Sustainable Development Agenda Pilot Zones Policy, Entrepreneurial Green Attention and Corporate Green Development

Abstract

Sustainable development represents a fundamental pathway for advancing high-quality economic and social transformation. Taking China’s Sustainable Development Agenda Pilot Zones Policy as a quasi-natural experiment and drawing on data from A-share listed firms from 2013 to 2022, this study constructs a difference-in-differences model to systematically assess the policy’s impact on corporate green development and the underlying mechanisms. The empirical results indicate that the policy significantly improves corporate green development and that entrepreneurial green attention exerts a significant positive moderating effect. The mechanism analysis shows that improvements in the digital–real integration, the strengthening of regional green innovation capability, and increases in media attention constitute the primary channels through which the policy takes effect. The heterogeneity analysis further reveals that the policy impact is more pronounced among non-state-owned enterprises, firms in non-heavily polluting industries, regions oriented toward modern urban development, and cities with higher levels of governmental environmental concern. Additional analyses suggest that, while fostering green development, the policy is also associated with a greater tendency toward inflation in green invention patents and a decline in the quality of environmental information disclosure. These findings deepen the understanding of the micro-level effects of differentiated environmental regulation and provide empirical evidence for improving the green governance system and promoting high-quality development in China.

1. Introduction

Sustainable development is a necessary requirement for advancing social productivity and technological innovation, and it represents a critical pathway for addressing global challenges such as resource constraints, ecological degradation, and social disparities. To implement the 2030 Agenda for Sustainable Development and explore development models suited to China’s national context, the State Council issued the Construction Plan for National Innovation Demonstration Zones for Implementing the 2030 Agenda in 2016, followed by three rounds of approvals establishing a total of eleven national pilot zones in 2018, 2019, and 2022. The Sustainable Development Agenda Pilot Zones Policy (SDA Policy) targets region-specific constraints related to resources, the environment, and social development; encourages local governments to devise tailored solutions [1]; strengthens the roles of technological and institutional innovation; and promotes collaborative governance among governments, enterprises, research institutions, and the public. By offering technical and financial support, the policy stimulates firms’ endogenous motivation for green innovation and transformation, thereby improving environmental quality and fostering sustainable economic growth. A systematic evaluation of this policy is therefore essential for assessing its performance and for refining Chinese governance system for sustainable development.

Enterprises play a central role in the green transition, as their level of green development reflects not only the degree of alignment between resource efficiency and environmental performance but also provides a micro-level indicator of the implementation effectiveness of national sustainable development strategies [2]. Existing studies widely recognize environmental regulation as an important external driver of corporate green development. Traditional regulatory instruments typically include command-and-control measures such as emission permits and environmental standards, and market-based incentives such as carbon trading, green credit, and environmental taxes, which shape firms’ environmental behavior through mandatory compliance requirements or economic incentives [3]. However, the literature has not yet reached consensus on whether environmental regulation leads to an innovation compensation effect or a compliance cost effect. Some research suggests that stricter regulation stimulates green technological innovation [4], whereas other studies argue that substantial compliance costs crowd out firms’ R&D investment [5]. These divergent findings can be attributed largely to the fact that traditional regulatory approaches often rely on relatively uniform policy designs that insufficiently account for firm-level heterogeneity, including ownership structure, pollution intensity, regional development disparities, and variations in local governments’ environmental governance priorities, thereby resulting in heterogeneous policy effects across different contexts [6].

In contrast to earlier policies characterized by unified planning and top-down implementation, the SDA Policy emphasizes problem-oriented and differentiated governance. It highlights institutional innovation and policy experimentation rooted in local conditions, thereby enabling a more precise alignment between policy tools and regional development bottlenecks and facilitating the formation of diverse green transition pathways. As an important institutional instrument for China’s sustainable development agenda, the policy not only reflects macro-level exploration of institutional innovation but also reshapes firms’ green production practices and governance behaviors. Because the SDA Policy bridges macro governance goals and micro-level corporate responses, assessing its actual impacts and the underlying mechanisms is of both theoretical and practical significance.

Drawing on data from A-share listed firms from 2013 to 2022 and exploiting the SDA Policy as a quasi-natural experiment, this study employs a difference-in-differences approach to evaluate its effects on corporate green development. After identifying the overall policy impact, the analysis proceeds to examine the moderating role of entrepreneurial green attention in the policy transmission process. It further investigates the mechanisms through which the policy functions, focusing on corporate digital–real integration, regional green innovation capacity, and media attention. Additional heterogeneity analyses are conducted with respect to ownership type, pollution intensity, regional development level, and local governments’ environmental concerns in order to deepen the understanding of differentiated policy effects and their contextual dependence, as well as to identify potential unintended consequences associated with policy implementation.

This study makes three primary contributions. First, it evaluates the impact of the SDA Policy on corporate green development from both theoretical and empirical perspectives. The results show that the policy significantly enhances the green development performance of listed firms, enriching empirical evidence on the environmental effects of comprehensive sustainable development policies and providing support for the rigorous assessment of hybrid environmental regulations. Second, it extends the research frontier by examining both the mechanisms and moderating factors through which the policy influences green development. By analyzing digital–real integration, regional green innovation capacity, and media attention, the study uncovers the core pathways through which the policy operates and incorporates entrepreneurial green attention into the analytical framework to reveal its reinforcing effect. Heterogeneity analyses across ownership type, pollution level, regional development, and government environmental concern further compensate for gaps in understanding the sources of variation in policy impacts. Third, the study identifies unintended governance effects: while the policy promotes green development, it may also contribute to inflated green invention patenting and declining environmental disclosure quality, suggesting a potential misallocation of resources between technological and governance activities under policy incentives. These findings expand the recognized boundaries of sustainable development policy effects and provide new evidence for understanding governance costs and improving environmental policy design.

2. Institutional Background and Theoretical Hypotheses

2.1. Policy Background

The concept of sustainable development emerged in the 1980s in response to increasing concerns over resource depletion, environmental pollution, and ecological degradation. The 1987 WCED report, Our Common Future, formally defined sustainable development as meeting present needs without compromising the ability of future generations. Since then, the concept has gradually evolved into an integrated framework encompassing economic, social, and environmental objectives.

In 2015, the adoption of Transforming Our World: The 2030 Agenda for Sustainable Development institutionalized global sustainable development efforts through 17 Sustainable Development Goals (SDGs). China actively responded by incorporating the 2030 Agenda into its national development strategy. In 2016, the State Council issued the Plan for Building Innovation Demonstration Zones to Implement the 2030 Agenda for Sustainable Development, establishing a policy framework that emphasizes innovation-driven and regionally differentiated approaches to sustainable development [1].

Following this framework, the State Council approved National Sustainable Development Agenda Innovation Pilot Zones in 2018, 2019, and 2022, covering 11 regions, including Taiyuan, Guilin, Shenzhen, and Ordos. The SDA Policy adopts a problem-oriented and technology-driven approach to explore sustainable development pathways under region-specific resource and environmental constraints [7]. At the national level, it promotes green and low-carbon transformation, while at the micro level, it is expected to strengthen firms’ green technological innovation and institutional support for sustainable development.

2.2. Theoretical Hypotheses

Enterprises are the primary actors in implementing sustainable development strategies. Their level of green development reflects the degree of coordination between resource utilization efficiency and environmental performance and also serves as a core indicator for assessing high-quality regional development [8]. As an important institutional innovation in the field of environmental governance, the SDA Policy differs from traditional uniform regulatory instruments by emphasizing precision and systemic governance. Through the combined effects of multiple mechanisms and guided by institutional constraints and resource incentives, the policy encourages firms to pursue green transformation under the dual objectives of environmental protection and economic growth.

From the perspective of institutional theory, the implementation of this policy reshapes the external institutional environment in which firms operate. Through mechanisms such as target assessment, performance evaluation, and accountability, green development has been integrated into the organizational legitimacy evaluation system, forming a binding institutional framework [9]. Under the legitimacy-oriented organizational logic, firms no longer merely passively respond to policy requirements but proactively adjust their production and operation models to align with new institutional expectations, thereby preserving organizational legitimacy and social reputation [10]. The transmission effect of institutional constraints drives firms to embed the concept of green production into their management processes, gradually forming endogenous green development routines [11].

From the perspective of stakeholder theory, the implementation of the policy has heightened the expectations of multiple entities, including the government, investors, consumers, and the public, regarding corporate environmental responsibility, forming a joint force of external oversight and reputational constraints [12]. When faced with multiple stakeholders, enterprises that neglect environmental performance may face regulatory penalties, capital constraints, or reputational losses [13]. Conversely, enterprises that actively respond to policy directives can gain advantages in resource allocation and policy recognition, generating positive spillover effects in areas such as financing, market competition, and brand value [14,15]. Consequently, under the dual influence of policy incentives and external pressure, enterprises gradually construct a green development logic based on a “responsibility–benefit” equilibrium.

According to the Porter Hypothesis, well-designed environmental policies can reduce the costs of green innovation and enhance long-term competitiveness through the “innovation compensation effect.” The SDA Policy provides institutional guarantees for corporate green innovation through resource-based measures such as fiscal subsidies, green credit, technical support, and platform building [16,17]. Within the policy environment that combines constraints and incentives, corporate innovation motivation is stimulated, leading to a significant increase in green R&D investment and the synergistic improvement of production efficiency and environmental performance. This “flexible regulation” model encourages firms to shift from passive compliance to proactive innovation, forming an internal driving force for green transformation [18].

Based on this, this paper proposes the following hypothesis:

Hypothesis 1 (H1). 

The SDA Policy has a positive impact on corporate green development.

The SDA Policy provides an external driving force for corporate green transformation through means such as capital investment, institutional supply, and signal transmission. However, the manifestation of the policy’s effects is not linear or automatic; its intensity and direction largely depend on the cognitive processing and strategic response of corporate decision-makers to the policy signals. Based on the attention-based view, entrepreneurial attention is a scarce and finite cognitive resource, and its allocation determines how a firm identifies, interprets, and responds to external stimuli in complex information environments [19]. If entrepreneurs focus their attention on green issues and sustainable development goals, they are more likely to keenly capture policy orientations and external institutional changes. Consequently, they can strategically and promptly identify the resources and opportunities provided by the policy, thereby promoting green technological innovation, optimizing production structures, and driving organizational change. This forms an internal amplification mechanism for the policy effects [20]. Conversely, when entrepreneurial attention is primarily directed toward short-term financial performance or market expansion, sensitivity to policy signals diminishes. This weakens the transmission chain of policy incentives, making it difficult for policy dividends to be effectively converted into drivers for green development [21,22,23].

From the perspective of internal corporate decision-making logic, entrepreneurial attention not only influences the extent to which firms recognize policy information but also affects their resource allocation and strategic orientation. When entrepreneurs place a high degree of attention on green issues, they are more inclined to proactively integrate green development objectives into corporate strategic planning, aligning policy directives with the firm’s long-term competitive advantages [24]. Simultaneously, positively interpreting and responding to external policy signals helps firms cultivate a positive environmental responsibility image in public discourse and capital markets, thereby strengthening trust relationships and enhancing reputation capital accumulation with the government and the public, which further amplifies the policy’s promoting effect on corporate green development [25].

Therefore, entrepreneurial green attention plays a crucial moderating role between the SDA Policy and corporate green development. A high level of attention allocation can strengthen the identification and absorption of policy signals and facilitate the effective transmission and transformation of policy effects within the firm, thereby amplifying the green incentive effect of the policy. Conversely, when entrepreneurial attention is dispersed or deviates from green issues, policy incentives may weaken due to cognitive lag or strategic misalignment, resulting in the policy effects not being fully realized. Based on this, this paper proposes the following hypothesis:

Hypothesis 2 (H2). 

Entrepreneurial Green Attention positively moderates the impact of the SDA Policy on corporate green development. The more importance entrepreneurs attach to green issues and sustainable development goals, the more pronounced the policy effect becomes.

Enterprises serve as the primary actors in policy response, with their internal resource allocation and technological capability upgrading constituting the micro-foundation for the realization of policy effects. SDA Policy not only provides external constraints and guidance through fiscal support and institutional incentives but also creates critical conditions for the digital–real integration within enterprises by establishing digital infrastructure and green technology innovation platforms. Digital–real integration (CEDRT) refers to the deep embedding of digital technology elements into enterprise production, management, and value-creation processes, achieving an organic integration of data resources and real economic activities. Its core lies in utilizing information technology to promote enhanced resource allocation efficiency, optimized energy usage, and the intelligent transformation of production processes [26], thereby becoming one of the critical pathways for corporate green development.

Enterprises are the central actors in responding to public policies, and their internal resource allocation and technological capability upgrading constitute the micro-foundation for the realization of policy effects. The SDA Policy not only provides external constraints and guidance through fiscal support and institutional incentives but also establishes digital infrastructure and green technology innovation platforms that create essential conditions for advancing CEDRT within enterprises. CEDRT refers to the deep embedding of digital technologies into firms’ production, management, and value-creation processes, enabling the organic alignment of data resources with real economic activities. Its core function lies in leveraging information technologies to enhance resource allocation efficiency, optimize energy utilization, and promote the intelligent transformation of production processes [26], thereby becoming a critical pathway for corporate green development.

From the perspective of the transmission mechanism, the CEDRT serves as a bridge between the policy and corporate green development. First, the digital infrastructure construction driven by the policy lowers the threshold for enterprises to adopt green digital technologies. This enables companies to utilize technological means such as big data, the Internet of Things, and cloud computing to achieve real-time monitoring and dynamic optimization of production processes [27], thereby significantly enhancing efficiency in areas such as pollution control, energy consumption management, and waste utilization. Second, the CEDRT strengthens enterprises’ capabilities in technology learning and knowledge absorption. It provides a data-driven decision-making foundation and algorithmic support for green innovation, reduces trial-and-error costs in the research, development, and application of green technologies, and improves the output rate of green innovation [28]. Third, the synergistic operation of digitalization and physical production allows enterprises to pursue the dual goals of “cost reduction and efficiency enhancement” and “green transformation” while balancing economic benefits and environmental performance. This, in turn, strengthens the internal drivers for enterprises to continuously engage in green investment and technological upgrading within the policy environment [29]. In summary, by promoting the CEDRT within enterprises, SDA Policy not only improves their production efficiency and technological structure but also stimulates their initiative in green innovation and environmental governance. The CEDRT plays a mediating role in the transmission of the policy effect, constituting a crucial mechanism for translating external policy incentives into internal sustainable capabilities within enterprises. Based on this, this paper proposes the following hypothesis:

Hypothesis 3a (H3a). 

The SDA Policy promotes corporate green development by enhancing the level of digital–real integration.

Corporate green development is not an isolated action but a dynamic process embedded within a specific regional innovation ecosystem. The implementation of the SDA Policy not only acts directly on enterprises through institutional incentives but also promotes the enhancement of regional green innovation capability at the meso-level by reshaping the regional innovation environment and optimizing the resource allocation structure. As an “experimental field” for institutional innovation, the SDA Policy, by guiding green technology investment, promoting collaborative innovation across industry chains, and facilitating the diffusion of green technologies, gradually fosters a sustainability-oriented regional innovation ecology, thereby providing institutional and technical external support for corporate green transition [30].

Regional green innovation capability reflects the overall level of a region in areas such as green technology R&D, achievement transformation, and innovation diffusion, embodying the degree of agglomeration of green innovation elements and the operational efficiency of the innovation system [31]. Its enhancement signifies more significant spillover effects of green technological knowledge, more complete supporting facilities of the green industry chain, and a higher degree of sharing of innovation resources [30]. When regional green innovation capability is strengthened, enterprises can acquire external innovation knowledge and technological achievements at lower search costs, and by leveraging regional technology service systems and collaborative networks, more efficiently achieve the absorption, digestion, and re-innovation of green technologies [32]. Simultaneously, the regional green innovation atmosphere can also reinforce corporate environmental awareness and institutional pressure, guiding enterprises to proactively integrate green innovation into their core strategies, thereby realizing a transformation from external policy-driven motivation to endogenous innovation impetus [33]. Consequently, the national SDA Policy can provide sustained knowledge spillover effects and external technological support for corporate green development by enhancing regional green innovation capability, thereby promoting green transformation and innovation performance at the corporate level [34]. Based on this, this paper proposes the following hypothesis:

Hypothesis 3b (H3b). 

The SDA Policy promotes corporate green development by enhancing regional green innovation capability.

At the macro level, the implementation of the SDA Policy not only provides an institutional orientation for China’s regional economic and ecological development but also generates a notable signaling effect within the public opinion sphere. As a key intermediary in information dissemination and the formation of social cognition, sustained media reporting on policy content, implementation progress, and environmental outcomes substantially shapes public expectations and evaluations of corporate green behavior. According to institutional theory and the legitimacy perspective, corporate strategic decisions are not solely driven by economic rationality but are also shaped by social norms and public opinion pressure [35]. When media attention (Media) increases, the public opinion environment in which firms operate becomes more transparent, accompanied by strengthened external scrutiny and heightened expectations for corporate social responsibility. This induces firms to respond to societal expectations by improving environmental information disclosure, increasing green investment, and enhancing production processes, thereby maintaining legitimacy and reputational capital [36].

In addition, Media exerts a distinct agenda-setting effect. Within the policy context of the Innovation Pilot Zones, positive and sustained media reporting reinforces and amplifies policy signals, facilitating the transformation of the green development concept from a government-initiated directive into a broader social consensus. This public opinion guidance effect enhances top managers’ understanding and recognition of policy objectives, thereby encouraging firms to engage proactively in green innovation and ecological transition [37]. Therefore, Media serves not only as an information transmission channel through which the SDA Policy influences corporate green development but also as an important macro-social mechanism that helps consolidate shared values and foster strategic corporate green transformation [38]. Based on this reasoning, this study proposes the following hypothesis:

Hypothesis 3c (H3c). 

The SDA Policy promotes corporate green development by enhancing media attention.

Theoretical analysis reveals multiple potential pathways through which the SDA Policy influences corporate green development. In order to integrate these perspectives and systematically elucidate the SDA Policy’s mechanisms of action, this study constructs a theoretical framework (as shown in Figure 1).

3. Research Design and Data

3.1. Samples and Data

In 2018, China officially established its first batch of Sustainable Development Agenda Pilot Zones, covering the cities of Guilin, Taiyuan, and Shenzhen. Considering that the first batch of pilot areas exhibited stronger characteristics of institutional innovation and policy impact effects during the initial stage of policy implementation, their governance practices can more fully reflect the initial effectiveness of the policy. Therefore, this paper selects the first batch of pilot zones as the research object to identify the causal impact of the SDA Policy on corporate green development. This study takes Chinese A-share listed companies from 2013 to 2022 as the initial research sample and constructs a quasi-natural experiment based on the implementation of SDA Policy to identify and estimate the policy effects. The sample screening process is as follows: (1) Excluding data from the financial and insurance industries. (2) Excluding enterprises that were suspended or delisted during the sample period. (3) Excluding listed companies with severe data deficiencies. (4) Excluding enterprises that changed their registered address or relocated their actual place of business during the sample period, as well as enterprises with branches or operational sites in multiple geographical locations, to ensure the geographical consistency of the sample data and the reliability of the analysis results. To mitigate the potential interference of outliers on the regression results, all continuous variables are winsorized at the 1% and 99% percentiles. The original corporate data are sourced from the CSMAR and CNRDS databases as well as annual reports of listed companies, while regional-level data are obtained from the China City Statistical Yearbook and the government work reports of prefecture-level cities.

3.2. Variable Specification

Corporate Green Development (GREEN) refers not to general corporate performance but to a firm’s capability to achieve coordinated economic viability, social responsibility, and environmental sustainability under green development constraints. It emphasizes whether firms can internalize environmental requirements into their production, operation, and governance processes while maintaining stable development trajectories.

Accordingly, drawing on Hongyi and Qiongwen [39], this study constructs a comprehensive indicator system for GREEN from three interrelated dimensions: economic profit, social value, and environmental benefit. Importantly, the economic profit dimension does not aim to capture profit maximization per se but rather reflects firms’ economic efficiency and cost-management capacity in the process of green transformation. Indicators such as Return on Total Assets, Net Profit Growth Rate, Inventory-to-Revenue Ratio, Operating Costs, Selling Expenses, and Administrative Expenses are employed to assess whether firms can absorb the additional compliance costs, innovation expenditures, and operational adjustments associated with green development without undermining economic sustainability.

The social value dimension includes indicators such as Earnings Per Share, Compensation Paid to Employees, and the Number of Employees, capturing firms’ ability to allocate economic gains toward stakeholders and employment stability during the green transition. The environmental benefit dimension directly measures firms’ environmental performance and sustainability orientation through indicators such as the number of green patent applications, environmental tax intensity, and ISO9001 [40] certification status.

By integrating economic, social, and environmental indicators within a unified framework, this composite index reflects green development as a constrained and multi-dimensional process rather than a simple proxy for overall corporate performance. The entropy method is employed to aggregate these indicators into a comprehensive GREEN index, allowing for an objective assessment of firms’ green development levels (as shown in

Table 1. Comprehensive Indicator System for Corporate Green Development.

Green Development	Dimensional Type	Indicator Type	Specific Calculation Method
Green Development	Economic Profit	Return on Total Assets (+)	The ratio of a company’s total net profit to its average total assets
		Net Profit Growth Rate (+)	The growth rate of the company’s current net profit compared to the previous period’s net profit
		Inventory-to-Revenue Ratio (+)	Inventory/Operating Revenue
		Net Fixed Assets (+)	Original value of fixed assets minus accumulated depreciation and impairment provisions
		Total Factor Productivity (+)	TFP is calculated using the Fixed Effects (FE) method
		Firm Size	The logarithm of the company’s total assets
		Operating Costs (−)	(Cost of main business) + (Cost of other businesses)
		Selling Expenses (−)	Sum of all sales expenditure costs
		Administrative Expenses (−)	Various expenses incurred by the enterprise’s administrative department for managing and organizing business activities, including company expenses, labor union dues, employee education expenses, labor insurance premiums, unemployment insurance premiums, board of directors fees, consulting fees, audit fees, etc.
	Social Value	Earnings Per Share (+)	Earnings Per Share = (Current Period Net Profit − Preferred Dividends)/Weighted Average Total Shares Outstanding for the Year
		Compensation Paid to Employees (+)	Total compensation paid to employees
		Number of Employees (+)	The natural logarithm of the number of employees
	Environmental Benefit	Number of Green Patent Applications (+)	The quantity of green patent applications identified according to the WIPO International Green Patent Classification List
		Environmental Tax Intensity (+)	The ratio of the logarithm of main business revenue to the natural logarithm of environmental tax
		ISO9001 Certification (+)	Assigned a value of 1 if the enterprise has passed ISO9001 certification, otherwise 0

Source: Own elaboration.

).

Explanatory Variable: SDA Policy (Pilot × Post). This variable is used to measure whether the SDA Policy has been implemented in the region where the enterprise is located. If the city where the enterprise is located is selected as a Sustainable Development Agenda Pilot Zone, the variable is assigned a value of 1 for the sample period starting from the pilot year and onwards; otherwise, it is assigned a value of 0.

Control Variables: To control for other factors that may influence GREEN, this study introduces a series of control variables at both the enterprise level and the city level. At the enterprise level, the study controls for variables including firm size (Size), leverage ratio (Lev), return on total assets (ROA), cash flow from operating activities ratio (Cashflow), Tobin’s Q (TobinQ), operating revenue growth rate (Growth), board size (Board), CEO duality (Dual), the shareholding ratio of the largest shareholder (Top1), and listing age (ListAge). At the city level, the study controls for variables including regional economic development level (GDP), financial development level (Finance), the proportion of secondary industry (GDP_two), and the degree of government intervention (Gov). The aforementioned control variables comprehensively account for various dimensions such as corporate financial conditions, corporate governance structure, growth characteristics, and the macroeconomic environment of the location. This approach aims to effectively mitigate omitted variable bias and ensure the robustness and reliability of the main conclusions.

Moderating Variable: Entrepreneurial Green Attention (Green Attention). Building on the research of Tian, J et al. and Marquez-Illescas et al., corporate annual reports, particularly the Management Discussion and Analysis (MD&A) section, can effectively reflect the level of attention and strategic inclination of entrepreneurs towards specific issues [41,42]. Based on the Attention-Based View, this paper conducts a textual analysis of the MD&A chapters in listed companies’ annual reports to construct an indicator for Green Attention allocation. The core logic of this method is that issues repeatedly emphasized by entrepreneurs’ reports reflect the strategic allocation of their cognitive resources and serve as an effective proxy variable for predicting future corporate actions [43].

First, a “green” seed dictionary is constructed. The base words are selected from green, innovation, and sustainable development-related vocabulary widely used in the current literature, including “green,” “environmental protection,” “low-carbon,” “energy saving,” “emission reduction,” “clean,” “sustainable development,” “circular,” “ecological,” “pollution prevention”, etc. For details, please refer to the word cloud in Figure A1 in the Appendix A.

Next, the CBOW model within Word2Vec is employed to train on the corpus of all A-share listed companies’ annual reports from 2010 to 2022, generating a word vector space. Based on this, similar word expansion is performed for the aforementioned seed words. The screening criteria are: (1) the word appears ≥1000 times in the annual report corpus and (2) its cosine similarity with any seed word is ≥0.30. This process yields expanded words, such as “carbon neutrality,” “dual carbon” (referring to China’s carbon peaking and carbon neutrality goals), “near-zero emissions,” “source reduction,” “renewable energy,” etc.

Finally, three scholars in the field of environmental economics and two practitioners with experience as board secretaries of listed companies are invited to conduct a manual review and cleansing of the machine-expanded vocabulary, forming the final “Green Attention Dictionary.”

Based on this dictionary, this paper uses the word frequency method to calculate, for each company in a given year, the proportion of the total frequency of green attention words in its MD&A section relative to the total number of words in that section, multiplied by 100, to obtain the core indicator: Green Attention. A higher value of this indicator indicates a greater level of strategic attention paid by entrepreneurs to green development issues. See the specific word cloud map in Figure A1. Detailed definitions of all variables are provided in

Table 2. Definitions and Calculation Methods of Key Variables.

Variable Type	Variable Name	Variable Symbol	Calculation Method
Explained Variable	Corporate Green Development	GREEN	Comprehensive index of corporate green development
Explanatory Variable	Sustainable Development Agenda Pilot Zones Policy	Pilot × Post	The policy effect is identified by constructing a policy treatment variable (Treat × Post). This variable takes the value of 1 for firms in the treatment group during the post-policy period, and 0 otherwise.
Moderating Variable	Entrepreneurial Green Attention	EGA	(Total word frequency of green attention vocabulary in the MD&A section/Total word frequency of the MD&A section) × 100
Control Variables	Firm Size	Size	Natural logarithm of total assets at year-end
	Asset–Liability Ratio	Lev	Total liabilities at year-end/Total assets at year-end
	Return on Total Assets	ROA	Net profit/Average balance of total assets
	Cash Flow Ratio	Cashflow	Net cash flow from operating activities/Total assets
	Operating Revenue Growth Rate	Growth	(Current year’s operating revenue/Previous year’s operating revenue) − 1
	Board Size	Board	Natural logarithm of the number of board directors
	CEO Duality	Dual	Equals 1 if the Chairman and the General Manager are the same person, otherwise 0
	Shareholding Ratio of the Largest Shareholder	Top1	Number of shares held by the largest shareholder/Total number of shares
	Tobin’s Q	TobinQ	(Market value of tradable shares + Number of non-tradable shares × Net assets per share + Book value of liabilities)/Total assets
	Listing Age	ListAge	Ln(Current year − IPO year + 1)
	Economic Development Level	GDP	Logarithm of per capita GDP
	Financial Development Level	Finance	Balance of various loans from financial institutions at year-end/GDP
	Industrialization Level	GDP_two	Proportion of value-added of the secondary industry in GDP
	Degree of Government Intervention	Gov	Local fiscal expenditure/GDP

Source: Own elaboration.

.

3.3. Model Specification

To accurately assess the impact of the SDA Policy on GREEN, this study employs a difference-in-differences model for analysis, based on the research by Baker et al. [44]. This model effectively utilizes the natural experiment characteristics of policy implementation to identify the causal effect of the policy by comparing differences before and after the policy implementation as well as between the treatment and control groups. The model is specified as follows:

$${GREEN}_{it}={\alpha }_0+{\alpha }_1{PilotPost}_{it}+{\alpha }_2{Controls}_{it}+{\mu }_i+v_t+{\epsilon }_{it}$$

(1)

Among them, ${GREEN}_{it}$ represents the comprehensive green development index of enterprise $i$ in year $t$. ${PilotPost}_{it}$ is the policy implementation treatment variable, indicating whether enterprise i is affected by the SDA Policy in year $t$. This variable takes the value of 1 for firms in the treatment group during the post-policy implementation period, and 0 otherwise. ${Controls}_{it}$ are the control variables, encompassing a total of 4 city-level and 10 enterprise-level characteristic variables, used to account for other potential influencing factors. ${\mu }_i$ and $v_t$ represent industry fixed effects and time fixed effects, respectively, controlling for firm-specific characteristics that do not change over time and annual effects common to all firms. ${\epsilon }_{it}$is the random error term, capturing the influence of other unobserved factors.

It should be noted that the primary objective of the difference-in-differences framework adopted in this study is to identify the causal impact of the SDA Policy rather than to maximize the overall explanatory power of the model. Consistent with standard practice in policy evaluation research, the inclusion of industry fixed effects and time fixed effects absorbs a substantial proportion of unobservable heterogeneity and common macro shocks. As a result, the R-squared values in firm-level regressions may appear relatively modest. However, this does not undermine the consistency or validity of the estimated policy effect, which remains the focus of this study.

4. Analysis of Empirical Results

4.1. Descriptive Statistics

Before conducting the empirical regression analysis, this study performed descriptive statistics on the main variables to comprehensively examine the sample characteristics and data distribution. The results are shown in

Table 3. Descriptive statistics of variables.

Variable	Obs	Mean	Std. Dev.	Min	Max
GREEN	11,250	0.033	0.044	0.005	0.411
Pilot × Post	11,250	0.130	0.336	0	1
Size	11,250	22.431	1.405	19.470	26.456
Lev	11,250	0.430	0.207	0.046	0.934
ROA	11,250	0.036	0.069	−0.415	0.255
Cashflow	11,250	0.044	0.068	−0.203	0.266
TobinQ	11,250	2.119	1.471	0.802	17.676
Growth	11,250	0.161	0.434	−0.673	4.429
Board	11,250	2.111	0.202	1.609	2.708
Dual	11,250	0.292	0.454	0	1
Top1	11,250	0.341	0.152	0.080	0.754
ListAge	11,250	2.239	0.791	0.693	3.401
GDP	11,250	11.710	0.453	9.799	13.056
GDP_two	11,250	32.570	10.509	15.83	67.45
Finance	11,250	5.219	1.577	1.185	8.777
Gov	11,250	0.180	0.046	0.081	0.312

Source: Own elaboration.

. The mean value of the core explained variable, GREEN, is 0.033, indicating that the overall green development level of listed enterprises during the research period was relatively low, and the practice of green production and sustainable operation was still in a phase of gradual deepening. The standard deviation is 0.044, with minimum and maximum values of 0.005 and 0.411, respectively. This reveals significant differences among enterprises in terms of resource allocation efficiency, accumulation of green technologies, and environmental governance capabilities. The substantial sample heterogeneity provides a solid empirical foundation for identifying the policy effects.

The mean value of the core explanatory variable, the policy dummy variable (Pilot × Post), is 0.130. This indicates that approximately 13% of the sample observations belong to regions and periods subject to the implementation of the SDA Policy, which is consistent with the policy’s institutional characteristic of being promoted in batches and gradually. The means and standard deviations of the other control variables are generally reasonably distributed, with no signs of abnormal concentration or dispersion, suggesting that the sample data possesses good representativeness and balance.

4.2. Baseline Regression

This paper employs a two-way fixed effects model to examine the impact of the SDA Policy on GREEN. The baseline regression results are presented in

Table 4. Baseline regression test results.

	(1)	(2)	(3)
	GREEN	GREEN	GREEN
Pilot × Post	0.0111 ***	0.0136 ***	0.0072 ***
	(0.001)	(0.002)	(0.002)
Size		0.0047 ***	0.0042 ***
		(0.000)	(0.001)
Lev		−0.0120 ***	−0.0051
		(0.003)	(0.003)
ROA		0.0044	0.0097
		(0.007)	(0.008)
Cashflow		0.0070	−0.0054
		(0.007)	(0.007)
TobinQ		−0.0005	0.0001
		(0.000)	(0.000)
Growth		−0.0023 **	−0.0020 ***
		(0.001)	(0.001)
Board		−0.0011	0.0033
		(0.002)	(0.004)
Dual		0.0031 ***	0.0019
		(0.001)	(0.001)
Top1		−0.0097 ***	−0.0042
		(0.003)	(0.005)
ListAge		−0.0070 ***	−0.0056 ***
		(0.001)	(0.001)
GDP		0.0002	−0.0000
		(0.001)	(0.002)
GDP_two		−0.0004 ***	−0.0002
		(0.000)	(0.000)
Finance		−0.0029 ***	−0.0022 ***
		(0.001)	(0.001)
Gov		−0.0256 **	0.0223
		(0.011)	(0.019)
Constant	0.0318 ***	−0.0161	−0.0399
	(0.000)	(0.017)	(0.042)
Observations	11,250	11,250	11,250
R-squared	0.007	0.037	0.099
Controls	NO	YES	YES
Ind FE	NO	NO	YES
Year FE	NO	NO	YES

Note: Standard errors are shown in parentheses; *** and ** indicate significance at the 1% and 5% levels, respectively.

. To ensure the robustness of the estimation results and a hierarchical interpretation, this paper analyzes the data by sequentially incorporating control variables.

First, under the model specification without controlling for any firm or regional characteristics or fixed effects, the estimated coefficient for the policy variable (Pilot × Post) is significantly positive, passing the significance test at the 1% statistical level. This indicates a significant positive relationship between the SDA Policy and GREEN. Second, after introducing control variables at both the firm and city levels—such as firm size, asset–liability ratio, profitability, and economic development level—the estimated coefficient for the policy variable remains significantly positive with minimal numerical change. This suggests that the policy effect is not systematically influenced by differences in firm characteristics, providing preliminary verification of result robustness.

Upon further controlling for industry and year fixed effects and adjusting for clustered robust standard errors at the city level, the baseline regression results remain robust. The core explanatory variable, Pilot × Post, consistently exhibits a positive influence at the 1% significance level. Although the R-squared values are relatively modest, this outcome is common in firm-level policy evaluation models with multiple fixed effects, where the emphasis lies on the consistency, significance, and economic meaning of the estimated policy coefficient rather than overall explanatory power.

Regarding control variables, the results indicate that the coefficient on financial development (Finance), measured by the ratio of total loans of financial institutions to GDP, is negatively associated with GREEN. One possible explanation is that in regions with deeper financial systems, credit resources may still be disproportionately allocated toward traditional, capital-intensive, and high-emission industries, thereby crowding out financing available for green and low-carbon projects. Moreover, higher financial depth does not necessarily imply greener financial allocation, particularly in the absence of sufficiently strong green finance screening mechanisms.

Similarly, the coefficient on government intervention (Gov), proxied by the ratio of fiscal expenditure to GDP, is weakly negative. This result does not contradict green regulation objectives but may reflect that fiscal expenditures are often directed toward short-term economic stabilization, infrastructure construction, or social welfare, rather than targeted environmental governance. In some cases, such expenditures may even indirectly support energy-intensive activities, thereby diluting their contribution to corporate green development. These results validate Hypothesis 1 (H1): The SDA Policy has a positive impact on corporate green development.

4.3. Endogeneity Test

To mitigate potential endogeneity biases arising from sample self-selection, this paper draws on the research approach of Dong and Wang [45], selecting the city’s long-term average rainfall as the instrumental variable for the policy dummy variable (the Pilot × Post interaction term) and employing the Two-Stage Least Squares (2SLS) method for identification.

The specification of the instrumental variable is based on the following two theoretical rationales. First, regarding the relevance condition, rainfall, as a key natural geographical feature determining regional environmental carrying capacity and pollution assimilation capacity, exhibits significant stability in its long-term average. Cities with more abundant precipitation generally possess stronger natural environmental self-purification capabilities and a better ecological foundation, making them more likely to meet the location selection criteria for comprehensive pilot policies oriented toward sustainable development, thereby increasing the probability of obtaining approval for the innovation pilot zone. Consequently, rainfall exhibits a strong correlation with the implementation of the pilot policy. Second, regarding the excludability condition, a city’s historical average rainfall is a completely exogenous natural endowment variable; it is neither driven by local government policies nor does it directly alter GREEN levels by influencing corporate operational decisions. Its primary potential pathway to affect GREEN is through influencing the national-level selection of pilot cities, thereby indirectly affecting the institutional environment in which firms operate, which aligns with the requirements of the excludability assumption.

The estimation results of the instrumental variable method are presented in

Table 5. Endogeneity Test Results.

	(1)	(2)
	First Stage	Second Stage
IV	0.4893 **
	(0.202)
Pilot × Post		0.0246 **
		(0.011)
Constant	−2.7742 ***
	(0.866)
Observations	11,250	11,250
R-squared	0.481	0.008

Note: Standard errors are shown in parentheses; *** and ** indicate significance at the 1% and 5% levels, respectively.

. In the first-stage regression, the coefficient of the instrumental variable is positive and significant at the 5% level, effectively ruling out the issue of a weak instrument. The results of the second-stage regression show that after correcting for endogeneity using the instrumental variable, Pilot × Post remains significantly positive, and its magnitude is largely consistent with the baseline regression, indicating that the conclusions of this paper are robust.

4.4. Robustness Tests

4.4.1. Parallel Trends Test

The identification validity of the Difference-in-Differences (DID) model relies on the “parallel trends” assumption, which posits that prior to the policy implementation, the explained variable should exhibit similar temporal evolution trends between the treatment group and the control group. If this assumption holds, the change in the difference between the two groups after the policy implementation can be regarded as the net result of the policy effect. To test this premise, drawing on the research of Song et al. [46], this paper employs the event study approach to construct the following dynamic effects model:

$${GREEN}_{it}={\gamma }_0+\sum _{k=-5}^{k=4}{\beta }_k{\times PilotPost}_{i,t+k}+{{\gamma }_{1}Controls}_{it}+{\mu }_i+v_t+{\epsilon }_{it}$$

(2)

Herein, $PilotPos{t}_{i,t+k}$ is a relative time dummy variable, where $k$denotes the years before or after the policy implementation year (2018, denoted as $k=0$). Considering that the sample period begins in 2013 and the policy was implemented in 2018, this study sets $k$ from −5 to 4, with the year prior to policy implementation ($k=-1$) serving as the baseline period. The coefficient ${\beta }_k$ captures the dynamic differences in the level of green development between the treatment group and the control group across each year, and the estimation results are presented in Figure 2.

As shown in Figure 2, in the five periods prior to policy implementation, the confidence intervals for each period’s coefficient β_k all straddle zero, with the estimated values being close to zero and statistically insignificant. This indicates no systematic difference in green development levels between the treatment and control groups before the policy’s implementation, verifying the validity of the parallel trends assumption and laying the groundwork for subsequent causal inference.

During the policy implementation year and the following year, although the coefficients were positive, they did not reach statistical significance, reflecting a certain lag in the release of the policy effect. This lag may stem from the transmission cycle from top-level policy design to local implementation and further to strategic adjustments at the enterprise level. The results show that the policy’s promoting effect on GREEN strengthens and stabilizes over time. It is noteworthy that between 2019 and 2021, the coefficients experienced a certain degree of decline, which may be related to the supply chain disruptions and operational pressures on enterprises caused by the COVID-19 pandemic [47], which temporarily suppressed GREEN in the short term. However, the coefficients remained positive during this period, indicating that the overall promoting effect of the policy was not reversed by the external shock.

4.4.2. Placebo Test

To rule out potential interference from unobservable omitted variables or stochastic factors on the baseline regression results and to further enhance the reliability of the research findings, this study conducted a placebo test. We randomly selected 143 non-treatment group samples to serve as a simulated treatment group and randomly assigned a policy implementation time point to each simulated treatment group, thereby constructing a virtual policy interaction term. Subsequently, while keeping the control variables and fixed effects specifications unchanged, we repeated the aforementioned baseline regression procedure 1000 times, obtaining the distribution of 1000 “pseudo-policy” estimated coefficients. Figure 3 presents the kernel density distribution of these randomly estimated coefficients and their corresponding p-value distribution. It can be clearly observed that all these randomly generated estimated coefficients are densely distributed around zero. Their kernel density curve exhibits an approximately normal distribution centered at zero, and they show a significant difference from the actual estimated value, further verifying the robustness of our research findings.

4.4.3. Other Robustness Tests

• Excluding Interference from Competitive Policies

During the sample period, the simultaneous implementation of other environmental regulation policies may have created confounding effects with the SDA Policy. To exclude interference from other policies, we selected two environmental policies: the “Low-Carbon City Pilot” and the “Carbon Emission Trading Pilot.” The primary reasons for selecting these two policies are as follows: First, their policy objectives overlap and intersect with those of the Sustainable Development Agenda Innovation Pilot Zones, as all are committed to promoting green and low-carbon transformation. Second, their implementation timelines coincide with the policy studied in this paper, and they cover some of the same provinces and cities, making policy superposition effects highly likely. Drawing on relevant research, we introduced dummy variables for these two policies (denoted as Pilot × Post1 and Pilot × Post2, respectively) into the baseline model simultaneously. The regression results (as shown in Column (1) of

Table 6. Robustness Test Results.

	(1)	(2)	(3)	(4)	(5)
	Competitive Policies	Special Period	Kernel Matching	Caliper Matching	1:1 Nearest Neighbor Matching
Pilot × Post	0.0051 ***	0.0080 ***	0.0066 ***	0.0071 ***	0.0047 **
	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
Pilot × Post1	0.0010
	(0.003)
Pilot × Post2	0.0060
	(0.004)
Constant	0.0055	−0.0341	−0.0514	−0.0473	−0.0529 **
	(0.019)	(0.039)	(0.048)	(0.050)	(0.021)
Observations	11,250	8474	10,960	10,300	2245
R-squared	0.100	0.103	0.098	0.100	0.136
Controls	YES	YES	YES	YES	YES
Ind FE	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES

Note: Standard errors are shown in parentheses; *** and ** indicate significance at the 1% and 5% levels, respectively.

) show that after controlling for these two competitive policies, the coefficient estimate for the core explanatory variable DID is 0.0051 and remains statistically significant at the 1% level, while the coefficients for the two competitive policy variables are both insignificant. This result strongly demonstrates that the GREEN enhancement effect identified in this paper indeed originates from the specific policy of the Sustainable Development Agenda Innovation Pilot Zones, rather than from the superposition of other environmental regulations implemented concurrently.

2.

Excluding the Impact of Exceptional Periods

Considering that the global outbreak of the COVID-19 pandemic from 2020 to 2021 imposed an unprecedented exogenous shock on China’s real economy and corporate behavior, potentially interfering with the normal implementation effects of the policy, we removed the observations from these two years from the sample and re-estimated the baseline regression model. The regression results (as shown in Column (2) of Table 6) indicate that after excluding the interference of this special pandemic period, the coefficient of the policy variable Pilot × Post remains statistically significant at the 1% level. This not only confirms the robustness of the baseline regression results but also suggests that the driving effect of the SDA Policy on GREEN can, to a certain extent, withstand the interference of major external negative shocks, demonstrating a degree of resilience in its policy effectiveness.

3.

PSM-DID

Although the Difference-in-Differences model can control for both observable and unobservable fixed effects, systematic differences in pre-treatment characteristics between the treatment and control groups may still exist—that is, a sample selection bias problem. To reduce this error, we adopt the Propensity Score Matching (PSM) method. Using all control variables from the year before policy implementation as covariates, we estimate the probability (i.e., propensity score) of a firm being assigned to the treatment group via a Logit model. Subsequently, we employ three matching approaches—kernel matching, caliper matching (caliper = 0.01), and 1:1 nearest neighbor matching (without replacement)—to identify the most similar control group firms for each treated firm. The results show that regardless of the matching method used, the policy variable (Pilot × Post) is statistically significant at least at the 5% level (as shown in columns 3–5 of Table 6). After matching, the pre-treatment characteristics between the treatment and control groups become more balanced. This suggests that the policy effect observed in this study does not stem from any inherent “pre-existing advantage” of the treatment group firms but rather reflects the genuine impact of the policy shock.

5. Results of Further Analysis

5.1. Moderating Effect Analysis

To further identify the boundary conditions of the policy effect, this study incorporates Green Attention into the moderating framework and includes its interaction term with the policy variable. The results in Column (1) of

Table 7. Results of Moderating Effect Analysis and Mechanism Analysis.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	GREEN	CEDRT	GREEN	Reg_GreenInnov	GREEN	Media	GREEN
Pilot × Post		0.1922 ***	0.0065 ***	0.6502 *	0.0063 ***	0.1199 ***	0.0068 ***
		(0.067)	(0.002)	(0.360)	(0.002)	(0.043)	(0.002)
c_DID	0.0064 ***
	(0.002)
c_EGA	−0.0011
	(0.001)
c.DID × c.EGA	0.0066 ***
	(0.001)
CEDRT			0.0035 ***
			(0.001)
Region_innov					0.0013 **
					(0.001)
Media							0.0033 ***
							(0.001)
Constant	−0.0384	−7.3562 ***	−0.0141	−15.0238 **	−0.0199	−6.7846 ***	−0.0174
	(0.041)	(1.488)	(0.036)	(6.277)	(0.040)	(0.573)	(0.036)
Observations	11,250	11,250	11,250	11,250	11,250	11,250	11,250
R-squared	0.101	0.304	0.110	0.690	0.100	0.461	0.103
Controls	YES	YES	YES	YES	YES	YES	YES
Ind FE	YES	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES	YES

Note: Standard errors are shown in parentheses; ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

show that the coefficient of the interaction term (c_DID × c_EGA) is 0.0066, which is significantly positive at the 1% level. This indicates that Green Attention can significantly strengthen the promoting effect of the Pilot Zones Policy on GREEN.

The empirical findings suggest that when entrepreneurs focus their limited attentional resources more on strategic issues such as green transformation and sustainable development, their ability to identify, understand, and respond to policy signals is enhanced. This, in turn, improves the firm’s efficiency in absorbing policy incentives. The prioritization of attention at the cognitive level by entrepreneurs can not only guide the allocation of internal resources towards green innovation but also facilitate organizational adjustments in strategic configuration, technological R&D, and governance processes that align with policy orientation, thereby amplifying the marginal effect of policy incentives. Conversely, if entrepreneurial attention is predominantly concentrated on short-term financial targets or capital operations, it may weaken sensitivity to policy information, diminish the effectiveness of policy transmission, and hinder the conversion of external incentives into corporate green practices. Consequently, Hypothesis H2 is validated.

5.2. Mechanism Analysis

5.2.1. Corporate Digital–Real Integration

The aforementioned theoretical analysis indicates that the implementation of SDA Policy, by constructing digital infrastructure and providing technological incentives, can effectively promote the deep integration of corporate digital technology with real-economy operations. In other words, the CEDRT constitutes a key technical foundation for the policy effect to materialize at the firm level. The implementation of SDA Policy has, to some extent, facilitated the CEDRT within enterprises.

To delve deeper into the internal mechanisms through which the SDA Policy affects GREEN, this paper constructs the following mediating effect model based on the baseline difference-in-differences model:

$$M_{it}={\beta }_0+{\beta }_1{PilotPost}_{it}+{\beta }_2{Controls}_{it}+{\mu }_i+v_t+{\epsilon }_{it}$$

(3)

$${GREEN}_{it}={\gamma }_0+{\gamma }_1{PilotPost}_{it}+{\gamma }_2{M}_{it}+{\gamma }_3{Controls}_{it}+{\mu }_i+v_t+{\epsilon }_{it}$$

(4)

Among them, $M_{it}$ represents a series of mediator variables. Corporate CEDRT is measured by drawing on the method of Guo et al. [48]. Based on corporate annual patent application data, the quantity of integrated patents is counted by identifying the “co-occurrence” of digital technology and real economy categories within their International Patent Classification (IPC) codes. This count is then taken as the natural logarithm after adding 1.

Columns (2) and (3) of Table 7 present the mediation effect test results for this mechanism. Column (2) shows that the estimated coefficient for the SDA Policy (Pilot × Post) on the level of Corporate CEDRT is 0.1922, which is statistically significant at the 1% level. This result suggests that the implementation of the SDA Policy is positively associated with a higher degree of digital–real integration within firms. Column (3) further incorporates both the mediator variable (CEDRT) and the policy variable into the regression. The results show that the impact coefficient of Corporate CEDRT on GREEN is 0.0035, also significant at the 1% level. This demonstrates that the CEDRT plays a significant mediating role in the policy transmission path. In other words, the policy enhances GREEN by improving firms’ digital operations and intelligent production capabilities. Consequently, Hypothesis H3a is validated.

This finding is consistent with prior studies emphasizing the role of digital infrastructure and digital transformation in enhancing firms’ green productivity and environmental performance by improving production efficiency and reducing resource misallocation [49]. Existing research suggests that digital technologies can facilitate cleaner production processes and promote green upgrading by strengthening the integration between information systems and real economic activities [50]. Extending this literature, the present study further demonstrates that such a digital–real integration channel is significantly reinforced under a comprehensive and place-based sustainable development policy framework, thereby providing new micro-level evidence on how policy-driven digitalization translates into corporate green development.

5.2.2. Regional Green Innovation Capability

The theoretical analysis above indicates that the implementation of the national SDA Policy not only directly incentivizes enterprises but also enhances the overall regional green innovation capability at the meso-level. In other words, regional green innovation capability represents the foundational supply of knowledge and technology through which the policy’s effects diffuse to enterprises at the meso-level. The enhancement of regional green innovation capability can generate significant knowledge spillover effects, enabling enterprises to access external green technological knowledge at lower costs, thereby conducting green development activities more efficiently.

To verify the mediating role of regional green innovation capability, drawing on the measurement method from the authoritative study by Chen et al. [51], this paper uses the natural logarithm of the number of green invention patent applications in the prefecture-level city where an enterprise is located plus one to measure regional green innovation capability (Reg_GreenInnov). This measurement is based on patent identification according to the “International Green Patent Classification List” published by the World Intellectual Property Organization (WIPO), with data sourced from the Chinese Research Data Services Platform (CNRDS).

The results concerning the mediating role of regional green innovation capability in the process through which the SDA Policy influences GREEN are presented in Table 7. Column (4) shows that the estimated coefficient of the policy variable (PilotxPost) on regional green innovation capability (Reg_GreenInnov) is 0.6502, which is significant at the 10% level. This finding indicates a positive association between the establishment of pilot zones and improvements in the regional green innovation environment. The results in Column (5) further demonstrate that, after controlling for the policy variable, the coefficient of regional green innovation capability’s impact on GREEN is 0.0013, which is significant at the 5% level. This suggests that regional green innovation capability plays a significant mediating role in the process through which the policy affects green development at the corporate level. Specifically, the improvement of the regional green innovation ecosystem can reduce the cost for enterprises to acquire green technologies and strengthen knowledge spillover effects, thereby encouraging enterprises to independently engage in green development activities. Consequently, hypothesis H3b is validated.

This result is consistent with the existing literature, which documents that regions endowed with a stronger green innovation ecosystem can significantly reduce firms’ innovation costs and improve their access to specialized green knowledge [52]. Extending this line of research, we find that the SDA Policy selectively reinforces this meso-level innovation environment, thereby amplifying the transmission of policy effects from the regional level to firm-level green development outcomes.

5.2.3. Media Attention

The theoretical analysis above indicates that the implementation of the SDA Policy not only increases the volume of media coverage on enterprises in pilot regions but, more crucially, strengthens the media’s supervisory and governance role. In other words, Media embodies an important governance mechanism through which the policy signal exerts external pressure and reputational constraints on enterprises at the macro level. Such supervision directly challenges organizational legitimacy, thereby compelling firms to respond through substantive actions such as improving environmental performance.

Drawing on the authoritative measurement method by Liu et al. [53], this study measures this variable using the natural logarithm of the total number of news reports on listed companies in mainstream media outlets plus one. The data is sourced from the Chinese Research Data Services Platform (CNRDS), ensuring the authority of the source and the continuity of the data.

The results of the mechanism analysis are shown in Column (6) of Table 7. The coefficient of the policy variable (Pilot × Post) on Media is 0.1199, which is significant at the 1% level. This result indicates that the SDA Policy is associated with increased media attention toward listed firms in pilot regions. Column (7) further shows that the coefficient of Media on GREEN is 0.0033, also significant at the 1% level. This suggests that Media serves as a crucial external governance channel for transmitting the policy effect into firms by strengthening external reputational pressure, raising transparency requirements, and elevating stakeholder expectations. Consequently, Hypothesis H3c is validated.

This evidence is in line with prior research emphasizing the governance role of media attention in shaping corporate environmental behavior by increasing transparency, reputational pressure, and stakeholder scrutiny [54]. Existing studies suggest that heightened media coverage can discipline firms’ environmental conduct and encourage substantive environmental actions [55]. The present study complements this literature by showing that under the SDA Policy, media attention serves as an important external governance channel through which policy signals are amplified, thereby strengthening the effectiveness of differentiated environmental regulation at the firm level.

5.3. Heterogeneity Analysis

5.3.1. Heterogeneity in the Nature of Ownership

This study divides the sample into two groups: state-owned enterprises (SOEs) and non-state-owned enterprises (non-SOEs). The regression results are presented in columns (1) and (2) of

Table 8. Heterogeneity tests.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	SOE	Non-SOE	Heavily Polluting	Non-Heavily Polluting	Modern Urban	Resource-driven	High Env. Concern	Low Env. Concern
Pilot × Post	0.0012	0.0075 ***	0.0012	0.0070 ***	0.0067 **	−0.0019	0.0091 ***	0.0056
	(0.004)	(0.001)	(0.004)	(0.001)	(0.002)	(0.005)	(0.001)	(0.004)
Constant	−0.0482	−0.0259	−0.0482	−0.0199	−0.0715	0.0824	−0.0511	−0.0344
	(0.060)	(0.022)	(0.060)	(0.022)	(0.058)	(0.079)	(0.046)	(0.042)
Observations	4357	6582	4357	6893	8653	2596	7309	3941
R-squared	0.126	0.100	0.126	0.097	0.111	0.111	0.104	0.100
Controls	YES	YES	YES	YES	YES	YES	YES	YES
Ind FE	YES	YES	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	0.0012	0.0070 ***	YES	YES	YES	YES

Note: Standard errors are shown in parentheses; *** and ** indicate significance at the 1% and 5% levels, respectively.

. The results indicate that the SDA Policy exhibits a more pronounced facilitative effect on the green development of non-SOEs. The distinct institutional foundations of the two types of enterprises in terms of governance structures and incentive mechanisms serve as important reasons for this differential impact. From an institutional and incentive-based perspective, SOEs are embedded in a multi-task governance framework, where environmental objectives are often fulfilled through administrative compliance rather than profit-driven behavioral adjustments, which weakens the marginal incentive effect of newly introduced policies. In practice, SOEs often shoulder dual objectives of economic efficiency and public functions. Their environmental investments and green behaviors are more significantly influenced by administrative tasks, performance evaluation requirements, and responsibility constraints [56]. Under the existing regulatory framework, new policies offer relatively limited additional incentive space for SOEs, making it difficult for the incremental policy effect to be fully translated into action at the firm level. In contrast, the operational logic of non-SOEs is more focused on market returns and cost–benefit considerations. Under a stronger residual claim and market-based incentive structure, policy-induced changes in expected returns and compliance costs are more likely to be internalized by non-SOEs’ management decisions. Their management structures are more flexible, enabling them to promptly adjust inputs and strategic direction based on anticipated policy benefits. When the establishment of the Sustainable Development Agenda Pilot Zones strengthens green governance requirements and raises the expected returns on green development, non-SOEs are more inclined to respond swiftly through measures such as technological upgrades, process optimization, or green investments. Consequently, the policy’s impact is more likely to translate into substantive effects for these firms.

5.3.2. Heterogeneity in Pollution Intensity

Based on the Guidelines for Industry Classification of Listed Companies by the China Securities Regulatory Commission and the Industry Classification Management List for Environmental Compliance Verification of Listed Companies issued by the Ministry of Environmental Protection, this study divides the sample into heavily polluting industries and non-heavily polluting enterprises for examination. The results are shown in columns (3) and (4) of Table 8. The regression results reveal that the SDA Policy primarily promotes the green development of non-heavily polluting enterprises, while its short-term impact on heavily polluting enterprises is not statistically significant. Differences in technological pathways and retrofit costs across industries constitute the core factors leading to the divergence in policy effectiveness. From an incentive compatibility perspective, when compliance costs are excessively high and adjustment margins are limited, policy signals may fail to generate sufficient marginal incentives to induce behavioral change in the short run. Heavily polluting industries are generally characterized by long equipment renewal cycles, fixed production process pathways, and substantial investment requirements for green transformation. Consequently, it is challenging for them to achieve significant environmental performance improvements in the short term through lightweight technological adjustments. Although the policy reinforces the orientation toward green development, for these enterprises, the capital and technological costs required for transformation make it difficult for the policy to rapidly yield substantial changes. Conversely, the production systems of non-heavily polluting industries are relatively flexible. The integration of green technologies, the introduction of digital management, and the optimization of production processes can all be implemented at a lower cost, making it easier to generate immediate green performance improvements. The incentive signals released by the SDA Policy are more readily convertible in these sectors, leading to more pronounced policy outcomes.

5.3.3. Heterogeneity in Regional Development

To examine the differential impact of the national SDA Policy across different regional development types, this study, following the approach of Cheng and Wang [7], categorizes the sample into two groups: modern urban construction-driven regions and natural resource-driven regions. These two types of regions exhibit significant differences in industrial structure, factor endowments, and governance capacity, which may lead to varying policy absorptive capacities and conditions for green transformation. The results of group regressions (see columns (5) and (6) of Table 8) based on regional type show that the policy significantly enhances the level of GREEN in modern urban construction-driven regions. Institutionally, these regions tend to possess stronger policy implementation capacity and more complete incentive-support systems, which reduce transaction costs and enhance firms’ expectations regarding policy credibility and continuity. These areas typically feature a higher degree of marketization, more abundant supplies of technological resources and green factors, and well-established supporting institutional systems. Policy signals can be transmitted rapidly within a smoother institutional environment, and enterprises find it easier to access resource support such as green finance, technical consultation, and public services. Consequently, a more positive and rapid response to the policy is fostered. In contrast, the policy effect is not statistically significant in natural resource-driven regions. The industrial structure in these areas is characterized by a pronounced reliance on natural resources, with limited alternative green pathways. Enterprises are more susceptible to technological path lock-in and capital constraints. Furthermore, local governance priorities and industrial development inertia may weaken the transmission of policy signals, making it difficult for the policy to translate into improved green performance at the corporate level in the short term. The rigidity of regional industrial structures and differences in governance capacity are the primary reasons for the significant divergence in policy effectiveness between the two types of regions.

5.3.4. Heterogeneity in Government Environmental Attention

This study further divides the sample into a high environmental concern group and a low environmental concern group based on the frequency of environment- and ecology-related keywords in the government work reports of each city. The results are presented in columns (7) and (8) of Table 8. The regression results indicate that the policy exhibits a stronger promoting effect in cities with higher government environmental concern. From an institutional incentive perspective, stronger environmental concern reflects higher regulatory credibility and enforcement intensity, which amplifies both compliance pressure and expected policy returns for enterprises. In cities with greater environmental concern, green development is often incorporated into core governance objectives. Local governments tend to strengthen regulatory system construction, improve the level of environmental information disclosure, and provide stronger support in terms of fiscal investment and institutional supply. These measures reduce the cost for enterprises to obtain policy information and comply with policy requirements, allowing policy signals to be transmitted efficiently. This, in turn, encourages enterprises to engage more actively in green transformation. In regions with lower environmental concern, issues such as inadequate regulatory enforcement, insufficient policy support, and imperfect supervision systems are common. These shortcomings result in a weaker policy transmission chain, providing enterprises with insufficient incentives and pressure from the policy, thereby limiting the effectiveness of green development.

5.4. Unintended Consequences

5.4.1. The Impact of the SDA Policy on Corporate Green Innovation Bubbles

The previous regression results indicate that the SDA Policy can significantly enhance the level of GREEN. Following the implementation of this policy, companies have demonstrated a more proactive inclination toward green transformation in areas such as energy conservation and emission reduction, green production management, and environmental protection investment. Existing literature points out that green innovation constitutes one of the core technological pathways through which enterprises achieve green development, as technological progress facilitates sustained improvements in resource utilization efficiency, pollution control, and clean production [57]. Therefore, improvements in GREEN are typically accompanied by an expansion of green innovation activities.

However, such expansion does not necessarily imply a synchronous improvement in both the quantity and quality of green innovation. Under strong policy incentives and performance evaluation pressure, enterprises may engage in behaviors such as resource reallocation, strategic adjustment of patent application portfolios, and symbolic innovation investment. These responses may lead to a rapid increase in the quantity of green patents, while substantive innovation quality improves at a relatively slower pace, thereby generating a structural imbalance. This “quantity-oriented rather than quality-oriented” response pattern is consistent with the literature on policy-induced distortions, which suggests that when governments rely heavily on observable indicators for performance assessment, firms tend to concentrate resources on signal-generating outputs at the expense of genuine efficiency improvements [58]. To more comprehensively identify the net impact of the SDA Policy on green innovation quality, it is therefore necessary to examine whether corporate green innovation exhibits characteristics of superficial growth that emphasize quantity over quality.

Based on this consideration, drawing on the approach of Geng et al. [59], this study applies Z-score standardization to the difference between the application volume and the grant volume of green patents to construct a green patent bubble indicator. This indicator captures the deviation between innovation scale and innovation quality and is used to identify whether the policy has induced structural “overheating” in green innovation activities. As shown in Columns (1) and (2) of

Table 9. Unintended Consequences.

	(1)	(2)	(3)
	Invention Patent Green Bubble	Utility Model Green Patent Bubble	Environmental Info. Disclosure Quality
Pilot × Post	0.0463 **	−0.0096	−0.0703 *
	(0.020)	(0.016)	(0.041)
Constant	−4.3525 ***	0.2511	−4.7242 ***
	(1.503)	(0.183)	(0.470)
Observations	11,250	11,250	11,250
R-squared	0.069	0.002	0.413
Controls	YES	YES	YES
Ind FE	YES	YES	YES
Year FE	YES	YES	YES

Note: Standard errors are shown in parentheses; ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

, the estimated coefficients of the SDA Policy on green invention patent bubbles are significantly positive at the 5% level, whereas its effect on green utility model patent bubbles is not statistically significant.

Invention patents typically embody higher technological content and stronger signaling value, and they are more likely to attract financing support, policy attention, and reputational benefits. Consequently, they become preferred targets for strategic applications under policy incentives. In this sense, the existence of policy rents renders invention patents a “rent-seeking signal”, motivating firms to accumulate applications rapidly in order to secure short-term compliance benefits and resource advantages, rather than committing to high-risk, long-cycle fundamental technological breakthroughs [60]. When patent applications surge in response to policy signals but firms’ technological accumulation and R&D capabilities fail to keep pace, a divergence emerges between application volume and grant outcomes, thereby forming a pronounced innovation bubble. By contrast, utility model patents feature lower technological thresholds and limited external signaling value, making them less responsive to policy incentives and less prone to bubble-like expansion.

Overall, the evidence indicates that although the SDA Policy significantly enhances firms’ overall green development performance, its incentive effect in the domain of green innovation primarily manifests as rapid growth in patent quantity rather than a synchronous improvement in substantive innovation quality. This finding highlights a potential risk of policy design that overly relies on quantifiable indicators while neglecting quality screening and dynamic evaluation mechanisms, which may induce strategic innovation behavior and ultimately undermine the long-term technological foundations of sustainable development [61]. Under strong policy signals and assessment pressure, firms tend to prioritize the expansion of invention patent applications with high signaling value instead of allocating scarce resources to genuinely transformative green technologies characterized by long development cycles and high uncertainty.

5.4.2. The Impact of the SDA Policy on the Quality of Corporate Environmental Information Disclosure

The preceding empirical results indicate that the SDA Policy significantly promotes GREEN, with companies generally showing positive changes in areas such as energy conservation, emission reduction, green operations, and sustainable business practices. However, progress in GREEN does not imply synchronous improvement in all environmental governance activities. To further delineate the policy’s extended effects within a firm’s internal governance system, this study expands the scope of analysis to the quality of corporate environmental information disclosure.

Drawing on the research of Kong et al. [62], this paper constructs an indicator for corporate environmental information disclosure quality (Eidq) utilizing the CSMAR Environmental Research Database. (See

Table A1. Environmental Information Disclosure Quality Indicator System.

Disclosure Type	Disclosure Item	Scoring Criteria
Environmental Management Disclosure	Environmental Philosophy	Disclosure: 2 points No Disclosure: 0 points
	Environmental Objectives
	Environmental Management System
	Environmental Education and Training
	Special Environmental Actions
	Environmental Incident Emergency Mechanism
	Environmental Honors or Awards
	“Three Simultaneities” System
Environmental Certification Disclosure	ISO 14001 [65] Certification	Yes: 2 points No: 0 points
	ISO 9001 Certification
Environmental Information Disclosure Medium	Annual Report	Disclosure: 2 points No Disclosure: 0 points
	Social Responsibility Report
	Environmental Report
Environmental Liability Disclosure	Wastewater Discharge	Quantitative and Qualitative Description: 2 points Qualitative Only: 1 point No Disclosure: 0 points
	COD Emissions
	SO2 Emissions
	CO2 Emissions
	Soot and Dust Emissions
	Industrial Solid Waste Discharge
	Waste Gas Emission Reduction and Treatment
	Wastewater Reduction and Treatment
Environmental Performance and Governance Disclosure	Dust and Soot Treatment
	Solid Waste Utilization and Disposal
	Noise, Light Pollution, Radiation Control
	Cleaner Production Implementation

for details.) The measurement of corporate environmental information disclosure quality follows the approach of Wiseman [63]. First, the indicator system is categorized into monetary information and non-monetary information. The former covers content quantifiable in monetary terms, such as environmental liabilities, environmental performance, and governance; the latter encompasses content difficult to quantify monetarily, such as environmental management, environmental certifications, and disclosure media. Second, a three-tier scoring system is applied to monetary information: simultaneous quantitative and qualitative disclosure receives 2 points, qualitative disclosure only receives 1 point, and non-disclosure receives 0 points. For non-monetary information, a two-tier scoring system is used: disclosure receives 2 points, and non-disclosure receives 0 points. According to these criteria, environmental information disclosure encompasses 25 evaluation items across five dimensions. This study scores each company’s disclosure status item by item and sums the scores, then applies a logarithmic transformation to the total score to form the corporate environmental information disclosure quality indicator (Eidq). This aims to comprehensively measure the sufficiency and standardization of corporate environmental information disclosure. Specific details of each evaluation item are provided in the Appendix A.

According to legitimacy theory, increased external institutional pressure drives firms to respond to stakeholder legitimacy demands by improving the quality of environmental information disclosure. However, as shown in column (3) of Table 9, the quality of corporate environmental information disclosure has significantly declined following the implementation of the SDA Policy. This decline primarily stems from the resource reallocation behavior induced by the SDA Policy. Firstly, the SDA Policy, with its core assessment focus on green technology innovation and substantive environmental performance, incentivizes companies to prioritize the allocation of limited R&D, management, and human capital towards areas directly observable and quantifiable by regulators, such as energy-saving and emission-reduction technology development and green patent output. In contrast, environmental information disclosure, as a typical governance input, exhibits significant short-term cost characteristics and lagging benefits, making it susceptible to marginalization or crowding out under resource constraints. Secondly, in the early stages of green transformation, corporate actual environmental performance often remains deficient. Significantly increasing the level of disclosure detail at this juncture may expose existing shortcomings in pollution control or environmental management, thereby inviting stricter regulatory scrutiny or negative market evaluations. Out of prudential considerations, management tends to temporarily reduce the depth and breadth of disclosure to avoid potential compliance and reputational risks. This pattern of “selective silence” is consistent with the literature on policy-induced strategic behavior, which suggests that when regulatory attention is concentrated on verifiable “hard indicators”, firms may deliberately curtail the disclosure of “soft information” to weaken external monitoring intensity, thereby generating a substitutive relationship between innovation scale expansion and disclosure quality deterioration [64].

The evidence above demonstrates that while the SDA Policy significantly promotes GREEN, it simultaneously produces an unintended effect of inhibiting the quality of environmental information disclosure. The policy’s strong incentives for technological innovation crowd out the resource inputs required for information disclosure in the short term and induce firms to adopt relatively conservative disclosure strategies. This finding cautions policymakers that, in the absence of synchronized evaluation mechanisms for disclosure quality, one-dimensional technology-oriented incentives may lead to an imbalance between “visible innovation” and “hidden information,” ultimately undermining the overall governance effectiveness of sustainable development policies.

6. Discussion

This study systematically examines the impact of the national SDA Policy on GREEN by constructing a multi-period difference-in-differences model and arrives at the following conclusions. First, the SDA Policy significantly improves the level of GREEN, underscoring the unique advantages of differentiated environmental regulation that integrates technological innovation orientation with regional coordination in advancing corporate green practices. Second, Green Attention exerts a positive moderating effect on the relationship between the SDA Policy and GREEN. Mechanism analysis indicates that the SDA Policy promotes GREEN through three pathways: enhancing the level of digital–real integration within enterprises, strengthening regional green innovation capability, and increasing Media. Third, the policy’s green development enhancement effect is more pronounced among non-state-owned enterprises, firms in non-heavily polluting industries, regions characterized by modern urban development, and areas with higher levels of government environmental concern. Fourth, despite its positive effects, the SDA Policy generates certain unintended consequences; while promoting green development, it is also associated with an increased tendency toward “bubbles” in green invention patents and a decline in the quality of environmental information disclosure.

Taken together, these findings suggest that incentive-based green policies may induce firms to adopt strategic or symbolic compliance behaviors, whereby observable and quantifiable indicators are prioritized over substantive improvements in environmental governance. This highlights a potential tension between policy-driven expansion of green activities and the depth and quality of firms’ internal green transformation.

In contrast to existing studies that primarily document the average effects of environmental or green innovation policies on firms’ innovation outputs, this study contributes by uncovering a nuanced tension between policy-induced green innovation expansion and the quality of environmental information disclosure at the firm level. Beyond their empirical relevance in the Chinese context, the unintended effects identified in this study offer broader insights into the design of sustainability-oriented public policies. Specifically, when policy incentives are strongly anchored to measurable outputs such as patent counts or formal disclosure requirements, firms may rationally reallocate resources toward indicators that are more visible to regulators, investors, and the public, while deprioritizing less observable but equally important dimensions of environmental governance. Such strategic responses do not necessarily undermine policy effectiveness in the short term, but they may weaken the long-term quality and credibility of green development outcomes if left unaddressed. This dynamic the importance of balancing incentive intensity with governance mechanisms that emphasize quality, transparency, and substantive environmental performance. While the empirical analysis is based on China, the underlying mechanism of firms’ strategic responses to incentive-based and indicator-oriented sustainability policies is likely generalizable to other institutional contexts where regulatory evaluation heavily depends on observable performance metrics.

Based on these conclusions, this study proposes the following policy implications. Within the institutional context of China’s SDA Policy and capital market system, first, the hierarchical and categorized implementation pathways of the pilot zone policy should be optimized to enhance policy alignment and the precision of institutional supply. To further strengthen policy effectiveness, policy portfolios should be implemented by gradient, industry, and scenario in accordance with regional differences in industrial bases, technological reserves, green governance capacity, and factor endowments, thereby improving the fit between policy design and regional development conditions. In regions with stronger governance capacity and more advanced green foundations. For regions with relatively mature governance capacity and stronger green foundations, institutional innovation pilots may be expanded to develop a replicable policy toolkit that can serve as a model for China’s broader green transition.

Second, the strategic leadership role of entrepreneurs in the green transition should be strengthened to improve policy absorption and internal transformation efficiency. For firms subject to capital market discipline and information disclosure requirements, governments should guide entrepreneurs to incorporate green objectives into their core strategic planning through policy communication, capacity-building programs, green performance evaluation systems, tax incentives, and support from multi-level capital markets, thereby enhancing their awareness of and responsiveness to sustainable development goals. Concurrently, enterprises should be encouraged to increase investment in digital–real integration, green process innovation, and low-carbon technology R&D, thereby improving the micro-level transmission efficiency of policy signals and enabling policy incentives to more effectively translate into corporate green performance.

Third, the quality supervision of green innovation and ecological governance should be reinforced to prevent green innovation “bubbles” and greenwashing behaviors. While policy incentives have effectively expanded the scale of green innovation among listed firms, the issue of uneven innovation quality requires urgent attention. The patent quality evaluation system should be refined, stringent green technology standards should be promoted, and oversight of science and technology resource allocation should be strengthened to ensure that innovation outputs possess substantive technological value. At the same time, the environmental information disclosure system should be improved to enhance the authenticity, completeness, and comparability of corporate disclosures, strengthening public oversight and capital market constraints and reducing strategic or formalistic disclosure practices.

Fourth, institutional coordination and external supervision systems should be improved to reduce governance costs during policy implementation and enhance policy sustainability. In the context of differentiated environmental regulation, coordination across governmental departments should be strengthened. Industrial, innovation, financial, and environmental governance policies should be jointly designed to establish cross-departmental and cross-regional information-sharing and collaborative regulatory mechanisms. In addition, increasing media attention, public participation, and third-party evaluations can enhance the transparency of policy implementation, improve the identification of corporate strategic behaviors, and reduce institutional distortions and policy enforcement costs.

7. Limitations and Future Research

Despite the rigorous empirical design and extensive robustness tests, several research boundaries should be acknowledged.

First, this study focuses on A-share listed firms, which are subject to relatively strict capital market regulation and information disclosure requirements. While this setting enhances data reliability and policy identification, the findings may be more representative of firms operating under formal governance and market discipline. Future research could extend the analysis to non-listed firms or small and medium-sized enterprises to further assess the generalizability of the conclusions.

Second, the empirical context is embedded in China’s SDA Policy, a place-based and institutionally coordinated policy framework. Although this institutional setting provides a valuable quasi-natural experiment, the effectiveness and transmission mechanisms of similar sustainability-oriented policies may vary across countries with different regulatory capacities and market environments. Cross-country or comparative studies would therefore be a promising avenue for future research.

Finally, corporate green development and entrepreneurial green attention are measured using observable proxies derived from multi-dimensional indicators and textual analysis. While these approaches are grounded in established literature and capture key aspects of firms’ green transformation, future research may further enrich the analysis by incorporating longer observation horizons, alternative measurement strategies, or more direct environmental outcome data, thereby deepening the understanding of firms’ strategic responses to sustainability-oriented policies.