How Tripartite Stakeholders Promote Green Technology Innovation of China’s Heavily Polluting Enterprises?

Abstract

Green technology innovation of heavily polluting enterprises is a critical way to alleviate environmental pressure and promote sustainable development. However, the ways in which the interaction of influencing factors affects heavily polluting enterprises in green technology innovation in China have not been fully addressed and need to be investigated in this field. This paper explored the impact of government environmental regulation, corporate social responsibility (CSR), and public attention on promoting green technology innovation of heavily polluting enterprises based on S-O-R (stimulus–organism–response) and stakeholder theories. The panel data of A-share listed companies in China’s heavily polluting industries from 2008 to 2020 was used to investigate their interactions by adopting POLS (Pooled Ordinary Least Square) method. The main results show that (1) environmental regulation has a significant positive effect on green technology innovation; it especially has a great impact on state-owned heavily polluting enterprises; (2) CSR plays a mediating role between environmental regulation and green technology innovation; (3) public attention has a moderating effect between CSR and green technology innovation and also moderates the indirect effect of environmental regulation on green technology innovation through CSR. The results illustrate that green technology innovation should be not only guided by governmental regulation, but also supplemented by enterprises’ internal driven force and public supervision, which can give implications for promoting the development of green technology innovation and optimizing environmental policy tools.

1. Introduction

In recent years, with the rapid development of China’s economy, China has confronted severe threats in environmental pollution and energy consumption. According to the 2022 global Environmental Performance Index (EPI) Report released by Yale University, China, as the largest energy consumption market and carbon emission country, ranks 160th out of 180 participating countries, with a performance score of only about 36% of the first-rank country [1]. Heavily polluting enterprises as the main polluters consume the most resources and produce the most environmental waste during their daily production and operation, which severely hinder the sustainable development of China [2]. Green technology innovation is one of the practical and direct ways for heavily polluting enterprises to solve a series of environmental problems such as energy consumption and environmental pollution [3]. However, without external intervention, heavily polluting enterprises are reluctant to carry out green technology innovation. Benefiting from market scale effect and productivity advantage of non-green technology, heavily polluting enterprises prefer to enjoy productive profits using current technology rather than developing or adopting green technology due to increased cost [4]. Therefore, it is critical to encourage heavily polluting enterprises to carry out green technology innovation through multiple forces.

Environmental regulation as an external force regulates the production activities of enterprises by taking environmental pollution into account to solve the environmental pollution problems through a series of policies or measures [5]. Public attention, an informal external force, has gradually attracted people’s attention and is playing an increasing role in green technology innovation. Besides external forces, internal driving forces also contribute to heavily polluting enterprises to promote green technology innovation activities. CSR as a non-market competitive internal force, requiring enterprises to fulfill not only relevant responsibility within the industry, but also corresponding social responsibility [6], that is, to meet the needs of external stakeholders to promote green technology innovation. However, the ways in which environmental regulation, CRS, and public attention interaction affect heavily polluting enterprises in green technology innovation in China have not been fully addressed and need further investigation. Although some scholars have put the government, enterprises, and the public together to explore this topic, they use the game theory approach to make strategic choices [7,8] or study their moderating roles [9], and the mechanism of mutual influence among them is still the lack of in-depth exploration.

In this paper, an empirical study on the panel data of A-share listed companies in China’s heavily polluting industries from 2008 to 2020 using POLS (Pooled Ordinary Least Square) method was conducted to further investigate the specific transmission paths and influence mechanisms among these factors. Moreover, to clarify the role of above-mentioned tripartite stakeholders in green technology innovation and to specify each stakeholder’s interest, the S-O-R theory was used to analyze their interactions. It was first proposed by Mechrabian and Russell; it is a model of human brain information processing in the field of cognitive psychology [10]. The route of this process is stimulus–organism–response. It processes the stimulus information according to the original cognitive mechanism so as to guide the response behaviors. Enterprises as the main “economic organism” receive external stimuli (environmental regulation) and process information and resources in all aspects into advantageous ones (committing CSR to establish a good reputation and tackle public pressure), then decide to carry out green behaviors as a response. This process is basically in line with the concept of the S-O-R model, making it a suitable model to analyze the influence mechanism in environmental regulation, CSR, public attention, and green technology innovation.

The main contributions of this paper are as follows: (1) from the enterprise perspective, this paper empirically analyses the impact mechanism of environmental regulation on green technology innovation at the micro-level by using the panel data of A-share listed companies in China’s heavily polluting industries; (2) based on the S-O-R theory, this paper discusses the role of CSR in environmental regulation and green technology innovation to probe conduction paths and specific conditions which may influence the relationship between environmental regulation and green technology innovation; (3) this paper also probes the effect of public attention in this process, which can enrich the theoretical basis; (4) the research results can provide theoretical guidance and practical advice for promoting the development of green technology innovation, optimizing environmental policy tools, and providing references for other developing countries.

The remaining parts of this paper are arranged as follows: Section 2 is the literature review regarding the relationship between environmental regulation, CSR, public attention and green technology innovation, respectively. Section 3 proposes research hypotheses and the model based on the literature review and theoretical analysis. Section 4 provides research design, including samples and data sources, variable selection, and model design. The empirical analyses are arranged in Section 5, including descriptive statistics, correlation analyses, regression analyses, robustness tests, and further analyses. The discussion is in Section 6, where the empirical results are explored. Section 7 provides research conclusions and policy implications. The explanation for the limitation of research is arranged in Section 8.

2. Literature Review

Extensive literature has studied the influence of environmental regulation on green technology innovation and resulted in multiple theories. The aim of environmental regulation is to internalize the environmental negative externality caused by the production activities of enterprises into their daily operation process and force them to carry out green technology innovation [11]. Based on the “Porter Hypotheses” [12], some scholars confirm that strict and reasonable environmental regulation has a positive and primary impact on green technology innovation [13,14]. The more stringent the environmental regulation is, the more incentive enterprises can obtain to carry out green technology innovation activities, such as implementing pollution control, increasing R&D investment, and improving green innovation abilities [15,16]. This may bring about an innovation compensation effect, as long as the innovation compensation effect is large enough; it can not only offset the extra costs due to compliance with the environmental regulation, but also promote green technology innovation, economic benefits, and social benefits [17,18,19]. Others hold different views. Preston et al. argued that environmental regulation leads to a decreasing trend in technology outputs which are not favorable to green technology innovation of enterprises [20]. Research by Triebswetter et al. supported the notion that environmental regulation is just one of the driving factors for enterprises to innovate [21]. Guo et al. proved that there is an inverted U-shaped relationship between environmental regulation and green technology innovation [22]. The inconsistency of findings suggests that there are other factors that influence the environmental regulation–green technology innovation process. Although studies have explored the moderating role of possible influencing factors such as the aspects of financial [23,24] and industrial agglomeration [25,26], relatively little exploration of mediating factors has been conducted. The transmission path from environmental regulation to green technology innovation is worth exploring.

CSR as an important factor in environmental protection has attracted widespread discussion in the academic field. Bocquet et al. discussed the fact that the impact of a combined strategy of innovation and CSR can generate positive firm performance [27]. Ruggier and Cupertino demonstrated that innovation is a critical factor in the relationship between financial performance and corporate social performance [28]. Liu et al. explored a combined effect of CSR and innovation on financial risk and discussed the fact that a higher level of innovation generates greater financial risk [29]. Albort et al. and Gu et al. believe that CSR can help enterprises’ technology innovation by improving their financing capacity, delivering effective information to investors, and reducing the cost for investors to obtain investment information based on the signal transmission theory and the information asymmetry theory [30,31]. In contrast, other researchers found that enterprises’ investment in CSR may reduce enterprises’ focus on technology innovation activities and consume resources that should be used for technological innovation [32,33]. However, these discussions mainly concentrate on the influence of CSR on enterprises’ economic value, and less so on the environmental value. In addition, few types of research unpack the interrelation between environmental regulation, CSR, and green technology innovation. As an important conductor in the process of environmental regulation and green technology innovation, CRS’s role in conduction has not been discussed in depth. In addition, relatively few studies have been explored on the CSR of heavily polluting enterprises.

From an informal perspective, public attention has gradually attracted the interestof scholars to its role in green technology innovation in heavily polluting enterprises. In fact, the public is the direct victim of environmental pollution, and their opinion and emotions can, to a certain extent, influence the actions of enterprises and governments [34,35]. Public attention is also called the “soft tool” of regulation, and it has the unique advantage of compensating for “government failure” and “market failure” [36]. Unexpectedly, public attention is a double-edged sword for green technology innovation [37]. Scholars who hold a positive attitude believe that public attention can promote green technology innovation in heavily polluting enterprises and urge them to solve environmental issues [38,39,40]. Others believe that too much public attention has a negative impact on green technology innovation [41]. Pan et al. found that over public attention causes “green-wash” behaviors (green false publicity, emission fraud, etc.) to offset the public opinion pressure, which inhibits green technology innovation [42]. Due to the late start, the development of the system related to public attention in environmental protection in China is still immature. Although scholars have explored this topic from different aspects, studies on the impact of public attention on green technology innovation are still relatively scarce when compared with related research on government regulation. In addition, most related literature focuses on the impact of public attention on government environmental governance and pollution reduction, and few probe its impact on enterprise behavior, especially green technology innovation of heavily polluting enterprises.

In summary, the promotion of green technology innovation is not determined by a single factor; instead, it is determined by the joint efforts of multiple stakeholders. According to the stakeholder theory [43], the stakeholders of green technology innovation refer to all relevant individuals or groups that can affect the achievement of green technology innovation goals. Therefore, specifying the role of tripartite stakeholders in green technology innovation and clarifying each stakeholder’s interest play key roles in promoting green technology innovation. The influential mechanism and specific conduction paths have not been discussed in depth yet. In addition, studies generally refer to all enterprises, and the specific impact mechanism and influencing conditions of heavily polluting enterprises have not been discussed extensively and deeply. Therefore, this study aims to fill these gaps by researching the relationship between tripartite stakeholders and green technology innovation and offer more specific suggestions for China’s heavily polluting enterprises’ green technology innovation.

3. Hypotheses and Model

3.1. The Impact of Environmental Regulation on Green Technology Innovation in China’s Heavily Polluting Enterprises

In China, the use of non-green technology has been a traditional productive means of heavily polluting enterprises for a long time. Along with the development of continuous expansion of heavily polluting enterprises, the disadvantage of traditional technology began to emerge. It neither follows the government’s orientation nor meets the green needs of the public. Green technology innovation has become the new way out to make competitive compensation to the costs and bring high profit in the long run [44]. Although some scholars believe that environmental regulation will increase the cost of human and material resources and reduce short-term profit growth [45], the economic growth brought by this innovation compensation can partially or completely offset these costs in the long run [12]. In addition to the mandatory regulation of the government, heavily polluting enterprises that respond to the green requirements can easily obtain government support, such as tax deductions, exemptions, and quick access to bank loans. All these factors can ensure the smooth progress of green technology innovation [46]. Based on the above, Hypothesis 1 was proposed:

Hypothesis 1 (H1). 

Environmental regulation has a positive impact on green technology innovation of heavily polluting enterprises.

3.2. The Mediating Role of Corporate Social Responsibility (CSR)

Because of environmental negative externality, heavily polluting enterprises obtain economic benefits at the expense of sacrificing environmental interests [47,48]. It is difficult for them to actively undertake CSR and disclose the relevant information without external stimulus. Kesidou and Demirel [49] reported that strict environmental laws and regulations will affect the environmental responsibility of enterprises, and then trigger the ecological innovation of enterprises. As an internal driving force of green technology innovation activities, CSR can help heavily polluting enterprises to combine economic benefits with environmental protection while ensuring the benign development of ecological environment [50,51]. This illustrates that CSR is not only important feedback on environmental regulation, but also the strategic requirement of green technology innovation. With the mandatory requirements of environmental regulation, heavily polluting enterprises will increasingly integrate the concept of actively undertaking social responsibility into enterprise strategy which will help them to form a new innovation mechanism and then promote green technology innovation. Based on the above, Hypothesis 2 was proposed:

Hypothesis 2 (H2). 

CSR has a mediating effect on environmental regulation and green technology innovation of heavily polluting enterprises.

3.3. The Moderating Role of Public Attention

As an informal institutional environment [52], public attention can urge enterprises to take positive and effective environmental rectification measures [53,54,55]. In order to establish a good enterprise image and safeguard the rights and interests of all stakeholders, heavily polluting enterprises will undertake social responsibility and carry out green technology innovation to obtain public trust and increase organizational competitive advantage [56]. Meanwhile, once they become victims of environmental pollution, the public will report to the environmental protection departments, resulting in changes in their purchase behavior decisions such as boycotting their products and also a large number of fines for enterprises. All these factors will eventually affect the final profits of enterprises [57,58]. To maintain sustained profitability, heavily polluting enterprises will cater to the needs of the public to take responsibility for the environment, such as carrying out green technology innovation and producing green products. These analyses reveal that public attention can impose pressure on the daily production and operation activities of heavily polluting enterprises, and supervise them to better commit CSR and carry out green technology innovation. Based on the above, Hypothesis 3 was proposed:

Hypothesis 3 (H3). 

Public attention plays a moderating role between CSR and green technology innovation of heavily polluting enterprises.

The effective implementation of environmental regulation is inseparable from the external supervision [59]. Public attention can not only influence the dynamic process of policy-making, but also contribute to government environmental management [60,61]. Therefore, in addition to the above arguments, the mediating role of CSR on environmental regulation and green technology innovation will also be affected by public attention. In other words, when the level of public attention becomes high, the indirect impact of environmental regulation on green technology innovation through CSR will be strengthened; conversely, when the level of public attention will lower, environmental regulation’s indirect impact through CSR on green technology innovation will be weakened. Therefore, a moderated mediating model has been constructed and Hypothesis 4 was put forward:

Hypothesis 4 (H4). 

Public attention moderates the indirect effect of environmental regulation on green technology innovation through CSR, that is, the indirect effect of high-level public attention on environmental regulation through CSR is stronger than that of low-level public attention.

Through above theoretical analyses, environmental regulation can stimulate heavily polluting enterprises to carry out green technology innovation, while CSR and public attention also have an important impact on green technology innovation. In order to more clearly show the relationship between environmental regulation, CSR, public attention, and green technology innovation, a modified conceptual model based on the S-O-R model was constructed (Figure 1).

4. Research Design

4.1. Samples and Data Sources

In this paper, the panel data of 680 A-share listed companies in heavily polluting industries of China’s 30 provinces from 2008 to 2020 were selected for empirical research (some seriously missing data have been eliminated). The data were collected from the Rankins CSR Ratings (RKS) Database, China Stock Market and Accounting Research (CSMAR) Database, Chinese Research Data Services (CNRDS) Database, China Patent Database of State Intellectual Property Office, China Urban Statistical Yearbook, China Industrial Statistical Yearbook, and China Environmental Statistical Yearbook.

4.2. Variable Selection

4.2.1. Independent Variable—ER

Based on Yuan et al.’s measurement method [62] of environmental regulation, industrial SO2 removal rate, industrial COD removal rate, comprehensive utilization rate of industrial solid waste, domestic sewage treatment rate, and harmless treatment of domestic waste were selected as main indicators. The score of environmental regulation was calculated using the entropy weight method, which is employed as the judgment indicator of the intensity of environmental regulation. The indicators are acquired from China Urban Statistical Yearbook, China Industrial Statistical Yearbook and China Environmental Statistical Yearbook.

4.2.2. Dependent Variable—GTI

For the measurement of green technology innovation, the widely used indicators include the productivity of scientific research activities, R&D investment, or the number of technology patents of enterprises to replace [63,64]. To better represent the results of green innovation activities, the adopted number of green technology patents was used instead of an overall number of technology patents. The data were acquired from the China Patent Database of State Intellectual Property Office.

4.2.3. Mediating Variable—CSR

Considering the availability and scientific accuracy of data, the social responsibility rating score from the RKS Database as the measurement index was used [65]. The rating data of the RKS Database comes from the information and data actively disclosed by the rated enterprise and the information and data disclosed by other authoritative websites. RKS adopts the ESG (environmental, social and governance) rating standard independently developed for analysis and produces the conclusion. The higher the score, the better the CSR performance of an enterprise.

4.2.4. Moderating Variable—Media Coverage (MC)

The approaches of measuring public attention are diverse and complex. Timothy and Violina [66] pointed out that there is a correlation between the number and attitude of the media to the enterprises’ news reports and the public’s attention and evaluation of enterprises. Therefore, this study chose MC as the alternative variable for public attention. Considering the availability and scientific accuracy of data, the number of company-related news from the CNRDS Database as the measurement index was used [67,68].

4.2.5. Control Variables

In addition to the above variables, other related enterprise-level data from the CSMAR Database were selected as control variables, which better correlates with China’s context [62,64,69]. They are market-to-book (MTB), financial leverage (LEV), return on assets (ROA), largest holder rate (Top1), TobinQ, enterprise age (Age), cash flow (Cflow), year, industry and province fixed effect.

4.3. Model Design

To examine the relationship between environmental regulation, CSR, green technology innovation and MC, the following models were estimated:

$$\begin{gathered} {\mathrm{G}\mathrm{T}\mathrm{I}}_1={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1\mathrm{E}\mathrm{R}+{\mathsf{\alpha }}_2\mathrm{M}\mathrm{T}\mathrm{B}+{\mathsf{\alpha }}_3\mathrm{L}\mathrm{E}\mathrm{V}+{\mathsf{\alpha }}_4\mathrm{R}\mathrm{O}\mathrm{A}+{\mathsf{\alpha }}_5\mathrm{T}\mathrm{o}\mathrm{p}1+{\mathsf{\alpha }}_6\mathrm{T}\mathrm{o}\mathrm{b}\mathrm{i}\mathrm{n}\mathrm{Q}+{\mathsf{\alpha }}_7\mathrm{A}\mathrm{g}\mathrm{e}+{\mathsf{\alpha }}_8\mathrm{C}\mathrm{f}\mathrm{l}\mathrm{o}\mathrm{w} \\ +\sum \mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}+\sum \mathrm{I}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{y}+\sum \mathrm{P}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{e}+{\mathsf{\epsilon }}_1+{\mathsf{\mu }}_1, \end{gathered}$$

(1)

$$\begin{gathered} \mathrm{C}\mathrm{S}\mathrm{R}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{E}\mathrm{R}+{\mathsf{\beta }}_2\mathrm{M}\mathrm{T}\mathrm{B}+{\mathsf{\beta }}_3\mathrm{L}\mathrm{E}\mathrm{V}+{\mathsf{\beta }}_4\mathrm{R}\mathrm{O}\mathrm{A}+{\mathsf{\beta }}_5\mathrm{T}\mathrm{o}\mathrm{p}1+{\mathsf{\beta }}_6\mathrm{T}\mathrm{o}\mathrm{b}\mathrm{i}\mathrm{n}\mathrm{Q}+{\mathsf{\beta }}_7\mathrm{A}\mathrm{g}\mathrm{e}+{\mathsf{\beta }}_8\mathrm{C}\mathrm{f}\mathrm{l}\mathrm{o}\mathrm{w} \\ +\sum \mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}+\sum \mathrm{I}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{y}+\sum \mathrm{P}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{e}+{\mathsf{\epsilon }}_2+{\mathsf{\mu }}_2, \end{gathered}$$

(2)

$$\begin{gathered} {\mathrm{G}\mathrm{T}\mathrm{I}}_2={\mathsf{\gamma }}_0+{\mathsf{\gamma }}_1\mathrm{C}\mathrm{S}\mathrm{R}+{\mathsf{\gamma }}_2\mathrm{M}\mathrm{T}\mathrm{B}+{\mathsf{\gamma }}_3\mathrm{L}\mathrm{E}\mathrm{V}+{\mathsf{\gamma }}_4\mathrm{R}\mathrm{O}\mathrm{A}+{\mathsf{\gamma }}_5\mathrm{T}\mathrm{o}\mathrm{p}1+{\mathsf{\gamma }}_6\mathrm{T}\mathrm{o}\mathrm{b}\mathrm{i}\mathrm{n}\mathrm{Q}+{\mathsf{\gamma }}_7\mathrm{A}\mathrm{g}\mathrm{e}+{\mathsf{\gamma }}_8\mathrm{C}\mathrm{f}\mathrm{l}\mathrm{o}\mathrm{w} \\ +\sum \mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}+\sum \mathrm{I}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{y}+\sum \mathrm{P}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{e}+{\mathsf{\epsilon }}_3+{\mathsf{\mu }}_3, \end{gathered}$$

(3)

$$\begin{gathered} {\mathrm{G}\mathrm{T}\mathrm{I}}_3={\mathsf{\delta }}_0+{\mathsf{\delta }}_1\mathrm{E}\mathrm{R}+{\mathsf{\delta }}_2\mathrm{C}\mathrm{S}\mathrm{R}+{\mathsf{\delta }}_3\mathrm{M}\mathrm{T}\mathrm{B}+{\mathsf{\delta }}_4\mathrm{L}\mathrm{E}\mathrm{V}+{\mathsf{\delta }}_5\mathrm{R}\mathrm{O}\mathrm{A}+{\mathsf{\delta }}_6\mathrm{T}\mathrm{o}\mathrm{p}1+{\mathsf{\delta }}_7\mathrm{T}\mathrm{o}\mathrm{b}\mathrm{i}\mathrm{n}\mathrm{Q}+{\mathsf{\delta }}_8\mathrm{A}\mathrm{g}\mathrm{e} \\ +{\mathsf{\delta }}_9\mathrm{C}\mathrm{f}\mathrm{l}\mathrm{o}\mathrm{w}+\sum \mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}+\sum \mathrm{I}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{y}+\sum \mathrm{P}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{e}+{\mathsf{\epsilon }}_4+{\mathsf{\mu }}_4, \end{gathered}$$

(4)

$$\begin{gathered} {\mathrm{G}\mathrm{T}\mathrm{I}}_4={\mathsf{\lambda }}_0+{\mathsf{\lambda }}_1\mathrm{C}\mathrm{S}\mathrm{R}+{\mathsf{\lambda }}_2\mathrm{M}\mathrm{C}+{\mathsf{\lambda }}_3\mathrm{C}\mathrm{S}\mathrm{R}\ast \mathrm{M}\mathrm{C}+{\mathsf{\lambda }}_4\mathrm{M}\mathrm{T}\mathrm{B}+{\mathsf{\lambda }}_5\mathrm{L}\mathrm{E}\mathrm{V}+{\mathsf{\lambda }}_6\mathrm{R}\mathrm{O}\mathrm{A}+{\mathsf{\lambda }}_7\mathrm{T}\mathrm{o}\mathrm{p}1+{\mathsf{\lambda }}_8\mathrm{T}\mathrm{o}\mathrm{b}\mathrm{i}\mathrm{n}\mathrm{Q} \\ +{\mathsf{\lambda }}_9\mathrm{A}\mathrm{g}\mathrm{e}+{\mathsf{\lambda }}_{10}\mathrm{C}\mathrm{f}\mathrm{l}\mathrm{o}\mathrm{w}+\sum \mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}+\sum \mathrm{I}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{y}+\sum \mathrm{P}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{e}+{\mathsf{\epsilon }}_5+{\mathsf{\mu }}_5. \end{gathered}$$

(5)

ER is environmental regulation, GTI is green technology innovation, CSR is corporate social responsibility, MC is media coverage (alternative variable of PA); MTB, LEV, ROA, Top1, TobinQ, Age, Cflowyear, industry and province fixed effect are control variables, α, β, γ, δ, λ are coefficients, ε is the error term, μ is the individual effect. Model (1) is used to test Hypothesis 1; Models (2)–(4) are used to test Hypothesis 2; Model (5) is used to test Hypothesis 3.

5. Empirical Results

5.1. Descriptive Statistics

The descriptive statistics of 11 variables are presented in

Table 1. Descriptive Statistics of Each Variable.

Variable	Variable Name	Obs	Mean	Std. Dev.	Min	Max
ER	Environment Regulation	680	0.847	0.120	0.338	0.997
GTI	Green Technology Innovation	680	16.319	43.177	0	440
CSR	Corporate Social Responsibility	680	40.368	13.198	18.340	85.499
MC	Media Coverage	680	179.985	293.380	0	1867
MTB	Market to Book	680	2.632	1.826	0.152	10.710
LEV	Financial Leverage	680	1.567	1.022	0.355	8.882
ROA	Return On Assets	680	0.053	0.045	−0.009	0.292
Top1	Largest Holder Rate	680	39.429	15.500	6.410	82.510
Tobin Q	TobinQ	680	1.749	0.949	0.805	8.002
Age	Enterprise Age	680	10.821	5.777	0	25
Cflow	Cash Flow	680	0.069	0.070	−0.470	0.347

and Figure 2. They demonstrate the mean, standard deviation, minimum and maximum values of independent variable environmental regulation, dependent variable GTI, mediating variable CSR, moderating variable media coverage (alternative variable of public attention), and control variables. A logarithm was used to transform the GTI, CSR, and MC data.

5.2. Correlation Analyses

The Pearson correlation matrix and Variance inflation factor test are shown in

Table 2. Pearson Correlation Matrix and Variance Inflation Factor.

	GTI	ER	CSR	MC	MTB	LEV	ROA	Top1	TobinQ	Age	C_Flow	VIF
GTI	1.000											-
ER	0.138 ***	1.000										1.04
CSR	0.525 ***	0.135 ***	1.000									1.31
MC	0.268 ***	0.069 *	0.178 ***	1.000								1.12
MTB	−0.334 ***	−0.060	−0.359 ***	0.007	1.000							4.23
LEV	0.113 ***	0.057	0.028	−0.132 ***	−0.136 ***	1.000						1.34
ROA	−0.155 ***	0.026	−0.103 ***	0.199 ***	0.358 ***	−0.463 ***	1.000					2.01
Top1	0.311 ***	−0.017	0.313 ***	0.174 ***	−0.179 ***	0.087 **	−0.058	1.000				1.15
TobinQ	−0.347 ***	−0.025	−0.320 ***	−0.006	0.863 ***	−0.192 ***	0.403 ***	−0.162 ***	1.000			4.30
Age	−0.076 ***	−0.031	−0.087 **	−0.080 **	0.013	0.106 ***	−0.011	0.021	0.102 ***	1.000		1.07
Cflow	−0.052	0.065 *	0.020	0.071 *	0.113 ***	−0.167 ***	0.516 ***	0.000	0.162 ***	0.069 *	1.000	1.41

Note: ***, **, * represent the significance at the level of 1%, 5%, 10%, respectively.

. It reflects that there is a significant correlation between the observed variables. The absolute values of all correlation coefficients are basically lower than 0.5 (5% statistical error allowed), which illustrates that there is no multi-collinearity problem in the empirical model. In addition, the VIF values are far smaller than 10, further indicating that there is no multi-collinearity problem in this research. These ensure the observed variables can effectively explain the dependent variable and also support the reliability and accuracy of empirical analyses.

5.3. Regression Analyses

Based on the models built in Section 4.3, this article first empirically examines the impact of environmental regulation on green technology innovation in the sample of the A-share listed of China’s heavily polluting enterprises. In this part, Pooled Ordinary Least Square (POLS) was used to estimate Model (1), which was built by way of Equation (1). In Model (1), year, industry and province fixed effects are controlled, as well as related variables. According to the regression results shown in

Table 3. The Results of Direct Effect.

	Model (0)	Model (1)
Variable	GTI	GTI
ER		1.706 *** (3.70)
MTB	−0.105 * (−1.71)	−0.084 (−1.38)
LEV	0.069 (1.11)	0.049 (0.79)
ROA	−0.213 (−0.12)	−0.476 (−0.28)
Top1	0.027 *** (7.27)	0.027 *** (7.46)
Tobin Q	−0.314 ** (−2.60)	−0.339 *** (−2.83)
Age	−0.018 * (−1.85)	−0.016 * (−1.67)
Cflow	0.143 (0.15)	−0.024 (−0.03)
Year, Industry, Province FE	Yes	Yes
Constant	1.135 *** (4.64)	−0.300 (−0.65)
N	680	680
R2	0.195	0.211
F	23.21	22.40

Note: *** p < 0.01, ** p < 0.05, * p < 0.1. The “t” value is in parentheses.

, Model (1) explored the impact of environmental regulations (ER) on green technology innovation (GTI); the coefficient is positive and statistically significant at 1% (1.706 ***). This indicates that environmental regulation has a positive impact on green technology innovation; thus, H1 was verified.

Based on the previous analysis, green technology innovation of heavily polluting enterprises may also be influenced by CSR and public attention. Thus, in this part, the mechanisms by which CSR and public attention influence green technology innovation were explored separately.

Stepwise regression analysis [70] was used to test Equations (2)–(4) aiming to find out the relationship among environmental regulation, CSR, and green technology innovation. In this process, this study focuses on the changes in direct effect that occurred after the addition of CSR and public attention (PA), respectively, and the regression results are shown in

Table 4. The Results of Mediating Effect and Moderating Effect.

	Model (2)	Model (3)	Model (4)	Model (5)
Variable	CSR	GTI	GTI	GTI
ER	0.302 *** (3.41)		1.095 ** (2.55)
CSR		2.081 *** (11.31)	2.020 *** (10.93)	1.705 *** (9.33)
MC				0.288 *** (8.28)
CSR × MC				0.447 *** (3.90)
MTB	−0.052 *** (−4.44)	0.012 (0.20)	0.021 (0.37)	−0.028 (−0.51)
LEV	−0.012 (−0.97)	0.086 (1.50)	0.072 (1.26)	0.102 * (1.89)
ROA	−0.322 (−0.99)	0.360 (0.23)	0.175 (0.11)	−1.364 * (−0.91)
Top1	0.005 *** (6.53)	0.017 *** (5.00)	0.018 *** (5.17)	0.010 *** (3.02)
TobinQ	−0.003 (−0.15)	−0.316 *** (−2.86)	−0.332 *** (−3.01)	−0.243 ** (−2.32)
Age	−0.004 ** (−2.01)	−0.010 (−1.07)	−0.009 (−0.96)	0.000 (0.03)
Cflow	0.297 * (1.66)	−0.537 (−0.63)	−0.624 (−0.73)	−0.233 (−0.29)
Year, Industry, Province FE	Yes	Yes	Yes	Yes
Constant	3.412 *** (38.73)	−6.497 *** (−9.14)	−7.193 *** (−9.48)	−6.168 *** (−8.85)
N	680	680	680	680
R2	0.210	0.324	0.330	0.401
F	22.25	40.14	36.70	44.68

Note: *** p < 0.01, ** p < 0.05, * p < 0.1. The “t” value is in parentheses.

. First, Model (2) showed that environmental regulation has a positive impact on CSR, the coefficient is positive and statistically significant at 1% (0.302 ***). Second, Model (3) showed that CSR has a positive impact on green technology innovation, the coefficient is positive and statistically significant at 1% (2.081 ***). Third, Model (4) showed that the impact of environmental regulations on green technology innovation weakened after the addition of CSR. From Model (1), it can be seen that environmental regulation has a significant positive impact on green technology innovation with a coefficient of 1.706 and statistical significance at 1%, while after the addition of CSR, the effect of environmental regulation on green technology is weakened with a coefficient of 1.095 and statistical significance at 5%. This indicates that CSR plays a partial mediating role between environmental regulation and green technology innovation (1.095 ** < 1.706 ***); thus, H2 was verified. The detailed results of mediating effect process are demonstrated in Figure 3.

Cohen’s (1983) method [71] was used to test Equation (5). An interaction term between CSR and media coverage (MC) was established to test the impact of public attention, and the regression results are shown in Table 4. Model (5) showed that the coefficient of CSR × MC is positive and statistically significant at 1% (CSR × MC = 0.447 ***), which illustrates that public attention moderates the relationship between CSR and green technology innovation; thus, H3 was verified. Figure 4 shows that compared with low-level media coverage, high-level media coverage can enhance the moderating effect of CSR on green technology innovation to a greater extent, which indicates that the more public attention there is, the more heavily polluting enterprises are motivated to act CSR and carry out green technology innovation.

In order to further test the moderated mediating effect, this study followed the method of [72,73,74] for reference. According to the views of the authors, four conditions should be met to test the existence of the moderated mediating effect. (1) Regressing GTI to ER and MC to test whether the effect of ER on GTI is significant. (2) Regressing CSR to ER and MC to test whether the effect of ER on CSR is significant. (3) Regressing GTI to ER, MC and CSR to test the mediating effect of CSR between ER and GTI is significant. (4) Regressing GTI to ER, MC, CSR and CSR × MC to test whether CSR × MC is significant. The corresponding results are shown in Models (6)–(9) in

Table 5. The Results of Moderated mediating Effect.

	Model (6)	Model (7)	Model (8)	Model (9)
Variable	GTI	CSR	GTI	GTI
ER	1.700 *** (3.89)	0.302 *** (3.42)	1.141 *** (2.79)	1.153 *** (2.85)
CSR			1.851 *** (10.46)	1.638 *** (8.94)
MC	0.335 *** (8.96)	0.022 *** (2.94)	0.294 *** (8.41)	0.289 *** (8.36)
CSR × MC				0.449 *** (3.94)
MTB	−0.134 ** (−2.32)	−0.056 *** (−4.72)	−0.032 (−0.58)	−0.018 (−0.33)
LEV	0.054 (0.92)	−0.011 (−0.94)	0.075 (1.37)	0.088 (1.63)
ROA	−2.521 (−1.56)	−0.459 (−1.40)	−1.671 (−1.11)	−1.567 (−1.05)
Top1	0.020 *** (5.67)	0.004 *** (5.75)	0.012 *** (3.70)	0.011 *** (3.20)
TobinQ	−0.251 ** (−2.20)	0.003 (0.11)	−0.255 ** (−2.42)	−0.260 ** (−2.49)
Age	−0.010 (−1.12)	−0.003 * (−1.80)	−0.004 (−0.47)	0.001 (0.16)
Cflow	0.350 (0.40)	0.322 * (1.81)	−0.246 (−0.30)	−0.324 (−0.40)
Year, Industry, Province FE	Yes	Yes	Yes	Yes
Constant	−1.439 *** (−3.19)	3.336 *** (36.48)	−7.615 *** (−10.52)	−6.898 *** (−9.33)
N	680	680	680	680
R2	0.295	0.220	0.394	0.408
F	31.12	20.93	43.48	41.79

Note: *** p < 0.01, ** p < 0.05, * p < 0.1. The “t” value is in parentheses.

. The results show that (1) in Model (6), the effect of ER on GTI is significant (1.700 ***); (2) in Model (7), the effect of ER on CSR is significant (0.302 ***); (3) in Model (8), the mediating effect of CSR between ER and GTI is significant (1.851 ***); (4) in Model (9), CSR × MC is significant (0.449 ***); all of the above four steps were tested. In summary, public attention moderates not only the relationship between CSR and green technology innovation, but also the indirect effect of environmental regulation on green technology innovation through CSR; thus, H4 was verified.

5.4. Robustness Analyses

In order to ensure the accuracy and robustness of empirical results, this study further tests the results by replacing the variable and the research method. In view of the different algorithms, this study introduces other variables to calculate the green input and output so as to measure the efficiency of green technology innovation efficiency, and also matches and processes related empirical data. (1) Replacement of the dependent variable: as green technology innovation efficiency is another key factor to test green technology innovation, it is used as the alternative variable of the number of green technology patents. By using the DEA method to calculate green technology innovation efficiency (GTIE), GTI is replaced with GTIE and tested again. (2) Replacement of the research method: Poisson Regression is adopted for another test. Except for very few values that are not significant, the robust test results are basically consistent with empirical results. The main results are shown in

Table 6. The Robust Test Results of Main Data by Replacing the Variable.

	Model (10)	Model (11)	Model (12)	Model (13)	Model (14)	Model (15)	Model (16)
Variable	GTIE	GTIE	GTIE	GTIE	CSR	GTIE	GTIE
ER	0.238 *** (2.67)	0.160 * (1.79)		0.215 ** (2.41)	0.526 *** (4.91)	0.136 (1.53)	0.139 (1.57)
CSR		0.149 *** (4.81)	0.145 *** (4.77)			0.150 *** (4.85)	0.137 (4.41)
MC			0.118 *** (3.09)	0.115 *** (2.93)	−0.004 (−0.08)	0.116 *** (2.99)	0.113 *** (2.94)
CSR × MC			0.319 *** (2.77)				0.321 *** (2.80)
Control Variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year, Industry, Province	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	680	680	680	680	680	680	680
R2	0.065	0.094	0.112	0.076	0.154	0.106	0.115
F	6.09	8.16	8.89	6.39	14.22	8.29	8.32

Note: *** p < 0.01, ** p < 0.05, * p < 0.1. The “t” value is in parentheses.

and

Table 7. The Robust Test Results of Main Data by Replacing the Research Method.

	Model (17)	Model (18)	Model (19)	Model (20)	Model (21)	Model (22)	Model (23)
Variable	GTI	GTI	GTI	GTI	CSR	GTI	GTI
ER	3.207 *** (0.117)	2.485 *** (0.122)		3.796 *** (0.136)	0.323 *** (0.053)	3.036 *** (0.135)	2.989 *** (0.135)
CSR		2.241 *** (0.038)	2.247 *** (0.048)			1.818 *** (0.039)	2.084 *** (0.047)
MC			0.306 *** (0.010)	0.362 *** (0.007)	0.033 *** (0.004)	0.260 *** (0.008)	0.311 *** (0.010)
CSR × MC			0.285 *** (0.026)				0.249 *** (0.026)
Control Variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year, Industry, Province	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	680	680	680	680	680	680	680

Note: *** p < 0.01. The “standard error” value is in parentheses.

.

5.5. Further Analyses

Different types of heavily polluting enterprises may be affected by environmental regulation and public attention to a different extent. In order to further verify the type of heavily polluting enterprises that is more affected by environmental regulation and public attention, a dummy variable SOE was generated as enterprise type (SOE = 1 means state-owned enterprises, SOE = 0 means non-state-owned enterprises), and then the sign of ER × SOE and CSR × MC × SOE coefficients was tested. If the coefficient is a plus sign, it illustrates that state-owned heavily polluting enterprises are affected more than non-state-owned heavily polluting enterprises. If the coefficient is a minus sign, it illustrates that non-state-owned heavily polluting enterprises are affected more than state-owned heavily polluting enterprises. The results are shown in

Table 8. The Further Test Results of Different Type of Heavily Polluting Enterprises.

	Model (24)	Model (25)
Variable	GTI	GTI
ER	1.010 * (1.74)
CSR	2.157 *** (9.73)	2.155 *** (8.40)
MC		0.267 *** (6.12)
ER × SOE	0.505 *** (2.68)
CSR × MC × SOE		−1.187 ** (−2.41)
Control Variables	Yes	Yes
Year, Industry, Province	Yes	Yes
N	680	680
R2	0.368	0.408
F	24.36	26.13

Note: *** p < 0.01, ** p < 0.05, * p < 0.1. The “t” value is in parentheses.

. Model (24) showed that the coefficient of ER × SOE is positive and statistically significant at 1% (0.505 ***), which indicates that environmental regulation has a stronger impact on state-owned heavily polluting enterprises than on non-state-owned heavily polluting enterprises. Model (25) showed that the coefficient of CSR × MC × SOE is negative and statistically significant at 1% (−1.187 ***), which indicates that public attention has a greater impact on non-stated-owned heavily polluting enterprises than state-owned heavily polluting enterprises.

6. Discussion

With the intensification of environmental pollution and energy consumption, green technology innovation has become an effective way to alleviate negative impacts of heavily polluting enterprises on the ecological environment. The promotion of green technology innovation is not a straightforward and simple process. It is determined by the joint efforts of multiple stakeholders.

Environmental regulation as a powerful governance means binds the survival and development of heavily polluting enterprises with environmental protection. The impact of environmental regulation on green technology innovation has been widely studied. The empirical results showed that environmental regulation can positively promote green technology innovation in heavily polluting enterprises, which is in line with the research of Cai et al., Behera and Sethi [14,44]. This positive performance may be explored in multiple ways. The punishment for damaging environment can constrain heavily polluting enterprises’ pollution behaviors and force them to carry out green technology innovation [44]. Environmental regulation can also encourage and support heavily polluting enterprises by policy incentives to conduct diverse innovation activities, such as green product design, clean process innovation, and energy conservation [12,13] that further incentivize them to carry out green technology innovation continuously.

The occurrence of green technology innovation depends on not only external stimulus but also internal driving motivation. This study proved that CSR can positively promote green technology innovation in heavily polluting enterprises, which is consistent with the findings of Albort et al. and Gu et al. [30,31]. The essence of its internal impact mechanism deserves to be discussed properly. On the one hand, unlike previous studies, this study confirms the mediating role of CSR in the transition from environmental regulation to green technology innovation, that is, whether heavily polluting enterprises fulfilling their responsibilities or not is the key to whether environmental regulation can actively promote green technology innovation. Subjected to internal factors such as profits, operating costs, and good reputation, heavily polluting enterprises need to be responsible for each stakeholder interest [33]. Therefore, as an internal driving motivation, CSR plays a key role in green technology innovation. The more CSR heavily polluting enterprises undertake, the more they can promote green technology innovation. On the other hand, although the nation and public expect heavy polluting enterprises to take on more social responsibility and to be spontaneous and responsible for their production operations and environmental protection, due to their “economic man” nature, they are bound to perform their duties with economic efficiency in mind. When a new environmental policy is released, heavily polluting enterprises need time to weigh their own capacities and resources, then determine whether to take responsibility and how to take action to respond to the national call. While it cannot be denied that some entrepreneurs possess good character, are environmentally conscious and can take the initiative to be environmentally responsible, most heavy polluters still make decisions at the expense of the environment in favor of economic gain. Therefore, this requires the government to regulate their production activities, take environmental protection into account and carry out green technology innovation. The positive role of CSR in promoting green technology innovation of heavily polluting enterprises tends to occur under the regulation of government. Only the combined action of external stimulus and internal driving motivation can urge heavily polluting enterprises to organize various resources and carry out green technology innovation activities.

Unlike in the past when the public could only passively accept the living products and environment, nowadays, consumers are more concerned about healthy products and a good living environment; their attention can even determine the development and survival of an enterprise. The findings demonstrated that public attention can actively promote green technology innovation of heavily polluting enterprises, which is agreeable with the results of Tang et al. and Wang et al. [35,40]. The moderating role of public attention illustrates that it plays a “soft” supervision role in environmental pollution behaviors of heavily polluting enterprises and it is also a complement to environmental regulation, which is in line with the findings of Zhao et al. [36]. Unlike previous studies, besides the complementary role of public attention, this study also proved the public’s amplifier role by conducting a moderated mediating test. This means that public attention influences the strength of the mediating role of CSR in environmental regulation–green technology innovation. The more public attention, the more CSR, and the influence of environment is relatively small. Due to the characteristics of the full scope, wide perspective and timeliness of modern media and negative news, the public can expose heavily polluting enterprises’ negative news and misconduct to society, and often companies suffer boycotts before the government penalizes them. These boycotts can be a disastrous blow to the development and economic profits of heavily polluting enterprises. Therefore, heavily polluting enterprises have to strive to improve green technology innovation, fulfil social responsibility and maintain a good public image in order to eliminate the negative impacts, obtain public trust, and increase organizational competitive advantage.

In addition, this study further explores the role of environmental regulation and public attention on green technology innovation of heavily polluting enterprises with different property rights. The results showed that environmental regulation has a stronger impact on state-owned heavily polluting enterprises than on non-state-owned heavily polluting enterprises. It could be that state-owned heavily polluting enterprises can better grasp policy trends in advance because of the advantage of the information asymmetry and take the lead in making improvement measures. They also receive easy access to related resources and policy support. Public attention has a greater impact on non-stated-owned heavily polluting enterprises than on state-owned heavily polluting enterprises. The reasons may be that in contrast to the naturally outstanding image and public trust of state-owned heavily polluting enterprises, non-state-owned enterprises need public attention as exposure to maintain good reputation. Meanwhile, once the environmental behaviors are off track, the negative reports will lead to distrust of the public which can bring a more fatal blow compared with that which can be dealt to state-owned heavily polluting enterprises.

7. Conclusions and Policy Implications

7.1. Conclusions

Environmental problems not only hinder the sustainable development of a nation, but also bring many social and economic problems. Heavily polluting enterprises as major polluters should take more social responsibility. However, the long-standing dilemma for heavily polluting enterprises is how to balance the relationship between corporate profits and environmental protection. Only vigorous promotion and implementation of green technology innovation can solve the dilemma and realize the “win–win” scenario of economic growth and environmental protection. The promotion of green technology relies on the joint efforts of all parties. Environmental regulation can internalize the environmental problems into the production and operation process of heavily polluting enterprises and force them to carry out green technology innovation. Along with the improvement of public awareness of environmental protection, public attention, as the “soft” regulation, also has a positive effect on promoting green technology innovation. In addition, the role of CSR as the intrinsic driver of green technology innovation implementation cannot be overlooked, either. Therefore, this paper aims to clarify the role of environmental regulation, CSR, and public attention in promoting green technology innovation of heavily polluting enterprises, explore the influence mechanism and find out the conduction paths through this process. Once the above issues are clarified, it will make more sense for environmental protection and sustainable development.

This study analyzed the crosstalk between environmental regulation, CSR, and public attention on promoting green technology innovation based on Stakeholder and S-O-R theories. The panel data of A-share listed companies in China’s heavily polluting industries from 2008 to 2020 was collected to implement empirical research. The main research conclusions are as follows. (1) Environmental regulation has a positive effect on green technology innovation of heavily polluting enterprises and has a greater impact on state-owned heavily polluting enterprises. It indicates that environmental regulation can effectively promote green technology innovation by issuing financial penalties and incentive policies. Environmental regulation has a stronger impact on state-owned heavily polluting enterprises than non-state-owned heavily polluting enterprises, indicating that state-owned heavily polluting enterprises have more information advantages and are more likely to receive policy support. (2) CSR has a mediating effect between environmental regulation and green technology innovation. It illustrates that the development of green technology innovation needs both external stimulus and internal driving motivation. As the “economic man”, heavily polluting enterprises are more concerned with financial profits than fulfilling their responsibilities. The implementation of green technology innovation and fulfilment of social responsibility are therefore dependent on government intervention. (3) Public attention moderates not only the relationship between CSR and green technology innovation but also the indirect effect of environmental regulation on green technology innovation through CSR. In addition, public attention has a greater impact on non-stated-owned heavily polluting enterprises. It indicates that public attention plays a significant “soft” supervision role in the era of rapidly development to information technology. It is not only an indispensable supplement to formal regulations but also an effective way to optimize heavily polluting enterprises’ green behaviors and cause them to commit to social responsibility. The promotion of green technology innovation by non-state-owned heavily polluting enterprises is more susceptible to public attention, indicating that non-state-owned heavily polluting enterprises rely more on public trust and good reputation to compete in the active market. In summary, the realization of the development of green technology innovation and a sustainable environment needs the joint actions and efforts of these tripartite stakeholders.

7.2. Policy Implications

Based on the research results, this study put forward several policy recommendations hoping that they can provide theoretical guidance and practical advice for promoting the development of green technology innovation, optimizing environmental policy tools, and providing references for other developing countries.

First, this study confirmed the positive effect of environmental regulation on green technology innovation, so government regulation of heavily polluting enterprises remains an important guarantee of environmental protection. If the regulation is not in place, enterprises will take the negative environmental externality as the price of profitability, which is not conducive to their green technology innovation. Therefore, the government could pay more attention to the implementation of environmental regulation and form a long-term institutionalized and normalized environmental regulation mechanism to regulate the behavior of heavily polluting enterprises with rigid economic constraints.

Second, given the bridging role of CSR, the government could also optimize incentives to encourage enterprises to fulfill their social responsibility. Heavily polluting enterprises face the dual pressure of environmental protection and profitability. Since the R&D cycle of green technology innovation is long and costly, the future innovation compensation may not balance the R&D cost, so the government could support the process of green technology innovation by appropriately relaxing the environmental policy, offering financial and tax reduction support to improve the R&D enthusiasm of heavily polluting enterprises and guarantee their normal operation.

Last, the government could motivate the public’s enthusiasm due to their complementary roles in environmental regulation. For example, the government can establish an online reporting platform to facilitate effective communication with the public and offer rewards to those who actively participate. Meanwhile, the government can actively advocate the importance of environmental protection to the public to strengthen its role in environmental regulation. This will help heavily polluting enterprises to recognize their responsibility for environmental protection and promote the construction of a corporate social responsibility system.

8. Research limitations

The measurement method and testing tools of environmental regulation and CSR have not been unified. Further research could conduct in-depth explorations in this direction to better measure the effectiveness of ER and better assess the CSR performance of enterprises. In addition, considering the application cycle and the time of registering patents, some suitable patents may not have been included.