Environmental Governance Innovation and Corporate Sustainable Performance in Emerging Markets: A Study of the Green Technology Innovation Driving Effect of China’s New Environmental Protection Laws

Abstract

Against the backdrop of the accelerated transition to sustainable development in global emerging markets, the synergistic mechanism between environmental governance innovation and corporate green transformation has become a key issue in realizing high-quality development. As the world’s largest emerging economy, China’s new Environmental Protection Law (EPL), implemented in 2015, has promoted green technology innovation and performance improvement of heavily polluting enterprises by strengthening environmental regulation. This paper takes Chinese A-share listed companies as samples from 2012–2023, treats the EPL as a quasi-natural experiment, and applies the DID method to explore the path of its impact on the performance of heavily polluting firms, with a focus on analyzing the mediating effect of green technological innovation and the moderating role of firm size and regional differences. The study revealed the following findings: the implementation of the EPL significantly improves the performance of heavily polluting enterprises, which verifies the applicability of “Porter’s hypothesis” in emerging markets; green technological innovation plays a partly intermediary role in the process of policy affecting enterprise performance, indicating that environmental regulation achieves win–win economic and environmental benefits by driving the innovation compensation mechanism; and there is significant heterogeneity in policy effects, with large-scale firms and firms in the eastern region experiencing more pronounced performance improvements, reflecting differences in resource endowments and institutional implementation strength within emerging markets. This study provides empirical evidence for emerging market countries to optimize their environmental governance policies and construct a “regulation–innovation–performance” synergistic mechanism, which will help green economic transformation and ecological civilization construction.

1. Introduction

Driven by the global Sustainable Development Goals (SDGs), emerging markets are undergoing a profound change from “growth prioritization” to “green transformation”. As the world’s largest cluster of emerging economies, emerging markets contribute more than 60% of global economic growth, but they also face severe environmental challenges—the traditional development model of high emissions and high energy consumption has led to ecological carrying capacity approaching its limit, and the lack of environmental governance capacity and lagging corporate sustainability practices have become the double bottleneck constraining their high-quality development. In this context, how to guide enterprises to adopt sustainable business practices through governance innovation and realize environmental impact and economic performance through green technology innovation has become the core issue facing emerging markets.

The report of the 20th CPC National Congress proposes to “actively and steadily push forward carbon peaking and carbon neutrality” and emphasizes that “achieving carbon peaking and carbon neutrality is a broad and profound economic and social systematic change”, providing a strategic blueprint for China’s green transformation [1].

In China, a typical emerging market, the EPL has been hailed as the “most stringent environmental protection law in history” as an important policy tool [2]. It signaled a major paradigm shift in environmental governance. By establishing a strict environmental regulatory system, the law for the first time incorporates environmental governance into the core considerations of corporate strategic decisions, providing an ideal quasi-natural experiment for exploring the transmission mechanism of “governance innovation–technological innovation–sustainable performance”. However, in the institutional environments specific to emerging markets, does the impact of environmental regulation on corporate sustainability practices show heterogeneity? What mediating role does green technology innovation play between governance policies and firm performance? The answers to these questions are of great theoretical and practical significance for optimizing environmental governance policies and building sustainable business ecology in emerging markets.

This paper empirically examines the policy effects of the EPL based on the data of A-share listed companies from 2012 to 2023 using the double-difference-in-differences (DID) method with Chinese heavy polluters as the research object. The study not only focuses on the direct impact of the policy on firms’ financial performance but also analyzes the intermediary mechanism of green technological innovation, as well as the moderating role of firm size and regional differences in policy transmission. By placing China’s experience under the framework of emerging market theory, this paper aims to provide a replicable “environmental-governance-innovation-driven sustainable development” path for global emerging economies and help to solve the dilemma of development and protection.

2. Literature Review

In order to systematically answer the above research questions and clarify the position of this paper in the existing knowledge system, we first systematically review and comment on the core literature related to environmental regulations, green technological innovation, and corporate performance.

Environmental regulatory policies in emerging markets are often both coercive and innovative, aiming to guide the transformation of corporate behavior through institutional change. China’s EPL, a typical governance innovation tool, has significantly strengthened the constraints on heavily polluting firms by establishing a strict environmental regulatory system (Zhang, 2023) [2]. Empirical studies have shown that such policies have a significant positive impact on corporate performance: using a quasi-natural experiment, Lin and Yuan (2025) found that the return on assets (ROA) of heavy polluters increased by 2% after the implementation of the EPL, verifying the promotional effect of environmental regulation on corporate financial performance [3,4]; Wang et al. (2021) further pointed out that the EPL, by reducing environmental pollution accidents, enhances the corporate social image and indirectly enhances the market competitiveness of listed industrial enterprises [5]. These findings support the applicability of Porter’s hypothesis in emerging markets, which states that environmental regulations can enhance productivity by stimulating firms’ innovative behavior. However, existing research in this field has obvious shortcomings. Most studies analyze a single policy tool in isolation [4,5,6], lacking integrated research on how the diversified governance system of emerging markets, comprising legal regulation, market incentives, and social supervision, promotes sustainable corporate practices, particularly overlooking the heterogeneity differences in policy responses among enterprises of different sizes and regions.

The mediating effect of green technology innovation between environmental regulation and firm performance has been preliminarily verified. Pan (2025), studying environmental tax policy, finds that tax incentives significantly improve firms’ environmental performance by promoting green technology innovation [7]. Tang et al. (2023), focusing on the EPL, point out that the law encourages firms to increase green patent applications by improving the environment of ecological rule of law, thus improving innovation-driven performance growth [8]. In the Chinese context, Dai (2023) finds that a 10% increase in green technological innovation investment by heavily polluting firms increases their financial performance by 1.5%, highlighting the central role of technological innovation in the transmission mechanism of environmental regulation [9]. However, there remains a pressing need to fill the gaps in this research area. Existing studies have not fully uncovered the unique pathways of green technological innovation in heavily polluting industries [7,8,9,10]. Unlike general manufacturing, technological innovation in heavily polluting enterprises may rely more on government subsidies and collaboration between industry, academia, and research institutions. The mechanisms by which such innovations contribute to environmental and economic performance are industry-specific and require targeted micro-level analysis. Moreover, in emerging markets, how government regulation can synergize with corporate independent innovation has yet to be adequately addressed [11,12].

Significant heterogeneity within emerging markets leads to a differentiation of the effects of environmental regulation. In terms of firm size, large-scale firms are more likely to transform environmental pressures into innovation incentives by virtue of stronger capital accumulation and technology absorption. This study finds that the performance-enhancing effect of the EPL is 12% stronger for large and heavily polluting firms than for SMEs, which is consistent with the findings of Opaluch and Jin (2005) in developed countries, reflecting the reinforcing effect of resource advantages on policy responses [13]. At the regional level, the policy effect is more pronounced in the eastern region due to higher enforcement and a well-developed innovation ecosystem, while the performance improvement in the central and western regions lags behind due to the traditional industrial structure and financing constraints. This difference echoes the “U-shaped relationship” between environmental regulation intensity and industrial competitiveness proposed by Fu et al. (2010), suggesting that regions with superior institutional environments are more likely to realize the virtuous cycle of “regulation–innovation–performance” [14]. However, existing research still falls short in providing a theoretical explanation for heterogeneity factors [15,16,17]. The moderating role of firm size and regional differences is often treated as a control variable rather than part of the theoretical framework in emerging market institutional settings. For instance, why do large firms respond more significantly to policies in the eastern region? This requires a deeper theoretical analysis from dimensions such as resource acquisition capabilities and institutional fit, which are currently lacking in relevant studies.

This paper is the first to construct a complete analytical framework of “governance innovation–green technological innovation–sustainable performance” in the context of emerging markets, revealing the mediating role of green technological innovation in policy transmission and enriching the application of Porter’s hypothesis in developing countries. By introducing enterprise size and regional differences as moderating variables, it has been found that the heterogeneity of policy effects is rooted in the differences in resource endowment and institutional implementation within emerging markets, which provides a new theoretical perspective for understanding the environmental response of enterprises in emerging markets. The methodology utilizes the DID method and propensity score matching (PSM-DID) to solve the problem of endogeneity and combines the analysis of mediating effects and heterogeneity with scientific assessment of the long-term impact of the EPL on heavy polluters; the research methodology is replicable and provides a methodological reference for similar emerging market research. The study finds that the policy effects of large-scale enterprises and the eastern region are more significant, which provides a basis for emerging market countries to formulate differentiated environmental policies. For example, financing support and technical assistance for small and medium-sized enterprises, as well as enforcement capacity building in the central and western regions, can help to narrow the policy effect gap and promote the overall green transition.

Based on the above literature review, it can be seen that although the impact of environmental regulations on corporate performance, the mediating role of green technological innovation, and issues of heterogeneity have been partially explored, there are still significant discrepancies and gaps in existing research that need to be filled. These theoretical controversies and research gaps provide a clear logical starting point for this paper to integrate the theoretical basis and propose research hypotheses.

3. Theoretical Foundations and Research Hypotheses

3.1. Theoretical Foundations

As an external environmental regulatory force, the EPL has a significant impact on the control and subsequent improvement of heavily polluting industries. Corporate behavior will be influenced by the EPL, an external environmental regulatory policy factor, which will ultimately be reflected in corporate performance.

Given the unique institutional dynamics and developmental stage of emerging markets, interpreting the complex relationship between stringent environmental regulations like the EPL and corporate performance requires a multi-dimensional theoretical perspective. We particularly focus on three competing yet complementary theories: the traditional hypothesis, Porter’s hypothesis, and the uncertainty hypothesis. These theories collectively encompass the various responses firms may adopt in the context of emerging markets, particularly in environments characterized by firm capability heterogeneity, regional differences, and evolving regulatory frameworks. Although they developed in different contexts, they provide essential conceptual tools for analyzing the complex and potentially divergent impacts of major environmental governance innovations like the EPL in emerging market environments.

3.1.1. Traditional Hypothesis

The traditional hypothesis suggests that the environmental regulatory effects of the new EPA will increase firms’ production costs and lead to a decline in firm performance. Neoclassical economic theory emphasizes that firms must internalize pollution control costs to meet environmental requirements, which directly increases operational burdens [18]. This process triggers a chain reaction, where increased pollution control costs weaken production efficiency [19], simultaneously reducing the market competitiveness of pollution-intensive products [20]. Highly polluting industries face resource competition challenges, as environmental regulations crowd out their productive investment space and suppress technological innovation investments [21,22], ultimately forming a theoretical closed loop of “rising compliance costs → distorted resource allocation → declining overall performance.” (Figure 1).

The relevance of the traditional hypothesis is particularly pronounced in emerging markets for the following reasons: (1) Businesses, especially small and medium-sized enterprises that dominate emerging markets, typically face lower profit margins and significant financing constraints, making it exceptionally burdensome and difficult to absorb substantial new compliance costs, particularly when compared to developed economies. (2) In emerging market economies, intensified resource competition may result in scarcer capital and skilled labor, and environmental compliance costs are more likely to crowd out productive investments critical to growth and basic technological upgrades, rather than merely cutting-edge innovation. (3) Initial stringent regulatory measures in emerging market economies, such as the EPL, often have a command and control nature, which can immediately generate potential compliance costs before support mechanisms are fully established to mitigate the impacts of the transition period.

3.1.2. Porter’s Hypothesis

Contrary to the traditional hypothesis, Porter’s hypothesis suggests that the implementation of the EPL will induce technological innovation in heavily polluting enterprises, triggering an “innovation compensation” effect that will increase their productivity and competitiveness. As environmental regulations tighten, firms will promote development through technological progress. Enterprises view pollutants as a reflection of underutilization of resources, and environmental regulations can promote technological upgrading, improve resource utilization, save energy, reduce emissions, lower costs, and increase revenues. Under the development trend of the green economy, enterprises take the lead in optimizing production technology and researching and developing new products to gain a “first-mover advantage”. Opaluch and Jin (2005) applied frontier production analysis to test the causal relationship between technological innovation and environmental regulatory policies, supporting Porter’s hypothesis [13]. In China, Li, Q. and Nie, R. (2009) empirically investigate the relationship between environmental regulation and technological R&D, and the results show that environmental regulation is positively related to technological R&D [15]. Huang, DC. and Liu, ZB. (2006) added technology coefficients to Robert’s model, confirming that environmental regulations can increase the cost of firms but also promote their innovation, and they make up for the increase in cost through the benefits of innovation [16]. Using data from industrial enterprises in 30 provinces and cities, Zhao, H. (2008) confirms that environmental regulation policies have a positive effect on firms’ technological innovation in the medium and long term [17].

The applicability of Porter’s hypothesis in emerging markets depends on specific conditions. Porter’s hypothesis posits that reasonable regulatory measures can stimulate an innovation offset effect. As a strict and credible regulatory signal in China, the EPL aims to create such stimulation by increasing pollution costs and signaling a long-term commitment to environmental goals, thereby altering companies’ strategic considerations regarding innovation. Additionally, in emerging market economies like China that are rapidly advancing green transitions, the potential for first mover’s advantages in green technology may be significant, driven by both domestic policies and global supply chain pressures. However, achieving Porter’s effect in emerging markets also faces challenges such as insufficient innovation capacity among firms, weak intellectual property protection, and high dependence on the quality of the institutional environment. The significant differences within emerging markets and the success of the EPL in triggering innovation provide empirical support for the potential validity of Porter’s hypothesis in emerging markets, although this validity depends on corporate capabilities and regional contexts.

3.1.3. Uncertainty Hypothesis

According to the “uncertainty hypothesis”, the effect of environmental regulation of the EPL on enterprise performance is uncertain due to factors such as industry attributes, internal and external environments, etc. Using data from the manufacturing industry, Fu, JY. et al. (2010) found a U-shaped relationship between environmental regulations and the international competitiveness of China’s polluting industries [14]. Through empirical analysis, Zhang, C. and Lu, Y. (2011) found a “U”-shaped relationship between environmental regulation intensity and technological innovation in the eastern and central regions [23].

The uncertainty hypothesis is undoubtedly the most intrinsically relevant theoretical perspective when analyzing the impact of the EPL on emerging markets. Its core assumptions align directly with the key characteristics of emerging markets: there is significant heterogeneity within the group of emerging market firms; there are significant differences in institutional quality and enforcement capacity across regions within large emerging markets, leading to uneven regulatory pressure and support; market conditions, infrastructure, and access to resources are highly uneven; and firms in heavily polluting industries may face differentiated challenges in terms of technology and costs.

Therefore, the net impact of unified national policies like the EPL is inherently uncertain and likely to exhibit significant heterogeneity, making the uncertainty hypothesis crucial for capturing the complex realities of regulatory environments in emerging markets.

3.2. Research Hypotheses

3.2.1. EPL and the Performance of Heavily Polluting Enterprises

“Porter’s Hypothesis” suggests that environmental regulation can induce firms to innovate and improve productivity and competitiveness. Under the strict requirements of the EPL, heavily polluting firms will increase their investment in technological innovation and optimize their production processes in order to comply, thus improving their performance. At the same time, the “uncertainty hypothesis” also suggests that the impact of environmental regulations varies according to the type of enterprise and the environment. For some heavily polluting enterprises, the EPL may become an opportunity for them to transform and upgrade their performance through technological upgrading and management optimization. In addition, the implementation of the EPL may also promote enterprise performance improvement through the following ways: first, improving the enterprise’s environmental management level, reducing environmental pollution accidents and fines, and lowering the enterprise’s environmental risks; secondly, enhancing the enterprise’s social image and brand value, strengthening the enterprise’s market competitiveness, and attracting more consumers and investors; and thirdly, promoting the enterprise’s relationship with the government, financial institutions, and other stakeholders to establish a better cooperative relationship and create a favorable external environment for the development of the enterprise. Therefore, this paper puts forward the following hypothesis:

H1. 

The EPL can improve the performance of heavily polluting enterprises.

3.2.2. EPL, Green Technology Innovation, and Heavily Polluting Enterprise Performance

The implementation of the EPL, as an important means of environmental regulation, has provided heavy polluters with an incentive to innovate in green technology. In order to cope with the regulatory requirements, enterprises will increase their R&D investment in green technology and develop more environmentally friendly production processes and products. This green technology innovation not only helps enterprises to reduce environmental costs but also enhances their market competitiveness and meets consumer demand for environmentally friendly products, which ultimately improves corporate performance. At the same time, green technology innovation can also enhance the social image of the enterprise and bring more business opportunities and economic benefits to the enterprise. In addition, green technology innovation can also promote enterprise performance through the following ways: first, improving resource utilization efficiency, reducing the consumption of raw materials and energy, and reducing production costs; secondly, optimizing the production process, improving production efficiency and product quality, and increasing the enterprise’s economic benefits; thirdly, developing new environmentally friendly products and services, expanding the market space, and creating new profit growth points for the enterprise. Therefore, this paper puts forward the following hypothesis:

H2. 

The implementation of the EPL promotes green technological innovation in heavily polluting firms, which ultimately improves firm performance.

4. Research Design

To visually conceptualize the research framework, Figure 2 presents the analytical model delineating the relationships between the explanatory variable, explained variable, mediating mechanism, and heterogeneity factors, which will be empirically tested in the subsequent models.

4.1. Sample Selection and Data Sources

Based on the principle of the ‘effective window’ of the DID method, policy anticipation effect avoidance, and data quality optimization, the sample data in this paper starts from 2012, avoiding too many years before the implementation of the EPL in order to introduce irrelevant shocks, anticipatory behaviors, and classification errors resulting in the weakening of the causal inferential validity of the study. In terms of the current research objectives, we select the data of China’s A-share listed companies from 2012 to 2023 as the sample data and carry out the following steps: (1) exclude ST, *ST, and PT samples; (2) exclude companies in the financial industry; (3) exclude individual samples in which the change in the treatment group and the control group occurs in the sample period; (4) exclude individual samples of the treatment group that are less than one period before the implementation of the policy; (5) exclude the single-sample observation values; and (6) perform 1–99% closing for continuous variables. After the above sample selection and screening, a total of 33,261 observations from 4983 listed companies are finally obtained. The relevant data of the research sample comes from the database of Cathay Pacific (CSMAR) and the annual reports of listed companies.

4.2. Variables Definitions and Explanations

4.2.1. Explained Variables

The explanatory variable in this paper is enterprise performance (Ep), referring to the study of Sun Chuanwang et al. (2022), using the net rate of total assets (ROA) as a measure of enterprise performance [24]. The net rate of total assets is the ratio of the net profit of the enterprise to the average total assets; that is, the level of profitability of the enterprise occupies the efficiency of the assets. A higher net rate of total assets indicates that the enterprise’s profitability is more efficient.

4.2.2. Explanatory Variables

The core objective of this research is to analyze the impact of the EPL on the performance of heavy polluters, so the difference-in-differences (DID) approach is used to construct the core explanatory variables of this paper. Referring to Wang, YP et al. (2021) and Pan, AL. et al. (2019), based on the industry classification standard of Chinese listed companies, the industry classification codes of A01, A02, A03, A05, B06, B08, B09, C17, C19, C22, C25, C26, C28, C29, C30, C31, C32, and D44 are set to be the heavily polluting enterprises [25,26], and if the enterprise is a heavily polluting enterprise, then the enterprise is set as a treatment group and Treat = 1; otherwise, Treat = 0. For the implementation of the EPL, i.e., 2015 and after, Post is quantified as 1, while for 2014 and earlier, Post is quantified as 0. We then construct the explanatory variable Did = Treat × Post.

4.2.3. Mediating Variables

Regarding the mediating variables, this paper chooses green technology innovation as the mediating variable. The State Intellectual Property Office (SIPO) has compiled statistics on four items of data, including the number of green invention GTI applications, the number of green invention GTI grants, the number of green utility GTI applications, and the number of green utility GTI grants by region. Referring to the method of defining green technological innovation in the study of Wang, YR et al. (2025), the total number of green GTI applications in the year, i.e., the sum of the number of green invention GTI applications and the number of green utility GTI applications, is selected as an indicator to measure the level of green technological innovation, and the larger the value, the higher the level of green technological innovation is [27].

4.2.4. Control Variables

In order to enhance the accuracy of the model, with reference to the study of Wang Liping et al., the control variables were selected mainly from the consideration of corporate characteristics [5], growth ability, development ability, etc. Specifically, the size of the board of directors, the proportion of independent directors, the size of the enterprise, the age of the enterprise, the status of cash flow, the concentration of shareholding, the balance sheet ratio, the growth of the enterprise, the capital tangible rate, and the intangible asset rate are used as control variables.

All the variables selected and defined in this paper are shown in

Table 1. List of variable definitions.

Variable Type	Variable Name	Variable Symbol	Variable Definition
Explained variable	Enterprise performance	Ep	Net profit/average total assets of the firm.
Explanatory variable	Policy effect	Did	Treat × Post, where Treat is the treatment group division; Treat = 1 if the firm is a heavy polluter, otherwise 0. Post is the policy implementation time variable; 2015 and after, Post = 1; 2014 and before, Post = 0.
Mediating variable	Green technology innovation	GTI	Logarithm of the sum of the total number of GTI applications of enterprises plus 1.
Control variable	Board size	Board	Number of directors + 1, taking the natural logarithm.
	Proportion of independent directors	Pid	Number of independent directors/total number of board members.
	Enterprise size	Size	Natural logarithm of total assets.
	Age of business	Age	The natural logarithm of the time of establishment of the enterprise.
	Shareholding concentration	Top10	Shareholding ratio of top ten shareholders.
	Balance sheet ratio	Lev	Total liabilities/total assets.
	Cash flow situation	Cashflow	Net cash flow from operations/total assets.
	Corporate growth	Growth	Revenue growth rate.
	Capital tangibility ratio	Tangible	Net fixed assets/total assets.
	Intangible assets ratio	Intangible	Net intangible assets/total assets.

.

4.3. Model Construction

Based on the policy implications, in order to test the effect of the law on the performance of heavy polluters and the mediating effect of green technology innovation, according to the analysis of the previous hypotheses, this paper constructs the following benchmark model:

Model 1: Impact of the EPL on business performance

$$\begin{array}{cc}\hfill {Ep}_{i,t}=& {\alpha }_0+{\alpha }_1{Did}_{i,t}+{\alpha }_2{Board}_{i,t}+{\alpha }_3{Pid}_{i,t}+{\alpha }_4{Size}_{i,t}+{\alpha }_5{Age}_{i,t}\hfill \\ & \hspace{1em}\hspace{1em}\hspace{1em}+{\alpha }_6{Top10}_{i,t}+{\alpha }_7{Lev}_{i,t}+{\alpha }_8{Cashflow}_{i,t}+{\alpha }_9{Growth}_{i,t}\hfill \\ & \hspace{1em}\hspace{1em}\hspace{1em}+{\alpha }_{10}{Tangible}_{i,t}+{\alpha }_{11}{Intangible}_{i,t}+{\lambda }_i+yea{r}_t+{\epsilon }_{i,t}\hfill \end{array}$$

(1)

In Equation (1), i,t denotes year t data for firm i, ${\alpha }_0$ is the intercept, ${\alpha }_1-{\alpha }_{11}$ is the coefficient of each variable, ${\lambda }_i$ is the individual fixed effect, $yea{r}_t$ is the year fixed effect, and ${\epsilon }_{i,t}$ is the randomized disturbance term.

Model 2: The mediating effect of green technology innovation in the impact of the EPL on firm performance

$$\begin{array}{cc}\hfill {GTI}_{i,t}=& {\alpha }_0+{\alpha }_1{Did}_{i,t}+{\alpha }_2{Board}_{i,t}+{ \alpha }_3{Pid}_{i,t}+{\alpha }_4{Size}_{i,t}+{\alpha }_5{Age}_{i,t}\hfill \\ & \hspace{1em}\hspace{1em}\hspace{1em}+{ \alpha }_6{Top10}_{i,t}+{\alpha }_7{Lev}_{i,t}+{\alpha }_8{Cashflow}_{i,t}+{\alpha }_9{Growth}_{i,t}\hfill \\ & \hspace{1em}\hspace{1em}\hspace{1em}+{\alpha }_{10}{Tangible}_{i,t}+{\alpha }_{11}{Intangible}_{i,t}+{\lambda }_i+yea{r}_t+{\epsilon }_{i,t}\hfill \end{array}$$

(2)

In Equation (2), the meaning of each symbol is the same as in Formula (1).

$$\begin{array}{cc}\hfill {GTI}_{i,t}=& {\alpha }_0+{\alpha }_1{Did}_{i,t}+{{\alpha }_2{Patent}_{i,t}+\alpha }_3{Board}_{i,t}+{\alpha }_4{Pid}_{i,t}+{\alpha }_5{Size}_{i,t}\hfill \\ & \hspace{1em}\hspace{1em}\hspace{1em}+{\alpha }_6{Age}_{i,t}+{\alpha }_7{Top10}_{i,t}+{\alpha }_8{Lev}_{i,t}+{\alpha }_9{Cashflow}_{i,t}\hfill \\ & \hspace{1em}\hspace{1em}\hspace{1em}+{\alpha }_{10}{Growth}_{i,t}+{\alpha }_{11}{Tangible}_{i,t}+{\alpha }_{12}{Intangible}_{i,t}+{\lambda }_i\hfill \\ & \hspace{1em}\hspace{1em}\hspace{1em}+yea{r}_t+{\epsilon }_{i,t}\hfill \end{array}$$

(3)

In Equation (3), ${\alpha }_1-{\alpha }_{12}$ is the coefficient of each variable, and the rest of the symbols have the same meaning as in Formula (1).

Based on the above research design, this section will report the descriptive statistics and correlation analysis results of the sample, the benchmark regression results, a series of robustness test results, and further mediation effect and heterogeneity analysis results.

5. Empirical Results and Analysis

5.1. Descriptive Statistical Analysis

Table 2. Descriptive statistical analysis.

VarName	Obs	Mean	Median	SD	Min.	Max.
Ep	33,261	0.035	0.036	0.079	−1.872	1.285
GTI	33,261	2.785	2.944	1.762	0.000	9.406
Treat	33,261	0.197	0.000	0.398	0.000	1.000
Post	33,261	0.816	1.000	0.387	0.000	1.000
Did	33,261	0.155	0.000	0.362	0.000	1.000
Board	33,261	2.228	2.303	0.177	0.000	2.944
Pid	33,260	0.378	0.364	0.056	0.143	0.800
Size	33,261	22.284	22.083	1.334	14.942	28.697
Age	33,253	7.600	7.600	0.003	7.580	7.609
Top10	33,261	0.576	0.583	0.152	0.229	0.901
Lev	33,261	0.424	0.415	0.204	0.058	0.904
Cashflow	33,261	0.047	0.046	0.068	−0.155	0.241
Growth	33,261	4.351	0.078	738.162	−1.309	135,000
Tangible	33,261	0.203	0.170	0.155	0.000	0.954
Intangible	33,261	0.047	0.033	0.060	0.000	0.895

shows the descriptive statistics of the main variables. Treat has a mean of 0.197, reflecting the fact that 19.7% of the sample is a treatment group sample, and Did has a mean of 0.155, reflecting the fact that 15.5% of the sample is affected by the policy.

5.2. Correlation Analysis

The correlation situation between the variables selected in this paper was further analyzed, and the results are shown in

Table 3. Correlation analysis.

	Top10	Lev	Cashflow	Growth	Tangible	Intangible
Top10	1
Lev	−0.094 ***	1
Cashflow	0.136 ***	−0.163 ***	1
Growth	0.004	0.005	0.000	1
Tangible	0.034 ***	0.062 ***	0.221 ***	0.004	1
Intangible	0.023 ***	0.021 ***	0.057 ***	0.006	0.069 ***	1
	Ep	GTI	Treat	Post	Did	Board	Pid	Size	Age
Ep	1
GTI	0.090 ***	1
Treat	0.009	−0.025 ***	1
Post	−0.040 ***	0.159 ***	−0.035 ***	1
Did	0.017 ***	0.019 ***	0.866 ***	0.203 ***	1
Board	0.030 ***	0.054 ***	0.102 ***	−0.103 ***	0.068 ***	1
Pid	−0.027 ***	0.023 ***	−0.047 ***	0.047 ***	−0.032 ***	−0.546 ***	1
Size	0.039 ***	0.326 ***	0.105 ***	0.098 ***	0.111 ***	0.270 ***	0.000	1
Age	0.048 ***	0.122 ***	−0.061 ***	0.113 ***	−0.039 ***	−0.134 ***	0.043 ***	−0.151 ***	1

Note: *** indicate significant correlation at the 1% levels, respectively, and values in parentheses are t-values.

. The results of the correlation analysis show that most of the explanatory variables, mediating variables, and control variables selected in this paper are significantly correlated with the explanatory variables at the 1% level, indicating that the variables in this paper are well selected for further analysis.

5.3. Multicollinearity Test

Linear regression models need to meet the premise that there is no serious multicollinearity between variables; otherwise, the regression results may be biased due to the multicollinearity problem, reducing the reliability of regression. Therefore, before regression analysis, this paper carries out the multicollinearity test to clarify the covariance situation between variables.

The results of the multicollinearity test in

Table 4. Multiple covariance test.

Variable	VIF	1/VIF
Treat	5.14	0.194428
Did	5.14	0.19461
Size	1.75	0.572384
Board	1.64	0.607939
Pid	1.48	0.67664
Lev	1.46	0.683576
Post	1.35	0.73973
Tangible	1.27	0.788533
GTI	1.19	0.842997
Cashflow	1.13	0.881891
Age	1.12	0.889332
Top10	1.09	0.91701
Intangible	1.01	0.987472
Growth	1	0.999604
Mean VIF	1.84

show that all the values of the variance inflation factor (VIF) test between variables are less than 10, indicating that there is no serious multicollinearity problem, which further develops the analysis.

5.4. Baseline Regression Analysis

Based on Model 1 constructed in this paper, the benchmark regression is determined, and the results are shown in

Table 5. Results of the baseline regression analysis.

	(1)	(2)
	Ep	Ep
Did	0.026 ***	0.020 ***
	(0.002)	(0.002)
Board		0.002
		(0.005)
Pid		−0.008
		(0.013)
Size		0.026 ***
		(0.002)
Age		0.012
		(0.001)
Top 10		0.068 ***
		(0.006)
Lev		−0.194 ***
		(0.008)
Cashflow		0.256 ***
		(0.010)
Growth		0.000 ***
		(0.000)
Tangible		−0.101 ***
		(0.006)
Intangible		−0.155 ***
		(0.017)
_cons	0.031 ***	−0.483 ***
	(0.001)	(0.038)
Firm	Yes	Yes
Year	Yes	Yes
N	33,261	33,252
R2	0.422	0.529

Note: *** denote significant correlation at the 1% levels, respectively; values in parentheses are t-values.

.

In the baseline regression in Table 5, Column (1) shows the regression results of the effect of the new Environmental Protection Act on firm performance without adding control variables, and the results show that the effect of Did on Ep is significantly positive. Column (2) shows the regression results with the inclusion of control variables, and the results show that the coefficient of the effect of Did on Ep is 0.020, which is significant at the 1% level, i.e., it shows that the implementation of the EPL leads to an increase in the level of performance of heavily polluting firms by 2%, controlling for the constancy of each of the control variables.

5.5. Robustness Tests

5.5.1. Parallel Trend Test

The premise that needs to be satisfied in order to identify the policy effect of the EPL by using the double-difference method is that there is a common trend between the treatment group and the control group before the implementation of the policy, so it is necessary to carry out the parallel trend test on the changes of the explanatory variables of the treatment group and the control group before the implementation of the policy. In this paper, we refer to the practice of Hu, J. et al. (2023) in their study to construct a dynamic double-difference model to carry out the parallel trend test [28], and the model is constructed as follows:

$$\begin{gathered} {Did}_{it}={\alpha }_0+\sum _{s=-3}^{-3}{\beta }_s^{pre}\left[{Treat}_i\times I\left(t-T_D=s\right)\right]+\sum _{s=0}^8{\beta }_s^{las}\left[{Treat}_i\times I\left(t-T_D=s\right)\right] \\ +{\gamma }_{j,i,t}Contro{l}_{j,i,t}+\sum Stkcd+\sum Year+{\epsilon }_{i,t} \end{gathered}$$

(4)

In Equation (4), ${\beta }_s^{pre}$ and ${\beta }_s^{las}$ denote the regression coefficients of the dummy variables in each period before and after the implementation of the EPL; ${Treat}_i$ is the identifying variable of whether the sample enterprises are treatment groups, and it has a value of 1 if the sample enterprises are affected by the policy during the data cycle and 0 otherwise; $I\left(·\right)$ is the schematic function; $t-T_D=s$ denotes the period before and after the implementation of the policy; and the rest of the symbols have the same meaning as in the baseline regression model. Meanwhile, referring to the base period setting method used in the studies by Chen, D.K. (2020), Jiang, L.D. et al. (2021), etc., the first period of the data cycle is used as the base period, and the test results are shown in Figure 3 [29,30].

Based on the results of the parallel trend test in Figure 1, it can be seen that the regression coefficient is close to 0 and insignificant in the period from −3 to −1 before the implementation of the EPL, indicating that before the implementation of the policy, there is no significant difference in the level of performance between the treatment group and the control group samples, i.e., the parallel trend test is passed. The regression coefficients of the current period after the implementation of the policy and the period 1–8 years after the implementation of the policy are positive, and the confidence intervals do not pass 0. This indicates that after the implementation of the policy, there is a difference between the performance level of the treatment group (enterprises in the heavy pollution industry) and the control group (enterprises in the non-heavy pollution industry), and the performance level of the enterprises in the heavy pollution industry is significantly higher than that of the non-heavy pollution enterprises, i.e., the test of the policy’s dynamic effect passes, and the performance level of the heavy pollution enterprises increases significantly after the implementation of the policy. That is, the policy dynamic effect test is passed, and after the implementation of the policy, the performance level of heavily polluting enterprises increases significantly.

5.5.2. Placebo Test

In order to test the impact of the EPL on business performance, which is generated by the implementation of the EPL, rather than formed by the impact of other external random factors, this paper adopts a placebo test to test the effect of the EPL on the performance of heavily polluting enterprises.

Referring to Hao, ZX. et al. (2024) in their study used in the random sampling of 500 times to construct pseudo-curative dummy variables [31], the placebo test, the resultant coefficients of the 500 samples, p-values, and kernel density curves are plotted in Figure 4.

As can be seen from the results of the placebo test in Figure 2, the interval of regression coefficients for the 500 random sample results are about [−0.0, 0.0], which is a large difference from the baseline regression result of 0.020. In addition, the majority of the random sample results have a significance level of greater than 0.1, which is insignificant, and the sample results basically obey the normal score centered on 0.

5.5.3. PSM-DID

In this paper, the sample proportion of the selected heavily polluting enterprises is 19.7%, the sample proportion of its control group of non-heavily polluting enterprises is 80.3%, and the sample proportion of heavily polluting enterprises in the treatment group is relatively low, which may result in the situation of sample self-selection bias, leading to the emergence of the endogeneity problem, which in turn causes bias in the regression results. Therefore, this paper adopts propensity score matching for processing and then conducts regression analysis to solve the endogeneity problem. In propensity score matching, this paper selects three types of methods for matching: mixed matching, year-by-year matching, and individual matching, respectively. Among them, mixed matching involved matching propensity scores on the full sample data; year-by-year matching involves matching propensity scores on each year’s sample year by year; and matching by individual involved matching propensity scores on the 2012–2014 sample data before the implementation of the policy by transforming the data into a wide-panel format and launching the propensity score matching, and the obtained matching result is the matching results of individual enterprises, which is an effective way to improve the mixed matching and the matching result caused by year-by-year matching, and to solve the problem of endogeneity. This method effectively improves the problem of discontinuous data of the control group enterprise samples caused by matching and year-by-year matching and provides a higher guarantee for the accuracy of DID regression after matching, and the results are shown in

Table 6. PSM-DID regression results.

	(1) Mixed Matching	(2) Matching on a Yearly Basis	(3) Individual Matching
	Ep	Ep	Ep
Did	0.0114 ***	0.0116 ***	0.00883 ***
	(0.003)	(0.002)	(0.002)
Board	0.00206	0.00850	−0.00347
	(0.007)	(0.008)	(0.008)
Pid	0.0177	0.0393 **	0.0408 **
	(0.017)	(0.018)	(0.019)
Size	0.0268 ***	0.0241 ***	0.0264 ***
	(0.003)	(0.003)	(0.003)
Top 10	0.0289 ***	0.0382 ***	0.0192 *
	(0.009)	(0.009)	(0.011)
Lev	−0.213 ***	−0.201 ***	−0.212 ***
	(0.011)	(0.010)	(0.012)
Cashflow	0.347 ***	0.343 ***	0.342 ***
	(0.017)	(0.015)	(0.019)
Growth	0.00693 ***	0.00615 ***	0.00319 ***
	(0.002)	(0.002)	(0.001)
Tangible	−0.0905 ***	−0.0810 ***	−0.0654 ***
	(0.009)	(0.009)	(0.009)
Intangible	−0.108 ***	−0.0885 ***	−0.102 ***
	(0.024)	(0.024)	(0.025)
_cons	−0.494 ***	−0.471 ***	−0.486 ***
	(0.064)	(0.060)	(0.069)
Firm	Yes	Yes	Yes
Year	Yes	Yes	Yes
N	12,310	12,385	8558
F	101.983 ***	106.886 ***	79.428 ***
R2	0.594	0.594	0.518

Note: *, **, and *** indicate significant correlation at the 10%, 5%, and 1% levels, respectively, and the values in parentheses are t-values.

.

As can be seen through the PSM-DID regression results, in the regression results after the matching of the three types of matching methods, the results show that Did has a significant positive effect on Ep; that is, it shows that after the solution of the sample self-selection bias, the effect of Did on Ep is still a significant positive effect, and the regression results in this paper have a high degree of reliability.

5.5.4. Capturing Provincial Policies, Capturing Urban Policies, and Eliminating 2020 Samples

Considering that the sample selected in this paper is a sample of A-share listed companies with a wide coverage of firms across provinces and cities, the methods of adding province and year interaction term fixed effects and city and year interaction term fixed effects are used, respectively, to capture the impact shocks of province-level and city-level policies on firms’ innovations in the heavy pollution industry during the sample period to ensure the reliability of the regression results. In addition, considering that the 2020 COVID-19 pandemic led to a large number of firms stopping work and production, which also impacted firms’ performance, the 2020 sample is further excluded from the robustness test, and the results are shown in

Table 7. Capturing provincial policies, capturing urban policies, and excluding 2020 sample robustness tests.

	(1) Capturing Provincial Policies	(2) Capturing Urban Policies	(3) Excluding the 2020 Sample
	Ep	Ep	Ep
Did	0.0192 ***	0.0189 ***	0.0197 ***
	(0.002)	(0.002)	(0.002)
Board	0.000652	0.00309	0.00103
	(0.005)	(0.006)	(0.005)
Pid	−0.00877	−0.000116	−0.00862
	(0.013)	(0.015)	(0.013)
Size	0.0255 ***	0.0256 ***	0.0248 ***
	(0.002)	(0.002)	(0.002)
Top 10	0.0651 ***	0.0663 ***	0.0691 ***
	(0.006)	(0.007)	(0.006)
Lev	−0.193 ***	−0.191 ***	−0.190 ***
	(0.008)	(0.008)	(0.008)
Cashflow	0.254 ***	0.245 ***	0.247 ***
	(0.010)	(0.011)	(0.010)
Growth	0.000000380 ***	0.000000375 ***	0.000000302 ***
	(0.000)	(0.000)	(0.000)
Tangible	−0.101 ***	−0.0964 ***	−0.0983 ***
	(0.006)	(0.007)	(0.007)
	−0.153 ***	−0.148 ***	−0.146 ***
	(0.017)	(0.019)	(0.017)
_cons	−0.473 ***	−0.488 ***	−0.463 ***
	(0.038)	(0.042)	(0.039)
Firm	Yes	Yes	Yes
Year	Yes	Yes	Yes
Province × year	Yes
City × year		Yes
N	33,260	33,260	30,206
F	224.765 ***	179.820 ***	211.938 ***
R2	0.539	0.588	0.532

Note: *** indicate significant correlation at the 1% levels, respectively, and the values in parentheses are t-values.

.

Based on the results of the robustness test in Table 7, it can be seen that in the regression results with the inclusion of the province and year interaction term fixed effects, the city and year interaction term fixed effects, and the exclusion of the 2020 sample, the results show that Did still has a significant positive effect on Ep, which further verifies the reliability of the regression results in this paper.

6. Further Analysis

6.1. Mediation Effects Test

Based on the mediation effect test model constructed in this paper, the regression analysis is launched, and the results are shown in

Table 8. Mediated effects test.

	(1)	(2)
	GTI	Ep
Did	0.237 ***	0.0193 ***
	(0.034)	(0.002)
GTI		0.000903 **
		(0.000)
Board	0.180 **	0.00138
	(0.071)	(0.005)
Pid	0.125	−0.00787
	(0.183)	(0.013)
Size	0.500 ***	0.0253 ***
	(0.017)	(0.002)
Age	0.011	0.013
	(0.001)	(0.001)
Top 10	0.316 ***	0.0678 ***
	(0.083)	(0.006)
Lev	−0.380 ***	−0.194 ***
	(0.060)	(0.008)
Cashflow	−0.194 **	0.256 ***
	(0.097)	(0.010)
Growth	0.000000566	0.000000302 ***
	(0.000)	(0.000)
Tangible	0.221 **	−0.102 ***
	(0.087)	(0.006)
Intangible	0.730 ***	−0.156 ***
	(0.213)	(0.017)
_cons	−8.939 ***	−0.475 ***
	(0.421)	(0.038)
Firm	Yes	Yes
Year	Yes	Yes
N	33,252	33,252
F	95.160 ***	213.465 ***
R2	0.798	0.529

Note: ** and *** indicate significant correlation at the 5% and 1% levels, respectively, and the values in parentheses are t-values.

.

Based on the results of the mediation effect test in Table 8, it can be seen that the regression results in Column (1) show that the effect of Did on GTI is 0.237, which is significant at the 1% level, indicating that the implementation of the EPL promotes the increase in the level of green technological innovation of heavily polluting enterprises; the results in Column (2) show that the effects of Did and GTI on Ep are both significantly positive, indicating that the mediation effect was established; thus, the implementation of the EPL will have a significant positive impact on the performance of heavy pollution enterprises by promoting enterprise green technology innovation, and hypothesis 2 was verified.

Although the coefficient value of GTI for Ep is relatively small (0.000903 **), it still holds significant economic and practical implications in the context of emerging markets. To quantify its economic impact, based on the data analysis shown in Table 2, we find that an increase of one standard deviation in GTI will raise Ep by approximately 0.00159 (1.762 × 0.000903), equivalent to 4.54% of the Ep’s mean (0.035). Given that the EPL accounts for a 2% increase in the total impact on Ep, GTI accounts for approximately 7.95% (0.00159/0.020) of this total impact. This partial mediating effect suggests that while direct regulatory pressure drives most of the short-term performance improvements, green innovation is a key channel for maintaining long-term competitiveness.

From a practical perspective, this partial mediating effect reflects the “innovation compensation” mechanism in Porter’s hypothesis as it operates in emerging markets. For heavily polluting firms, investing in green technology not only alleviates regulatory penalties but also generates cumulative efficiency improvements. The significance of this pathway lies not only in its immediate financial returns but also in its strategic transformation of compliance costs into future competitive advantages.

6.2. Heterogeneity Analysis

6.2.1. Heterogeneity in Enterprise Size

Based on the heterogeneity of enterprise size, the median enterprise size was used to classify enterprises into large-scale enterprises and small and medium-sized enterprises for heterogeneity analysis, and the results are shown in

Table 9. Heterogeneity in firm size.

	(1) Large-Scale	(2) Small and Medium-Sized
	Ep	Ep
Did	0.0177 ***	0.0170 ***
	(0.002)	(0.004)
Board	−0.00280	0.00775
	(0.005)	(0.010)
Pid	0.00824	−0.0294
	(0.013)	(0.025)
Size	0.0247 ***	0.0355 ***
	(0.002)	(0.004)
Top 10	0.0454 ***	0.0968 ***
	(0.006)	(0.013)
Lev	−0.217 ***	−0.202 ***
	(0.008)	(0.014)
Cashflow	0.237 ***	0.233 ***
	(0.011)	(0.015)
Growth	0.0000566	0.000000329 ***
	(0.000)	(0.000)
Tangible	−0.104 ***	−0.103 ***
	(0.008)	(0.011)
Intangible	−0.120 ***	−0.236 ***
	(0.019)	(0.033)
_cons	−0.438 ***	−0.699 ***
	(0.041)	(0.090)
Firm	Yes	Yes
Year	Yes	Yes
N	16,491	16,397
F	148.563 ***	92.187 ***
R2	0.610	0.532

Note: *** indicate significant correlation at the 1% levels, respectively, and the values in parentheses are t-values.

.

Based on the results of the analysis of the heterogeneity of enterprise size, it can be seen that Did has an impact coefficient of 0.0177 on the Ep of large-scale enterprises, which is significant at the 1% level, and it has a coefficient of 0.0170 on the Ep of small and medium-sized enterprises, which is significant at the 1% level. The same test of difference in coefficients between groups was used to analyze the difference in coefficients between the two groups of regression results, and the results are shown in

Table 10. Tests of coefficient differences between firm size groups.

Variables	Small–Medium Scale − Large Scale	Freq.	p-Value
Did	−0.001	56	0.00
Board	0.011	13	0.00
Pid	−0.038	94	0.06
Size	−0.011	0	0.00
Lev	0.051	0	0.00
Cashflow	−0.014	20	0.2
Growth	−0.005	64	0.36
Tangible	0.002	80	0.2
Intangible	0.001	46	0.46
_cons	−0.116	100	0.00

.

Based on the results of the intergroup coefficient difference test for enterprise size heterogeneity in Table 10, it can be seen that of the 100 random sampling results, all 100 results show that the DID regression coefficient of small and medium-sized enterprises is smaller than that of large-scale enterprises, i.e., it indicates that the impact of the EPL on the performance of large-scale heavy polluters is significantly stronger than that of the innovations of small and medium-sized heavy polluters. This situation may be due to the fact that, firstly, large-scale enterprises usually have stronger capital accumulation capacity, technological reserves, and human resource advantages [32], enabling them to integrate sustainable business practices more effectively by investing in green technology innovation and adopting advanced environmental management systems, which are critical for achieving both environmental compliance and performance improvement in emerging markets. They are able to quickly complete the upgrading of pollution control facilities or the replacement of clean technology through large-scale investment, thus internalizing the environmental compliance cost as a driving force for productivity improvement. Conversely, small and medium-sized enterprises (SMEs), with higher marginal cost sensitivity and limited by financing constraints and technology absorption capacity, often resort to end-of-pipe governance or passive production cuts—short-term strategies that are inconsistent with sustainable business practices. This highlights a key challenge in emerging markets: how to enhance SMEs’ ability to adopt innovative environmental strategies rather than relying on reactive measures, which is essential for promoting inclusive sustainable development.

Secondly, environmental regulatory agencies, subject to administrative resource constraints in emerging markets, tend to use the “catch the big and let go of the small” law enforcement logic, focusing high-frequency monitoring and high-intensity penalties on large-scale enterprises that account for a high proportion of pollutant emissions and have strong social visibility. This governance approach reflects the practical constraints of emerging markets in allocating regulatory resources, where prioritizing large enterprises can achieve greater environmental impact with limited resources. In contrast, SMEs, due to their scattered distribution and high monitoring costs, often become law enforcement blind spots, weakening the innovative forcing mechanism they face. This heterogeneity in governance effectiveness underscores the need for targeted policy designs in emerging markets to ensure that regulatory pressure drives sustainable practices across all enterprise sizes.

Thirdly, by virtue of their dominant position in the industry, large-scale enterprises can transfer environmental costs to upstream and downstream enterprises through green supply chain management, environmental certification premiums [33], and even obtain capital market favor through ESG (environmental, social, and governance) ratings. These strategies exemplify sustainable business practices that leverage innovation in governance and market mechanisms to enhance competitiveness, a model that is increasingly important for emerging markets seeking to integrate into global sustainable value chains. In contrast, SMEs with weak bargaining power in the industry chain struggle to convert green technological innovation achievements into market competitiveness, creating an asymmetric cycle of “innovation input–benefit”. This highlights the role of institutional support, such as policy incentives and capacity-building programs, in helping SMEs overcome barriers to sustainable practice adoption in emerging market environments.

6.2.2. Regional Heterogeneity

China’s large land area, the provincial administrative regions with different resources, location conditions, business environment, policies, and other differences lead to different levels of regional economic development, different leading industries, and different rates of development, which in turn affect the role of the EPL on the performance of heavily polluting enterprises. According to the National Bureau of Statistics, there are three major economic development regions in China, namely, the East, the Middle East and the West, and due to the different levels of development, the implementation of the EPL has different effects.

The regression results in

Table 11. Analysis of regional heterogeneity.

	(1) Eastern Region	(2) Midwest Region
	Ep	Ep
Did	0.0196 ***	0.0175 ***
	(0.002)	(0.003)
Board	0.00298	0.000949
	(0.006)	(0.009)
Pid	−0.00852	−0.000726
	(0.016)	(0.022)
Size	0.0274 ***	0.0229 ***
	(0.002)	(0.003)
Top 10	0.0787 ***	0.0366 ***
	(0.007)	(0.012)
Lev	−0.193 ***	−0.190 ***
	(0.009)	(0.014)
Cashflow	0.246 ***	0.273 ***
	(0.012)	(0.017)
Growth	0.0000181	0.000000283 ***
	(0.000)	(0.000)
Tangible	−0.108 ***	−0.0871 ***
	(0.008)	(0.011)
Intangible	−0.173 ***	−0.112 ***
	(0.021)	(0.026)
_cons	−0.528 ***	−0.407 ***
	(0.046)	(0.070)
Firm	Yes	Yes
Year	Yes	Yes
N	23,955	9289
F	145.907 ***	88.182 ***
R2	0.530	0.542

Note: *** indicate significant correlation at the 1% levels, respectively, and the values in parentheses are t-values.

show that the coefficient of influence of Did on the Ep of heavy polluters in the eastern region is 0.0196, which is significant at the 1% level, and the coefficient of influence on the Ep of heavy polluters in the central and western regions is 0.0175, which is also significant at the 1% level. The same test of difference in coefficients between groups was conducted, and the results are shown in

Table 12. Tests for differences in coefficients between groups for regional heterogeneity.

Variables	Non-Eastern Region − Eastern Region	Freq.	p-Value
Did	−0.002	67	0.03
Board	−0.002	58	0.02
Pid	0.008	46	0.46
Size	−0.004	75	0.25
Top 10	−0.042	100	0.00
Lev	0.002	52	0.48
Cashflow	−0.028	13	0.13
Growth	−0.001	71	0.29
Tangible	0.021	16	0.16
Intangible	−0.061	5	0.05
_cons	−0.121	15	0.00

.

Based on the test results in Table 12, showing differences in coefficients between groups for regional heterogeneity, it can be seen that in 100 samples, the results show that the influence coefficient of Did on Ep of heavily polluting enterprises in the central and western regions is smaller than that of heavily polluting enterprises in the eastern regions; i.e., it shows that there is a significant difference between the coefficients of the two groups, and that the intensity of the influence of the EPL on the performance of heavily polluting enterprises in the eastern regions is significantly higher than that of heavily polluting enterprises in the central and western regions. The reasons for this situation may be the following: Firstly, there are regional differences in the strength of the implementation of environmental regulations. In the eastern region, the level of economic development is higher, and local governments emphasize and invest relatively more in environmental governance, enforce the law more vigorously, and regulate heavily polluting enterprises more strictly. This forces heavily polluting enterprises in the eastern region to increase environmental protection investment and carry out technological transformation and upgrading in order to meet stricter environmental standards, which in turn has a greater impact on the performance of enterprises. Secondly, there are differences in the level of economic development and industrial structure [34]. The economy of the eastern region is more developed, and the scale of enterprises is larger, so they are more capable of environmental protection investment and technological upgrading. At the same time, the industrial structure of the eastern region is relatively optimized; enterprises can more easily cope with environmental regulations through industrial transfer, industrial chain upgrading, etc. In contrast, the central and western regions exhibit a relative lag in economic development, with an industrial structure of traditional high-pollution, high-energy-consumption industries, making transformation more difficult, resulting in the EPL having a relatively small impact on its performance. Thirdly, there are differences in technological innovation and financing environment. Enterprises in the eastern region have an advantage in technological innovation, and it is easier to reduce environmental costs and improve production efficiency through technological innovation. In addition, the financial market in the eastern region is more developed, and the financing channels of enterprises are more open, which makes it easier for them to obtain financial support to meet environmental protection requirements. On the other hand, the central and western regions are relatively weak in technological innovation and financing environment, which restricts the transformation and development of enterprises under the environmental regulation, and makes the EPL’s effect on the improvement of their performance not obvious.

7. Conclusions and Practical Implications

Based on the above empirical analysis results, we systematically summarize the core findings of this study, clarify its theoretical contributions and practical implications, and provide policy recommendations for environmental governance and corporate sustainability in emerging markets.

7.1. Research Conclusions

Focusing on heavy polluters in China, this study reveals the mechanism by which the EPL affects firm performance and finds that environmental governance policies enhance firm performance by driving green technological innovations and that the effects show significant differences in the scale and regional dimensions, providing new perspectives on sustainable business practices in emerging markets.

To further contextualize the findings of this study and clarify their marginal contributions, we have specifically analyzed the core findings within the context of the existing literature:

Firstly, regarding the direct impact of the EPL on corporate performance, we found that, compared to Wang, Y.Z. (2025)’s conclusion that the EPL may have a negative impact on corporate financial performance in the short term [35], this study, based on a longer time span, revealed that the impact of the EPL on corporate performance is more complex and heterogeneous, particularly in terms of long-term performance. This deepens our understanding of the complexity of the economic consequences of the EPL.

Secondly, regarding the mediating role of green technological innovation, the core contribution of this study lies in empirically testing and confirming that green technological innovation is a key mediating channel through which the EPL influences corporate performance. This significantly distinguishes our findings from Yuan, L. et al.’s research, which primarily focuses on the threshold effect of formal environmental regulation intensity on the performance of heavily polluting industries, and complements Yu, X.H. (2023)’s theoretical mention of the bridging role of green innovation, which lacked empirical support [36]. Our findings provide more direct and robust micro-level evidence for the applicability and causal pathways of the “Porter hypothesis” in the context of the EPL.

Finally, regarding the advancement of research perspectives and methods: Unlike most existing studies that focus on the macro-level [4,5,6], this paper focuses on heavily polluting enterprises, sets up control and experimental groups, and deeply analyzes the internal mechanisms through which the EPL influences performance via green innovation, as well as the differences between heavily polluting and non-heavily polluting enterprises. This provides a richer and more detailed micro-level evidence chain, thereby advancing the depth and breadth of scientific analysis in this field. Through the above comparison, we have more clearly articulated how the core findings of this study validate, supplement, or challenge existing research.

Based on the in-depth analysis from this micro-level perspective, our study confirms that strict environmental regulations significantly improve the financial performance of heavily polluting firms by forcing firms to increase their investment in green technology innovation. Green technology innovation plays a partially mediating role in this process, suggesting that the policy not only enhances firms’ environmental compliance but also translates into actual economic benefits through technology upgrading [37,38,39,40,41,42]. The identified partial mediation confirms that green technology innovation, while not the dominant short-term driver, is a structurally important pathway for converting regulatory pressure into sustainable performance gains. This highlights the necessity for emerging market policymakers to couple stringent regulations with innovation incentives to amplify this channel. It is worth noting that the effects of the policy are significantly differentiated among enterprises of different sizes and regions: large-scale enterprises are more likely to convert environmental costs into innovation incentives and even obtain additional premiums through green supply chain management and market certification by virtue of their capital reserves, technological absorptive capacity, and industrial chain bargaining advantages, while small and medium-sized enterprises are limited by financing constraints and insufficient coverage of law enforcement and have comparatively limited policy dividends. At the regional level, firms in the eastern region are more likely to improve their performance due to stronger enforcement and a mature innovation ecosystem, highlighting the critical role of government governance structures in mediating institutional environments and resource allocation. Specifically, the effectiveness of environmental regulation hinges on a well-designed governance framework that aligns regulatory stringency with regional institutional capacity and firm capabilities.

Regarding the generalizability of the research findings and their applicability to other emerging markets, we believe that while the empirical evidence in this study primarily comes from China, the core mechanisms identified provide valuable insights with potential applicability for other emerging economies facing similar challenges.

The findings of this study can provide replicable pathways for coordinating environmental protection and economic growth in the context of developing countries. Additionally, the heterogeneity observed in this study highlights the widespread challenges of uneven institutional capacity and resource allocation disparities in emerging markets. This heterogeneity is not unique to China, it resonates with the situations in economies such as India, Brazil, Vietnam, and Indonesia, which exhibit significant differences in enforcement capabilities, financing channels, technological infrastructure, and regional development levels. Therefore, the effectiveness of policies largely depends on corporate capabilities and regional institutional environments. This core conclusion may have broad applicability. Meanwhile, the intermediary role of green technology innovation provides a key lever for policymakers in emerging markets, emphasizing that regulatory design should not only impose restrictions but also actively stimulate and support innovation. However, the extent of specific impacts and the optimal design of support mechanisms necessarily require customization based on each country’s unique institutional, economic, and industrial contexts.

Therefore, this study presents a “regulation → innovation → performance” framework and highlights key enabling factors. Other emerging economies can utilize these factors to diagnose their own environments and design more effective and targeted environmental governance strategies to achieve the win–win goal of sustainability and competitiveness.

7.2. Research Limitations

Notwithstanding the potential applicability discussed above, it is crucial to explicitly acknowledge a key limitation arising from the study’s empirical foundation. The analysis relies exclusively on data from Chinese firms. While the core mechanisms identified may offer valuable insights for other emerging economies facing similar sustainability challenges, China’s unique institutional and cultural specificities, such as its distinctive regulatory evolution, enforcement patterns, state–market dynamics, and socio-cultural context, may constrain the direct generalizability of the findings to other national settings. We emphasize that these contextual factors inherent to the China sample could limit the extent to which the results are applicable in different emerging market contexts.

7.3. Practical Implications for Emerging Markets

Bearing in mind the aforementioned limitation concerning generalizability, the framework and insights derived from this study can still inform policy design in other emerging economies. Policymakers aiming to replicate the observed win–win outcomes between environmental regulation and corporate performance may consider the following implications, ensuring they are adapted to their specific national contexts:

In terms of policy design, it is crucial to implement differentiated measures that account for the heterogeneity of enterprises. For small and medium-sized enterprises, governments should employ special subsidies and innovation vouchers to reduce the threshold for green technology research and development, thereby alleviating financing constraints [42,43,44,45,46,47]. Regarding the central and western regions, enhancing the transparency of environmental law enforcement and establishing an inter-regional technology-sharing platform are essential steps to narrow the policy effect gap caused by disparities in innovation capacity. Among the various policy orientations, the hybrid model policy orientation has emerged as the most effective in promoting corporate green innovation [48,49,50,51,52,53,54,55].

The market-driven policy orientation relies on market mechanisms to incentivize corporate green innovation. For instance, carbon emission trading markets enable enterprises to trade carbon emission rights, thereby stimulating the development and adoption of energy-saving and emission reduction technologies [18,19]. However, the effectiveness of market-driven policies is constrained by factors such as market imperfections and information asymmetry. In emerging markets, where market mechanisms are underdeveloped, enterprises may lack the motivation and capacity to participate in carbon trading and other market activities. Consequently, solely relying on market-driven policies may not be sufficient to effectively promote corporate green innovation.

The state-driven policy orientation focuses on using administrative measures and direct interventions to boost corporate green innovation, such as establishing stringent environmental regulations, setting emission standards, and offering R&D subsidies [20,22]. This orientation provides a clear direction and momentum for green innovation, reducing uncertainty for enterprises. For example, R&D subsidies can mitigate the financial risks associated with green innovation, fostering the development and application of new technologies. Nevertheless, excessive reliance on state-driven policies may undermine enterprises’ autonomous innovation capabilities, leading to path dependence on government support. In emerging markets, where government resources are limited, overreliance on state-driven policies can result in uneven policy implementation and inefficient resource allocation.

The hybrid model policy orientation combines the strengths of market mechanisms and government interventions. By providing market incentives alongside appropriate government support and regulation, governments can encourage enterprises to engage in green innovation while meeting environmental standards [55]. For example, governments can introduce environmental taxes and R&D subsidies while strengthening environmental regulation. This approach helps overcome the limitations of single-policy orientations, promoting the sustainable development of green innovation. In emerging markets, the hybrid model policy orientation can fully leverage the dual roles of governments and markets, better balancing enterprises’ social responsibilities and economic benefits and fostering the sustainable development of corporate green innovation.