1016/j.physa.2019.01.126, 2-s2.0-85061395792.
10.1016/j.physa.2019.01.126
Web of Science® Google Scholar
35 Hong X. J., Lin X., and Chen L. F., Study on the Effect of Local Green Credit Discount Policy: Based on the Perspective of Fiscal and Financial Policy Coordination, Chinese Industrial Economy. (2023) 9, 80–97.
Google Scholar
36 Yan W., Chen S., and Tian Z., Audit Pricing Regulation and Transaction Surplus—A Natural Experiment of the Failure of Audit Pricing Regulation in 2010, Statistical Research. (2020) 37, no. 7, 93–103.
Google Scholar
37 Ma W. T. and Zhang P., Implicit Government Guarantee, Marketization Process and Credit Allocation Efficiency, Public Finance Research. (2021) 8, 91–106.
Google Scholar
38 Lim C. Y., Wang J., and Zeng C., China’s ’Mercantilist’ Government Subsidies, the Cost of Debt and Firm Performance, Journal of Banking & Finance. (2018) 86, 37–52, https://doi.org/10.1016/j.jbankfin.2017.09.004, 2-s2.0-85033493166.
10.1016/j.jbankfin.2017.09.004
Web of Science® Google Scholar
39 Ma Y. C., Sha Y. Z., Wang Z. L., and Zhang W. J., The Effect of the Policy Mix of Green Credit and Government Subsidy on Environmental Innovation, Energy Economics. (2023) 118, https://doi.org/10.1016/j.eneco.2023.106512, 106512.
10.1016/j.eneco.2023.106512
Web of Science® Google Scholar
40 Takalo T. and Tanayama T., Adverse Selection and Financing of Innovation: Is There a Need for R&D Subsidies?, The Journal of Technology Transfer. (2010) 35, no. 1, 16–41, https://doi.org/10.1007/s10961-009-9112-8, 2-s2.0-76049103579.
10.1007/s10961-009-9112-8
Web of Science® Google Scholar
41 Meuleman M. and De Maeseneire W., Do R&D Subsidies Affect SMEs’ Access to External Financing?, Research Policy. (2012) 41, no. 3, 580–591, https://doi.org/10.1016/j.respol.2012.01.001, 2-s2.0-84857051377.
10.1016/j.respol.2012.01.001
Web of Science® Google Scholar
42 Xie Y., Has Interest Rate Marketization Improved the Bargaining Power of Corporate Loans?—An Empirical Study Based on the Bilateral Stochastic Frontier Model, Shanghai Finance. (2023) 6, 17–29, 45.
Google Scholar
43 Xu Z., Meng L., He D., Shi X., and Chen K., Government Support’s Signaling Effect on Credit Financing for New-Energy Enterprises, Energy Policy. (2022) 164, no. 164, https://doi.org/10.1016/j.enpol.2022.112921, 112921.
10.1016/j.enpol.2022.112921
Web of Science® Google Scholar
44 Song M., Xie Q., and Shen Z., Impact of Green Credit on High-Efficiency Utilization of Energy in China Considering Environmental Constraints, Energy Policy. (2021) 153, no. 22, https://doi.org/10.1016/j.enpol.2021.112267, 112267.
10.1016/j.enpol.2021.112267
Web of Science® Google Scholar
45 Lin B. Q. and Pan T., How Does Environmental Regulation Affect the Development of Green Credit?, China Population, Resources and Environment. (2022) 32, no. 8, 50–61.
CAS Google Scholar
46 Chen H., Zhang Y., and Liu D. X., Government Subsidies, Tax Incentives and Corporate Innovation Performance: An Empirical Study at Different Life Cycle Stages, Nankai Business Review. (2019) 22, no. 3, 187–200.
CAS Google Scholar
47 Liu J. K., Liu J. X., and Chao Y., Green Credit and Low-Carbon Transformation: Capital Consolidation or Technology Innovation?, Journal of Quantitative & Technological Economics. (2024) 41, no. 6, 151–171.
Google Scholar
48 Kochiyama T. and Nakamura R., Debt Covenants in Japanese Loan Markets: In Comparison With the Traditional Relationship Banking, Accounting & Finance. (2021) 61, no. 1, 305–334, https://doi.org/10.1111/acfi.12568.
10.1111/acfi.12568
Web of Science® Google Scholar
49 Zhang Y. G., Dou R. Y., and Bai Y. J., Measurement on China’s Green Low-Carbon Circular Developing Economic System Construction, Quantitative Economics and Technical Economics Research. (2020) 37, no. 8, 83–102.
Google Scholar
50 Guo Y., Signal Transmission Mechanism of Government Innovation Subsidy and Enterprise Innovation, China’s Industrial Economy. (2018) 9, 98–116.
Google Scholar
51 Xu J. B., Qi Y. F., and Xu C. X., The Synergistic Effect and Mechanism of Government Procurement and Subsidies on Enterprise Innovation, Reform. (2024) 4, 144–163.
Google Scholar
52 Feng T. and Zhu Z. Y., A Study on the Impact of Heterogeneous Government Grants on Firm Resilience, Journal of Southwest University (Social Science). (2024) 50, no. 1, 144–155.
CAS Google Scholar
53 Yin S. and Zhao Y., Digital Green Value Co-Creation Behavior, Digital Green Network Embedding and Digital Green Innovation Performance: Moderating Effects of Digital Green Network Fragmentation, Humanities and Social Sciences Communications. (2024) 11, no. 1, https://doi.org/10.1057/s41599-024-02691-5, 228.
10.1057/s41599-024-02691-5
Google Scholar
54 Yin S. and Zhao Y., An Agent-Based Evolutionary System Model of the Transformation From Building Material Industry (BMI) to Green Intelligent BMI Under Supply Chain Management, Humanities and Social Sciences Communications. (2024) 11, no. 1, https://doi.org/10.1057/s41599-024-02988-5, 468.
10.1057/s41599-024-02988-5
Google Scholar

All articles

Impact of Green Credit and Heterogeneous Government Subsidies on New Energy Enterprises’ Bargaining Power in China’s Credit Market

Abstract

1. Introduction

2. Literature Review and Theoretical Analysis

2.1. Literature Review

2.2. Theoretical Analysis and Hypothesis

3. Research Design

3.1. Measurement of Bargaining Power and Analysis of Influencing Factors

3.1.1. Measurement of Bargaining Power

3.1.2. Analysis of Factors Influencing Bargaining Power

3.2. Sample, Variables, and Data

3.2.1. Sample Selection

3.2.2. Variable Selection

3.2.2.1. Variables for Measuring Bargaining Power

3.2.2.2. Variables for Analyzing Policy Factors Influencing Bargaining Power

3.2.3. Data Sources

4. Empirical Analysis

4.1. Descriptive Statistics of Variables

4.2. Measurement Estimation Results: Bargaining Power of NEEs and Banks

4.2.1. Estimation Results of the Credit Bargaining Model

4.2.2. Analysis of Temporal Trends in Bargaining Power and Surplus

4.3. Analysis of the Impact of Policy Factors on Bargaining Power and Surplus

4.3.1. Individual Impact of Policies

4.3.1.1. Individual Impact of GCPs

4.3.1.2. Individual Impact of Subsidies

4.3.2. Synergistic Impact of Policies

4.4. Robustness Test

4.5. Endogeneity Treatment

5. Discussion

5.1. Theoretical Contributions

5.2. Practical Implications

5.3. Limitations and Future Research Directions

6. Conclusions

Nomenclature

Conflicts of Interest

Author Contributions

Funding

Acknowledgments

Open Research

Data Availability Statement

References

Figures

References

Related

Information