How Does a Major Corporate Customer’s ESG Performance Drive the Supplier’s Green Innovation?

Abstract

In the current climate of intensifying global demands for sustainability, the influence of major corporate customers in driving environmental initiatives across supply chains has emerged as a pivotal focus of academic research and practical application. This study investigates how these influential customers can drive green innovation along their supply chains, which remains under-explored in the previous literature. The study focuses on critical indicators such as the green patent application to measure green innovation outcomes. Leveraging a panel dataset of paired Chinese listed firms from 2009 to 2022, we examine the impact of customers’ environmental, social, and governance (ESG) performance on suppliers’ green innovation activities. Our model is integrated with fixed effects for both industry and year. Our analysis reveals that customers with firm ESG commitments significantly enhance their suppliers’ green innovation outcomes, particularly when they possess greater bargaining power and advanced green innovation capabilities. Additionally, we identify pressure and collaboration as critical mechanisms through which these effects are realized. These findings extend the discourse on sustainable supply chain management by highlighting the crucial role of customer-driven ESG initiatives in fostering upstream innovation, offering valuable insights for scholars and practitioners.

1. Introduction

With the worsening situation of global environmental pollution, natural disasters, and climate change, stakeholders such as investors, regulators, and consumers increasingly emphasize the role of companies in sustainability. This growing emphasis has made investments in environmental, social, and governance (ESG) initiatives strategically fundamental for firms. ESG initiatives are crucial not only for mitigating ecological risks and complying with regulatory requirements but also for enhancing long-term competitiveness through pathways such as improved resource efficiency and strengthened stakeholder relationships. These initiatives ultimately contribute to financial and non-financial performance, reinforcing their strategic importance.

Among the various initiatives to advance ESG, green innovation (GI) has emerged as a critical strategic initiative for companies aiming to achieve sustainable development, owing to its long-term competitive advantages and capacity to address global environmental challenges. Unlike traditional innovation, which primarily seeks economic gains, green innovation involves transforming or generating new ideas, technologies, practices, products, services, processes, or management systems to enhance resource and energy efficiency, reduce environmental risks and pollution, and contribute to sustainable development. It plays a critical role in addressing ecological issues [1,2].

Although green innovation benefits a company’s long-term growth, it is not conducive to short-term performance [3,4]. Companies’ willingness to invest in green innovation tends to decrease [5] due to the double externalities of economic and environmental spillover effects [4,6]. This phenomenon is observed because green innovation typically requires substantial upfront investments in new technologies, processes, and training. These investments can strain a company’s financial resources in the short term, especially before the green innovations yield significant benefits. Enterprises also face agency problems in pursuing green innovation [3,4,7]. Decision-makers may reduce green innovation efforts in pursuit of short-term interests.

Moreover, the market adoption of green products and technologies can be gradual, meaning that the revenue benefits of such innovations may take time to materialize. Therefore, while the long-term benefits of green innovation, such as enhanced resource efficiency, compliance with regulatory standards, and improved brand reputation, are well documented [3,4], these advantages often accrue over an extended period, leading to a disparity between short-term costs and long-term gains. Therefore, how to enhance firms’ performance in green innovation has been a widely discussed and important topic in recent years [8,9,10,11].

Previous research has investigated the antecedents of corporate green innovation from firm-specific and external environmental factors such as corporate internal culture [12] and leadership [13,14], technological progress [8,15,16], market demand [15,17], regulatory pressure [15], and policy support [18,19], etc. Due to the dual externalities and the complex process of green innovation, current research has primarily focused on the driving forces related to various stakeholders [20], including the government, research institutions, and consumers’ environmental awareness [21,22,23,24,25]. While these studies provide valuable insights, limited research specifically examines the role of corporate customers, who are among the most influential stakeholders within supply chains.

In practice, corporate customers profoundly impact supplier firms’ investment strategies. The situation becomes particularly critical when a customer constitutes a substantial proportion of the suppliers’ overall revenue [26]. Given that corporate customers have significant leverage over their suppliers, understanding how they can facilitate green innovation among these suppliers is crucial.

However, the current research on the effect of major customers’ characteristics and suppliers’ investment in sustainability is limited in scope, with inconclusive empirical findings that warrant investigation. Some research finds that companies with higher customer concentration have a notable adverse effect on their suppliers’ CSR performance [17] and engagement in ESG activities [26]. On the other hand, some research finds that if customers engage in CSR activities, it can positively influence their suppliers’ CSR performance [27]. Additionally, when the government becomes a major customer, it can facilitate improvements in supplier environmental responsibility [28] and ESG performance [29].

This study addresses this gap by focusing on the typical corporate sustainability development investment, green innovation, which is essential for corporate sustainability development and contributes to long-term competitive advantages and sustainability [30]. We mainly investigate whether the major corporate customer’s ESG performance can facilitate its supplier’s green innovation and the mechanism through this process.

To examine our research question, we utilize data from the China Stock Market and Accounting Research (CSMAR) database, constructing a comprehensive dataset of 1033 unique customer–supplier dyads from 2009 to 2022. Our sample in the analyses consists of 1033 unique supplier–customer pairs. The supplier’s green innovation capability is measured using green patent applications sourced from the China Research Data Services (CNRDS) database, while ESG performance indicators are collected from the Sino-Securities Index Information Service (Shanghai) Co., Ltd. (Shanghai, China).

To analyze the relationship between the major corporate customer’s ESG performance and its supplier’s green innovation, we employed a fixed-effects regression model, controlling for potential time and industry-fixed effects. This approach allows us to account for unobserved heterogeneity across different years and industries, thereby mitigating potential biases in the regression results that could arise from these factors. Moreover, to ensure the robustness of our findings, we conducted a series of robustness checks, including alternative variable measurements, extending the lag period, the instrumental variable method, the Heckman two-stage procedure, and a falsification test utilizing the Propensity Score Matching (PSM) approach. These additional analyses reinforce the validity of our results.

This paper makes three contributions to the current literature. First, it enriches the literature on the impact of the supplier–customer nexus on corporate decisions by investigating whether and how the major customers’ sustainability development commitment can enhance their suppliers’ green innovation, an essential decision a firm makes regarding investment in sustainability. Second, it explores the antecedents of improved corporate green innovation from a demand-side perspective. Previous research has primarily focused on the antecedents of green innovation from the supply side, such as firm-specific characteristics and external environmental factors, overlooking the factors from their demand side. This paper expands the research on the antecedents of green innovation by focusing on the major customer stakeholder group. Third, we contribute by providing novel evidence on how major customers’ ESG performance can improve suppliers’ green innovation. We demonstrate that pressure and collaboration from major corporate customers drive suppliers’ green innovation, contributing to ESG and green innovation literature by integrating demand- and supply-side perspectives to explain why firms engage in sustainability.

The remaining sections of the paper are structured as follows: Section 2 presents the theoretical background and hypotheses. Section 3 outlines the methodology. Section 4 presents the empirical findings. Section 5 explores potential underlying mechanisms. Section 6 conducts a heterogeneity analysis based on industry, firm nature, and geographical distance. Section 7 discusses and Section 8 concludes.

2. Theoretical Background and Hypotheses

2.1. Major Corporate Customers’ Power through Sustainable Supply Chain Management

Given the escalating global environmental degradation and heightened environmental consciousness, sustainable supply chain management has emerged as a critical framework for integrating ecological considerations into supply chain operations. It has become an inevitable choice for the long-term development of enterprises [31]. Grounded in the broader framework of the sustainable concept, which emphasizes the need for firms to operate in ways that are not only economically viable but also environmentally and socially responsible [32], sustainable supply chain management encompasses a wide range of practices to reduce the environmental impact of supply chain activities. These practices include green procurement, environmentally friendly technologies, production processes, transportation, and waste management [33].

Empirical evidence has supported the effectiveness of these practices. For instance, research found that firms implementing green procurement practices significantly reduced their environmental impact, particularly regarding waste generation and resource use [34]. A recent meta-analysis found that firms adopting sustainable production technologies reduced their carbon footprints and enhanced their operational efficiency and cost savings [35].

Additionally, sustainable supply chain management practices can strengthen firms’ competitiveness. For example, a longitudinal study reported that firms integrating environmental practices into their supply chains can improve their innovation and market performance [36].

Major corporate customers are essential stakeholders in sustainable supply chain construction. As important stakeholders, downstream companies are motivated to engage in sustainable supply chain management because poor environmental performance by upstream suppliers can negatively impact these companies [37,38]. For example, customer companies may experience revenue declines or reputation damage due to suppliers’ irresponsible behavior [39]. Companies implementing sustainable supply chain management practices can bolster the firm’s reputation and customer loyalty, providing an advantage in the market. Additionally, these practices can help companies experience cost savings from reduced waste and increased resource efficiency [40]. Major customers, particularly those with higher ESG commitments, are pivotal in driving sustainable supply chain management initiatives. By setting high environmental and social standards, these customers demand sustainable practices throughout the supply chain [41], ensuring that the entire supply chain contributes to the company’s sustainability goals.

However, it is essential to recognize that the primary customer’s strong ESG performance may also impose specific challenges on the supplier’s green innovation efforts. Firstly, complying with stringent ESG standards set by major customers can increase operational costs and resource allocation for suppliers, especially for small and medium-sized enterprises with limited capabilities. This financial pressure might hinder their ability to invest adequately in innovative green technologies and processes [42]. Secondly, the mismatch between the major customer’s ESG expectations and the supplier’s existing operational practices can result in implementation difficulties and inefficiencies, potentially leading to delays and reduced competitiveness. Lastly, excessive reliance on significant customers’ ESG directives may limit suppliers’ autonomy and creativity in developing tailored green innovations that suit their specific contexts and markets [32]. These challenges suggest that while major customers’ ESG performance can drive suppliers towards greener practices, the impact on innovation may vary depending on the suppliers’ resources, capabilities, and contextual factors.

Based on the above two aspects, in the following theoretical hypothesis, we mainly explore how the ESG performance of major customer companies can promote the green innovation of suppliers.

2.2. The Major Customer’s ESG Performance and Supplier’s Green Innovation

Previous research found that suppliers’ environmental management issues can significantly harm buyers’ environmental reputations [43]. As an essential stakeholder, major customers with high ESG orientation may influence suppliers by exerting both pressure and providing support, thereby encouraging them to prioritize sustainability and engage in green innovation.

First, major customers with strong ESG commitments often set rigorous environmental and social standards for their suppliers. These standards act as external pressures, referred to as “green procurement pressure”, that compel suppliers to adopt greener practices and develop innovative solutions to meet these expectations [42]. For instance, large multinational corporations frequently require their suppliers to adhere to sustainability criteria that align with global environmental standards, thus pushing suppliers toward green innovation [41].

Second, these ESG-oriented customers also play a supportive role by providing critical resources and engaging in collaborative efforts, known as “green innovation collaboration”. This support can include technical assistance, financial incentives, or direct involvement in joint innovation initiatives [44]. By leveraging these resources and collaborating with their customers, suppliers are better equipped to overcome barriers such as high initial costs and technological uncertainties associated with green innovation [45]. This collaboration not only facilitates compliance with current sustainability standards but also fosters a culture of ongoing innovation aligned with the evolving demands of sustainability [46].

Thus, we posit the following:

Hypothesis 1.

The major customer’s ESG performance positively impacts its supplier’s green innovation (GI) capability.

2.3. The Moderating Role of the Major Customer’s Bargaining Power

The relationship between a major corporate customer’s ESG practice and a supplier’s green innovation can be significantly influenced by the customer’s market share, which reflects the customer’s bargaining power in the industry. Bargaining power enables these customers to exert substantial control over their suppliers, meaningfully shaping their behaviors and practices.

First, customers with a higher market share can impose more stringent ESG requirements that suppliers must follow due to the significant business opportunities. The high market share means non-compliance could result in a considerable loss of business for suppliers, creating a strong incentive for them to adopt green practices and innovate to meet these standards [43]. The heightened pressure to conform to ESG expectations is more pronounced when the customer holds a dominant market position, as suppliers cannot afford to lose such a critical business partner [41]. Suppliers are often willing to invest in green innovations if they believe it will secure their relationship with a powerful customer who can offer long-term contracts or other significant business advantages [41]. Thus, the customer’s bargaining power acts as a catalyst, motivating suppliers to align their innovation efforts with the ESG goals of their influential customers.

Second, customers with high bargaining power can foster an environment of increased scrutiny and accountability. Suppliers may face regular audits and performance evaluations related to their ESG practices, which creates continuous pressure to innovate and improve [44]. This ongoing oversight ensures that suppliers consistently work towards higher environmental performance and innovation standards, driven by the need to meet the stringent expectations of their influential customers.

Third, the reputation impact of a major customer with a high market share can amplify the importance of complying with ESG standards for suppliers. Suppliers know that maintaining a relationship with a prominent industry player can enhance their market standing and open further business opportunities. Therefore, they are more motivated to innovate and improve their green practices to align with the customer’s high ESG standards [46]. This reputation effect motivates suppliers to pursue green innovation, knowing that success can lead to broader recognition and new business prospects. Thus, we posit the following:

Hypothesis 2.

The major corporate customer’s bargaining power amplifies the positive relationship between its ESG performance and its supplier’s GI capability.

2.4. The Moderating Role of the Major Customer’s Green Innovation Capability

The above ESG–GI nexus can also be significantly influenced by the major corporate customer’s green innovation (GI) capability. When a major customer has a high ability for green innovation, it can provide more effective support and exert more substantial influence, thereby enhancing the positive impact of its ESG performance on suppliers’ GI efforts.

First, major customers with higher green innovation capabilities are better equipped to provide their suppliers with advanced technical support and resources. This includes sharing green technologies, best practices, and specialized knowledge to help suppliers overcome green innovation’s technical and financial barriers [20]. The superior technical assistance these customers provide reduces the complexity and uncertainty of implementing green innovations for suppliers, making it easier and more feasible for them to meet the high ESG standards set by the customer.

Second, customers with solid green innovation capabilities can more effectively demonstrate the benefits and feasibility of green innovations. When suppliers see tangible examples of successful green innovations implemented by their major customers, they reduce perceived risks and increase their confidence in adopting similar practices [47]. This demonstration effect can motivate suppliers to engage in green innovation.

Third, customers with robust green innovation capabilities can foster a culture of innovation and continuous improvement across their supply chain. These customers often initiate joint innovation projects, provide platforms for collaborative R&D, and facilitate ongoing interactions between suppliers and other stakeholders [48]. This collaborative environment enhances the exchange of ideas and technologies and creates a supportive network that encourages and sustains green innovation efforts among suppliers.

Therefore, a major customer’s high green innovation capability amplifies its ability to influence suppliers by providing more substantial support, reducing barriers, and fostering an innovative culture. This results in a stronger positive ESG–GI nexus. Thus, we propose the following hypothesis:

Hypothesis 3.

The major customer’s GI capability strengthens the positive relationship between its ESG performance and its supplier’s GI capability.

3. Methodology

3.1. Sample and Data

We identified customer–supplier linkages using data on the top five customers paired with suppliers from the China Stock Market and Accounting Research (CSMAR) database. This database provides extensive information on company relationships, enabling us to accurately map the connections between major customers and their suppliers. Following previous research [49], we construct the supplier–customer year dataset. For instance, in the current year (2009), the supplier (S1) may have multiple customers (C1, C2). Therefore, we create the observations S1-C1-2009 and S1-C2-2009.

We collected yearly green patent applications for the firms in our sample from the China Research Data Services (CNRDS) database from 2009 to 2022 to measure green innovation. Green patents are a widely recognized indicator of a firm’s commitment to environmental sustainability. Despite challenges such as the potential non-representation of all types of innovation, patent data are valued as an intermediate indicator and meaningful tool for innovation measurement due to their public accessibility and international standardization [50].

Our ESG performance data are sourced from the Sino-Securities Index Information Service (Shanghai) Co., Ltd. database, which provides ratings based on three primary dimensions: environmental, social, and governance (ESG). Its rating system includes 16 secondary indicators, 44 tertiary indicators, over 70 quaternary indicators, and more than 300 underlying data points. Through its detailed and comprehensive categorization of topics and issues, the China Securities ESG Database offers companies a scientific and systematic framework for evaluating sustainable development. This framework assists investors and stakeholders in gaining a holistic understanding of a company’s non-financial performance and its capacity for sustainable development. It combined global mainstream methods with Chinese capital market characteristics, widely used in previous research to study Chinese sustainability problems [29,51]. Higher ESG scores indicate better environmental, social, and governance performance. These scores provide a standardized and reliable measure of the companies’ commitment to sustainability.

Data on green procurement and green innovation collaboration were gathered through content analysis of the companies’ annual and CSR/ESG reports. Appendix A shows the detailed method and vocabulary. This qualitative analysis allowed us to capture the extent and nature of the companies’ green initiatives within their supply chains.

The data on firm size, age, and nature were collected from the Wind Economic database, and the financial data on financial leverage, return on assets, R&D intensity, corporate growth, Tobin’s Q, percentage of ownership held by the largest shareholders, and CEO duality were collected from the CSMAR database. The data on the proportion of independent directors and the percentage of government subsidies to operating income were collected from the CSMAR database and were manually computed. By integrating multiple data sources and methodologies, we aim to provide a nuanced understanding of the factors influencing green innovation within the context of the supply chain.

Based on data availability, we selected only paired samples of companies where both customers and suppliers are listed entities. This criterion ensures the robustness and reliability of our dataset, as listed companies must adhere to stringent reporting standards, thereby providing comprehensive and verifiable data. In addition, we excluded (1) firms subjected to special treatment, (2) suppliers in the financial sector, and (3) observations with missing or extreme outlier data. All continuous variables were winsorized at the 1% level to mitigate the impact of extreme values on our regression results. We chose winsorizing because it allows us to mitigate the influence of extreme values without completely discarding them. This method has also been used in most previous studies as it retains all observations within the dataset, thereby minimizing the potential bias that could arise from excluding outliers altogether. By adjusting the extreme values to a specific percentile, we ensure that the data remain representative of the original sample while reducing the undue impact of outliers on the results. The final sample includes 1033 pairs of observations spanning from 2009 to 2022.

3.2. Measurements

3.2.1. Dependent Variable

Suppliers’ green innovation (LnGreenInv^S). The green innovation is measured by the number of green patents a firm applies for each year. In this study, we employ patent applications as the primary indicator of green innovation for multiple reasons. First, patent applications are filed at a relatively early stage of the innovation process, providing a timely reflection of a firm’s innovation activities. This immediacy is crucial for capturing the effects of strategic initiatives of the major customer’s ESG practices on the supplier’s timely response and effort in green innovation. Second, this metric involves verification and validation by a public institution to certify the process, making it more rigorous compared to some other innovation metrics [50]. Third, patent applications offer a comprehensive view of a firm’s innovation intentions, including exploratory innovations that may not ultimately lead to published patents but are essential for understanding the full scope of a firm’s innovative efforts. While published patents provide valuable insights, particularly for tracing the impact of innovations through citations, they often reflect innovations that have undergone additional legal scrutiny and may not capture the full spectrum of a firm’s innovative activities. Finally, previous research commonly uses patent applications, allowing for more robust and consistently available analysis in our research context [23,24,25].

Like many other count variables, patent counts are often highly skewed, with a small number of firms holding a disproportionately large number of patents. This skewness can lead to issues in statistical modeling, such as non-normality and heteroscedasticity. Log transformation is a widely used method to address skewness by compressing the range of the data, which makes the distribution more normal and stabilizes the variance across observations. Following the previous research [23,24], we use the logarithm of the count of green patents plus one as our proxy variable. Some other variables, such as customers’ ESG scores, firm size, and firm age, are transformed for the same reason.

We also construct an alternative measure of green innovation focusing on green innovation efficiency, which is discussed in robustness checks. Considering the time lag effect of innovation, we estimated the regression using a one-year time lag, a method widely applied in the previous literature [23,24,25]. We also added regression data with lags of two and three periods based on the previous literature [24,49].

3.2.2. Independent Variable

Corporate customers’ ESG performance (LnESG^C). The ESG performance of listed companies has become a crucial indicator of their sustainable competitiveness. In this study, we utilize the ESG score data published by the Sino-Securities Index Information Service (Shanghai) Co., Ltd., where higher scores denote superior ESG performance.

3.2.3. Moderating Variables

Bargaining power of corporate customer (Bar_power^C). Drawing from previous studies [52], bargaining power is closely linked to switching costs and the relative strength of suppliers and customers within their respective industries. To measure Bar_power^C, we capture the switching costs faced by suppliers when seeking alternative customers by calculating the customers’ market share in their sector. It takes the client’s sales revenue proportion to the overall industry sales revenue within a specific Standard Industrial Classification (SIC) code category [53]. A higher customer market share indicates greater supplier dependence on that customer. The data are sourced from the CSMAR database.

Green innovation capability of corporate customer (LnCuGreInv^C). The measurement method used for corporate customers’ green innovation capability is identical to that used for our dependent variable, focusing on the number of green patent applications. We also use the logarithm of the count of green patents plus one as our proxy variable.

3.2.4. Control Variables

We also control several variables that may impact corporate green innovation. Precisely, we control for firm size (Size^S), firm age (FirmAge^S), financial leverage (Lev^S), return on assets (ROA^S), R&D intensity (RD^S), corporate growth (Growth^S), Tobin’s Q (TobinQ^S), the proportion of independent directors (Indep^S), percentage of ownership held by the largest shareholders (Top1^S), the percentage of government subsidies to operating income (Gov_subsidy^S), CEO duality (Dual^S), and firm nature (SOE^S). In addition, the suppliers’ ESG scores (LnESG^S) were also controlled. Detailed measurements of these variables are provided in

Table 1. Measurements of variables.

Variables	Measurements
LnGreenInvS	The natural logarithm of the count of suppliers’ green patent applications plus one.
LnESGC	The natural logarithm of the corporate customer’s ESG scores plus one.
LnCuGreInvC	The natural logarithm of the count of customers’ green patent applications plus one.
Bar_powerC	Customer enterprises operating income/total operating income of all enterprises in the same industry.
LnGreenpurchaseC	The natural logarithm of the score index of corporate customer’s green procurement plus one. For detailed specific vocabulary, please see Appendix A.
CollaborationC	The score index calculated by content analysis. For detailed measurement and specific vocabulary, please see Appendix A.
LnESGS	The natural logarithm of the suppliers’ ESG scores plus one.
SizeS	The natural logarithm of the firm’s total assets in year t.
FirmAgeS	The natural logarithm of one plus firm’s age.
LevS	The year-end ratio of the firm’s total liabilities to total assets.
ROAS	Operating income before depreciation divided by the book value of the firm’s total assets at the end of the fiscal year.
RDS	The ratio of R&D expenditures and total sales.
GrowthS	The growth rate of the company’s operating income.
TobinQS	(Market value of equity + Total book liabilities)/(Book value of total assets)
IndepS	The ratio of independent directors to the total board size of the firm.
Top1S	The percentage of ownership held by the firm’s largest shareholder.
Gov_subsidyS	The ratio of government subsidies and operating income.
DualS	A dummy variable that equals 1 if the dual role of the firm’s board chairman and 0 otherwise.
SOES	A dummy variable that equals 1 if the firm is a state-owned enterprise and 0 otherwise.

.

Among the various control variables used in our study, we recognize that there is ongoing debate regarding the best practices for calculating Tobin’s Q. This metric, intended to capture the ratio of a firm’s market value to the replacement cost of its assets, has been subject to different interpretations and calculation methods.

Specifically, we used the following formula:

$$\mathrm{Tobin}’\mathrm{s} \mathrm{Q}={\displaystyle \frac{\mathrm{Market} \mathrm{Value} \mathrm{of} \mathrm{Equity}+\mathrm{Total} \mathrm{Book} \mathrm{liabilities}}{\mathrm{Book} \mathrm{Value} \mathrm{of} \mathrm{Total} \mathrm{Assets}}}$$

This approach is widely used due to its practicality and data availability. It allows us to capture the market’s valuation of the firm relative to the replacement cost of its assets. However, we recognize that this method may not fully capture the theoretical concept of Tobin’s Q, particularly concerning the accurate valuation of liabilities and assets. While Tobin’s Q is a secondary control variable in our analysis, we acknowledge that alternative calculations could yield different results.

3.3. Models

In this study, we constructed Equation (1) to explore the impact of major customers’ ESG performance on their suppliers’ green innovation. Our model is integrated with fixed effects for both industry and year, which are crucial for addressing potential unobserved heterogeneity and ensuring the robustness of our findings. The following regression model was estimated to examine Hypothesis 1:

$${{\mathrm{LnGreenInv}}^{\mathrm{S}}}_{i,t/t+1}=\alpha +{\beta }_1 {{\mathrm{LnESG}}^{\mathrm{C}}}_{i,t}+\gamma {\mathrm{Controls}}_{i,t}+\sum {\mathrm{Year}}_t+\sum {\mathrm{Industry}}_i+{\epsilon }_{i,t}$$

(1)

Equation (1) represents the baseline relationship between the major corporate clients’ performance in ESG and suppliers’ outcomes in green innovation. LnGreenInv^S_i,t/t+1 represents the supplier’s green innovation. LnESG^C_i,t refers to the ESG scores of the focal corporate customer. Controls_i,t includes a set of control variables. Our regression models also include a constant term (α), year dummies (Year_t), industry-specific effects (Industry_i), and a random error term (ε_i,t).

Industry fixed effects are included in the model to control for unobserved characteristics common across firms within the same sector but vary across different industries. By doing so, we aim to isolate the impact of customers’ ESG performance on suppliers’ green innovation from industry-specific factors, such as regulatory environments, industry norms, or competitive pressures, that could influence this relationship.

We also include year-fixed effects to account for temporal shocks, such as macroeconomic fluctuations, policy changes, or technological advancements, that might affect all firms simultaneously within a given year. This control helps ensure that the observed effects are not confounded by time-specific events or trends that could otherwise skew the results.

To further verify the moderating effect of the bargaining power of major corporate customers and major corporate customers’ green innovation capability on the primary impact, thereby examining Hypotheses 2 and 3, we established Equation (2) as follows:

$${{\mathrm{LnGreenInv}}^{\mathrm{S}}}_{i,t/t+1}=\alpha +{\beta }_1 {{\mathrm{LnESG}}^{\mathrm{C}}}_{i,t}+{\beta }_2 {\mathrm{Moderator}}_{i,t}+{\beta }_3 {{\mathrm{LnESG}}^{\mathrm{C}}}_{i,t}\times {\mathrm{Moderator}}_{i,t}+\gamma {\mathrm{Controls}}_{i,t}+\sum {\mathrm{Year}}_t+\sum {\mathrm{Industry}}_i+{\epsilon }_{i,t}$$

(2)

In Equation (2), Moderator_i,t represents our moderating variables, including the bargaining power of major corporate customers and major corporate customers’ green innovation capability. Our regression models also included a constant term (α), year dummies (Year_t), industry-specific effects (Industry_i), and a random error term (ε_i,t).

Additionally, to ensure the robustness and validity of our findings, we have employed several additional checks discussed in later sections of the paper. These include alternative measures, extending the lag period, and applying instrumental variable methods to address potential endogeneity concerns. We also implement the Heckman two-stage procedure to correct for any potential sample selection bias and conduct falsification tests to rule out spurious correlations. Moreover, we delve deeper into the underlying mechanisms by examining green procurement pressure and innovation collaboration. We analyze heterogeneity to explore how the effects vary across contexts. These additional tests strengthen the credibility of our empirical results.

4. Results

4.1. Descriptive Analysis

Winsorizing was applied to specific variables in our dataset, including Gov_subsidy^S and TobinsQ^S, due to the presence of extreme values that could potentially skew the results. These variables, by their nature, are susceptible to outliers because of the variability in government subsidies across firms and the fluctuations in market valuation ratios like Tobin’s Q. The purpose of winsorizing was to reduce the influence of these extreme values, thereby ensuring that a small number of outliers does not disproportionately drive our results.

Table 2. Descriptive statistics.

VarName	Obs	Mean	SD	Min	Max
LnGreenInvS	1033	0.683	0.863	0.000	4.331
LnESGC	1033	4.328	0.068	4.029	4.490
Bar_powerC	1033	0.084	0.120	0.0002	0.728
LnCuGreInvC	1033	2.108	1.837	0	7.242
LnESGS	1033	4.290	0.068	3.965	4.441
SizeS	1033	21.770	1.223	19.790	25.110
FirmAgeS	1033	2.691	0.377	1.386	3.466
LevS	1033	0.379	0.207	0.039	0.891
ROAS	1033	0.049	0.054	−0.144	0.191
RDS	1033	5.316	6.621	0.002	36.880
GrowthS	1033	0.156	0.309	−0.643	1.662
TobinQS	1033	2.130	1.436	0.886	9.199
IndepS	1033	0.367	0.050	0.333	0.571
Top1S	1033	0.348	0.137	0.114	0.692
Gov_subsidyS	1033	3.239	4.596	0.019	26.44
DualS	1033	0.269	0.444	0	1
SOES	1033	0.325	0.469	0	1
LnGre_purchaseC	953	2.622	0.596	0.000	3.807
CollaborationC	953	3.052	2.157	0.000	10.000

presents the descriptive statistics of our variables. The dataset contains 1033 sample pairs of customer–supplier listed companies. The primary variable, suppliers’ green innovation (LnGreenInv^S), has a mean value of 0.683 and a standard deviation of 0.863, indicating significant variability in green innovation performance among supplier firms. We tested the Variance Inflation Factor (VIF) for all variables. We found that all VIF values were below the critical threshold, indicating no multicollinearity concerns among the independent variables. It allowed us to include the explanatory variables in the models simultaneously. For the interactive variables in the moderating analysis, we mean-centered the first-order variables before incorporating them into the interaction terms to mitigate potential collinearity issues.

4.2. Regression Results of the Baseline Model

The baseline regression results are presented in

Table 3. Baseline regression estimates of major customers’ ESG and suppliers’ green innovation.

Variables	(1)	(2)	(3)	(4)
	LnGreeInvSt	LnGreeInvSt	LnGreeInvSt+1	LnGreeInvSt+1
LnESGC	0.991 **	1.032 ***	1.646 ***	1.602 ***
	(2.55)	(2.80)	(3.57)	(3.57)
LnESGS		1.131 ***		0.679
		(2.85)		(1.47)
SizeS		0.251 ***		0.204 ***
		(7.74)		(5.45)
FirmAgeS		0.049		0.133
		(0.59)		(1.38)
LevS		0.182		0.547 ***
		(0.99)		(2.61)
ROAS		0.099		0.983
		(0.16)		(1.41)
RDS		0.010 **		0.008
		(1.99)		(1.33)
GrowthS		0.109		−0.058
		(1.25)		(−0.59)
TobinQS		0.036		0.037
		(1.57)		(1.43)
IndepS		−0.309		−0.803
		(−0.62)		(−1.43)
Top1S		0.150		0.296
		(0.78)		(1.35)
Gov_subsidyS		0.018 ***		0.025 ***
		(3.06)		(3.60)
DualS		0.021		−0.060
		(0.35)		(−0.88)
SOES		0.041		−0.086
		(0.63)		(−1.14)
Year FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Constant	−3.607 **	−14.468 ***	−6.365 ***	−14.093 ***
	(−2.15)	(−6.12)	(−3.18)	(−5.01)
Observations	1032	1032	954	954
R-squared	0.147	0.266	0.118	0.209

Note: Robust t-statistics in parentheses; *** p < 0.01, ** p < 0.05.

. Columns (1) and (2) show the outcomes for the current year, while columns (3) and (4) show the results with a one-year lag. These results indicate that the coefficient of LnESG^C is positive and statistically significant (p < 0.01), regardless of whether control variables are included in the models. The results in Table 3 support Hypothesis 1, demonstrating that major corporate customers’ performance in ESG positively enhances their suppliers’ outcomes in green innovation.

In addition to the primary variable of interest, several control variables also show statistical significance, indicating their relevance in explaining variations in green innovation among suppliers. For example, firm size (Size^S) shows significant effects, and their positive signs were consistent with our expectations, as larger firms are more likely to engage in green innovation due to their greater resources and R&D capacity.

The variable for financial leverage (Lev^S) is significant in the lagged models, indicating that firms with higher leverage might have more constrained resources in the short term but may prioritize green innovation in the long run to improve their market positioning and operational efficiency. They may hope to send signals of their long-term development and sustainable operations through green innovation, enhancing the confidence of external investors and creditors in their debt repayment ability.

The positive and significant coefficients for government subsidies (Gov_subsidy^S) across multiple models suggest that government support plays a critical role in enabling firms to engage in green innovation and further enhance their green innovation outcomes.

However, R&D intensity (RD^S) only shows significance in one model, indicating that the relationship between R&D investment and green innovation is unstable. This suggests that companies that invest more in R&D may not mainly pursue green innovation. The control variables generally behave as expected, with their signs and significance levels consistent with prior research, confirming the validity of our model specifications.

4.3. Regression Results of the Moderating Effects

The findings regarding the moderating effects of the major customer’s bargaining power and green innovation capability are presented in

Table 4. Regression estimates of moderating effects.

Variables	(1)	(2)	(3)
	LnGreInvSt+1	LnGreInvSt+1	LnGreInvSt+1
LnESGC	1.602 ***	1.955 ***	1.459 ***
	(3.57)	(4.04)	(3.14)
Bar_powerC		−0.023
		(−0.09)
LnCuGreInvC			0.050 ***
			(3.16)
LnESGC ×Bar_powerC		12.707 ***
		(2.69)
LnESGC × LnCuGreInvC			0.525 **
			(2.03)
LnESGS	0.679	0.697	0.572
	(1.47)	(1.51)	(1.25)
SizeS	0.204 ***	0.203 ***	0.207 ***
	(5.45)	(5.40)	(5.56)
FirmAgeS	0.133	0.131	0.130
	(1.38)	(1.36)	(1.36)
LevS	0.547 ***	0.521 **	0.483 **
	(2.61)	(2.49)	(2.31)
ROAS	0.983	0.781	1.017
	(1.41)	(1.11)	(1.47)
RDS	0.008	0.008	0.009
	(1.33)	(1.28)	(1.48)
GrowthS	−0.058	−0.061	−0.046
	(−0.59)	(−0.62)	(−0.47)
TobinQS	0.037	0.036	0.040
	(1.43)	(1.37)	(1.53)
IndepS	−0.803	−0.733	−0.643
	(−1.43)	(−1.30)	(−1.15)
Top1S	0.296	0.308	0.289
	(1.35)	(1.41)	(1.33)
Gov_subsidyS	0.025 ***	0.025 ***	0.024 ***
	(3.60)	(3.59)	(3.46)
DualS	−0.060	−0.053	−0.067
	(−0.88)	(−0.78)	(−0.98)
SOES	−0.086	−0.098	−0.089
	(−1.14)	(−1.31)	(−1.19)
Year FE	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes
Constant	−14.093 ***	−15.683 ***	−13.221 ***
	(−5.01)	(−5.24)	(−4.59)
Observations	954	954	954
R-squared	0.209	0.216	0.223

Note: Robust t-statistics in parentheses; *** p < 0.01, ** p < 0.05.

. The significance of the coefficients of the control variables is consistent with the above Table 3, indicating that larger firms are more likely to engage in green innovation. The coefficients of financial leverage and government subsidies are also positive and significant. The interaction term between LnESG^C and Bar_power^C is positive and significant (coefficient = 12.707, p < 0.01) in column (2), indicating the major corporate customers’ bargaining power magnifies the positive effect of their ESG performance on suppliers’ subsequent outcomes in green innovation. Customers’ higher bargaining power promotes suppliers to prioritize the environmental protection needs conveyed by their customers’ ESG commitments. It also stimulates upstream companies to explore innovative solutions and actively engage in sustainable practices. Consequently, this promotes supplier companies to implement green innovation actively. Hypothesis 2 is supported. Column (3) examines the moderating effect of customers’ green innovation capability (CuGreInv^C). The interaction term between LnESG^C and CuGreInv^C is positive and significant (coefficient = 0.525, p < 0.05), indicating that customers’ green innovation capabilities amplify the positive impact of customers’ ESG performance on suppliers’ green innovation. When a customer possesses more substantial green innovation capabilities, it can facilitate the transmission and sharing of targeted knowledge, information, and resources with the suppliers and more effectively guide supplier enterprises in their green innovation efforts. Hypothesis 3 is supported.

4.4. Robustness Checks

To confirm the robustness of our results, we performed several robustness checks.

4.4.1. Alternative Measurement of Suppliers’ Green Innovation

We utilize green innovation efficiency as our dependent variable to verify the consistency of our findings. Following the procedure from previous research [54,55], we measured green innovation efficiency using the proportion of green innovation output relative to its input. Given the limited availability of specific green innovation input data, the firm’s yearly R&D expenditure was utilized as an approximation proxy for the input of green innovation. The natural logarithm of the count of green patent applications plus one quantifies green innovation output.

The results in

Table 5. An alternative measurement of suppliers’ green innovation.

Variables	(1)	(2)	(3)
	GreInv_EfficiencySt+1	GreInv_EfficiencySt+1	GreInv_EfficiencySt+1
LnESGC	0.051 **	0.061 ***	0.048 **
	(2.51)	(2.67)	(2.25)
LnESGC × Bar_powerC		0.488 **
		(1.99)
LnESGC * LnCuGreInvC			0.022 *
			(1.82)
Bar_powerC		0.013
		(1.08)
LnCuGreInvC			0.002 **
			(2.45)
LnESGS	0.029	0.027	0.025
	(1.42)	(1.28)	(1.22)
SizeS	0.005 ***	0.004 **	0.005 ***
	(2.66)	(2.52)	(2.73)
FirmAgeS	0.006	0.005	0.005
	(1.29)	(1.20)	(1.27)
LevS	0.023 **	0.021 **	0.020 **
	(2.37)	(2.22)	(2.13)
ROAS	0.106 ***	0.100 ***	0.107 ***
	(3.38)	(3.17)	(3.40)
RDS	−0.000	0.000	0.000
	(−0.05)	(0.02)	(0.09)
GrowthS	−0.004	−0.004	−0.003
	(−0.93)	(−0.93)	(−0.81)
TobinQS	0.000	0.000	0.000
	(0.21)	(0.13)	(0.24)
IndepS	−0.045 *	−0.041	−0.038
	(−1.78)	(−1.63)	(−1.50)
Top1S	0.007	0.007	0.007
	(0.73)	(0.75)	(0.74)
Gov_subsidyS	0.001 ***	0.001 ***	0.001 ***
	(3.50)	(3.21)	(3.37)
DualS	0.002	0.002	0.002
	(0.58)	(0.63)	(0.52)
SOES	−0.003	−0.004	−0.003
	(−0.92)	(−1.09)	(−0.96)
Year FE	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes
Constant	−0.440 ***	−0.465 ***	−0.416 ***
	(−3.44)	(−3.35)	(−3.16)
Observations	905	905	905
R-squared	0.129	0.136	0.139

Note: Robust t-statistics in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

confirm that major corporate customers’ positive ESG performance significantly impacts suppliers’ green innovation efficiency. These effects are consistent with those observed in Table 3, reinforcing the robustness of the findings concerning the impact of the major corporate customers’ ESG commitments on suppliers’ engagement in green innovation practices. The ROA in this regression is positive and significant, indicating that companies with higher ROA will also produce more green innovations. The significance of other control variables remains the same as before.

4.4.2. Alternative Measurement of ESG Dimensions and ESG Indicators

In

Table 6. Robust checks using customers’ ESG pillars and alternative ESG indicators.

Variables	(1)	(2)	(3)	(4)	(5)
	LnGreeInvSt+1	LnGreeInvSt+1	LnGreeInvSt+1	LnGreeInvSt+1	LnGreeInvSt+1
LnEC	0.545 **
	(2.50)
LnSC		0.428 **
		(2.17)
LnGC			0.639
			(1.59)
LnESGC_BBG				0.253 **
				(2.04)
LnESGC_Syn					0.617 **
					(2.22)
LnESGS	0.664	0.692	0.654	0.759	0.996
	(1.44)	(1.50)	(1.41)	(1.38)	(1.07)
SizeS	0.207 ***	0.203 ***	0.202 ***	0.092 **	0.261 ***
	(5.49)	(5.40)	(5.35)	(2.05)	(3.48)
FirmAgeS	0.123	0.131	0.133	0.079	0.306
	(1.27)	(1.35)	(1.37)	(0.64)	(1.08)
LevS	0.561 ***	0.608 ***	0.572 ***	1.094 ***	0.530
	(2.67)	(2.89)	(2.71)	(4.23)	(1.13)
ROAS	1.103	1.120	1.129	2.010 **	1.756
	(1.58)	(1.60)	(1.61)	(2.39)	(1.30)
RDS	0.008	0.008	0.008	0.007	0.010
	(1.32)	(1.43)	(1.37)	(1.09)	(1.20)
GrowthS	−0.062	−0.068	−0.066	−0.060	−0.330 *
	(−0.63)	(−0.69)	(−0.67)	(−0.50)	(−1.92)
TobinQS	0.045 *	0.043 *	0.042	0.031	0.123 ***
	(1.73)	(1.67)	(1.60)	(1.05)	(3.09)
IndepS	−0.672	−0.734	−0.686	−0.514	1.750
	(−1.19)	(−1.30)	(−1.22)	(−0.79)	(1.28)
Top1S	0.279	0.336	0.301	0.358	1.577 ***
	(1.27)	(1.53)	(1.37)	(1.41)	(3.10)
Gov_subsidyS	0.025 ***	0.025 ***	0.024 ***	0.027 ***	0.026 *
	(3.66)	(3.64)	(3.48)	(3.26)	(1.66)
DualS	−0.069	−0.071	−0.074	−0.105	−0.460 ***
	(−1.00)	(−1.04)	(−1.08)	(−1.29)	(−3.31)
SOES	−0.093	−0.097	−0.098	−0.042	−0.112
	(−1.23)	(−1.29)	(−1.31)	(−0.46)	(−0.66)
Year FE	Yes	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes	Yes
Constant	−9.463 ***	−9.125 ***	−9.888 ***	−6.158 **	−12.774 ***
	(−4.21)	(−4.07)	(−3.70)	(−2.50)	(−3.20)
Observations	954	954	954	642	219
R-squared	0.204	0.202	0.200	0.245	0.451

Note: Robust t-statistics in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

, we separately examined the effects of the scores for the three components of ESG, environmental, social, and governance, on suppliers’ green innovation. Our analysis reveals that the coefficients for the environmental and social components align with previous findings, though the significance level of the social dimension’s coefficient has diminished. Interestingly, the governance pillar’s coefficient is positive but lacks statistical significance. These results suggest that major corporate customers’ environmental commitment and performance are suppliers’ primary drivers of green innovation.

Additionally, the differences in ESG ratings provided by various evaluators could influence the results of our study, particularly if these evaluators emphasize different aspects of ESG performance. Such variations could affect the robustness of our findings if the conclusions are sensitive to the choice of ESG data source. Thus, we conducted robustness tests using ESG ratings from Bloomberg and SynTao Green Finance evaluators to re-conduct our primary effect analyses. The results in columns (4) and (5) show that our findings are consistent and robust with the alternative ESG rating indicators.

4.4.3. Two- and Three-Period Lagged Effect and Accumulation Innovation Tests

To further validate the robustness of our findings, we conducted tests using data lagged by two periods, three periods, and the accumulation of patents applied for the last three years. Specifically, first, we incorporated the key variables lagged by two and three periods into our model to examine whether their effects on the dependent variable remain significant over longer time horizons. Second, we calculate the natural logarithm of one plus the number of patents applied for during the last three years.

The results in

Table 7. Baseline results for two- and three-year lag and accumulation of green innovation.

Variables	(1)	(2)	(3)	(4)	(5)
	LnGreeInvSt+2	LnGreeInvSt+2	LnGreeInvSt+3	LnGreeInvSt+3	LnGreeInvSstock
LnESGC	2.076 ***	1.988 ***	1.661 ***	1.404 **	2.421 ***
	(3.90)	(3.81)	(2.78)	(2.41)	(3.54)
LnESGS		1.199 **		0.660	0.794
		(2.20)		(1.08)	(1.12)
SizeS		0.218 ***		0.217 ***	0.244 ***
		(4.95)		(4.44)	(4.26)
FirmAgeS		0.289 ***		0.164	0.205
		(2.63)		(1.37)	(1.45)
LevS		0.243		0.773 ***	1.016 ***
		(0.98)		(2.83)	(3.15)
ROAS		0.507		2.108 **	2.529 **
		(0.60)		(2.23)	(2.27)
RDS		0.008		0.023 ***	0.020 **
		(1.17)		(2.93)	(2.19)
GrowthS		0.247 **		0.251 *	0.338 **
		(2.05)		(1.94)	(2.19)
TobinQS		0.042		0.020	0.033
		(1.43)		(0.58)	(0.84)
IndepS		−0.300		0.181	0.058
		(−0.46)		(0.25)	(0.07)
Top1S		0.214		0.085	−0.011
		(0.86)		(0.31)	(−0.03)
Gov_subsidyS		0.014 *		0.014	0.028 ***
		(1.80)		(1.55)	(2.75)
DualS		−0.132		−0.155 *	−0.135
		(−1.63)		(−1.72)	(−1.28)
SOES		−0.153 *		−0.115	−0.111
		(−1.76)		(−1.21)	(−1.00)
Year FE	Yes	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes	Yes
Constant	−8.091 ***	−18.563 ***	−6.201 **	−13.735 ***	−18.848 ***
	(−3.51)	(−5.60)	(−2.40)	(−3.77)	(−4.43)
Observations	861	861	795	795	789
R-squared	0.118	0.193	0.082	0.176	0.226

Note: Robust t-statistics in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

indicate that the findings with the two-period lagged, three-period lagged, and the accumulation data are consistent with those observed with the one-period lagged data. The key variables continue to exhibit significant effects in the regressions. This consistency suggests that our conclusions are robust and that the impact of the major customer’s ESG performance remains substantial not only in the short term but also over extended periods.

Our analysis of the impact of significant customers’ ESG practices on their suppliers’ green innovation across different periods reveals an interesting pattern. The effect peaks at 1.988 (p < 0.01) in the second year and then decreases to 1.404 (p < 0.05) in the third year. This trend suggests that the influence of ESG practices may intensify as suppliers initially respond to these pressures, with the second year showing the most substantial effect. This could be due to the time required for suppliers to fully integrate ESG practices into their operations, leading to a more pronounced impact in the second year. These findings underscore the importance of sustained and well-timed ESG initiatives. They suggest that while the influence of such practices can grow over time, there may be a peak after which the returns on innovation begin to taper off. This highlights the need for continuous innovation strategies to maintain momentum beyond the initial response period.

The results in column (5) indicate that the major customers’ ESG performance has a more substantial effect (coefficient = 2.421, p < 0.01) on the accumulation of green patents over the last three years, which suggests that the stock innovation may better explain the innovative effect than flow results (patents year by year) [56,57]. These studies reinforce the idea that measuring patent accumulation over time (stock innovation) provides a more comprehensive understanding of a firm’s innovation capacity, making it a more robust explanatory variable for long-term innovation outcomes [56,57].

4.4.4. Endogenous Analysis

Identifying a causal effect of the major clients’ performance in ESG on suppliers’ green innovation is challenging due to potential endogeneity and the influence of unobservable variables on both factors. We employed an instrumental variable (IV) method to address this issue. Previous studies have utilized industry or regional ESG levels as IVs for enterprise ESG [58], which often meet the correlation requirements but raise concerns regarding the exclusivity hypothesis [59]. Our study leverages the recent rise of ESG-themed public funds in China as an exogenous event, using the stock market value (FV) of enterprises held by these ESG funds as an IV for the enterprises’ ESG performance to mitigate the above concerns.

The shareholding data of ESG funds is a suitable IV because public funds, particularly those with an ESG focus, can significantly influence corporate governance through mechanisms such as “voting with their feet” [60]. This relationship underscores a positive correlation between ESG fund holdings and corporate ESG performance. Notably, the shareholding data of ESG funds is unlikely to directly affect suppliers’ green innovation for two reasons: (1) the establishment and scale of ESG funds are exogenous, determined by fund companies, and not directly related to suppliers’ green innovation; (2) fund managers make the portfolio decisions and changes in the ESG funds and do not directly impact suppliers’ green innovation. Therefore, the shareholding information of ESG funds meets the exclusivity requirement.

Column (1) of

Table 8. Endogenous analysis: instrumental variable results.

Variables	(1)	(2)
	First-Stage	Second-Stage
	LnESGC	LnGreInvSt+1
LnESGC		3.864 ***
		(2.87)
IV	0.002 ***
	(10.18)
LnESGS	−0.040	0.725
	(−1.17)	(1.47)
SizeS	−0.001	0.210 ***
	(−0.36)	(4.92)
FirmAgeS	−0.002	0.155
	(−0.34)	(1.49)
LevS	0.020	0.465 **
	(1.25)	(1.97)
ROAS	0.113 **	0.632
	(2.29)	(0.82)
RDS	0.000	0.007
	(0.10)	(1.17)
GrowthS	−0.005	−0.043
	(−0.80)	(−0.44)
TobinQS	0.004 **	0.030
	(2.20)	(1.10)
IndepS	0.084 **	−0.912
	(2.25)	(−1.51)
Top1S	−0.006	0.283
	(−0.42)	(1.26)
Gov_subsidyS	0.000	0.024 ***
	(0.09)	(3.21)
DualS	−0.008	−0.041
	−1.60	(−0.56)
SOES	−0.006	−0.071
	−1.14	(−0.87)
Year FE	Yes	Yes
Industry FE	Yes	Yes
Constant	4.408 ***	−24.622 ***
	(29.86)	(−4.01)
Observations	941	941
Partial R2	0.208	0.194

Note: Robust t-statistics in parentheses; *** p < 0.01, ** p < 0.05.

indicates the first-stage regression results for the instrumental variable of the market value of ESG fund holdings. The results indicate that the IVs are significantly positive, demonstrating that higher market value for ESG fund holdings corresponds to higher ESG performance of enterprises. Additional tests show the F value in the first stage is above 10, confirming the absence of weak IV issues. The partial R² of the first stage is 0.208, indicating the IVs have substantial explanatory power. The coefficient of LnESG^C is positive and statistically significant (coefficient = 3.864, p < 0.01) in column (2). These findings affirm that, after addressing endogeneity, the conclusion that customer enterprise ESG performance promotes suppliers’ green innovation remains robust.

4.4.5. Heckman Two-Stage Method

The China Securities Regulatory Commission promotes listed companies to voluntarily disclose precise details regarding their top five clients, although such disclosure is not mandatory. Consequently, this voluntary disclosure could lead to a sample selection problem in our analysis. Following the previous research [61], we employed the Heckman two-stage method to address this potential sample selection bias.

In the first-stage model, we use a Probit model for regression. The explained variable indicates if the supplier discloses details about their main customers (a binary variable that takes the value of one if disclosed and zero if not). We introduce supplier firm transparency as an instrumental variable in this model because the higher the company transparency, the more likely it is to choose to disclose more information, such as customer information. However, this will not affect the company’s green innovation outcome. From this model, we calculate an inverse Mills ratio (IMR), which is the ratio of the probability density function of the predicted values of the dependent variable to the cumulative distribution function of a distribution of the expected values of the dependent variable based on the first-stage Probit model.

In the second-stage model, the estimated IMR from the first stage is utilized as a control variable with other control variables. The results in column (2) of

Table 9. Heckman two-stage method.

Variables	(1)	(2)
	First-Stage	Second-Stage
LnESGC		1.553 ***
		(3.48)
TRANS	0.285 **
	(2.16)
LnESGS	−1.033 ***	0.562
	(−3.42)	(1.22)
SizeS	−0.135 ***	0.016
	(−4.33)	(0.23)
FirmAgeS	0.033	0.203 **
	(0.36)	(2.07)
LevS	0.268 *	0.927 ***
	(1.67)	(3.91)
ROAS	−1.056 ***	−0.611
	(−3.68)	(−0.73)
RDS		0.007
		(1.11)
GrowthS	−0.005	−0.047
	(−1.63)	(−0.48)
TobinQS	0.008	0.039
	(0.42)	(1.52)
IndepS	−0.518	−1.699 ***
	(−1.25)	(−2.74)
Top1S	−0.684 ***	−0.693 *
	(−3.75)	(−1.92)
Gov_subsidyS	−0.177	0.026 ***
	(−0.73)	(3.85)
DualS	−0.110 **	−0.189 **
	(−2.27)	(−2.44)
SOES	−0.213 ***	−0.394 ***
	(−3.06)	(−3.38)
IMR		5.172 ***
		(3.43)
Year FE	Yes	Yes
Industry FE	Yes	Yes
Constant	9.160 ***	−9.666 ***
	(6.46)	(−3.13)
Observations	22,424	953
Partial R2	0.056	0.220

Note: Robust t-statistics in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

show that the impact of LnESG^C remains robust (coefficient = 1.553, p < 0.01) even after controlling for the IMR. The Heckman two-stage method effectively mitigates the sample selection bias problem’s influence on whether the suppliers disclose detailed information on their customers. By doing so, we ensure that our estimates of the impact of major customers’ ESG performance on suppliers’ green innovation are not biased due to non-random sample selection, thereby enhancing the validity and reliability of our findings.

4.4.6. Falsification Tests

We are concerned that there may be an unobserved positive correlation between the significant client’s performance in ESG and its supplier’s subsequent outcomes in green innovation rather than this relationship being solely driven by their economic partnership. To address this concern, we employ a falsification test. If the major clients’ performance in ESG enhances their suppliers’ subsequent green innovation, this effect must occur through the real supplier–customer pairs. In our research setting, for each customer–supplier pair observed in the data, we take the customer as given and utilize the Propensity Score Matching (PSM) approach to create a virtual supplier company for each customer. These matched companies have corporate fundamentals like actual supplier companies but lack an economic connection through the customer–supplier relationship.

We expect the client’s performance in ESG to have a limited impact on the subsequent green innovation outcomes of those matched pseudo-suppliers. Specifically, for each pair of supplier S and customer C, we calculate supplier S’s propensity score and match it with a similar company, S’, which does not have a direct economic link to customer C. This process results in a matched pair with a customer–pseudo-supplier relationship. Our sample’s propensity score is the supplier’s predicted probability based on our control variables.

We then estimate the regression models of the customers and their pseudo-suppliers. Consistent with our expectations, the coefficients for LnESG^C in columns (1) and (2) in

Table 10. Falsification test results with fictitiously assigned matched suppliers.

	(1)	(2)
Variables	LnGreInvSt+1	LnGreInvSt+1
LnESGC	0.982 (1.56)	0.952 (1.59)
Controls	No	Yes
Year FE	Yes	Yes
Industry FE	Yes	Yes
Constant	−3.506 (−1.29)	−9.211 ** (−2.34)
Observations	534	534
R-squared	0.150	0.253

Note: z-statistics in parentheses; ** p < 0.05.

are both not statistically significant. The findings indicate that the observed impact of a customer’s performance in ESG on a supplier’s green innovation is primarily due to their economic linkages rather than other confounding factors.

5. Possible Mechanism Analysis

In this part, we explore the potential mechanisms by which the significant clients’ ESG performance can facilitate their suppliers’ green innovation. We test two plausible mechanisms: green procurement pressure and green innovation collaboration.

Investigating the pressure and collaborative mechanism poses a challenge in our research due to the limited observability and measurability of the available secondary data. In this section, we draw on the coding criteria of previous research [62] using content analysis to quantify the two proxy variables of green procurement and green innovation collaboration. Seed words from the survey items of the existing literature were selected as coding items for our content analysis. We use Python 3.10.12 software to crawl companies’ annual reports, CSR reports, ESG reports, and sustainability reports. Dummy variables were applied, assigning a score of one if all the keywords of an item were found within 40 consecutive words in any of the above reports and 0 if not. If there is a quantized value, assign it a value of two. The value of green procurement and innovation collaboration was determined by the total score of all the items. For specific coding guidelines and details, see Appendix A.

5.1. Customers’ Green Procurement Pressure Mechanism

Previous research indicates that customers can improve suppliers’ environmental capabilities through supply chain management [41,43,44]. Major customers, especially if they pay attention to green development, will push upstream green innovation through green procurement. Following previous research [63], we measured green procurement in two dimensions: product-based and process-based green procurement. Appendix A lists the specific keywords and references for the two dimensions. We use the total score of each sub-indicator as our proxy variable for measuring green procurement pressure and a three-stage method to test the mediating effect.

Table 11. Channel analysis: customers’ green purchase pressure mechanism.

Variables	(1)	(2)	(3)
	LnGreeInvSt+1	LnGre_purchaseCt	LnGreeInvSt+1
LnESGC	1.604 ***	3.040 ***	1.279 ***
	(3.58)	(11.02)	(2.68)
LnGre_purchaseC			0.107 **
			(1.99)
LnESGS	0.686	0.921 ***	0.587
	(1.49)	(3.25)	(1.27)
SizeS	0.203 ***	0.019	0.201 ***
	(5.41)	(0.83)	(5.36)
FirmAgeS	0.137	0.095	0.127
	(1.42)	(1.59)	(1.31)
LevS	0.553 ***	0.157	0.536 **
	(2.63)	(1.21)	(2.55)
ROAS	0.968	0.095	0.958
	(1.38)	(0.22)	(1.37)
RDS	0.008	0.004	0.007
	(1.33)	(0.98)	(1.27)
GrowthS	−0.060	−0.008	−0.059
	(−0.61)	(−0.13)	(−0.60)
TobinQS	0.037	0.033 **	0.034
	(1.44)	(2.07)	(1.30)
IndepS	−0.804	−0.060	−0.797
	(−1.43)	(−0.17)	(−1.42)
Top1S	0.305	0.439 ***	0.258
	(1.39)	(3.25)	(1.17)
Gov_subsidyS	0.025 ***	0.003	0.024 ***
	(3.60)	(0.62)	(3.57)
DualS	−0.059	−0.009	−0.058
	(−0.86)	(−0.22)	(−0.84)
SOES	−0.086	−0.019	−0.084
	(−1.14)	(−0.40)	(−1.12)
Year FE	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes
Constant	−14.133 ***	−15.431 ***	−12.481 ***
	(−5.01)	(−8.90)	(−4.25)
Observations	952	952	952
R-squared	0.209	0.215	0.213

Note: Robust t-statistics in parentheses; *** p < 0.01, ** p < 0.05.

examines the mediating role of customers’ green procurement pressure. The significance of the coefficients of the firm size, financial leverage, and government subsidies are also positive and significant in columns (1) and (3), which is consistent with the above results. The positive coefficient of major corporate customers’ ESG performance in column (2) is statistically significant (coefficient = 3.040, p < 0.01), indicating that high ESG performance by major customers may enhance their green procurement practices. The results in column (3) show that LnGre_purchase^C is statistically significant (coefficient = 0.107, p < 0.05) in the regression model. Additionally, the coefficient for LnESG^C is lower than that in column (1), suggesting that green procurement mediates between customers’ ESG performance and suppliers’ green innovation. This finding implies that major corporate customers’ adoption of ESG strategies exerts pressure on suppliers to adopt sustainable development practices to achieve green procurement requirements from their major customers. As a result, suppliers are more willing to engage in green innovation-related activities to meet these demands.

5.2. Customers’ Green Innovation Collaborative Mechanism

Table 12. Channel analysis: customers’ green innovation collaborative mechanism.

Variables	(1)	(2)	(3)
	LnGreeInvSt+1	CollaborationCt	LnGreeInvSt+1
LnESGC	1.625 ***	6.168 ***	1.456 ***
	(3.65)	(5.44)	(3.22)
CollaborationC			0.027 **
			(2.11)
LnESGS	0.577	2.280 *	0.515
	(1.26)	(1.95)	(1.12)
SizeS	0.209 ***	0.071	0.207 ***
	(5.60)	(0.75)	(5.56)
FirmAgeS	0.133	0.378	0.122
	(1.38)	(1.55)	(1.28)
LevS	0.535 **	0.356	0.525 **
	(2.57)	(0.67)	(2.52)
ROAS	1.015	0.342	1.006
	(1.46)	(0.19)	(1.45)
RDS	0.008	−0.003	0.008
	(1.33)	(−0.22)	(1.35)
GrowthS	−0.062	−0.100	−0.059
	(−0.63)	(−0.40)	(−0.61)
TobinQS	0.038	0.061	0.036
	(1.47)	(0.93)	(1.41)
IndepS	−0.924 *	−0.754	−0.904
	(−1.65)	(−0.53)	(−1.61)
Top1S	0.290	0.691	0.271
	(1.33)	(1.25)	(1.24)
Gov_subsidyS	0.024 ***	0.030 *	0.023 ***
	(3.55)	(1.73)	(3.43)
DualS	−0.060	−0.169	−0.055
	(−0.87)	(−0.97)	(−0.81)
SOES	−0.091	−0.193	−0.086
	(−1.22)	(−1.01)	(−1.16)
Year FE	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes
Constant	−13.820 ***	−36.198 ***	−12.828 ***
	(−4.94)	(−5.08)	(−4.53)
Observations	952	952	952
R-squared	0.209	0.097	0.213

Note: Robust t-statistics in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

examines the mediating role of customers’ green innovation collaboration. The significance of the coefficients of the firm size, financial leverage, and government subsidies are also consistent with the above results. The positive coefficient of LnESG^C in column (2) is statistically significant (coefficient = 6.168, p < 0.01), indicating that higher ESG performance of major customers may enhance more green innovation collaboration with their suppliers. The results in column (3) show that Collaboration^C is statistically significant (coefficient = 0.027, p < 0.05) in the regression model. Additionally, the coefficient for LnESG^C is lower than that in column (1), suggesting that green innovation collaboration mediates between customers’ ESG performance and suppliers’ green innovation. This finding implies that major corporate customers’ adoption of ESG strategies can assist suppliers in adopting sustainable development practices to achieve green innovation.

6. Heterogeneity Analysis

To further explore the robustness of our findings and understand how the impact of major corporate customers’ ESG performance on suppliers’ green innovation might vary across different contexts, we conducted a heterogeneity analysis by categorizing customers based on their industry type, technological intensity, ownership structure, and geographic proximity to suppliers.

We first examined whether the industry type of customer influences the strength of the relationship between ESG performance and suppliers’ green innovation. As shown in columns (1) and (2) of Panel A of

Table 13. Heterogeneity analysis.

Panel A: Industrial Heterogeneity
Variables	(1)	(2)	(3)	(4)
	ManufactureC	Non-ManufactureC	High-TechC	Non-High-TechC
	LnGreenInvSt+1	LnGreenInvSt+1	LnGreenInvSt+1	LnGreenInvSt+1
LnESGC	1.527 *** (2.66)	2.165 *** (2.93)	1.621 *** (2.76)	2.407 *** (3.35)
Controls	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Constant	−14.642 *** (−4.18)	−17.815 *** (−3.62)	13.987 *** (−3.87)	22.109 *** (−4.73)
Observations	599	349	569	381
R-squared	0.209	0.424	0.193	0.422
Panel B: Firm Nature and Customer–Supplier Distance Heterogeneity
Variables	(1)	(2)	(3)	(4)
	SOEC	Non-SOEC	Short Distance	Long Distance
	LnGreenInvSt+1	LnGreenInvSt+1	LnGreenInvSt+1	LnGreenInvSt+1
LnESGC	2.111 *** (3.63)	0.424 (0.54)	2.802 *** (2.82)	−0.472 (−0.31)
Controls	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Constant	−14.642 *** (−4.18)	−17.815 *** (−3.62)	−10.860 (−1.27)	−9.946 (−1.20)
Observations	659	291	530	417
R-squared	0.244	0.247	0.318	0.335

Note: Robust t-statistics in parentheses; *** p < 0.01.

, the effect of customers’ ESG performance is more pronounced for non-manufacturing customers (coefficient = 2.165, p < 0.01) compared to manufacturing customers (coefficient = 1.527, p < 0.01). This may be because the focus and pressures on environmental innovation differ between manufacturing and non-manufacturing industries. Manufacturing industries focus more on energy-saving and emission-reducing technologies within production processes, whereas non-manufacturing industries might concentrate more on green procurement and sustainable development practices. Therefore, non-manufacturing customers’ ESG performance may emphasize sustainability practices within the supply chain, thus more significantly promoting green innovation among suppliers, an effect less evident in manufacturing.

Second, we investigated the role of technological intensity by distinguishing between high-tech and non-high-tech customers. Our results indicate that the influence of ESG performance on green innovation is stronger for non-high-tech customers (coefficient = 2.407, p < 0.01) than for high-tech customers (coefficient = 1.621, p < 0.01). Firms in high-tech industries are often driven by technology performance and market speed. Environmental regulatory pressure may not be as important as technical performance and market competitiveness in these industries. Market pressure in these industries comes more from innovation speed and technological leadership than environmental sustainability goals. Therefore, these customers may not consider ESG requirements as a core criterion for cooperation with suppliers, resulting in lower investment in green innovation by suppliers. Therefore, even if these large customers have good ESG performance, they are still mainly driven by technology-driven innovation. Suppliers are more concerned about meeting technical rather than environmental requirements. In contrast, large customers in non-high-tech industries may be more pressured by environmental requirements from consumers and regulators and, therefore, more strongly push suppliers to carry out green innovation.

Third, we also considered the ownership structure of the customer firms, specifically comparing state-owned enterprises (SOEs) with non-state-owned enterprises (non-SOEs). Previous research indicates that ESG activities are not costless, and the benefits to stakeholders from ESG activities may come at the expense of shareholders’ benefits [64]. Research has also found that if major customers are governments, they can mitigate operational risks and ease financial constraints for their suppliers through stable and enduring contracts, providing them with additional resources for ESG investments [29]. In our context, if a client’s performance in ESG can enhance a supplier’s subsequent outcomes of green innovation through collaboration (such as the transfer of green knowledge, know-how, and resources from customer to supplier), we propose this effect would be more substantial if the major customers are state-owned enterprises (SOEs) with political missions and more resources. SOE customers can also assist their suppliers in creating a favorable image with banks and investors [53]. It enables them to secure funding from the capital market, enhancing their capacity for green innovation investment [29]. As depicted in columns (1) and (2) of Panel B of Table 13, the results reveal that the ESG performance of SOEs has a significant and positive effect on suppliers’ green innovation (coefficient = 2.111, p < 0.01). In contrast, the effect is not substantial for non-SOEs (coefficient = 0.424, p > 0.1). This finding aligns with the notion that SOEs are more effective in driving sustainability practices across their supply chains due to their alignment with governmental policies and greater emphasis on social responsibility. SOEs, which often face stricter government mandates and social responsibilities, may not only exert greater green procurement pressure but also provide more substantial collaborative support, thus enhancing the overall impact on suppliers.

Finally, we examined the role of geographic proximity between customers and suppliers. Previous research found that the customer–supplier geographic proximity enhances the exchange of soft information [49]. We calculate our sample’s average customer–supplier geographic distance and divide them into two groups. It is a long-distance group if the distance exceeds the average distance. Otherwise, it is a short-distance group. The analysis shows that when the distance between the customer and supplier is short, the ESG performance of customers has a significant positive impact on suppliers’ green innovation (coefficient = 2.802, p < 0.01). However, this effect diminishes and becomes insignificant when the geographic distance exceeds the sample average (coefficient = −0.472, p > 0.1). It suggests closer geographic proximity facilitates better communication and more effective implementation of ESG-related initiatives, enhancing their impact on suppliers’ green innovation.

7. Discussion

This study examined how a major corporate customer’s ESG performance can drive the supplier’s green innovation. In our theoretical framework, major corporate customers play an important role in the supply chain. According to the theoretical framework, the major corporate customer has the motivation to manage their supply chain. In particular, if the major corporate customers perform well in ESG practice, they can leverage their suppliers’ efforts to invest and achieve good outcomes in green innovation. Thus, we predicted and found that major clients’ performance in ESG practice positively enhances their suppliers’ performance in the subsequent green innovation. This positive relationship is moderated by the customers’ bargaining power and green innovation capability. We further find that green purchase pressure and green innovation collaboration are two potential channels through this relationship. These findings have important theoretical and practical implications.

7.1. Theoretical and Practical Implication

First, this study was motivated by the theoretical puzzle regarding the role of significant customers in the sustainable development of their supply chain. Based on both Chinese and US contexts, one stream of literature found that companies with higher customer concentration tend to have relatively low pro-social behavior motivation [17,26]. They view ESG activities as a corporate strategy to attract customers and investors. This leads to the hypothesis that suppliers may reduce their ESG activities if they have stable principal customers. However, this perspective primarily considers the supply side, leaving the theoretical gap on the demand side unaddressed. This study fills this gap by examining whether major customers can push their suppliers to engage in pro-social behavior from the customer’s perspective. We suggest that the characteristics of the major customers play an essential role in suppliers’ pro-social behavior. Major customers greatly influence supplier decision-making due to their large purchase volumes. Our findings suggest that major customers prioritizing sustainable development can influence their suppliers to enhance green innovation practices. However, it is essential to recognize that these results are context-dependent and may vary based on industry and firm characteristics. Our heterogeneity tests revealed some preliminary differences. On the one hand, the effect is more pronounced in non-manufacturing and non-high-tech industries, where environmental pressures and procurement practices may focus more on sustainability. Additionally, state-owned enterprises (SOEs) have a more substantial influence on suppliers’ green innovation than non-SOEs, likely due to their alignment with governmental policies. The impact also varies with geographic proximity, with closer customer–supplier distance yielding more substantial effects. These preliminary findings highlight that the influence of ESG performance may vary depending on industry type, technological intensity, ownership structure, and geographic proximity. In this way, our research also enriches the current research on the effects of the customer–supplier nexus on corporate decisions. Previous research primarily focuses on the impacts of the customer–supplier nexus on corporate finance decisions [65,66,67] but overlooks its effects on corporate investment decisions [49], particularly investments in sustainability development. While Chu et al. (2019) extend the literature by examining the impacts of supplier–customer relationships on corporate technological innovation, the effect on corporate investment in sustainability development remains underexplored [49]. Unlike traditional technological innovation, which pursues potential financial benefits, green innovation represents a distinct type of investment decision in sustainability due to its double externalities. The motivations for enterprises to engage in these innovation activities are markedly different. Our study contributes to the customer–supplier nexus and green innovation literature by investigating whether and how major corporate customers’ sustainability development commitments, specifically their ESG performance, can enhance their suppliers’ green innovation, a crucial decision regarding investment in sustainability.

Second, we explore the antecedents of improved corporate green innovation from a new perspective. Previous research has primarily focused on firm-specific characteristics, such as corporate environmental orientation [68], corporate governance [7], political capital [69], managerial perceptions [70], managerial characteristics [71], and strategic market goals [72]. However, existing research overlooks an essential driving force: stakeholder pressure and collaboration. Our research broadens this scope by focusing on the unique customer–supplier nexus and investigating whether pressure and collaboration from major corporate customers can motivate suppliers’ performance in green innovation. Our study builds on the work of Wei et al. (2023), who documented that those primary clients with sustainable development concepts will improve their suppliers’ performance in green innovation [73]. However, we extend this literature by examining how the sustainability commitment of major corporate customers, specifically their ESG performance—an essential decision reflecting a firm’s actual investment in sustainability—can enhance their suppliers’ green innovation. The present study finds that if the significant corporate customer has more bargaining power and higher green innovation capability, their effects on suppliers’ green innovation outcomes are more pronounced. On the one hand, whether suppliers take the ESG orientation of their major customers into strategic consideration depends on their relative power and the absolute power of the client in its industry. If the customer has more market share in its industry, it can exert more pressure on its suppliers to embrace environmental demands. Suppliers must meet these demands and invest in green innovation to maintain major customer relationships. As expected, the results show that the positive correlation strengthens if customers have more bargaining power. On the other hand, if suppliers’ investment in green innovation is motivated by their customers, they must meet these requirements. Therefore, assistance and guidance from customers are essential. To ensure that suppliers can complete the green transition according to the requirements, customers will also collaborate to innovate and provide suppliers with the information, knowledge, and technical know-how required for the transformation. As expected, the positive relationship becomes stronger if customers have more green patents.

Third, we explore the underlying mechanisms through which the performance of major corporate customers in ESG practices can enhance their suppliers’ subsequent outcomes in terms of green innovation. The present study suggests that suppliers prioritize green innovation when they face higher green purchase pressure and can access more green innovation spillover from their primary customers. Chu et al. (2019) found that the geographic proximity of customers and suppliers can enhance suppliers’ innovation through feedback and demand channels [49]. Huang et al. (2023) identified demand expansion and resource support as fundamental influence mechanisms [29]. According to our theoretical framework, we examine pressure and collaboration mechanisms. Due to the limited observability and measurability of the available data on mediating variables, we provide suggestive evidence to enhance our comprehension of these two channels. We use the content analysis method to construct the green purchase variable for the pressure channel and green innovation collaboration variable for the collaboration channel. Our findings indicate that major corporate customers with solid ESG commitments may set high environmental and social performance standards, creating a market demand for sustainable products and services. It, in turn, pressures suppliers through green procurement, compelling them to improve their green innovation capabilities. For the collaboration channel, we propose that major corporate customers can provide critical resources, such as technical support and financial incentives, which enable suppliers to overcome barriers to green innovation.

Fourth, the results from our control variables reveal essential insights into the innovation process, particularly the inherent uncertainty that not all investments lead to immediate or successful innovation outcomes [50]. For instance, while firm size (Size^S) and government subsidies (Gov_subsidy^S) consistently show a positive and significant impact on green innovation, R&D intensity (RD^S) presents a more complex picture. The significance of RD^S in only one model suggests that higher R&D investment does not always equate to increased green innovation. This variability highlights the unpredictable nature of innovation, where the same input (e.g., R&D expenditure) can lead to different outcomes depending on various contextual factors such as market conditions, firm capabilities, and the specific nature of the R&D activities undertaken. These findings underscore the need for a more strategic approach to innovation investments, where firms must consider not just the quantity of investment but also the quality and alignment with broader innovation goals. Moreover, they reinforce the importance of continuous and adaptive innovation strategies to navigate the uncertainties inherent in the innovation process.

Finally, our research also has some practical implications. For managers, the findings suggest that integrating ESG considerations into procurement and partnership strategies can encourage a ripple effect of green innovation among suppliers and enhance their environmental footprint. For policymakers, these findings highlight the potential of using ESG standards as a regulatory tool to promote environmental innovation across industries. Understanding the differential impact of ESG practices across industries and ownership structures can help companies tailor their sustainability strategies more effectively, maximizing their influence on suppliers’ innovation activities. Governments can catalyze green innovation throughout the supply chain by encouraging or mandating more substantial ESG commitments from key industry players, contributing to national and global sustainability goals.

7.2. Limitation and Future Research

However, our study has several limitations. First, using a single indicator—patent applications—to measure green innovation may not capture the full scope of sustainable practices adopted by suppliers. While patent applications provide timely and comprehensive insights into a firm’s innovation efforts, they do not fully capture all dimensions of green innovation. For instance, published patents, which reflect innovations that have passed through legal and technical scrutiny, hold significant value for large companies, as they often leverage previous patents to enhance subsequent innovations, which could also serve as an essential measure of innovation outcomes. Future research could improve the robustness of green innovation measurement by incorporating multiple indicators, including input, intermediate, and output indicators. For example, Taques et al. (2021) seek to identify 26 indicators relevant to account for innovation [50]. Although the scope of this study focuses on patent applications, it is essential to acknowledge that these additional indicators exist and can be valuable tools for managers and entrepreneurs. They offer opportunities to assess and enhance innovation strategies from multiple dimensions. This multi-faceted approach would provide a more holistic understanding of green innovation and its drivers [50].

Second, another limitation of this study is the use of SIC codes for industry classification. While these codes provide a standardized method for segmenting industries, they may not fully capture the sectoral differences within broader industry categories, particularly in the services sector. Given that information is often more readily available for manufacturing companies than for services firms [50], the design of innovation metrics may lead to distorted estimates between these two sectors, potentially introducing biases in our results, especially when making industry or sector comparisons [50]. Future research could consider using alternative classification systems that offer more detailed sectoral distinctions and may provide additional insights.

Additionally, this study is focused on the Chinese context, and our sample is limited to specific firms and regions, which may affect the generalizability of the findings. The impact of customer-driven sustainability initiatives likely varies depending on factors such as the size of the companies, their sector, and the specific economic activities in which they are engaged. Therefore, while our study provides evidence of the positive influence of major customers on suppliers’ green innovation, further research is needed to explore the extent to which these findings apply in different contexts and consider a broader range of geographic areas.

Finally, while we examined the role of customers’ ESG performance, other factors, such as market conditions and regulatory environments, may also play a critical role in shaping suppliers’ innovation strategies. Future research can continue to explore other factors that influence green innovation.

8. Conclusions

In summary, this study finds that major corporate customers with higher ESG performance can positively influence their suppliers’ green innovation activities. This relationship is moderated by the customers’ bargaining power and green innovation capabilities and is facilitated through mechanisms such as green procurement pressure and collaborative innovation efforts. While our findings contribute to the literature on sustainable supply chain management, further research is needed to explore additional green innovation indicators and examine these dynamics in a broader range of industries and geographic contexts.