Business Environment Optimization, Digital Transformation, and Enterprises’ Green Innovation

Abstract

The optimization of the business environment makes regional economic development more dynamic. Whether this can be used as an influencing factor to promote the green innovation of enterprises is worthy of attention. Based on the concept of digitalization and green collaborative transformation, combined with the panel data of A-share listed companies in China’s capital market from 2016 to 2020 and the corresponding business environment index of each province, this paper empirically tests the impact of business environment optimization on enterprises’ green innovation and its channel mechanism by using a two-way fixed-utility model. The findings are as follows: Firstly, business environment optimization has a significant positive impact on enterprises’ green innovations, and this conclusion is still valid after a series of endogenous and robustness tests. Secondly, business environment optimization can foster conditions conductive to green innovation by promoting enterprises’ digital transformation. Through empirical tests, a chain-channel mechanism is identified whereby business environments can promote enterprises’ green innovation via digital transformation, innovation ability, and a willingness to pursue environmental management. Finally, further analysis reveals significant heterogeneity in the positive impact of business environment optimization on green innovation, with a more pronounced positive impact in state-owned enterprises, high-tech enterprises, and low-polluting enterprises. The research findings of this paper provide a specific reference and basis for the Chinese government to further optimize the business environment and effectively promote the green innovation of enterprises.

1. Introduction

After more than four decades of reform and opening up, China’s economic development has yielded remarkable outcomes. By 2024, it had maintained its position as the world’s largest manufacturing nation for 14 consecutive years. Nevertheless, despite the rapid growth of the macro-level economy, a range of unsustainable factors are gradually surfacing at the micro level, including low resource-conversion efficiency, outdated technology, and severe environmental pollution. As a result, there is a pressing need to shift the paradigm of economic growth [1]. In the contemporary era, the Fifth Plenary Session of the 18th Central Committee of the Communist Party of China introduced the five major development concepts, emphasizing innovation and environmental sustainability for the first time. The report of the 19th National Congress of the Communist Party of China advocated for the implementation of the Five Development Concepts with a focus on high-quality development for the first time, and the report of the 20th National Congress of the Communist Party of China has outlined explicit requirements for green transformation to facilitate the high-quality development of enterprises. Looking to the future, amid the pressures of a global economic downturn and a diminishing “demographic dividend” alongside efforts to achieve the “dual carbon” goal, the exploration of a viable path for green transformation has become imperative in both theoretical and practical spheres [2]. Green transformation necessitates an industrial structure that features high technology, low energy consumption, and minimal emissions. Nonetheless, traditional industries in China constitute a significant portion of the economy, with considerable energy demands, and they encounter challenges such as outdated technological infrastructure and inadequate capital investment, rendering the path to green transformation in China fraught with obstacles. Green innovation, serving as a pivotal driver of green transformation, not only enhances economic competitiveness but also facilitates reductions in pollution and emissions. It is crucial and necessary for China’s economy to transition towards high-quality development [3].

In light of the current reality, addressing the challenge of enhancing green innovation while ensuring the stability of enterprise development has become a critical issue demanding urgent attention. In terms of capabilities, enterprises in China exhibit a relatively low proficiency in green technology development, necessitating increased investment in funds, talent, and time for green innovation. Sole reliance on internal resources proves insufficient to fulfill the prerequisites for green innovation [4,5]. Regarding willingness, enterprises are hindered by traditional extensive development ideologies, and there persists a viewpoint that stringent government enforcement of environmental regulations will inevitably impede the pace of economic progress. Despite improvements in enterprises’ green innovations, they will encounter the challenge of inadequate market demand, thereby impeding genuine ecological progress [6].

Scholars have attempted to explore the driving factors of green innovation capabilities and willingness in enterprises from various perspectives. Some scholars have analyzed the impact of internal driving forces such as managers [7], corporate governance [8], enterprise size [9], and digital technology [10] on green innovation in enterprises. Similarly, scholars have also analyzed external factors such as the role of policies and regulations [11], research and development subsidies [12], and financial services [13] in promoting green innovation in enterprises. However, the above studies focus on individual driving factors, and whether the business environment as a comprehensive system can affect green innovation in enterprises has not received sufficient attention. Especially given the current imbalanced regional development in China, it is clearly unsuitable to ignore the influence of the regional business environment on the green innovation performance of enterprises [1]. In recent years, the Chinese government has attached great importance to the business environment, successively introducing a series of policies to enhance and optimize it. In 2020, the Global Business Environment Report was released, and China’s ranking in the global business environment has risen to 31st [14]. It is apparent that as the business environment undergoes continuous optimization, the level of green innovation in enterprises will accordingly increase.

Whether the level of green innovation in enterprises can be improved with the optimization of the business environment also depends on the ability of enterprises to respond. According to research by different scholars, when the “green” pressure and support coexist in the business environment, enterprises will focus on solving the green innovation problem and promote the improvement of the green innovation level [15]. However, in the absence of relevant coping abilities, enterprises will be tired of coping with the ever-changing law enforcement pressure, “hesitate” on the road to green development, and fail to properly use the financial resources provided by the government to solve the problem of the shortage of green innovation resources [16,17]. The rise of digital technologies has brought new opportunities to the relationship between the business environment and green innovation. On February 27, 2023, the Overall Layout Plan for the Construction of Digital China, jointly issued by the Central Committee of the Communist Party of China and the State Council, proposed the idea of the collaborative transformations of digitalization and green enterprise. In this context, with the improvement of regional digital infrastructure and the improvement of government digital services, enterprises will be encouraged to carry out green innovation. According to relevant scholars’ research, digital transformation’s role in promoting enterprises’ green innovation is reflected in reducing costs and expanding financing channels [10]. In addition, enterprises undergoing digital transformation quickly attract more attention from governments, media, and investors. In this case, whether an enterprise actively carries out green innovation is not only related to its image and competitive market position but also has a profound impact on the incentive of green innovation. Therefore, taking digital transformation as an opportunity to deeply explore the impact and mechanism of business environment optimization on enterprises’ green innovation will help enterprises and governments pay targeted attention to the business environment. This is not only of great significance to promote the implementation of environmental policies and promote the green development of enterprises, but also the key to the coordinated development of the national economy and the environment. Based on the above analysis, this study selects China’s A-share listed companies as research samples, uses relevant business environment data of various provinces as explanatory variables, and further explores the impact of business environment optimization on enterprises’ green innovation through empirical analysis, aiming to provide valuable references for academic research and practical application.

This paper may have the following research contributions: First, in terms of research content, the existing research on green innovation mainly focuses on the internal strategy of enterprises or external supervision and incentives and often ignores the impact of the macroeconomic development environment on enterprises’ green innovation. This study focuses on examining the impact of business environment optimization on enterprises’ green innovation, expands the research scope of influencing factors in green innovation, and provides policy basis and inspiration for promoting the realization of the “double carbon” goal. Second, this study has specific reference significance for the subsequent research on how the soft macro-environment affects the micro-enterprise behavior. In the past, scholars have conducted relatively sufficient studies on the macro and micro levels, respectively, and this paper attempts to integrate the two. By exploring the positive effects at the macro level, this paper reveals the specific mechanism of the business environment affecting the green innovation of enterprises at the micro level. Thirdly, this paper researches digital transformation, which also has specific reference significance for enterprises’ digital transformation. In the context of the digital economy, this paper profoundly studies how to draw out the definite connection between the business environment and enterprises’ green innovation through digital transformation, which provides a new way to understand the relationship between government and enterprise.

2. Literature Review and Research Hypothesis

2.1. Literature Review

Scholars have extensively researched corporate green innovation, which can be broadly categorized into two groups. The first focuses on the form and value of such innovation. Based on various innovation motives, green innovation can be classified into symbolic green innovation and substantive green innovation [17], where the former emphasizes quantity and speed, while the latter prioritizes quality [18]. Liu et al. classified green invention patents as high-quality green innovations and green utility model patents as low-quality green innovations based on different categories of green patents [19]. When examining the value of green innovation, it is often observed that enterprise-wide green innovation encompasses both forms [20]. Scholars have investigated the favorable effects of green innovation on corporate financial performance [19], equity financing [21], energy efficiency [22], and eco-efficiency [23]. As green practices gain popularity across various sectors of society, comprehending the types and value of green innovation aids in determining the nature and rationale behind green innovation efforts. The second category examines the influencing factors of green innovation. Within the firm, the magnitude of corporate governance capabilities notably influences corporate green innovation [8]. In corporate governance capabilities, executives, pivotal as decision-makers within enterprises [7], and digital transformation, a critical endeavor for enterprises to curtail costs and bolster efficiency [24], can also exert a notable influence on corporate green innovation. Externally, as viewed by enterprises, governmental subsidization of environmental protection or the implementation of inclusive green finance prompts active engagement in green innovation by enterprises seeking environmental protection subsidies [13]. Simultaneously, in response to government-imposed environmental regulations, firms opt to mitigate pollution emissions and undertake green innovation endeavors to evade penalties [25].

Whether optimizing the business environment can influence the green innovation of enterprises remains to be investigated. Nevertheless, numerous useful clues have been provided by previous scholarly studies [17]. Firstly, some scholars have examined the design and selection of indicators for the business environment, while the World Bank has developed an evaluation system comprising 11 general indicators and 43 sub-indicators. In conjunction with the current situation in China, certain scholars have conducted a more comprehensive evaluation of the business environment in Chinese cities or regions, considering dimensions such as public services, human resources, and market conditions [26]. Other scholars have analyzed how the business environment influences the economy and corporate behavior. As an institutional soft environment, the business environment is linked to the investment decisions, development potential, and business conditions of local enterprises [27,28]. With regard to whether the business environment can influence green innovation in enterprises, Chrisman et al.’s study on innovation management in family firms provides valuable insights and emphasizes the two crucial conditions necessary for successfully achieving a goal or task: ability and willingness [29]. In the “double synergy” between digitalization and environmental sustainability, actively integrating green innovation and digital transformation may enhance enterprises’ capacity and motivation for green innovation [16]. Is the business environment also a factor influencing these abilities and willingness? Luo et al. demonstrated that optimizing the business environment attracts more capable managers and increases government subsidies and incentives for firms to increase R&D investment, which in turn facilitates firms’ digital transformation [30].

The above analysis of the existing literature shows that, on the one hand, green innovation has received extensive attention from academics, but whether the business environment as an integrated system can affect corporate green innovation has not received sufficient attention; on the other hand, given China’s status as the world’s largest manufacturing country, it is imperative to examine the impact of the regional business environment on corporate green innovation in the context of digital transformation. The research in this article will provide valuable policy insights on how the country can promote green and high-quality development.

2.2. Hypotheses Development

To formulate the primary research hypothesis, this study adopts the following analytical framework. The initial stage involves analyzing whether the macro-level business environment can impact the digital transformation of enterprises. Subsequently, it examines at the micro level whether digital transformation can affect the capacity and inclination for green innovation. Furthermore, it investigates whether the business environment can impact enterprises’ green innovation through this capacity and inclination, as depicted in Figure 1 of the theoretical framework.

2.2.1. Innovation Ability

With the progressive implementation of the Overall Layout Plan for Digital China Construction, a favorable business environment can enhance the efficient flow of information, talent, and other resources in the market, consequently providing solid institutional guarantees and accelerating enterprises’ digital transformation.

Financial constraints are a key concern. As the “streamlining administration and delegating power” reform continues to deepen, government departments have transitioned from managerial roles to service-oriented ones. This transformation has simplified the process for entering and exiting the market, leading to a spillover effect and invigorating capital flow within the market. Simultaneously, optimizing the business environment enhances the judicial and commercial credit environments, diminishes market intervention, and strengthens property rights protection [31]. This transformation creates an environment characterized by greater economic freedom, openness, and transparency, which in turn attracts more enterprises to invest. Meanwhile, in tandem with improvements to the business environment, the government has set up a strong multi-level capital market. This has increased the proportion of direct financing available to enterprises, expanded their financing options, and, as a result, reduced their financing costs, thereby alleviating financial burdens related to digital transformation [32].

When it comes to the level of urbanization, the theory of economies of scale suggests that a complete industrial chain and efficient resource allocation can only emerge when the economic scale attains a certain threshold, ultimately leading to an increase in marginal benefits. Consequently, a strong correlation exists between a region’s business environment and its urbanization level. Specifically, urbanization offers a stage for the concentration of population and industries, facilitating optimal resource allocation. On the contrary, by enhancing the business environment, more investment and enterprises are attracted, which further boosts the agglomeration of population and industry. This, in turn, not only generates higher market demand for the digital transformation of enterprises but also equips them with comprehensive infrastructure.

Concerning the labor force, the 20th National Congress of the Communist Party of China emphasized the vital role of talent in national development and stressed the need to fully implement a strategy to fortify the nation with skilled talent while also fully valuing labor. Talent policies, as pivotal elements of the business environment, exert a substantial influence on the labor force level and regional economic development potential. Specifically, enhancing the regional business environment can generate additional employment opportunities, attract labor market participation, and bolster labor market dynamics. Moreover, it can effectively cultivate the enthusiasm and creativity of the workforce by refining institutional mechanisms, strengthening labor security, and fostering a positive cultural environment. These efforts are conducive to facilitating the seamless advancement and subsequent maintenance of digital transformation.

Based on the above analysis, this article contends that optimizing the business environment will promote the digital transformation of enterprises.

2.2.2. Environmental Management Willingness

Existing research indicates that the extent of a company’s green innovation capability hinges on the magnitude of its R&D investment. In contrast, the environmental commitment of the management team dictates whether the company’s innovation efforts will steer towards a “green” trajectory. Accordingly, this paper delves deeper into the interplay between business environment optimization, digital transformation, and enterprises’ green innovation across the following two dimensions.

Innovation capability. Innovation is a primary driver for enterprise development, helping micro-enterprises adapt and thrive in the macro-environment. By enhancing innovation capability, enterprises can secure a competitive advantage in technology development, product design, and production processes. This enables them to meet the market demand for environmentally friendly products effectively. As a result, enterprises’ capacity for green innovation is significantly enhanced. As an aspect of the institutional soft environment, optimizing the business environment also enhances enterprises’ innovation capabilities. Firstly, the 14th Five-Year Plan for National Economic and Social Development of the People’s Republic of China and the Long-Range Objectives for 2035 explicitly call for increasing enterprises’ investment in research and development to boost their technological innovation capabilities. In addition, existing research suggests that improving the business environment can encourage enterprises to engage in digital transformation actively, lowering costs and enhancing their confidence [33]. Simultaneously, digital transformation within enterprises can drive innovation by accelerating product development, adaptation, and reorganization; refining innovative models; establishing new organizational structures; and nurturing innovation in digital business models [34]. Based on the previous discussion, this paper contends that optimizing the business environment can enhance enterprises’ innovation capacity and promote green innovation.

Environmental management willingness. Since the Fifth Plenary Session of the 18th Central Committee of the CPC, innovation has been highlighted as crucial for national development, and green development has become a fundamental national policy. As the business environment is continuously optimized, the market’s focus on high pollution emissions and related issues is also intensifying. Consequently, enterprises that persist in employing high-emission production methods will inevitably encounter heightened social pressure. Previous studies have demonstrated that the academic qualifications of managers and the involvement of environmentally conscious investors influence a company’s tendency toward environmental protection. Moreover, improving the regional business environment helps attract highly educated professionals and eco-friendly investors. Furthermore, digital transformation enables enterprises to identify potential environmental concerns and formulate effective environmental protection strategies accordingly [35]. Therefore, this paper argues that as the business environment continues to be optimized, the tendency of enterprises to protect the environment will also continue to increase, which in turn will positively impact the enhancement of enterprises’ green innovation level.

After the above analysis, we propose the following hypotheses:

H0: 

Optimizing the business environment has no significant impact on enterprises’ levels of green innovation.

H1: 

Optimization of the business environment can significantly enhance the level of enterprises’ green innovation.

3. Economic Modeling and Data Sources

3.1. Data

First, this paper selects China’s A-share listed companies from 2016 to 2020 as research samples. The reasons for this are as follows: (1) On 12 May 2015, the State Council, for the first time, put forward the concept of “regulation service” reform, marking the beginning of broader attention to business environment issues. In October 2015, the Fifth Plenary Session of the 18th Central Committee of the CPC put forward five development concepts, including green and innovation, for the first time, so this study takes 2016 as the starting year of the sample. (2) Considering the outbreak of COVID-19 and its significant impact on the global economy, this study selects data up to 2020 to explore the research question more comprehensively.

Secondly, this paper follows the conventional treatment of sample companies: (1) Select according to enterprise scale, listing years, and other indicators to ensure the diversity and balance of samples; (2) Remove samples in abnormal trading states such as ST and ST*; (3) Delete the sample of listed companies in the financial and insurance industry; (4) Exclude samples newly listed in the same year; (5) To reduce the impact of outliers on the research results, we conducted tailing processing of up or down 1% for some financial data. In the end, 10,209 companies—annual observations—were obtained.

Finally, in terms of data sources, this study’s green innovation data are taken from the number of green patents of listed companies each year, and the business environment data are from the market indexes of various provinces in China. The relevant financial data of the research design are calculated through the annual reports of listed companies and the CSMAR database. The statistical analysis software used in this study is Stata16.

3.2. Variables

3.2.1. Dependent Variables

Green innovation. This study utilizes the methodology proposed by Wang Xin and Wang Ying [26], which employs the total count of green patent applications as an indicator of enterprises’ green innovation [35]. In order to address the issue of right-skewed distribution, 1 is added to the total count of green patents, and the natural logarithm is applied. The specific calculation procedure is outlined in

Table 1. The definitions of variables.

Variable Type	Variable	Description	Definition
Explained variables	$PAT$	Green innovation	The natural logarithm of number of green patent applications plus 1
Explanatory variables	$busen$	Business environment	Market index of various provinces in China
Control variables	$tobinQ$	Tobin Q value	The ratio of enterprise market value to enterprise reset cost
	$size$	Enterprise size	The natural logarithm of assets
	$lev$	Leverage ratio	The ratio of liability to assets
	$roa$	Return on total assets	The ratio of net profit to total assets
	$eoc$	Operating capacity	The ratio of operating income to average current assets
	$em$	Enterprise maturity	The natural logarithm of the company’s establishment period
	$top$	Top 10 shareholding ratio	The ratio of sum of the top 10 shareholders’ shareholdings to total shareholdings
	$grow$	Enterprise growth	The ratio of revenue growth to previous period revenue

.

3.2.2. Independent Variables

Business environment. Considering the diverse operational scope of listed companies, they are influenced not only by the local business environment where the headquarters are situated but also by the business climate of the regions where they conduct operations. For instance, according to the Enterprise Income Tax Law of the People’s Republic of China, enterprises conducting business activities in jurisdictions other than their headquarters must adhere to local tax withholding and payment regulations. Consequently, this research adopts the provincial marketization index from the China Provincial Market Index Report [31] as the benchmark for assessing the business environment. The report systematically assesses and evaluates the progression of marketization in 31 provinces nationwide, where a higher marketization index correlates with a more favorable provincial business environment.

3.2.3. Control Variables

In order to enhance this study’s accuracy, this research incorporates the relevant literature and selects several control variables, including Tobin Q value ($tobinQ$), enterprise size ($size$), leverage ratio ($lev$), return on total assets ($roa$), enterprise growth ($grow$), enterprise maturity ($em$), equity concentration ($top$), and operating capacity ($eoc$). The selection and calculation methods of specific variables are presented in Table 1.

3.2.4. Models

To examine the influence of the business environment on fostering green innovation among enterprises, we construct the following econometric model (1) to test Hypothesis H1.

$${PAT}_{i,t}={\alpha }_0+{\alpha }_1{busen}_{i,t}+{\alpha }_2{Controls}_{i,t}+{year}_t+{ind}_i+{\epsilon }_{i,t}$$

(1)

The subscripts $i$ and $t$ represent the enterprise and year, respectively. ${PAT}_{i,t}$ serves as the explanatory variable, indicating the level of green innovation of enterprise $i$ in year $t$. ${busen}_{i,t}$ stands as the key explanatory variable in this paper, representing the business environment index of the province where enterprise $i$ is located in year $t$. $Controls$ denotes a series of control variables. In model (1), both the annual fixed effect ${year}_t$ and the industry fixed effect ${ind}_i$ are controlled, while ${\epsilon }_{i,t}$ represents the random perturbation term. Assuming H1 to be true, implying that optimization of the business environment can foster the green innovation of enterprises, ${\alpha }_1$ should exhibit a significantly positive relationship. Additionally, robustness corrections of the standard error of the model are implemented. Variables are defined in Table 1.

4. Results of Empirical Analysis

4.1. Descriptive Statistics

Table 2. Descriptive statistics for variables.

Variable	Max	Min	Mean	median	Sd	N
$PAT$	4.875	0.000	1.013	0.693	1.246	10,209
$busen$	8.990	3.510	6.969	7.070	1.188	10,209
$tobinQ$	8.307	0.124	1.639	1.201	1.480	10,209
$roa$	0.210	−0.275	0.032	0.032	0.066	10,209
$em$	3.611	2.398	3.033	3.045	0.243	10,209
$lev$	0.884	0.064	0.442	0.439	0.197	10,209
$size$	26.497	20.171	22.652	22.49	1.279	10,209
$top$	0.895	0.241	0.555	0.552	0.145	10,209
$eoc$	5.638	0.127	1.293	1.034	0.983	10,209
$grow$	1.811	−0.595	0.127	0.084	0.332	10,209

displays the descriptive statistical results of the primary variables. Specifically, the mean value of green innovation ($PAT$) was 1.01, the median was 0.69, and the maximum and minimum values were 4.88 and 0, respectively. This suggests significant disparities in green innovation levels among firms, with the sample firms exhibiting relatively low degrees of green innovation, which is consistent with the statistical conclusions of previous studies. The mean value of the business environment ($busen$) is 6.97, and the maximum and minimum values are 8.99 and 3.51, respectively, reflecting substantial regional disparities in the business environment, consistent with China’s uneven economic development. Regarding control variables, the standard deviation of enterprise size $\left(size\right)$ is 1.28, and the maximum and minimum values are 26.5 and 20.17, respectively, indicating significant differences in enterprise scale. Additionally, the standard deviation of the Tobin Q ($tobinQ$) value is 1.48, and the maximum and minimum values are 8.31 and 0.12, respectively, suggesting substantial discrepancies between market and actual values, consistent with China’s capital market dynamics. The ranges of other variables were generally consistent with the previous relevant literature.

4.2. Benchmark Regression Results

Expand the control variables ($Controls$) in Model (1) to obtain Model (2):

$$\begin{array}{c}{PAT}_{i,t}={\alpha }_0+{\alpha }_1{busen}_{i,t}+{\alpha }_2{tobinQ}_{i,t}+{\alpha }_3{roa}_{i,t}+{\alpha }_4{em}_{i,t}+{\alpha }_5{lev}_{i,t}+{\alpha }_6{size}_{i,t}\\ \hspace{1em}+{\alpha }_7{top}_{i,t}+{\alpha }_8{eoc}_{i,t}+{\alpha }_9{grow}_{i,t}+{year}_t+{ind}_i+{\epsilon }_{i,t}\end{array}$$

(2)

The regression analysis results on model (2) are shown in

Table 3. Benchmark regression results.

Variable	$\left(1\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(2\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(3\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(4\right)\mathit{P}\mathit{A}\mathit{T}$
$busen$	0.074 *** (0.010)	0.082 *** (0.010)	0.093 *** (0.009)	0.106 *** (0.008)
$tobinQ$			0.039 *** (0.008)	0.041 *** (0.009)
$roa$			−0.189 (0.171)	−0.234 (0.167)
$em$			−0.537 *** (0.049)	−0.370 *** (0.048)
$lev$			0.192 *** (0.067)	0.471 *** (0.065)
$size$			0.479 *** (0.014)	0.516 *** (0.013)
$top$			−1.004 *** (0.083)	−0.697 *** (0.078)
$eoc$			−0.030 *** (0.011)	−0.045 *** (0.011)
$grow$			−0.008 (0.032)	0.039 (0.030)
_cons	0.496 *** (0.066)	−0.173 ** (0.083)	−8.405 *** (0.333)	−10.975 *** (0.322)
$year$	no	yes	no	yes
$ind$	no	yes	no	yes
N	10,209	10,209	10,209	10,209
r2_a	0.005	0.096	0.210	0.327

Note: The robust standard error of clustering is indicated in parentheses. *** and ** respectively indicate significance at the 1% and 5% levels.

. In column (1), only univariate regression was carried out, and the results showed that the regression coefficient of $busen$ to $PAT$ was 0.074 and significant at a 1% level, which initially proved that business environment optimization had a significant positive impact on enterprises’ green innovation. (2) As the regression results after the control year and industry show, the regression coefficient of Busen to $PAT$ is still significantly positive at a 1% level; (3) is listed as the regression result of adding control variables, and Busen’s regression coefficient is still positive and significant at a 1% level. Finally, (4) is listed as the regression result after adding control variables and fixed effects of year and industry. The regression coefficient of Busen is 0.106 and significant at a 1% level. The above results significantly reject the original hypothesis H0 and unanimously confirm the H1 hypothesis of this study, that is, optimizing the business environment will significantly improve enterprises’ green innovation level.

In the analysis results of the control variables in column (4), the regression coefficients of $tobinQ$, $lev$, and $size$ are strongly and significantly positive at the 1% significance level, indicating that enterprises exhibit superior performance in the financial market. Moreover, high asset–liability ratios and larger enterprise scales positively influence the likelihood of engaging in green innovation. The regression coefficients of $em$, $top$, and $eoc$ are strongly and significantly negative at the 1% significance level. This implies that younger enterprises with a more dispersed shareholding structure and weaker operational capabilities, compared to well-established enterprises with higher equity concentration and better operating capabilities, tend to exhibit stronger environmental awareness and higher innovation momentum, potentially leading to higher levels of green innovation. Furthermore, there exists a negative correlation between $roa$ and $PAT$, suggesting that firms with higher return on assets prioritize production and operational aspects over technological research and development. This finding aligns better with real-world scenarios. Lastly, the positive correlation between $grow$ and $PAT$ implies that companies with stronger growth tendencies are more inclined to engage in green innovation [36].

Overall, the regression results in Table 3 provide good evidence of the hypothesis that optimizing the business environment will significantly improve the level of green innovation of enterprises, that is, H1 is confirmed. Additionally, the results of the control variables are largely consistent with our expectations, thus further strengthening the robustness of the research conclusions.

5. Endogeneity Test and Robustness Test

5.1. Endogeneity Test

To address potential endogeneity concerns, this study incorporates t + 1 green innovation and t − 1 business environment as explanatory variables in model (1) alongside the utilization of the instrumental variable method and the Heckman two-stage regression model. These approaches contribute to bolstering the reliability of the research findings.

5.1.1. Adjusting the Time Series

To tackle potential reverse causality issues, this study applies a one-period lag to the explanatory variables, resulting in $L.busen={busen}_{t-1}$. Meanwhile, the dependent variable is shifted forward by one period, resulting in $F.PAT={PAT}_{t+1}$, which facilitates a clearer understanding of the relationship between the business environment and green innovation of enterprises. To revalidate our hypothesis H1, column (1) of

Table 4. Endogeneity test.

Variable	$\left(1\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(2\right)\mathit{F}.\mathit{P}\mathit{A}\mathit{T}$	$\left(3\right)\mathit{b}\mathit{u}\mathit{s}\mathit{e}\mathit{n}$	$\left(4\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(5\right)\mathit{P}\mathit{A}\mathit{T}\_\mathit{d}\mathit{u}\mathit{m}$	$\left(6\right)\mathit{P}\mathit{A}\mathit{T}$
$busen$		0.113 *** (0.010)		0.239 *** (0.024)		0.102 *** (0.012)
$L.busen$	0.110 *** (0.010)
$RWT$			0.161 *** (0.004)		0.036 *** (0.004)
$Imr$						0.326 * (0.196)
$Controls$	yes	yes	Yes	yes	yes	yes
$year$	yes	yes	Yes	yes	yes	yes
$ind$	yes	yes	Yes	yes	yes	yes
N	8107	8107	10,209	10,209	10,209	5336
r2_a	0.335	0.327	0.190	0.191		0.212

Note: The robust standard error of clustering is indicated in parentheses. *** and * respectively indicate significance at the 1% and 10% levels.

presents the regression results of $L.busen$ on $PAT$, revealing a significantly positive regression coefficient at the 1% significance level. Furthermore, column (2) displays the regression results of $busen$ on $F.PAT$, also exhibiting a significantly positive regression coefficient at the 1% significance level, with minimal deviation from the previous results. Collectively, these results further substantiate the robustness of the research conclusions of this study, even when potential reverse causality concerns are taken into account.

5.1.2. Instrumental Variable Method

The instrumental variable method is crucial for solving endogeneity problems in regression models [37]. The ordinary least squares (OLS) estimates produce biased and inconsistent results when the explanatory variable is related to the error term. Instrumental variables must meet two key conditions: first, they are highly correlated with endogenous explanatory variables; second, they are unrelated to the error term. By introducing such instrumental variables, the bias caused by endogeneity can be substantially mitigated, and the regression results can be more reliable. In this study, a good business environment will promote enterprises’ green innovation, but enterprises may also actively adapt to the environment and move to areas with a better business environment. Therefore, this paper draws on the research of He et al. and selects the harmless treatment rate of domestic waste ($RWT$) as an instrumental variable to alleviate the possible endogenous problems [38]. The rationale for this choice is that the government’s efficient handling of logistics and sanitation reflects its ability to provide public services, affecting the business environment. However, the government’s garbage disposal behavior has no direct impact on the green innovation of listed companies, so RWT meets the correlation and homogeneity assumptions required by the instrumental variables. Column (3) of Table 4 shows the first-stage regression results with $RWT$ as the instrumental variable. The results show that the $RWT$ regression coefficient is significantly positive at 1%, which indicates that the government’s harmless treatment rate of domestic waste substantially improves the business environment, thus verifying the correlation requirement of this instrumental variable. Column (4) shows the regression results of the second stage, and the business environment optimization still has a significant positive impact on enterprises’ green innovation. This further confirms the robustness of the conclusions in this paper.

5.1.3. Heckman Two-Stage Regression Model

In many empirical studies, due to the existence of non-random sampling, the samples are not entirely randomly included in the research, which leads to the possibility of bias in the estimation results of ordinary least squares (OLS), and the Heckman two-stage regression model can effectively deal with such problems. A selection equation is typically constructed in the first stage, typically using the Probit model. In the second stage, the inverse Mills ratio obtained in the first stage is added to the central regression equation, and then OLS is used for regression estimation. This way, the coefficient estimation in the central regression equation can control the sample-selection bias to obtain more accurate and reliable regression results. This paper further employs the Heckman two-stage regression model to mitigate the potential issue of sample self-selection. In the first stage, the following probit equation is formulated:

$${PAT\_dum}_{i,t}={\alpha }_0+{\alpha }_1{RWT}_{i,t}+{\alpha }_2{Controls}_{i,t}+{year}_t+{ind}_i+{\epsilon }_{i,t}$$

(3)

Among them, if the green innovation level of enterprise $i$ in year $t$ exceeds the median value of 0.693, the value of the variable $PAT\_dum$ is set to 1; otherwise, it is set to 0. $RWT$ represents the instrumental variable referred to earlier. The remaining components of model (3) are consistent with the benchmark model (1). In the first stage, the inverse Mills ratio ($Imr$) is obtained by estimating model (3). In the second stage, the inverse Mills ratio ($Imr$) is included in the regression model (1) as a control variable. The regression results are presented in columns (5) and (6) of Table 4. Regression analysis reveals a significant estimation coefficient (at the 10% level) for $Imr$, suggesting the presence of sample self-selection bias. However, even after accounting for the sample-selection problem and including $Imr$ in the regression model, the coefficient of $busen$ for $PAT$ remains significantly positive at a 1% significance level. This finding further strengthens the robustness of the paper’s conclusion, highlighting the significant positive impact of a favorable business environment on corporate green innovation.

5.2. Robustness Test

To further enhance the reliability of the empirical results in the previous paper, this study adopts the substitution of business environment variables, the substitution of green innovation variables, and the exclusion of alternative explanations to ensure the robustness of the conclusions.

5.2.1. Variable Substitution

To control the possible deviation of the measurement error of key variables to the research conclusion, the explanatory variables and the explained variables in model (1) were replaced, respectively. Take the variable measuring green innovation as an example. If the original variable is the number of green patent applications of the enterprise, we may choose the proportion of green patents as the replacement variable. The number of green patent applications reflects enterprises’ absolute number of green innovation achievements. In contrast, the proportion of green patents (number of green patents/total patent number) reflects the relative proportion of green innovation achievements of enterprises’ overall innovation achievements. Both are closely focused on the green innovation achievements of enterprises. The proportion of green patents can more intuitively reflect the degree and relative level of enterprises’ focus on green innovation from the perspective of relative proportion, which is similar to the number of green patent applications in nature and function and can be used to test whether the research results will change due to different ways of measuring the green innovation achievements. The explained variable is replaced by the green patent proportion ($REP$). This measurement method refers to the research of Qi Shaozhou et al. [39]. It selects the proportion of green patents in all patents of sample enterprises in the current year as the measurement method for enterprises’ green innovation [14]. In terms of explanatory variables, this paper adopts the method adopted by Zhou et al. [40] and selects Chinese provinces’ business environment index ($MTI$) to replace the market index. This index covers 31 provinces in China and contains 8 aspect indexes and 30 sub-indexes, making it a suitable substitute for market-oriented indexes. The regression results are shown in columns (1), (2), and (3) of

Table 5. Robustness test.

Variable	$\left(1\right)\mathit{R}\mathit{E}\mathit{P}$	$\left(2\right)\mathit{R}\mathit{E}\mathit{P}$	$\left(3\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(4\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(5\right)\mathit{P}\mathit{A}\mathit{T}$
$busen$	0.006 *** (0.001)			0.145 *** (0.012)	0.107 *** (0.009)
$MTI$		0.005 *** (0.001)	0.086 *** (0.007)
$Controls$	Yes	yes	yes	yes	yes
$year$	Yes	yes	yes	yes	yes
$ind$	Yes	yes	yes	yes	yes
N	10,209	10,209	10,209	5301	8193
r2_a	0.083	0.083	0.327	0.285	0.281

Note: The robust standard error of clustering is indicated in parentheses. *** indicates significance at the 1 level.

. The regression coefficients of $busen$ to $REP$, $MTI$ to $REP$, and $MTI$ to $PAT$ are all significantly positive at 1%. The above results consistently indicate that the conclusion that business environment optimization can promote the level of green innovation of enterprises is still robust even if the measurement method of variables is replaced.

5.2.2. Exclude Alternative Interpretations

This paper excludes non-manufacturing sample enterprises and sample enterprises registered in municipalities directly under the central government, and the regression analysis is carried out again. The implementation of this step is based on the following considerations: First, given that China is the world’s first manufacturing country, the manufacturing industry plays an important role in economic development and energy consumption, so the analysis of the manufacturing industry is not only of great practical significance but also helps to eliminate the impact of outliers or extreme values. Secondly, as a region under the direct jurisdiction of the central government, although municipalities have the same political and economic rights as provinces, they may enjoy superior policies and welfare and bear more responsibilities, which may lead to a particular sample self-selection problem of listed companies in municipalities. Thirdly, excluding non-manufacturing sample enterprises enables us to focus more on the manufacturing industry, which is significantly affected by the correlation between the business environment and green innovation, and reduce the interference factors brought by industry differences to more accurately analyze the relationship between the two [41]. Finally, the exclusion of enterprises registered in municipalities directly under the central government helps control the influence of unique factors, such as regional policies and resource endowments, on the research results so that the results are more universal and representative. The regression results are shown in columns (4) and (5) of Table 5. In column (4), $busen$’s regression coefficient for PAT is 0.145, which is higher than before and significantly positive at a 1% level, which is in line with the basic expectation of this paper. In the regression results of $busen$ vs. $PAT$ in column (5), $busen$’s coefficient is still significantly positive at the 1% level. These results further confirm the positive impact of business environment optimization on enterprises’ green innovation, thus enhancing the robustness of the conclusions of this study.

6. Impact Path Analysis

According to the above empirical results, the optimization of the business environment has played a significant role in promoting green innovation in enterprises, and its mechanism is as follows: firstly, optimizing the business environment helps to improve the level of digital transformation of enterprises and promote green innovation in enterprises; secondly, optimizing the business environment helps to increase R&D investment and promote green innovation in enterprises; and thirdly, optimizing the business environment helps to enhance environmental protection willingness and promote green innovation in enterprises [39]. It is worth mentioning that optimizing the business environment can also affect the digital transformation of enterprises, thereby affecting their innovation ability and environmental protection willingness, ultimately affecting their green innovation.

To validate the theoretical deduction, this paper develops models (4) through (6):

$${DT}_{i,t}={\alpha }_0+{\alpha }_1{busen}_{i,t}+{\alpha }_2{Controls}_{i,t}+{year}_t+{ind}_i+{\epsilon }_{i,t}$$

(4)

$${rd}_{i,t}={\beta }_0+{\beta }_1{busen}_{i,t}+{{\beta }_{2}DT}_{i,t}+{\beta }_3{Controls}_{i,t}+{year}_t+{ind}_i+{\epsilon }_{i,t}$$

(5)

$${PAT}_{i,t}={\gamma }_0+{\gamma }_1{busen}_{i,t}+{\gamma }_2{DT}_{i,t}+{\gamma }_3{rd}_{i,t}+{\gamma }_4{Controls}_{i,t}+{year}_t+{ind}_i+{\epsilon }_{i,t}$$

(6)

In model (4), the dependent variable ${DT}_{i,t}$ represents the degree of digital transformation of the sampled enterprise $i$ in year $t$. This study adopts the methodology proposed by Wu et al. [41], utilizing the frequency of digital-related terms in the annual reports of sampled enterprises to gauge the extent of digital transformation. A higher index value signifies a greater degree of digital transformation. In model (5), the variable ${rd}_{i,t}$ represents the innovation capability of enterprise $i$ in year $t$. This study employs the ratio of R&D expenditure to operating income as a proxy variable for measuring a company’s innovation capability ($rd$). A higher value of this indicator suggests a stronger innovation capability of the company. ${\beta }_1$ and ${\beta }_2$ measure the effect of the business environment ($busen$) and digital transformation ($DT$) on the company’s innovation capability ($rd$), respectively. In model (6), ${\gamma }_1$ through ${\gamma }_3$ respectively revealed the influence of business environment, digital transformation, and innovation capability on green innovation in enterprises. The settings of other variables in models (4) through (6) are consistent with the benchmark model (1). Similarly, this study refers to the research method of Zhang et al. [42] and uses $iso14001$ [43] (environmental management system certification) as a proxy variable to measure a company’s willingness to engage in environmental management. Here, dummy variables are used, with a value of 1 denoting active participation in environmental management system certification during the current year and a value of 0 indicating otherwise. By replacing the $rd$ in models (5) and (6) with the $iso14001$, the effect of business environment ($busen$) and digital transformation ($DT$) on environmental management willingness ($iso14001$) can be obtained, as well as the influence of business environment, digital transformation, and environmental management intention on green innovation in enterprises.

The regression results of the above model are shown in

Table 6. Analysis of impact mechanisms.

Variable	$\left(1\right)\mathit{D}\mathit{T}$	$\left(2\right)\mathit{r}\mathit{d}$	$\left(3\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(4\right)\mathit{i}\mathit{s}\mathit{o}14001$	$\left(5\right)\mathit{P}\mathit{A}\mathit{T}$
$busen$	0.137 *** (0.010)	0.2 *** (0.000)	0.075 *** (0.008)	0.014 *** (0.003)	0.085 *** (0.008)
$DT$		0.005 *** (0.000)	0.116 *** (0.009)	0.005 (0.003)	0.142 *** (0.009)
$rd$			5.277 *** (0.319)
$iso14001$					0.084 *** (0.026)
_cons	−3.331 *** (0.322)	−0.010 (0.009)	−10.446 *** (0.312)	−0.113 (0.116)	−10.491 *** (0.317)
year	yes	yes	yes	yes	yes
ind	yes	yes	yes	yes	yes
N	10209	10209	10209	10209	10209
r2_a	0	0.422	0.364	0.061	0.346
direct			0.075 ***		0.085 ***
indirect (${\alpha }_1\times {\gamma }_2$)			0.016 ***		0.195 ***
indirect (${\beta }_1\times {\gamma }_3$)			0.011 ***		0.001 **
indirect (${\alpha }_1\times {\beta }_2\times {\gamma }_3$)			0.084 ***		0.002 ***

Note: The robust standard error of clustering is indicated in parentheses. *** and ** respectively indicate significance at the 1% and 5% levels.

. In column (1), the regression coefficient of $busen$ to $DT$ is significantly positive at the 1% level, indicating that optimizing the business environment helps promote the digital transformation of enterprises. In column (2), the regression coefficients of $busen$ and $DT$ are significantly positive at the 1% level, indicating that there is a partial mediation between the optimization of the business environment and innovation capability, that is, the optimization of the business environment can directly affect the innovation capability of enterprises and can also affect the innovation capability of enterprises through digital transformation. In column (3), the regression coefficients of $busen$, $DT$, and $rd$ are all significantly positive at the 1% level, indicating the existence of a chain-mediated effect. Column (4) shows the regression coefficients of $busen$ and $DT$ for $iso4001$, where the regression coefficients of $busen$ are significantly positive at the 1% level, while the regression coefficients of $DT$ are positive but not significant. This may be due to a potential correlation between the business environment and digital transformation, leading to biased results. The regression coefficients of $busen$, $DT$, and $iso4001$ in column (5) are all significant at the 1% level, indicating the existence of a chain-mediated effect of optimizing the business environment by influencing the digital transformation of enterprises and enhancing their willingness to manage the environment, ultimately affecting the green innovation in enterprises. To further enhance the robustness of the above test results, this study conducted 500 bootstrap tests on the regression, and the results showed that both direct and indirect effects were significant. The above results collectively prove the existence and rationality of digital transformation and innovation capabilities, as well as the willingness mechanism for digital transformation and environmental management.

7. Further Analysis of Heterogeneity

The previous analysis has examined and confirmed the mechanism for enhancing the business environment to foster green innovation within enterprises. However, the variations in property rights, technological characteristics, and pollution attributes among enterprises might influence the association between the business environment and green innovation. To further explore this matter, this article constructs the following model (7):

$$\begin{array}{l}{PAT}_{i,t}={\alpha }_0+{\alpha }_1{busen}_{i,t}+{\alpha }_2{busen\times Z}_{i,t}+{\alpha }_3{Z}_{i,t}+{\alpha }_4{Controls}_{i,t}+{year}_t\\ \hspace{1em}\hspace{1em} +{ind}_i+{\epsilon }_{i,t}\end{array}$$

(7)

Among them, $Z_{i,t}$ represents the moderating variable, ${busen\times Z}_{i,t}$ represents the interaction between the business environment and the moderating variable, and the remaining variables are identical to those in the benchmark regression model (1).

7.1. Heterogeneity Analysis Based on Property Rights

The optimization of the business environment may have a varying impact on the level of green innovation among enterprises with different property rights. Firstly, state-owned enterprises face relatively less pressure from market competition and talent demand in comparison with non-state-owned enterprises, and their level of digitalization is lower [41]. Secondly, state-owned enterprises often face strong social regulations and tend to proactively assume social responsibility, which can contribute to the enhancement of green innovation. Based on the previous analysis, this study investigates the influence of state-owned property rights on the relationship between the business environment and green innovation among sampled enterprises. Consequently, a dummy variable ($soe$) representing state-owned enterprises is constructed and included as a moderating variable in the regression model (7) for examination. The results in column (1) of

Table 7. Heterogeneity test.

Variable	$\left(1\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(2\right)\mathit{P}\mathit{A}\mathit{T}$	$\left(3\right)\mathit{P}\mathit{A}\mathit{T}$
$busen$	0.110 *** (0.008)	0.072 *** (0.010)	0.114 *** (0.010)
$soe\_busen$	0.091 *** (0.017)
$soe$	0.072 *** (0.023)
$tech\_busen$		0.082 *** (0.017)
$tech$		−0.178 (0.118)
$poll\_busen$			−0.061 *** (0.018)
$poll$			0.191 (0.125)
$Controls$	yes	yes	yes
Empirical p-value	0.040	0.008	0.034
$year$	yes	yes	yes
$ind$	yes	yes	yes
N	10209	10209	10209
r2_a	0.330	0.347	0.332

Note: The robust standard error of clustering is indicated in parentheses. *** indicates significance at the 1% level.

indicate that the regression coefficients for the interaction term ($soe\_busen$) and $busen$ to $PAT$ are both significantly positive at the 1% level. Moreover, a grouped regression analysis is conducted on samples comprising state-owned and non-state-owned enterprises, followed by Fisher’s combination test. Following 500 sampling tests to assess the coefficient variance between groups, the empirical p-value was 0.04, signifying significance at the 5% level. The findings suggest that state-owned property rights amplify the favorable influence of the business environment on enterprises’ green innovation.

7.2. Heterogeneity Analysis Based on Scientific and Technological Attributes

Enterprises with varying technological attributes exhibit disparities in their technological R&D capabilities and sensitivity to the business environment, potentially resulting in divergent effects of the business environment on green innovation within these enterprises. To examine the impact of the scientific and technological attributes of the sampled enterprises on the relationship between the business environment and green innovation, this study introduces the variable indicating whether they are classified as high-tech enterprises ($tech$) and incorporates it as a moderating factor into the regression model (7). The regression outcomes are presented in column (2) of Table 7, revealing that the regression coefficients of the cross-term $tech$_$busen$ term ($tech\_busen$) and $busen$ for $PAT$ are both significantly positive at the 1% significance level, signifying that high-tech attributes amplify the favorable influence of the business environment on enterprises’ green innovation. To further confirm these results, group regressions are conducted on both high-tech and non-high-tech enterprise samples, followed by Fisher’s combination test. Following 500 sampling tests assessing the coefficient disparities between groups, the empirical p-value was found to be 0.008, which is significant at the 1% level. This outcome reaffirms that optimizing the business environment has a significant effect on promoting green innovation within high-tech enterprises.

7.3. Heterogeneity Analysis Based on Industry Attributes

Whether an industry is polluting may have different impacts on the relationship between the business environment and green innovation of enterprises. On the one hand, with the introduction of environmental protection policies, heavily polluting enterprises will inevitably face a series of political and financing pressures [44]. On the other hand, in order to obtain support from the country’s green credit policy, enterprises may actively reduce pollution or engage in green innovation. Due to the characteristics of some industries, some sectors inevitably generate pollution during production and operation. According to the notice issued by the Ministry of Ecology and Environment of the People’s Republic of China, and drawing on the research method of He et al. [45], this paper defines the mining industries, such as coal and oil, and the production and supply of electricity and heat as heavily polluting industries. In this study, $poll$ is introduced into the model (7) as a moderating variable. The regression results, presented in column (3) of Table 7, show that the regression coefficient of the cross-term $poll$_$busen$ term ($poll\_busen$) to $PAT$ is significantly negative at the 1% level, indicating that the positive impact of business environment optimization on corporate green innovation is more pronounced in non-heavily polluting enterprises. This may be attributed to the challenges heavily polluting companies encounter in pursuing green innovation. In an optimized business environment, such companies might prioritize pollution reduction through production transformation or reduction rather than investing in green innovation. To enhance result robustness, this study conducts group regression analysis on samples from both high-tech and non-high-tech enterprises, followed by Fisher’s combination test. Following 500 sampling tests of the coefficient difference between groups, the empirical p-value was 0.034, indicating significance at the 5% level. This finding reaffirms that optimizing the business environment significantly fosters green innovation among non-heavy polluting enterprises.

8. Conclusions and Policy Implications

8.1. Conclusions

Green innovation is pivotal not only for the transformation of enterprises towards sustainability but also for the realization of the national green development agenda. Nevertheless, at the macroeconomic level, amidst the dual challenges of economic downturn and the imperative for pollution and emission reduction, further investigation into the impact and mechanisms of optimizing the business environment on green innovation within enterprises remains imperative. Consequently, this study employs Chinese A-share listed companies from Shanghai and Shenzhen as the research subjects covering the period from 2016 to 2020, conducting empirical analysis on the effects and mechanisms of optimizing the business environment across different provinces on enterprises’ green innovation. The results of this study indicate that (1) the optimization of the business environment significantly contributes to enhancing green innovation within enterprises. This conclusion remains valid even after subjecting it to robustness tests, including substituting business environment variables, eliminating alternative explanations, and employing lag, instrumental variable method, and Heckman two-stage regression to mitigate endogeneity concerns. (2) The mechanism by which the optimization of the business environment affects green innovation in enterprises is primarily manifested in three aspects. Specifically, the optimization of the business environment can facilitate the digital transformation of enterprises, thereby promoting green innovation. Secondly, the optimization of the business environment fosters green innovation in enterprises by boosting their innovation capabilities and willingness to engage in environmental protection. Finally, the optimization of the business environment not only facilitates the digital transformation of enterprises but also boosts their innovation capabilities and environmental management willingness, further promoting green innovation. (3) In the expansion test, this study additionally conducted a heterogeneity analysis across three dimensions: property rights, technology, and industry, revealing that optimizing the business environment has a more pronounced impact on fostering green innovation in state-owned enterprises, high-tech enterprises, and non-heavily polluting enterprises.

8.2. Policy Implications

Based on the above research, this paper draws the following policy implications:

(1)

Government perspective

First, improve the policy system and strengthen incentives and guidance; the government should formulate and improve a series of green innovation support policies. On the one hand, increase financial subsidies, give direct financial support to enterprises that carry out green innovation projects, and ease the financial pressure on enterprises to innovate. On the other hand, implementing preferential tax policies, such as corporate income tax reduction for enterprises with significant green innovation achievements, reduces the cost of enterprise innovation and encourages enterprises to participate in green innovation actively.

Secondly, strengthen infrastructure construction and provide a resource guarantee; increase investment in green innovation infrastructure, build green technology research and development centers, public testing platforms, etc., to provide hardware support for enterprises’ green innovation. At the same time, it is necessary to integrate innovative resources such as universities and scientific research institutions, build platforms for industry–university research cooperation, promote the transformation of scientific and technological achievements, and provide a technical and human resources guarantee for green innovation in enterprises.

Finally, improve the regulatory mechanism and create a fair environment; establish a sound green innovation regulatory mechanism, formulate strict environmental protection standards and green innovation norms, strengthen the supervision and inspection of the production and operation process of enterprises, and impose strict penalties on enterprises that do not meet green standards. At the same time, it strengthens the protection of intellectual property rights, cracks down on infringement, protects the legitimate rights and interests of green innovative enterprises, creates a fair, competitive market environment, and encourages enterprises to carry out green innovation actively.

(2)

From the perspective of enterprise managers

First, seize policy opportunities and formulate green strategies. The optimized business environment is often accompanied by more green-related policy support. Enterprise managers should pay close attention to policy trends and formulate enterprise green innovation strategies according to policy orientation. For example, they could use the government’s preferential tax policies for green industries, increase investment in green technology research and development, clarify the proportion of green products in the next few years, and guide enterprises to transform in the green direction.

Second, strengthen cooperation and exchanges to enhance innovation capacity. A good business environment will promote exchanges and cooperation between enterprises and between enterprises and scientific research institutions. Managers can actively organize and participate in industry green innovation seminars to share experiences and exchange ideas with peers. At the same time, establish industry–university research cooperation with universities and research institutes, introduce external advanced green technologies and concepts, and enhance the enterprise’s green innovation ability.

Finally, cultivate green talents and build innovative teams. In a high-quality business environment, the flow and training of talents will be smoother. Business managers should develop attractive talent policies to attract professionals in green innovation to join the company. In addition, strengthen the green training of internal employees and regularly offer training courses on green technology, environmental protection concepts, and other aspects of training courses to create a high-quality green innovation team.

(3)

Corporate investor perspective

First, channel funds to support green projects. Optimizing the business environment can provide investors with more information and investment channels. Investors can invest more in green innovation projects through investment decisions. For example, we will prioritize investing in enterprises committed to the research and development of new environmentally friendly materials and clean energy utilization and provide financial support to these enterprises to promote the development and application of green technologies.

Second, participate in corporate governance and promote green change. After investing in enterprises, investors can actively participate in corporate governance, transmit the concept of green development to the management of enterprises, and encourage enterprises to formulate green innovation strategies. For example, proposals such as increasing the green research and development budget and setting green production indicators are put forward to the board of directors to encourage enterprises to pay more attention to green innovation in the operation process.

Finally, encourage green investment to form industrial agglomeration. Investors can join other investors to launch green investment funds to attract capital into green innovation. Through investment, we will guide relevant enterprises to form a cluster effect in the region, promote the improvement and development of the green industry chain, and create an industrial ecological environment conducive to green innovation.

8.3. Research Limitations and Prospects

Sample selection level: Although we excluded non-manufacturing sample enterprises and enterprises registered in municipalities directly under the central government to make the research more targeted, it also resulted in a relatively narrow sample scope. Focusing only on specific types of enterprises cannot fully reflect the differences in the relationship between business environment and green innovation among enterprises in all industries and regions. For example, non-manufacturing enterprises may have unique green innovation models and influencing factors. In contrast, the driving factors of green innovation in enterprises in municipalities directly under the central government may be quite different from those in other regions due to the advantages of policies and resources. The information contained in these excluded samples cannot be reflected in the research, which limits the universality of the research conclusions. In addition, the sample data period is only from 2016 to 2020, which may not cover the impact of long-term factors such as economic cycle fluctuations and significant policy changes on corporate green innovation. This leads to research results limited in the time dimension.

Research methods: Although the endogeneity problem is addressed to some extent by combining Heckman’s two-stage regression model, the instrumental variable method, and multiple robustness testing methods, these methods rely on strict assumptions. For example, in selecting instrumental variables, although we try our best to find qualified instrumental variables, it is difficult to fully guarantee the correlation between instrumental variables and endogenous variables and independence from the error term in actual operation. The estimation results may be biased once the assumed conditions are not met. At the same time, although the robustness testing methods are diverse, they cannot exhaust all the factors that may affect the research results. For some potential interfering factors, the existing testing methods may not be able to identify and exclude them, thus affecting the reliability of the research results.