Decentralization or Cooperation? The Impact of “Government–Market” Green Governance Synergy on Corporate Green Innovation: Evidence from China

Abstract

The partnership between government and market plays a crucial role in allocating green resources and fostering collaboration across organizations and departments. It integrates diverse knowledge types into the green innovation process and offers multifaceted insights into enterprises’ responses to green governance decisions. However, existing research predominantly examines the interplay among government green governance instruments, with insufficient exploration of the synergistic impacts of government and market in green governance. This study constructs a capacity coupling coefficient model to measure the synergy degree of “government–market” green governance (GMGG). Exploiting a balanced dynamic panel of 28,451 firm-year observations for 3807 Chinese listed companies from 2010 to 2020, we estimate the causal effect of GMGG synergy on corporate green innovation (CGI) and further dissect the underlying transmission mechanisms as well as the moderating channels through which the effect operates. Empirical results reveal that the effect of GMGG synergy on CGI is subject to diminishing marginal returns, with the effect being significantly more pronounced for substantive green innovation. Heterogeneity analysis indicates that non-state-owned firms, eastern-region firms, and those in non-heavy-polluting industries respond with markedly greater sensitivity. Mechanism analysis further demonstrates that the extent of marketization serves as a mediating channel, whereas an elevated level of digital-economy development mitigates the impact of GMGG synergy on CGI. This study delineates the effective boundary of GMCC synergy in stimulating CGI, providing empirical benchmarks for the synergistic implementation of effective government and efficient market actions in green governance. It further corroborates the positive roles of marketization and the digital economy as novel governance instruments, thereby offering critical policy insights for the coordinated advancement of the “dual-carbon” goals and high-quality economic development.

1. Introduction

The respective roles of the market and government in allocating resources have long been central to discussions in economic theory. Pigou [1] pointed out that market failure can lead to various inefficiencies and may require government intervention. The New Keynesian school also proposed that the market itself is unstable, and therefore requires the intervention of government management functions to solve the problem of market failure caused by insufficient effective demand [2,3]. Environmental and natural resources represent distinctive public goods characterized by significant externalities, necessitating a complex integrated governance framework involving multiple actors such as government, market, and enterprises [4]. This framework aims to achieve strategic objectives of sustainable economic and social development, pollution reduction, enhanced social responsibility, and competitive advantage [5,6]. Ansell and Gash [7] contend that collaborative governance hinges on negotiated agreements, information exchange, and joint action among government, market, and civil society actors. Such cooperation is indispensable for resolving collective action dilemmas in the provision of public goods. Concurrently, the Porter Hypothesis posits that well-designed environmental regulations not only do not hinder competitiveness but can also stimulate green innovation among firms, leading to a “win-win” scenario [8,9]. Building on this foundation, the quality of institutions and regulatory frameworks determines whether regulatory signals can be effectively recognized and converted into innovation investments by enterprises. A transparent, stable, and predictable institutional environment can reduce uncertainty associated with innovation and amplify the effects of policy incentives [10,11]. Therefore, green governance should be viewed as a multidimensional framework that integrates collaborative governance, innovation incentives, and institutional quality. Its strategic objectives focus on pollution reduction and the enhancement of ecological carrying capacity, and on the internalization of externalization via collaborative governance mechanisms [12]. These efforts collectively promote the sustainable evolution of economic–social–ecological systems.

Within the framework of green governance, the government functions as the primary provider of energy and environmental policies. It dominates the enactment of environmental laws and regulations, the formulation of green development strategies, and the oversight of policy implementation. Social organizations within the market, meanwhile, are significant participants in green governance, contributing auxiliary functions such as monitoring, evaluation, and feedback. Throughout the evolution of green governance, the government has played a pivotal leadership role [13], while market-based green governance acts as a crucial complement to government efforts, forming a dual structure known as GMGG. This architecture enhances the efficiency of ecological-resource use and conservation. At the same time, it heightens environmental awareness and oversight among market actors, motivating their active participation in green governance. Consequently, governmental governance costs decline, and both the efficiency and efficacy of green governance are markedly enhanced.

China’s market economy system has been continuously refined in recent years. Consequently, a new coupling of an effective government with an efficient market has emerged as the dominant feature of contemporary green governance [14,15]. The symbiosis and win-win between the government and the market has become the main theme of current research. In the process of green development, the government and the market constitute two pivotal mechanisms for resource allocation. They are both interconnected and complementary, jointly advancing the strategic goal of green, low-carbon, and high-quality development [16].

The term “collaboration” can be traced back to the group theory of Bentley in the 1940s and intergovernmental cooperation studies in the 1960s. Subsequently, Ansell and Gash [7] first systematized the concept of collaborative governance. They defined it as “a collective decision-making process in which one or more public institutions engage with non-governmental stakeholders to formulate or implement public policies, manage public projects, or oversee public assets.” Emerson et al. [17] further expanded this definition. They describe collaborative governance as “an institutional arrangement and process that enables actors to cross the boundaries of public institutions, governmental levels, and public, private, and civic domains to achieve common purposes.”

Institutional theory posits that collaboration significantly reduces transaction costs and overcomes collective action dilemmas in the context of green innovation by reshaping formal and informal institutional arrangements. Williamson’s transaction cost framework [18] indicates that high asset specificity and uncertainty can lead to failures in both market and hierarchical governance. North [19] further emphasizes that effective institutions can lower information, negotiation, and enforcement costs. Feiock’s model of institutional collective action [20] argues that collaborative mechanisms internalize the externalities of cross-boundary green innovation and curb opportunism. By lowering search, contracting, and regulatory costs, the model offers an actionable framework for measuring collaboration intensity and evaluating governance structures.

This paper focuses on the collaborative governance of green initiatives between the government and the market. It examines the ongoing institutional interactions between these two entities concerning green innovation and low-carbon transition. The government provides regulations and incentives, while the market contributes technology, capital, and oversight. Through mechanisms such as environmental regulations, fiscal subsidies, green finance, and media supervision, the externalities of green innovation behaviors are internalized, resulting in dual enhancements of environmental and economic performance.

Unlike existing research on green governance and green innovation, this study will actively investigate how synergistic efforts between the effective government and the efficient market influence CGI behaviors. It aims to explore the potential impact mechanisms or pathways: Are there synergistic effects between GMGG instruments in promoting CGI? How does the GMGG synergy influence CGI? Moreover, does regional or industry-specific heterogeneity affect these relationships? Answering these questions will help clarify how GMGG synergy affects green innovation at the firm level and reveal additional impacts of green governance on business behavior. This research will offer nuanced policy recommendations to policymakers, facilitating the enhancement of green governance’s top-level design and leveraging collaborative governance effectiveness among market entities.

The marginal contribution of this study has two aspects:

(1)

The extant literature has largely focused on the isolated effects of individual governmental green governance instruments, leaving the systemic implications of GMGG influence on CGI behaviors underexplored. This study clarifies how GMGG synergy affects CGI and identifies the underlying transmission channels. It advances scholarly understanding of the economic consequences of green governance and extends both the conceptual breadth and theoretical depth of the field.

(2)

Scholars have primarily analyzed the drivers of green innovation within enterprises from internal and external perspectives, yet research on the coordinated influence of green governance from governmental and market perspectives remains scarce. This study introduces a capacity coupling coefficient model to quantify the synergy between governmental and market green governance systems. Empirical testing of the model evaluates the impact of GMGG synergy on CGI, thereby enhancing understanding of the factors driving green innovation.

The framework of this paper is outlined as follows: Section 2 provides a literature review, Section 3 presents the research hypotheses, Section 4 describes the research design, Section 5 shows the empirical results, Section 6 offers the discussion, and Section 7 concludes with conclusions and recommendations.

2. Literature Review

2.1. Research on the Relationship Between Government and Market in the Green Governance System

In the long-term development of environmental and ecological governance system, dynamics and uncertainty are crucial factors influencing governance decision outcomes [21]. The social environment and ecosystems are dynamically evolving, leading to nonlinear and complex interactions between social and environmental domains. Simultaneously, green governance processes led by governments with shared market participation face numerous unknown variables. These include information asymmetry about the causes or consequences of specific environmental issues, as well as varying attitudes of market stakeholders toward environmental issues [22]. To address these uncertainties, scholars have introduced the concept of collaborative governance into the field of green governance [4], actively exploring the profound impacts of collaborative governance by market stakeholders on corporate behavior. Collaborative governance in green development facilitates the sharing of decision-making authority and ecological information across multiple organizations [23,24], and reduces government green governance costs through systematic coupling effects [25].

Global evidence demonstrates that efficient market participation can substantially lower the fiscal burden of green governance. Nevertheless, proactive government intervention remains indispensable for ensuring effective implementation. Decentralized governance models have proven impractical. This underscores the institutional constraints of government-led green governance, which must maximize public interests, safeguard the rights of vulnerable groups [26], prevent abuses of power and resource over-exploitation, mitigate market failures and externalization [27], and compel stakeholders to consider environmental impacts and sustainable development factors. Despite shifting towards diverse market-oriented approaches, government enforcement and regulatory bodies continue to play a pivotal role [28]. Given the complexity of public and long-term interests involved in green governance, compounded by conflicts of interests, market mechanisms tend to prioritize private and short-term gains, while societal and public awareness of green, environmental, and low-carbon concerns remains limited [29]. Consequently, reliance solely on market forces and individual awareness proves inadequate in addressing challenges in green development, necessitating government leadership in the current phase of green governance. This involves guiding and regulating market behaviors to ensure activities occur within reasonable bounds [30].

Under GMGG frameworks, the shift to green production and the creation of sustainable value chains require heightened collaboration and oversight among diverse stakeholders. Consequently, coordinated market action becomes an indispensable complement to government-led green governance [31,32]. In terms of selection and competition logic, market involvement in green governance can overcome the boundary constraints of government-led initiatives by leveraging market instruments such as pricing, finance, transactions, and communication. This enhances the capacity for green resource allocation and improves the social benefits of green public goods and services, thereby promoting cross-sectoral spatial governance cooperation. The synergy between government and market can provide holistic solutions for green governance, encompassing management systems, governance paradigms, and delineation of responsibilities between government and market. GMGG thereby enables cross-organizational and inter-departmental collaboration and coupling, fusing diverse knowledge streams into green-innovation processes and illuminating how firms respond to green-governance decisions from multiple perspectives [33]. It enables the systematic evaluation and monitoring of corporate environmental behavior, effectively reducing the costs of government green governance while enhancing its efficiency and efficacy. Hence, it incentives firms to address environmental and sustainable development challenges through green innovation.

2.2. Research on the Synergistic Effects of Government Green Governance Instruments in Influencing Green Innovation

In recent years, the role of government green regulations in promoting CGI has garnered increasing attention. Ashford [34] posited that stable implementation schedules combined with well-calibrated design standards for environmental policy instruments can facilitate the formulation of green technology innovation strategies at the firm level. Building on this, Porter and van der Linde [9] elucidated the underlying mechanism by which environmental regulation stimulates innovation. They found that appropriately designed policy instruments enhance firm-level productivity, thereby increasing market profitability and offsetting the additional costs compliance imposed by regulation. Jennings and Zandbergen [35] further argued that escalating regulatory pressure and rising prices for environmentally preferable inputs increase the expected cost of non-compliance, incentivizing firms to pursue green innovation as a pathway to sustained competitiveness. Building on this framework, Cole et al. [36] demonstrated that environmental policy instruments exert a positive effect on CGI, and that this effect is especially pronounced in jurisdictions characterized by stringent environmental regulation. Conversely, Zhao [37] advanced the view that environmental policy instruments can crowd out R&D expenditures by inflating firms’ environmental compliance costs, thereby exerting a negative influence on green technology innovation.

Recent literature underscores the dynamic nature of the regulation–innovation nexus. Exploiting German firm-level data, Bitat [38] revealed that only long-term regulatory regimes complemented by market-based incentives generate statistically significant positive effects on green innovation, whereas traditional command-and-control regulations remain ineffective at the corporate level. Du et al. [39] documented a regime shift whereby the marginal effect of environmental regulation transitions from inhibitory to stimulatory as economic development advances. Leveraging a panel of 31 OECD economies, Hassan [40] demonstrated that energy taxes exert positive short- and long-run impacts on CGI. Consistent with these findings, Mahmood [41], also utilizing OECD data, showed that eco-taxes not only foster CGI but also contribute to broader sustainable-development trajectories. Employing unique data from the China–Pakistan Economic Corridor, Raza [42] documented that environmental regulations enhance agricultural green technology innovation and, consequently, improve green supply chain performance.

Given the specificity of ecological and environmental issues, governments typically employ a variety of green governance policies. The judicious selection of policy instruments is crucial to ensuring policy effectiveness [43]. However, the externalities and high-risk nature of green innovation activities discourage companies from engaging in green innovation [44,45]. To address this issue, many countries have implemented incentive-based green governance policies centered on fiscal subsidies, rewards, and tax rebates on the foundation of regulatory green governance [46], aiming to overcome internal resource constraints within enterprises [47].

Some scholars argue that the combination of government coercive and incentive-based green governance instruments leads to an “innovation synergy effect,” generating additional innovation incentives [48,49,50]. For instance, He [51] constructed a model of coexistence of incentive and constraint mechanisms for green innovation induction. The model suggested that the dual externalities of green innovation can only be addressed by integrating government incentives and constraints to create complementary effects. Yuan and Zheng [48] developed a comprehensive model to analyse the interaction between government subsidies and environmental regulations on firms technological innovation. The results confirmed that the appropriate use of government subsidies and other incentive measures can mitigate the negative impact of environmental regulations and encourage firms to innovate technologically in advance. Based on this, Yu et al. [44] also believe that government constraints and incentives in the environmental field will have a synergistic effect on green innovation in resource-based enterprises. Liu et al. [45] validated in the same model the coupling effects of environmental regulations and government green subsidies on CGI. The findings reveal that environmental regulations have a threshold effect on the impact of green process innovation in the presence of government subsidies. Thus, the intensity of government regulatory constraints and green subsidies should be increased. Yang [52] also found that as government R&D subsidies increase, the promotion effect of environmental regulations on the green total factor productivity of enterprises is enhanced. Wang and Li [53] also reached a similar conclusion, suggesting that the combination of environmental regulations and government R&D subsidies can better enhance substantive CGI. These research results all indicate that different green governance instruments may have a synergistic effect on CGI, but the realization of this synergy requires specific institutional environments and conditions.

Some point also argue that government regulatory constraints and incentives may sometimes have a reverse effect on promoting green innovation. For example, Yi et al. [54] found that government incentives, primarily in the form of subsidies, had no significant positive impact on the efficiency of environmental regulations and green innovation in the manufacturing industry. Zhang and Zhao [55] suggest that environmental constraints from the government weaken the U-shaped impact of subsidy incentives on green innovation in heavy-polluting firms in China, with a more pronounced weakening effect as environmental regulations become stronger. Additionally, some perspectives suggest that the synergistic effects of government incentives and policy constraints are not evident. For instance, Li et al. [56] argue that while environmental constraints significantly promote innovation in enterprises, government incentives do not generate a synergistic effect.

Scholars have conducted extensive research into the roles of government and the market in green governance systems, as well as the synergistic effects of government green governance instruments. These studies have yielded diverse conclusions from various perspectives, laying the groundwork for our research. However, these studies primarily focus on the roles of government and market in green governance systems. The selection and measurement of variables in empirical processes used to assess government and market green governance remains uncertain. Moreover, little research has been conducted into the impact of GMGG synergy on CGI. Although a few studies have examined synergies among governmental green governance instruments, the existence and CGI consequences of synergies between these instruments and market-based green governance remain largely unexplored.

Therefore, this study constructs a conceptual model to elucidate the logical relationship between public and personal environmental governance (Figure 1).

3. Research Hypotheses

3.1. The Impact of the GMGG Synergy on CGI

As global environmental issues and sustainable development become the focus of international attention, the participation of an efficient market in green governance systems led by an effective government has become increasingly important. From the perspective of market-level green governance, various market entities such as media, banks, insurance companies, securities firms, trusts, and funds engage in green governance through media supervision and participation in green finance. They can leverage mechanisms such as supervision, transactions, pricing and evaluation to make environmental regulations more efficient and effective, promote environmental tax reform and set up a market-oriented green product certification system. This will help to remove bottlenecks in corporate green development. Accordingly, markets can optimise the allocation of green resources by reflecting changes in supply and demand, and by transmitting this information. This reduces government green governance costs, improves efficiency, and promotes the development of CGI.

The synergy between government and market mechanisms in green governance can effectively leverage the “visible hand” of the government and the “invisible hand” of the market to drive CGI, thereby enhancing the willingness of firms to implement such innovations. However, excessive collaboration between government and market in green governance may suppress the intrinsic motivation for CGI, leading to diminishing marginal effects. This phenomenon of institutional fatigue has received theoretical support within the field of economics. For instance, grounded in the theory of government regulatory failure, North [19] posits that when external regulations become overly complex, the growth rate of supervision and compliance costs can exceed the marginal benefits derived from coordination. Additionally, the resource-based view suggests that over-reliance on government resources may crowd out the development of firms’ internal innovation absorption capabilities, thereby undermining the motivation for green innovation [57]. Consequently, when the degree of government–market synergy exceeds a critical threshold, firms may become overly dependent on government and market regulatory and evaluative functions, resulting in a loss of intrinsic motivation to engage in green innovation activities. This decline in environmental responsibility awareness may further diminish the enthusiasm for green innovation, leading to a decrease in the marginal effects of the coupling and coordinated development of government–market green governance on CGI. For example, when the proportion of government subsidies allocated to a firm’s green investments exceeds a certain threshold, the firm tends to rely on technology acquisition and equipment purchases to obtain the subsidies, rather than engaging in independent research and development. This phenomenon results in “subsidy dependence” and “crowding out of R&D” [58]. Reliance on end-of-pipe governance technologies or external technology acquisition over a prolonged period may cause firms to adopt fixed technological trajectories. This diminishes their motivation for exploratory green learning and ultimately inhibits breakthrough green innovations [59]. Therefore, we propose a “critical point” hypothesis:

Hypothesis 1a. 

The marginal utility of the impact of GMGG synergy on CGI exhibits a diminishing trend.

Government and market coupling and coordination in green governance can leverage governmental policy guidance in environmental regulation and fiscal incentives, while also mobilizing the active participation of financial institutions and media in the market. This dual influence helps synergistically enhance the efficiency and effectiveness of green governance. Faced with the combined impact of governmental and market-driven green governance, enterprises can engage in substantive green innovation more effectively to respond to coordinated regulation from both sectors. This enables enterprises to reduce environmental compliance costs in the long term, utilise resources more efficiently and enhance business performance. This establishes sustainable business models and gives enterprises a competitive advantage in the market. Therefore, the following hypothesis is proposed:

Hypothesis 1b. 

The impact of GMGG synergy on substantive CGI is more sensitive.

3.2. The Intermediary Effect of Marketization Level

Based on the theory of collaborative governance, the implementation of green governance requires the coordinated efforts of government and market. Coordinating governance policies between the government and the market helps to direct the attention of regional markets towards green development. This involves leveraging market mechanisms such as information disclosure and price systems to optimise resource allocation efficiency and enhance market fairness. Coupling and coordinating government and markets in green governance can improve the environment for green development within markets. It can also increase market entities’ identification with green development concepts and guide markets towards cleaner, more sustainable practices. This enhances the level of regional marketisation. A higher level of marketization signifies a stronger role played by the market in resource allocation [60]. Enterprises can capitalize on unique market opportunities and scale advantages through the development of green technologies and services [61], thereby promoting corporate green technology research and product innovation. Additionally, market-oriented regulatory mechanisms can strengthen the enforcement of corporate environmental policies, achieve energy conservation, emission reduction, and environmental compliance through green innovation. This assist enterprises in achieving sustainable development [62]. Therefore, this paper proposes the following hypothesis:

Hypothesis 2. 

The marketization level acts as an intermediary between the GMGG synergy and CGI.

3.3. The Regulatory Effect of Digital Economic Development Levels

The digital economy, through technologies such as big data and cloud computing, is grounded in the theory of information asymmetry and signalling. It reduces information asymmetry between firms and external investors, thereby improving the market’s ability to identify and reward corporate green innovation initiatives. This enhanced transparency increases the efficiency of resource allocation for innovation, and facilitates data collection and sharing [63], which in turn amplifies the green-innovation effect of GMCC synergy. Under the coordinated green governance of government and markets, digital economic development enables enterprises to better understand and access market demands and consumption trends, promoting optimized allocation of innovation resources in the market [64,65]. This facilitates more effective green innovation and product development by enterprises, thereby mitigating the impact of GMGG synergy on businesses. Additionally, digital regulatory measures can precisely monitor enterprises’ environmental efficiency and achievements. Data-based regulatory methods can enhance the efficiency and credibility of government supervision, stimulate enterprises’ environmental enthusiasm and sense of responsibility, enabling them to better cope with increased environmental compliance costs resulting from GMGG synergy. Therefore, the following hypothesis is proposed:

Hypothesis 3. 

The level of digital economic development helps mitigate the impact of GMGG synergy on CGI.

Grounded in theoretical analysis and research hypotheses, this study conceptualizes the influence of GMGG on CGI within a three-layer framework of direct, mediating, and moderating effects (Figure 2). Specifically, GMGG not only exerts a direct impact on CGI but also transmits its effect indirectly via the mediating channel of marketization. Furthermore, the development of the digital economy moderates this relationship, amplifying the positive influence of GMGG on CGI.

4. Research Design

4.1. Data Sources

For comparative purposes, the sample comprises Chinese A-share listed firms from 2010 to 2020. Observations for 2021–2023 are omitted to avoid potential distortions arising from the pronounced impact of the COVID-19 pandemic on corporate innovation behavior. CGI is computed based on relevant data from the CNRDS database (https://www.cnrds.com (accessed on 12 May 2024)). Calculation of GMGG synergy relies on data from the CSMAR database (https://data.csmar.com (accessed on 18 May 2024)), China Environmental Statistics Yearbook, and China Statistical Yearbook, among others. Provincial levels of marketization are sourced from the China Provincial Marketization Index database (https://cmi.ssap.com.cn (accessed on 25 May 2024)). Digital economic development levels are computed using data from CSMAR, Wind databases, and Beijing University Digital Finance Research Center. Control variables data are sourced from CSMAR and Wind databases. To ensure data reliability, the study excludes samples from ST, *ST, and financial enterprises. Furthermore, truncation is applied to main variables at the 1st and 99th percentiles. Following these procedures, the study employs a dynamic panel dataset comprising 28,451 observations from 3807 listed companies.

4.2. Variable Selection

4.2.1. The Core Explanatory Variable

The core explanatory variable in this study is the GMGG synergy. Essentially, GMGG constitutes a complex system comprising multiple subsystems. Therefore, determining representative subsystems to measure the synergy between government and market green governance systems is a critical unresolved issue.

This study selects two subsystems, environmental regulation and green subsidies, at the government level to represent constraint-based green governance and incentive-based green governance, respectively. Drawing from studies of Lanoie et al. [66], this study calculates environmental regulation by taking the logarithm of the ratio of industrial pollution control investment to the added value of the industrial secondary industry in each province (municipalities and autonomous region) to measure the strength of environmental regulation. A higher ratio indicates stronger environmental regulation. As for green subsidies, this study aggregates the amounts of financial subsidies, incentives, and tax refunds provided by the government for green development projects on an annual basis and processes this data by taking the logarithm.

Within the existing green governance system, the deep integration of digital technology and new, integrated media has led to an exponential increase in the speed at which information is disseminated. This change has gradually made media supervision a critical factor in influencing corporate reputation and profitability [67], and has had a profound impact on CGI [68]. Green finance, involving multiple market participants such as banks, insurance, securities, trust, and funds, can fully reflect the trend of price mechanism changes and influence the allocation efficiency of green innovation resources and green sustainable development [69], making green financial governance another important factor affecting CGI [70,71]. Therefore, based on the current research trends and key issues in market green governance systems, this study selects media supervision and green financial support as representative subsystems of market green governance, respectively, serving as proxy variables for market constraint-based and incentive-based green governance.

In order to provide a more accurate measurement of media governance, this study adopts the methodology employed by Jia et al. [72]. This involves obtaining annual news reports of listed companies on the Shanghai and Shenzhen stock exchanges from the CSMAR database, and extracting the number of negative keywords related to the environment and resources reported by the media. This number is then taken as the logarithm and used as a proxy variable for media supervision. Referring to the research of Zhang and Mei [73] and Xu et al. [74], this study selects five indicators—green credit, green investment, carbon finance, green securities, and green insurance—and employs the entropy method to calculate the comprehensive index of green finance.

Table 1. Calculation of proxy indicators for various subsystems of GMGG.

Items	Proxy Variables	Explanation of Indicator Calculation
Government constraint-based green governance	Environmental regulations	The natural logarithm of the ratio of industrial pollution control investment to the added value of the industrial secondary industry in each province (city, district) for the current year.
Government incentive-based green governance	Green subsidies	The natural logarithm of the total amount of fiscal rewards, tax refunds, and other green subsidies related to green development.
Market constraint-based green governance	Media supervision	The natural logarithm of the number of negative keywords related to corporate environment and resources in media reports.
Market incentive-based green governance	Green finance support	Selection of five indicators: green credit, green investment, carbon finance, green securities, and green insurance, and calculation of a comprehensive index using the entropy method.

presents the descriptions of the specific components used to construct the GMCC.

Figure 3 illustrates the composition of GMCC and the relationships among its various indicators.

Due to the different dimensional units of indicators across various subsystems of green governance, this study initially standardizes the subsystem indicators to enhance comparability. Specifically, positive indicators are standardized as shown in Equation (1), while negative indicators are standardized according to Equation (2).

$$X_{nij}={\displaystyle \frac{x_{nij}-min\left(x_{nij}\right)}{max\left(x_{nij}\right)-min\left(x_{nij}\right)}}$$

(1)

$$X_{nij}={\displaystyle \frac{max\left(x_{nij}\right)-x_{nij}}{max\left(x_{nij}\right)-min\left(x_{nij}\right)}}$$

(2)

Apart from constraint-based green governance, incentive-based green governance, media scrutiny governance, and green financial governance are all considered positive indicators.

In this context, C_i represents the standardized value of environmental regulation, S_i represents the standardized value of green subsidies. M_i represents the standardized value of media supervision, and F_i represents the standardized value of green finance. The coupling degree C_i of the four subsystems is calculated using Equation (3).

$$C_i={\displaystyle \frac{4\sqrt[4]{G_i\times S_i\times M_i\times F_i}}{G_i+S_i+M_i{+F}_i}}$$

(3)

In Formula (3), a larger value of C_i indicates a lower degree of disparity among subsystems within the GMGG, implying better coupling and more orderly system operation. Conversely, a smaller C_i value suggests greater disparity among subsystems, poorer coupling, and a tendency towards disorderly development.

Because both government and market systems are dynamic and often imbalanced, current methods for measuring their interaction still face limitations, particularly in accounting for positive and negative couplings between subsystems, potentially leading to errors in assessing overall system functionality or comprehensive benefits. To mitigate these shortcomings, scholars have introduced the capacity coupling coefficient model to objectively measure the extent of synergy and complementary advantages among subsystems [75]. Thus, this study introduces the Coupling Coordination Index as an indicator for assessing the GMGG synergy, aiming to address the challenges posed by dynamic changes and imbalance within the system. The specific calculation formulas are as follows:

$$Q_{ni}=W_{n1}{G}_i+W_{n2}{S}_i+W_{n3}{M}_i{+W_{n4}F}_i$$

(4)

$$D_i=\sqrt{C_i\times Q_i}$$

(5)

In Formulas (4) and (5), D_i represents the GMGG synergy, Q_ni denotes the comprehensive evaluation index reflecting the overall level of green governance subsystems, and W_nj stands for the weighting coefficients of each subsystem.

Based on the research by Zou et al. [76], this study employs an improved multi-objective decision-making weight calculation method that combines the Analytic Hierarchy Process (AHP) with the Entropy Weight Method (EWM). This approach allows for the incorporation of both expert subjective judgments and objective information regarding data dispersion, effectively overcoming the issues of single subjective bias and sensitivity to extreme values. As a result, it yields more robust and reliable judgment outcomes.

First, the indicator data were standardized according to Equations (1) and (2), and the resulting standardized values were substituted into Equation (6) to obtain the share of indicator j for province (municipalities and autonomous region) i in year n.

$$P_{nij}={\displaystyle \frac{X_{nij}^{\prime }}{{\sum }_{n=1}^n{\sum }_{i=1}^m{X}_{nij}^{\prime }}}$$

(6)

Secondly, the information entropy of indicator j in year n was calculated using Equation (7) based on the outcomes derived from Equation (6). Information entropy quantifies the uncertainty or randomness of information. The greater the uncertainty, the larger the entropy.

$$E_{nj}=-{\displaystyle \frac{1}{\mathrm{l}\mathrm{n}\left(m\right)}}{\sum }_{n=1}^n{\sum }_{i=1}^m{X}_{nij}^{\prime }·\mathrm{l}\mathrm{n}(P_{nij})$$

(7)

Third, the coefficient of variation for indicator j in year n was computed according to Equation (8). A higher coefficient indicates that indicator j conveys more effective information.

$$G_{nj}=1-E_{nj}$$

(8)

Subsequently, the coefficient obtained from Equation (8) was then inserted into Equation (9) to derive the entropy weight of indicator j in year n.

$$W_{nj}={\displaystyle \frac{G_{nj}}{{\sum }_{j=1}^m{G}_{nj}}}$$

(9)

Fourth, the weights under the AHP, denoted as W_r, were obtained using the SPSSAU platform. The final composite weights, W_nj*, were determined by assigning equal (i.e., 50%) importance to the entropy and AHP weights. Equation (10) was subsequently applied to calculate the weight of subsystem j in year n.

$${W_{nj}}^{\ast }=\left({\displaystyle \frac{G_{nj}}{{\sum }_{j=1}^m{G}_{nj}}}+W_r\right)/2$$

(10)

Finally, leveraging the yearly weights of GMGG together with the standardized indicator values, the GMGG synergy index for province (municipalities and autonomous region) i in year n was obtained via Equation (11).

$${\mathrm{Q}}_{ni}={W_{nj}}^{\ast }\times X_{nij}$$

(11)

4.2.2. Explained Variables

The first aspect concerns the quantity of CGI. Given the difference in time between green patent applications and grants, this study uses the number of green patent applications to measure the amount of CGI. Furthermore, in line with the perspectives of Lin et al. [77] and Xu et al. [78], green invention patents are subject to substantive examination and typically involve a higher level of technological content. Consequently, this study utilizes the number of green invention patent applications as a measure of substantive green innovation. In contrast, green utility model patents require only formal examination and thus have a lower threshold for innovation. Therefore, the quantity of green utility model patents is employed to gauge the extent of strategic green innovation. Recognizing the lag effect of GMGG synergy on CGI, this study processes CGI data with a one-period lag. This approach also partially mitigates any residual autocorrelation in the data.

The second aspect pertains to the quality of CGI. This study standardises the quantities of substantive and strategic green innovation by dividing them by the total volume of green innovation. This normalization yields indicators of the proportion of substantive and strategic green innovation, which serve to measure the quality of green innovation. Thus, a higher preference for substantive green innovation indicates higher quality of green innovation by the enterprises.

4.2.3. The Mediating Variable

The mediating variable is the level of marketization. This study utilizes the Marketization Index database of China by province (https://cmi.ssap.com.cn (accessed on 28 May 2024)) to measure this variable. The index is a widely adopted indicator at present for assessing the relative progress of marketization across provinces (cities, districts). Specific indicators include the development level of government–market relations, product markets, factor markets, non-state economy development, development of market intermediaries, and legal institutional environment. To enhance comparability of data across different years, this paper applies technical adjustments to the Marketization Index, enabling longitudinal comparisons across years.

4.2.4. The Moderator Variable

The moderator variable of this study is digital economy index. Drawing on methodologies from Zhao et al. [79], five indicators of internet penetration rate, number of internet related practitioners, internet related output, number of mobile internet users and digital inclusive finance index are selected, and the EWM is used to calculate the digital economic index of each region. The specific calculation process is shown in

Table 2. The measurement basis of digital economy.

Primary Indicators	Secondary Indicators	Indicator Calculation Basis	Indicator Attributes
Digital economy index	Internet penetration rate	Internet users per 100 people	+
	Number of Internet related practitioners	The proportion of computer service and software professionals to the total population of the region	+
	Internet related output	Per capita total telecommunications services	+
	Number of mobile Internet users	Number of mobile phone users per 100 people	+
	Digital Inclusive Finance Index	Digital Inclusive Finance Index released by Peking University	+

.

4.2.5. Control Variables

Referring to existing research, this study controlled for variables related to CGI, including the scale of enterprises, establishment time, profitability, growth ability, debt paying ability, board size, the proportion of independent directors, equity concentration, and enterprise value. To mitigate issues arising from autocorrelation and heteroskedasticity, the scale of enterprises, establishment time, and board size were log-transformed prior to analysis.

Table 3. Main variables and interpretation.

Type of Variables	Name of Variables		Symbol of Variables	Interpretation of Variables
Explained variables	The number of CGI	Total number of CGI	CGI	The natural logarithm of the total number of green patent applications plus 1 and lagged by one period
		The number of substantial CGI	CGI_S	The natural logarithm of the number of green invention patent applications plus 1 and lagged by one period
		The number of strategic CGI	CGI_C	The natural logarithm of the number of green utility model patent applications plus 1 and lagged by one period
	The proportion of CGI	The proportion of substantial CGI	CGI_SL	The proportion of green invention patent applications to the total number of green patents and lags behind by one period
		The proportion of strategic CGI	CGI_CL	The proportion of green utility model patent applications to the total number of green patents and lags behind by one period
The core explanatory variable	GMGG synergy		GM	Calculation using capacity coupling coefficient model
The mediating variable	Marketization		Market	Using marketization index for calculation, representing the relative progress of marketization in each province (city, district)
The moderator variable	Digital economy		Digital	Digital economy index calculated by EWM
Control variables	The scale of enterprises		Scale	The natural logarithm of a company’s total assets
	Establishment time		Age	The natural logarithm of the establishment time of a company
	Profitability		ROA	Net profit/total assets
	Growth ability		Growth	Revenue growth/opening revenue
	Debt paying ability		Debt	Total ending liabilities/total assets
	Board size		Board	The natural logarithm of the number of board members
	The proportion of independent directors		Indep	The proportion of independent directors to all directors
	Equity concentration		Top10	The shareholding ratio of the top 10 shareholders
	Enterprise value		TQ	The value of Tobin Q

shows the Main variables.

4.3. Model Construction

This paper conducts a Hausman test on the data, and the results indicate that a fixed-effects model should be adopted. Considering that the synergy of GMGG may vary due to time and industry factors, this study incorporates time and industry fixed effects. Meanwhile, to address potential heteroskedasticity and autocorrelation, we estimated all regressions with standard errors clustered at the province level. This procedure relaxes the conventional assumption of independent and identically distributed errors, allowing for arbitrary within-cluster correlation and variance structures while maintaining consistency of the point estimates. Consequently, inference remains valid even in the presence of region-specific heteroskedasticity or serial correlation in the disturbances.

To investigate the influence of GMGG synergy on CGI, this study constructs the following model:

$${CGI}_{i,t+1}={\alpha }_0+{\alpha }_1{GM}_{i,t}+{\alpha }_2{{GM}_{i,t}}^2+{\alpha }_3{Controls}_{i,t}+{year}_t+{industry}_i+{\epsilon }_{i,t}$$

(12)

To examine the mediating effect of the level of marketization between the GMGG synergy and CGI, this study constructs the following models:

$${Market}_{i,t}={\beta }_0+{\beta }_1{GM}_{i,t}+{\beta }_2{GM}^2+{\beta }_3{Controls}_{i,t}+{year}_t+{industry}_i+{\epsilon }_{i,t}$$

(13)

$$C{GI}_{i,t+1}={\gamma }_0+{\gamma }_1{Market}_{i,t}+{\gamma }_2{GM}_{i,t}+{\gamma }_3{{GM}_{i,t}}^2+{\gamma }_4{Controls}_{i,t}+{year}_t+{industry}_i+{\epsilon }_{i,t}$$

(14)

To test the moderating effect of the digital economy, this study constructs the following model:

$${CGI}_{i,t+1}={\delta }_0+{\delta }_1{GM}^2\times {Digital}_{i,t}+{\delta }_2{GM}_{i,t}\times {Digital}_{i,t}+{\delta }_{3}GM+{\delta }_4{{GM}_{i,t}}^2+{{\delta }_{5}Digital}_{i,t}+{{\gamma }_{6}Controls}_{i,t}+{year}_t+{industry}_i+{\epsilon }_{i,t}$$

(15)

In models (12)–(14), CGI_i,t+1 represents the lagged term for the total amount of CGI, the application volume and proportion of strategic and substantive CGI, GM_i,t denotes the GMGG synergy, Market_i,t signifies the mediating variable of the level of marketization, Digital_i,t stands for the moderating variable of the digital economy, year_t represents the time fixed effect, and industry_i represents the industry fixed effect. Meanwhile, α₀, β₀, γ₀, and δ₀ are intercept terms; α_1−n, β_1−n, γ_1−n, and δ_1-n are correlation coefficients; and ε_i,t denotes the random disturbance term.

4.4. Descriptive Statistics

Table 4. Descriptive statistics of main variables.

Symbol of Variables	N	Min	Max	Mean	SD
CGI	28,451	0	7.319	0.846	1.167
GM	28,451	0.013	0.928	0.631	0.220
Market	28,451	2.330	12	8.518	1.826
Digital	26,825	0.0770	0.982	0.457	0.194
Scale	28,451	17.43	28.64	22.16	1.328
Age	28,451	0	4.143	2.842	0.371
ROA	28,451	−0.366	0.206	0.0360	0.0670
Growth	28,451	−0.789	10.29	0.437	1.272
Debt	28,451	0.0490	0.980	0.425	0.210
Board	28,451	1.386	2.944	2.245	0.179
Indep	28,451	0.125	0.800	0.375	0.0560
Top10	28,451	0.837	4.627	4.049	0.295
TQ	28,451	0.515	4.814	1.055	0.358

presents the descriptive statistics of the main variables. The minimum and maximum values of total green innovation are 0 and 7.319, respectively, with an average of 0.846, indicating significant variation in green innovation levels among Chinese listed companies. The minimum and maximum values of GMGG synergy are 0.013 and 0.928, respectively, with a mean exceeding the moderate coordination level (0.6), suggesting substantial differences among regions. There are varying degrees of difference among other variables as well, and the selected variables have variance inflation factors (VIF) all below 10, indicating no high collinearity among variables, thus suitable for subsequent empirical analysis in this study.

From the kernel density curve of GMGG synergy from 2016 to 2020 shown in Figure 4, a unimodal distribution is evident, indicating a clear trend of concentration. Furthermore, the highest peak is shifting to the right, suggesting that the GMGG synergy among provinces (municipalities and autonomous region) in China is continuously increasing.

From a regional perspective, the kernel density curve of GMGG synergy in the eastern region has the lowest peak (Figure 5), indicating a more balanced level of collaboration, while the northeastern region exhibits the highest peak, suggesting an uneven distribution and a trend of high concentration. Upon analysing the reasons for this, it can be concluded that the higher level of marketisation and the wide range of industries in the eastern region make it difficult for the government to implement a ‘one-size-fits-all’ approach across all sectors using a single policy tool. The differing demands of multiple stakeholders regarding green financing, technology pathways, and emission reduction standards force the government to adopt differentiated collaborative strategies, leading to a more dispersed and balanced GMGG synergy within the region. In contrast, the northeastern region is dominated by heavy industry and state-owned enterprises, with a relatively simple market structure and severe lock-in of industrial technology pathways. Consequently, the effects of GMGG synergy policies are limited to a few large enterprises, making it challenging to diffuse these effects to small and medium-sized entities, resulting in a trend of highly concentrated collaboration.

5. Empirical Results

5.1. Baseline Regression

Table 5. Baseline regression.

Variables	(1)	(2)	(3)	(4)	(5)
	CGI	CGI_S	CGI_C	CGI_SL	CGI_CL
GM2	−0.5653 ***	−0.4230 **	−0.3355 **	−0.1543 ***	−0.1006 **
	(0.1589)	(0.1358)	(0.1294)	(0.0384)	(0.0382)
GM	0.4004 **	0.2587 *	0.2148	0.1265 ***	0.0979 **
	(0.1514)	(0.1303)	(0.1236)	(0.0365)	(0.0358)
Controls	YES	YES	YES	YES	YES
Cons	−7.7208 ***	−6.8603 ***	−5.3655 ***	−1.2985 ***	−0.6094 ***
	(0.2312)	(0.2072)	(0.1872)	(0.0521)	(0.0487)
Time FE	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES
N	28,451	28,451	28,451	28,451	28,451
R2	0.273	0.239	0.254	0.144	0.151
Utest (Slope-U)	−0.6485	−0.5261	−0.4077	−0.1598	−0.0887
Utest (Slope-L)	0.3862	0.2481	0.2064	0.1226	0.0954
Utest (Extreme Point)	0.3541	0.3058	0.3202	0.4099	0.4867

Note: Robust standard errors clustered to province in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001.

presents the results of the impact of GMGG synergy on CGI. According to the empirical results in column (1), the coefficients of the quadratic term (GM2) and the linear term (GM) on CGI are −0.5653 and 0.4004, respectively, significant at the 1% and 5% levels.

To identify potential non-linearities, we implemented the U-test proposed by Sasabuchi [80] and refined by Lind and Mehlum [81]. The U-test result indicates a critical value of 0.3541, within the range of coupling coordination values [0.013, 0.928], rejecting the null hypothesis at the 1% significance level. Additionally, the negative slope within this interval suggests an inverted “U” shaped effect of GMGG synergy on CGI, confirming Hypothesis 1. This finding indicates that moderate levels of GMGG synergy significantly promote CGI, but sustainability of this promotion diminishes beyond a critical threshold, confirming Hypothesis 1a.

From the empirical results in columns (2) and (4) of Table 5, it is evident that the quadratic term (GM2) of GMGG synergy has significant effects on substantive and strategic green innovation with coefficients of −0.4230 and −0.1543, respectively, significant at the 5% level. The U-test results indicate critical values of 0.3058 and 0.4099, demonstrating that these critical points fall within the range of observed values. The negative slope values within the interval suggest an inverted “U” shaped relationship between GMGG synergy and both substantive and strategic CGI. From the results in columns (3) and (5), the quadratic term (GM2) of GMGG synergy significantly affects the proportion of substantive and strategic CGI with coefficients of −0.3355 and −0.1006, respectively, significant at the 1% and 5% levels, and confirmed by U-test. In terms of the magnitude and significance of these coefficients, GMGG synergy exhibits greater sensitivity to substantive CGI, validating Hypothesis 1b.

To visually demonstrate the dual impact of GMGG synergy on CGI, this study plotted a fitted curve representing their relationship (Figure 6). The findings corroborate that, within the dataset examined, GMGG synergy exhibits a distinct “double-edged sword” effect on CGI. Furthermore, the figure highlights that the peak enhancement of CGI occurs at a GMGG synergy level of 0.35.

5.2. Robustness Test

5.2.1. Robustness Test with a Lag of Two Periods

Furthermore, this study conducted robustness tests using lagged two-period data of the explanatory variables, with results presented in

Table 6. Robustness test (1).

Variables	(1)	(2)	(3)	(4)	(5)
	CGI_2	CGI_S2	CGI_C2	CGI_SL2	CGI_CL2
GM2	−0.4970 ***	−0.4122 ***	−0.2793 **	−0.1324 ***	−0.0613
	(0.1627)	(0.1383)	(0.1317)	(0.0388)	(0.0381)
GM	0.3492 **	0.2537 *	0.1827	0.1019 ***	0.0654 *
	(0.1541)	(0.1320)	(0.1245)	(0.0369)	(0.0354)
Controls	YES	YES	YES	YES	YES
Cons	−6.6479 ***	−5.9730 ***	−4.5755 ***	−1.1159 ***	−0.4894 ***
	(0.2345)	(0.2070)	(0.1888)	(0.0521)	(0.0493)
Time FE	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES
N	28,451	28,451	28,451	28,451	28,451
R2	0.225	0.197	0.209	0.120	0.125
Utest (Slope-U)	−0.5730	−0.5111	−0.3356	−0.1437	−0.0484
Utest (Slope-L)	0.3367	0.2433	0.1756	0.0985	0.0638
Utest (Extreme Point)	0.3513	0.3078	0.3270	0.3849	0.5328

Note: Robust standard errors clustered to province in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001.

. The empirical results from columns (1)–(3) indicate that the quadratic term of the GMGG synergy has a significantly negative impact on the total volume of CGI, as well as on substantial and strategic CGI. The Utest conclusions remain valid, thereby affirming Hypothesis 1a. Analysis of the empirical results in columns (4) and (5) reveals that the quadratic term of the GMGG synergy has a significant negative impact on the proportion of substantial CGI, while its impact on the proportion of strategic CGI is not significant. This further underscores the higher sensitivity of GMGG synergy to substantial CGI, confirming Hypothesis 1b and indicating the robustness of the study’s findings.

5.2.2. Robustness Test for Replacing the Dependent Variables

Table 7. Robustness test (2).

Variables	(1)	(2)	(3)	(4)	(5)
	CGIA	CGIA_S	CGIA_C	CGIA_SL	CGIA_CL
GM2	−0.2998 **	−0.2282 ***	−0.1943	−0.0660 **	−0.0409
	(0.1351)	(0.0807)	(0.1213)	(0.0282)	(0.0468)
GM	0.1895	0.1306 *	0.1389	0.0436	0.0423
	(0.1280)	(0.0776)	(0.1148)	(0.0273)	(0.0439)
Controls	YES	YES	YES	YES	YES
Cons	−3.9965 ***	−1.8819 ***	−3.1368 ***	−0.3433 ***	−0.7066 ***
	(0.1879)	(0.1117)	(0.1677)	(0.0362)	(0.0613)
Time FE	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES
N	28,130	28,130	28,130	28,130	28,130
R2	0.182	0.098	0.159	0.051	0.129
Utest (Slope-U)	−0.3668	−0.2928	−0.2216	−0.0790	−0.0336
Utest (Slope-L)	0.1820	0.1248	0.1340	0.0419	0.0413
Utest (Extreme Point)	0.3160	0.2861	0.3575	0.3297	0.5168

Note: Robust standard errors clustered to province in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001.

reports the robustness test results obtained by replacing the dependent variables with the natural logarithm of the granted green patents, the natural logarithm of granted green invention patents and utility models separately, and the share of granted patents, while additionally lagging these measures by two periods. It is evident that the quadratic term of GMGG synergy continues to significantly negatively impact the CGI. The conclusions from the U-test remain valid. However, the coefficients for the number and proportion of strategic CGI are not significant. This further underscores the heightened sensitivity of GMGG synergy to substantial CGI, thereby validating the robustness of the study’s findings.

5.2.3. Robustness Test Using an Alternative Measure of the Dependent Variables

The number of post-grant citations a green patent subsequently receives is a better indicator of how widely the technology is recognised, adopted, disseminated and diffused. This provides a direct measure of its impact and socio-economic value. Therefore, we conduct a robustness check by using the natural logarithm of (i) the total citations received by a firm’s granted green patents and (ii) the citations separately received by granted green invention patents and utility models. The empirical results reported in

Table 8. Robustness test (3).

Variables			(1)
	CGI_2	CGI_S2	CGI_C2
GM2	−0.8378 ***	−0.8205 ***	−0.6205 ***
	(0.1819)	(0.1300)	(0.1650)
GM	0.6748 ***	0.7664 ***	0.4010 *
	(0.1741)	(0.1228)	(0.1589)
Controls	YES	YES	YES
Cons	−4.8969 ***	−3.3881 ***	−4.5520 ***
	(0.2819)	(0.2136)	(0.2631)
Time FE	YES	YES	YES
Industry FE	YES	YES	YES
N	28,451	28,451	28,451
R2	0.159	0.134	0.132
Utest (Slope-U)	0.6537	0.3854	0.7458
Utest (Slope-L)	−0.8797	−0.7502	−0.7560
Utest (Extreme Point)	0.4027	0.3232	0.4671

Note: Robust standard errors clustered to province in parentheses, * p < 0.05, *** p < 0.001.

confirm that our main findings remain robust.

5.3. Endogeneity Test

To address potential endogeneity issues that may bias empirical results, this study employs instrumental variable (IV) and two-stage least squares (2SLS) regression. Drawing on approaches by Hering & Poncet [82], this study selects the air circulation coefficient as the instrumental variable. This coefficient is considered a crucial factor influencing environmental pollution dispersion, correlated with natural climatic conditions of the region and exogenous to other variables affecting CGI. Therefore, it exhibits exclusivity and exogeneity. Regression results using the IV approach are presented in

Table 9. Endogeneity test of 2SLS.

Variables	(1)	(2)	(3)	(4)	(5)
	The First Stage		The Second Stage	The Second Stage	The Second Stage
	GM	GM2	CGI	CGI_SL	CGI_CL
IV2	−0.4268 ***	−0.4331 ***
	(0.0099)	(0.0087)
IV	5.9992 ***	6.1099 ***
	(0.1402)	(0.1230)
GM2			−5.1444 ***	−1.1492 ***	−0.7353 ***
			(0.9839)	(−4.9280)	(0.2346)
GM			5.7446 ***	1.2274 ***	0.8766 ***
			(0.9635)	(5.3523)	(0.2308)
Controls	YES	YES	YES	YES	YES
Cons	−19.8236 ***	−20.5460 ***	−9.8863 ***	−1.6908 ***	−0.9587 ***
	(0.5005)	(0.4394)	(0.3515)	(0.0801)	(0.0782)
Time FE	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES
N	28,451	28,451	28,451	28,451	28,451
Kleibergen-Paap rk LM			777.290
Cragg-Donald Wald F			362.436
Stock-Yogo weak ID (10%)			7.03

Note: Robust standard errors clustered to province in parentheses, *** p < 0.001.

. Tests based on the Kleibergen-Paap rk LM statistic and the Cragg Donald Wald F statistic confirm the validity of weak instrument assumptions and instrument relevance, thus justifying the selection of the air circulation coefficient as an instrumental variable. Empirical findings from Table 9 indicate that after incorporating the instrumental variable for air circulation coefficient, the second-stage impact of GMGG synergy on CGI continues to demonstrate an inverted U-shaped relationship. Furthermore, the absolute value of its effect on the proportion of substantial CGI is significantly larger than that on strategic CGI, underscoring greater sensitivity to substantial CGI and validating the consistency of the study’s primary conclusions.

Drawing on the research by Arellano and Bond [83] and Blundell and Bond [84], the two-step generalized method of moments (GMM) effectively mitigates the interference of reverse causality by introducing appropriate instrumental variables and considering heteroscedasticity, thereby addressing the endogeneity issues in dynamic panel data. Consequently, this paper employs this method for further validation. The empirical results presented in

Table 10. Endogeneity test of two-step GMM.

Variables	(1)	(2)
	CGI	CGIA
GM2	−3.614 **	−3.038 **
	(−2.083)	(−2.070)
GM	4.394 ***	3.485 **
	(2.621)	(2.460)
Controls	YES	YES
Cons	−10.397 ***	−5.877 ***
	(−18.625)	(−12.893)
Time FE	YES	YES
Industry FE	YES	YES
N	11,748	11,535
R2	0.282	0.198

Note: Robust standard errors clustered to province in parentheses, ** p < 0.01, *** p < 0.001.

indicate that the diminishing marginal effect of GMGG synergy on CGI remains significant.

5.4. Transmission Mechanism

Firstly, based on the empirical results in

Table 11. Testing results of the transmission mechanism.

Variables	(1)	(2)	(3)	(4)
	Market	CGI	CGI_S	CGI_C
Market		0.0530 ***	0.0453 ***	0.0372 ***
		(0.0037)	(0.0030)	(0.0030)
GM2	−3.8755 ***	−0.3598 *	−0.2475	−0.1913
	(0.2503)	(0.1585)	(0.1353)	(0.1292)
GM	2.4970 ***	0.2680	0.1457	0.1219
	(0.2242)	(0.1508)	(0.1297)	(0.1232)
Controls	YES	YES	YES	YES
Cons	8.1681 ***	−8.1539 ***	−7.2301 ***	−5.6695 ***
	(0.3203)	(0.2328)	(0.2094)	(0.1888)
Time FE	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES
N	28,451	28,451	28,451	28,451
R2	0.293	0.278	0.244	0.258

Note: Robust standard errors clustered to province in parentheses, * p < 0.05, *** p < 0.001.

, the coefficient of the quadratic term of GMGG synergy on the level of marketisation is −3.8755, which is significant at the 1% level. This indicates a typical inverted U-shaped relationship. Secondly, the marketisation level has a significant positive impact on CGI, with a coefficient of 0.0530 at the 1% level. This highlights its promotional effect on CGI and validates Hypothesis 2. Further tests confirm that marketization level positively influences both substantial and strategic CGI, indicating robust results.

5.5. The Regulatory Effect of the Digital Economy

Table 12. Testing results of moderating effect.

Variables	(1)	(2)	(3)
	CGI	CGI_S	CGI_C
DE × GM2	0.8136 ***	0.4824 **	0.7469 ***
	(0.2581)	(0.2273)	(0.2053)
DE × GM	−0.1580 ***	−0.0933 *	−0.1340 ***
	(0.0548)	(0.0485)	(0.0442)
GM2	−0.8866 ***	−0.6687 ***	−0.5152 ***
	(0.1799)	(0.1514)	(0.1460)
GM	0.9940 ***	0.7341 ***	0.5952 ***
	(0.1784)	(0.1499)	(0.1452)
DE	0.0797 ***	0.0676 ***	0.0620 ***
	(0.0080)	(0.0069)	(0.0065)
Controls	YES	YES	YES
Cons	−7.8703 ***	−7.0196 ***	−5.4373 ***
	(0.2378)	(0.2137)	(0.1932)
Time FE	YES	YES	YES
Industry FE	YES	YES	YES
N	26,825	26,825	26,825
R2	0.277	0.243	0.258

Note: Robust standard errors clustered to province in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001.

presents the moderation effect of digital economic level on CGI. The coefficient for the interaction term between digital economic level and the quadratic term of GMGG synergy (DE × GM2) is 0.8136, and it is significantly positive for both substantial and strategic CGI with coefficients of 0.4824 and 0.7469, respectively. It is evident that the advancement in digital economic development mitigates the impact of GMGG synergy on CGI. This validates Hypothesis 3.

5.6. Heterogeneity Analysis

To examine the heterogeneous impact of GMGG synergy on different types of CGI and provide targeted policy recommendations for enhancing green governance and incentivizing the innovation vitality, this study further conducts heterogeneity tests by grouping firms according to the property right, the region, and whether they are heavy-polluting firms.

5.6.1. Group by the Property Right

Furthermore, we conduct a heterogeneity analysis by partitioning the sample into state-owned and non-state-owned firms based on ownership type. As shown in columns (1)–(3) of

Table 13. Group by property right.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	State-Owned Group			Non-State-Owned Group
	CGI	CGI_S	CGI_C	CGI	CGI_S	CGI_C
GM2	−0.0299	−0.0604	0.0104	−0.9299 ***	−0.2104 ***	−0.1619 **
	(0.2568)	(0.0596)	(0.0588)	(0.1999)	(0.0513)	(0.0516)
GM	−0.1309	0.0521	−0.0023	0.8437 ***	0.1784 ***	0.1570 **
	(0.2408)	(0.0551)	(0.0539)	(0.1914)	(0.0499)	(0.0491)
Controls	YES	YES	YES	YES	YES	YES
Cons	−8.1366 ***	−1.2715 ***	−0.6246 ***	−6.5213 ***	−1.2699 ***	−0.5284 ***
	(0.3698)	(0.0857)	(0.0795)	(0.3291)	(0.0763)	(0.0723)
Time FE	YES	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES	YES
N	10,179	10,179	10,179	17,776	17,776	17,776
R2	0.371	0.199	0.198	0.206	0.116	0.132

Note: Robust standard errors clustered to province in parentheses, ** p < 0.01, *** p < 0.001.

, the coefficient of GMGG synergy on CGI is insignificant for the state-owned group. In contrast, it is significant across all specifications for the non-state-owned group. It indicates that non-state-owned firms are more responsive to GMGG synergy in driving CGI, a result consistent with their stronger competitive orientation and higher propensity for innovation.

5.6.2. Group by Whether They Are Heavy-Polluting Firms

As shown in

Table 14. Group by whether they are heavy-polluting firms.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Heavy-Polluting Group			Non-Heavy-Polluting Group
	CGI	CGI_S	CGI_C	CGI	CGI_S	CGI_C
GM2	−0.3237	−0.2373 **	0.0133	−0.4376 *	−0.1094 *	−0.1152 **
	(0.3012)	(0.0744)	(0.0831)	(0.1865)	(0.0449)	(0.0429)
GM	0.1160	0.2249 **	−0.0177	0.3315	0.0856 *	0.1177 **
	(0.2987)	(0.0729)	(0.0809)	(0.1752)	(0.0422)	(0.0397)
Controls	YES	YES	YES	YES	YES	YES
Cons	−8.5146 ***	−1.4507 ***	−0.9601 ***	−7.4267 ***	−1.2680 ***	−0.4227 ***
	(0.4202)	(0.0942)	(0.1015)	(0.2753)	(0.0623)	(0.0556)
Time FE	YES	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES	YES
N	7652	7652	7652	20,689	20,689	20,689
R2	0.277	0.117	0.140	0.295	0.161	0.172

Note: Robust standard errors clustered to province in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001.

, GMGG synergy has a significant positive effect on substantive CGI only in the heavy-polluting group, whereas all other coefficients are statistically insignificant. In contrast, for non-heavy-polluting group, GMGG synergy significantly influences both substantive and strategic CGI, with the absolute magnitude of the coefficient being larger for strategic CGI. It indicates that GMGG synergy primarily incentivizes heavy polluters to pursue substantive green innovation, while more strongly inducing strategic green innovation among non-heavy polluters.

5.6.3. Group by the Region

Table 15. Testing results of group by area.

Variables	(1)	(2)	(3)	(4)
	Eastern	Central	Western	Northeastern
	CGI	CGI	CGI	CGI
GM2	−0.7880 ***	−0.1802	−0.7273	5.3449
	(0.1930)	(0.5266)	(0.4495)	(5.5364)
GM	0.6122 ***	0.6381	0.3288	−6.6766
	(0.1792)	(0.5258)	(0.4733)	(7.9709)
Controls	YES	YES	YES	YES
Cons	−8.4176 ***	−7.8010 ***	−4.6966 ***	−5.6718
	(0.2877)	(0.5831)	(0.5855)	(3.0299)
Time FE	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES
N	19,647	4032	3383	1389
R2	0.304	0.259	0.231	0.207

Note: Robust standard errors clustered to province in parentheses, *** p < 0.001.

presents the regression results disaggregated by region. The estimates suggest that GMGG synergy has a significant impact on CGI only among firms in eastern China, with coefficients for the central, western and north-eastern regions being statistically insignificant. This pattern is plausibly attributable to eastern regions’ superior resource agglomeration and higher levels of economic development.

6. Discussion

This study employs a dual fixed-effects model and empirical testing to find that the GMGG synergy has diminishing marginal effects on CGI, which is similar to the study of Shang et al. [85] and Zhang et al. [86]. When the collaboration between government and market in green governance is within a reasonable range, government efforts can effectively guide the behavior of market stakeholders. By using functions such as supervision and evaluation, the government can fully leverage the stakeholders’ subjective initiative, thus achieving multi-agent collaborative governance. This not only addresses the scarcity of government green governance resources but also enhances governance efficiency, effectively reducing government investment in green governance while improving its effectiveness.

However, when the synergy between government and market in green governance surpasses a critical threshold, firms may become overly dependent on external regulatory systems, reducing their internal motivation to pursue green innovation independently. This dependence may not only diminish firms’ sense of environmental responsibility but also suppress their enthusiasm for spontaneously engaging in green innovation activities. Ultimately, as the GMGG synergy deepens, the marginal benefits it provides for fostering CGI are expected to diminish gradually, potentially impacting the sustained momentum of green transformation.

Further empirical tests reveal that GMGG synergy has a more significant impact on firms’ substantial CGI, consistent with the research of Zhang et al. [87]. This underscores that under the dual drivers of governmental and market governance environments, enterprises can effectively reduce environmental compliance costs over the long term by implementing deep-seated green innovation strategies. Such strategies steadily enhance resource efficiency and operational performance. They not only strengthen enterprises’ collaborative capabilities in meeting governmental regulations and market demands but also significantly foster the emergence of green innovation achievements, thereby laying a robust foundation for sustainable development.

Mediation analysis results indicate that the level of marketization mediates between GMGG synergy and CGI, aligning with the study of Zhang et al. [60]. The synergistic effects of governmental and market roles in green governance greatly enhance market acceptance of green development concepts, driving markets towards cleaner and more sustainable trajectories and accelerating regional market maturity. High marketization implies a dominant role of markets in optimizing resource allocation, creating unique market opportunities for developing green technologies and services, thereby endowing firms with scalable competitive advantages. This process robustly promotes firms’ engagement in green technology research and product innovation, achieving dual economic and environmental benefits.

Empirical findings further suggest that the development of the digital economy helps mitigate the impact of GMGG synergy on CGI, affording firms more preparation time to respond to environmental policy changes, a viewpoint that is similar to that of Dai et al. [88] and Qiao et al. [89]. Within the framework of collaborative green governance by governments and markets, digital economic advancement provides unprecedented convenience to firms, enabling them to swiftly grasp market demands and consumer trends. This transformation significantly facilitates the efficient allocation and flow of innovation resources in the market, motivating firms to advance green innovation and product development with greater precision and efficiency. In this way, firms not only effectively mitigate challenges and impacts from the GMGG process but also progress steadily towards sustainable development through the path of green transformation.

The impact of GMGG synergy on CGI in non-state-owned firms exhibits heightened sensitivity, aligning with findings by Zhao et al. [90] and Liu & Li [91]. Non-state-owned firms demonstrate superior competitiveness and heightened innovation consciousness compared to their state-owned counterparts, facilitated by more adaptable decision-making frameworks capable of agile responses to market dynamics and environmental policy shifts. In the context of green governance policies, these firms can promptly realign their research and development focus and resource allocation strategies to effectively meet emerging environmental standards.

Moreover, GMGG synergy primarily incentivizes heavy-polluting firms to pursue substantive green innovation, while more strongly inducing strategic green innovation among non-heavy-polluting firms. This may be attributed to heavy-polluting firms facing stricter emission standards and environmental inspections, where the costs of non-compliance are significantly high. The policy signals conveyed by GMGG can directly reduce the marginal costs of process innovation or end-of-pipe treatment, thereby enhancing the motivation for substantial innovation in these firms. The effect of GMGG synergy on CGI is markedly stronger among eastern-region firms, reflecting their mature factor markets, comprehensive green-development infrastructure, and pronounced economies of scale in industrial clusters. These advantages allow policy signals to be swiftly converted into low-cost capital, substantially amplifying CGI.

Nevertheless, certain limitations of this study merit attention.

(1)

This study is confined to A-share listed firms, whose large size, high information transparency, and relatively loose financing constraints may limit the generalizability of our findings. Small and medium-sized enterprises (SMEs) and unlisted firms—often resource-constrained and heavily reliant on external finance—could exhibit markedly different sensitivities and response trajectories to GMGG. Consequently, caution is warranted when extrapolating our results to these populations. Future research should expand the sample to include SMEs, specialized and innovative “little giant” enterprises, and regional equity market entrants to test the external validity of our conclusions.

(2)

Although the sample spans multiple industries and provinces, data limitations preclude the incorporation of industry-specific fiscal incentives and environmental regulations, nor do we account for heterogeneous regional policies. Future research should employ difference-in-differences designs to estimate the green-innovation effects of targeted interventions.

(3)

Moreover, our reliance on two-way fixed effects captures only contemporaneous or one-period-lagged average effects, leaving the dynamic feedback between GMGG synergy and CGI, as well as potential path dependence and long-run equilibrium relationships, unexplored. Static specifications also fail to identify spatial or supply chain spillovers arising from green-technology interactions within regions or across upstream–downstream industries. Dynamic panel, spatial-econometric, or structural-equation models could address these shortcomings in future work.

(4)

Drawing on China’s green governance practices, this study has developed the GMGG synergy index and verified its diminishing marginal effects on CGI. The estimated inflection point reflects the optimal range under China’s specific institutional and market conditions only. Substantial cross-country heterogeneity in government capacity and firm-level innovation implies that the identified threshold cannot be directly extrapolated to other economies. Future research should recalibrate the inflection point in light of institutional differences and development levels across countries to test the global applicability of the GMGG framework.

(5)

Although the combined use of the AHP and EWM mitigates the biases inherent in purely subjective or objective weighting schemes, it remains susceptible to researcher-dependent judgments and entropy polarization in high-dimensional or sparse data. Future research should therefore incorporate dynamic weighting mechanisms to enhance robustness and temporal adaptability.

(6)

Following the existing literature, we employ the air circulation coefficient as an instrumental variable to mitigate endogeneity concerns. However, the effectiveness of the coefficient varies markedly across heterogeneous topographies and climatic zones, introducing non-negligible estimation errors. Future research could enhance identification by integrating higher-frequency pollution monitoring data with more refined meteorological simulations.

7. Conclusions and Recommendations

7.1. Conclusions

(1)

Empirical results reveal a significant Inverted-U relationship between GMGG synergy and CGI. The effect is positive and increasing up to a synergy level of 0.35, beyond which the impact diminishes monotonically, indicating pronounced diminishing marginal returns. Accordingly, guided by a ±0.1 “green-band” around the inflection point, the dynamic range for GMGG synergy should be tightened to 0.25–0.45, thereby preventing both under-synergy and over-synergy. Once GMGG synergy exceeds 0.45, authorities should progressively scale back fiscal subsidies and prescriptive regulations, while expanding market-based instruments and digital-economy policies to forestall a surge in strategic green innovation driven by excessive incentives.

(2)

This study empirically confirms the mediating role of marketization. It illustrates that GMGG synergy affects CGI by promoting market-oriented reforms, enhancing fair market competition, and optimizing resource allocation. Furthermore, the paper examines the moderating role of digital economic development in this process. It indicates that with the vigorous development of the digital economy, characterized by high penetration, integration, and innovation, it effectively mitigates the inverted U-shaped impact of GMGG synergy on CGI, providing new sources of momentum for green innovation growth.

(3)

Additionally, the study conducts detailed heterogeneity analysis, revealing differentiated responses of CGI behaviors among different economic entities to the influence of GMGG synergy. Particularly, non-state-owned firms, eastern-region firms, and those in non-heavy-polluting industries exhibit a higher responsiveness to changes in GMGG synergy.

These findings furnish policymakers with actionable insights for simultaneously pursuing economic growth and environmental protection, and offer firms strategic guidance for advancing green and sustainable development. Nevertheless, this study is subject to several limitations regarding sample selection, model specification, the generalizability of the findings, methodological diversity, and the choice of instrumental variables. These issues represent priority areas for future research.

7.2. Recommendations

(1)

Given the overlapping and converging boundaries of government and market in green governance, they are inseparable and share mutual interests. Achieving the driving effect of green innovation through green governance necessitates the rational enhancement of synergies between government and market. Diverse entities such as government departments and market organizations should prioritize information sharing, resource exchange, and joint actions to jointly promote energy conservation and environmental protection. It is crucial to strengthen collaborative mechanisms among government departments and between government and market entities, break down organizational barriers, and promote the establishment of functionally sound, coordinated, and efficient flat governance structures. Such efforts will enhance cooperation among different stakeholders and reduce administrative and communication costs. On the other hand, considering the marginal diminishing effect of GMGG synergy on CGI incentives, it is imperative for government and market entities to exercise prudence in the implementation intensity of policies during collaborative green governance processes. This helps prevent excessive reliance of enterprises on external regulatory and evaluation functions, thereby preserving the internal driving force for implementing green innovation activities. Drawing upon the empirical findings, the optimal range of GMGG synergy for effectively stimulating CGI is 0.25–0.45. At present, Beijing, Sichuan, and Chongqing in China have already positioned themselves within this threshold and should therefore maintain their current levels. In contrast, Sichuan, Jiangxi and Shanxi remain below the lower bound. Targeted policy interventions are required to increase their synergy. Conversely, Hainan, Hubei, and Shandong significantly exceed the upper bound, and a moderate downward adjustment is recommended to avert efficiency losses associated with excessive synergy.

(2)

To prevent excessive synergy, a negative list should delimit green technologies and market segments in which the government abstains from direct intervention—e.g., green-patent trading and carbon-financial-product pricing. Any government actions exceeding this list must be subject to automatic sunset clauses, ensuring their expiration and forestalling prolonged overreach. Direct subsidies for green technologies should be phased out and supplanted by ex-post tax credits tied to verified carbon-reduction performance, thereby mitigating governmental displacement of firms’ micro-level decisions. Mandatory environmental disclosure standards should be imposed on large and medium-sized firms. Their sustainability reports must be audited by independent third parties and released to the market at regular intervals to strengthen external oversight.

(3)

During the process of conducting green innovation activities, enterprises should leverage the constraints and incentive policies of government and market green governance to enhance their own quality of green innovation and achieve sustainable development. Firstly, green innovative enterprises should adhere strictly to government green environmental regulations, reducing compliance costs through innovative green technologies to enhance their own capabilities for sustainable development. Secondly, enterprises should actively participate in relevant green-subsidy programs by applying for and evaluating them, strategically planning and investing according to policy guidance to enhance their green orientation and operational compliance. They should also select and nurture projects with innovative potential and market prospects in order to obtain support from fiscal and taxation incentive policies [92]. Thirdly, enterprises should increase positive media attention on their green innovation activities, reducing information asymmetry to shape a positive corporate image and expand the market influence of green innovative products. Fourthly, enterprises should utilize green financial policies to alleviate financing constraints, increase investment in green innovation, and thereby enhance the quality of green innovation.