The Impact Mechanism of Digitalization on Green Innovation of Chinese Manufacturing Enterprises: An Empirical Study

Abstract

With the rapid development of the digital economy, promoting green innovation through digitalization has become an important means for manufacturing enterprises to improve their core competitiveness. However, the existing studies focus more on enterprise green technology innovation than green innovation, and the empirical tests mostly use regional-level data rather than enterprise-level data. This paper empirically examines the impact effect and mechanism of digitalization on green innovation in manufacturing enterprises using a sample of Chinese A-share listed manufacturing enterprises from 2013–2019. It is found that: digitalization significantly promotes the improvement of green innovation level in manufacturing enterprises; digitalization promotes green innovation more prominently in labor-intensive industries and manufacturing enterprises in central China than in capital- or technology-intensive industries and enterprises in eastern China; and digitalization can influence green innovation in manufacturing enterprises through three intermediary channels: promoting enterprise value chain upgrading, empowering industrial structure optimization, and enhancing technological innovation.

1. Introduction

In recent years, green innovation has become increasingly important as a way to address environmental challenges such as climate change, pollution, and natural resource depletion. Manufacturing is a significant contributor to greenhouse gas emissions and resource depletion, and sustainable practices must be adopted to reduce environmental impacts and improve competitiveness. Many governments have recognized the importance of green innovation and have implemented policies and initiatives to promote it, such as the EU’s Circular Economy Package and China’s policies and programs to promote green manufacturing, and the implementation of green practices by companies is essential to achieving a balance between economic growth, environmental protection, and social development [1].

The manufacturing industry is the foundation of a strong country. China’s manufacturing sector has experienced rapid growth over the past few decades, becoming the largest manufacturing country in the world. However, this growth has also led to significant environmental challenges. Public data shows that the manufacturing industry accounts for 36% of China’s total carbon emissions, and facing the current problems of high energy consumption, high pollution, and high emissions in China’s manufacturing industry, the development of green innovation in manufacturing enterprises has become an inevitable requirement and trend for China’s balanced economic development and resource environmental improvement. To achieve sustainable development in manufacturing, it then requires consideration of multiple dimensions such as economic, environmental, and social factors [2], and to address these challenges, green innovation is seen as a key driver to reduce the negative impact of manufacturing activities on the environment [3,4].

Meanwhile, the digital economy is the core driver of high-quality development within the manufacturing industry, and green innovation has become an inevitable requirement and trend for balanced economic development and resource and environmental improvement. The enhancement of digital technology can reduce the input cost of enterprise innovation activities, promote the transformation of enterprise technology development and management modes [5], improve the innovation capability and adaptability of enterprises, and then promote the development of green innovation. However, existing studies still need to strengthen empirical research on the impact of digitization on enterprise green innovation.

Therefore, this paper aims to explore the impact and mechanism of digitalization on green innovation in manufacturing enterprises through an empirical study based on the micro level. To achieve this goal, this paper is divided into three main research contents: first, to study the role of digitalization in promoting green innovation in manufacturing enterprises and to analyze its effect empirically. Second, to explore the heterogeneity of digitalization in promoting green innovation in manufacturing firms. Third, the impact mechanism of digitalization on green innovation in manufacturing enterprises is explored through three intermediate channels (value chain upgrading, industrial structure optimization, and technological innovation). The findings of this paper contribute to an in-depth understanding of the role and mechanism of digitalization on green innovation in manufacturing enterprises and provide valuable references and lessons for enterprises and policymakers.

The main marginal contributions of this paper are as follows: First, the current research mostly studies the influence of digitalization on green innovation of manufacturing enterprises from a macro perspective, and this paper discusses the relationship between them from a micro perspective, which provides a new research perspective for this research field; second, the comprehensive evaluation of green innovation of manufacturing enterprises is a supplement and improvement to the research only from the perspective of green technology; third, empirically test the influence of digitalization on green innovation of manufacturing enterprises from a micro perspective and explore and analyze three channels of action: value chain upgrading, industrial structure optimization, and technological innovation.

The following sections of this paper are organized as follows: Section 2 includes a comprehensive review of the literature; Section 3 presents theoretical analysis and research hypotheses; Section 4 outlines the research design, including model construction, variable design, and data sources; Section 5 consists of empirical results and analysis; and the final section introduces the conclusion and policy implications.

2. Literature Review

2.1. The Development Status of Digitalization

In this paper, the research on digital development at home and abroad is roughly divided into three stages: the germination period (before 2003), the development period (from 2003 to 2017), and the prosperity period (from 2018 to the present). In its infancy, the Internet, as a typical digital technology, achieved rapid development and began to be commercialized on a global scale, which promoted global economic progress and social development. With the rapid development of new network technology, the external environment of enterprises has changed greatly, which drives enterprises to carry out digital transformation. Tso et al. put forward an intelligent GM service system that uses computer networks and information technology, based on the reasoning ability of rules, to transform the work request from the client into a basic task, and intelligently coordinate and monitor the constant changes in customer requirements [6]. Jorgensen’s research shows that information technology has a sustained impact on productivity growth and also promotes economic growth [7].

In the development stage of digitalization, digital technology has gradually merged with economic development, and all countries have begun to vigorously promote the development of the digital economy. With the rapid development of digital technology and the changing policy and business environment, more and more scholars began to study digital business strategy and focused on the impact of digital technology innovation on industrial change. Lucas et al., taking Kodak Company as the research object, expanded the theory of disruptive innovation by using the method of case study and analyzed how the enterprise responded to the challenge that digital transformation technology threatened its historical business model [8]. Mithas et al. discussed the influence of IT duality on the company’s digital business strategy, determined the difference between digital business strategy and traditional business strategy, and further clarified the concept of digital business strategy [9].

During the boom period, the research related to digital development began to increase greatly, and the attention of researchers has shifted from simple conceptual issues to specific implementation issues, which has spawned many different research directions, such as digital empowerment technology innovation, enterprise transformation and upgrading, and high-quality development. Yan Tao et al. conducted an empirical analysis through provincial panel data. The research shows that digitalization can promote high-quality development through two intermediary channels: improving the level of scientific and technological innovation and promoting the upgrading of industrial structure, which provides a practical basis for the specific path to achieve high-quality economic development [10]. Wu L et al. found that resource integration is a bridge between digital capability and open innovation through a multi-case analysis of different types of manufacturing enterprises. Enterprises promote the identification and acquisition, resource matching, and utilization of digital resources through digital capabilities and further empower open innovation [11].

2.2. Green Innovation

In this section, 525 SSCI English journal papers from 2014–2021 are selected as research samples and explored by using bibliometric and knowledge mapping analysis methods, aiming to explore the current hot frontier topics of academic research in the field of green innovation around the world and the shortcomings of existing research so as to provide theoretical support for subsequent research. The purpose of this study is to explore the hot topics and shortcomings of current research in the field of green innovation and to provide theoretical support for subsequent research.

As can be seen from the analysis of Figure 1, green innovation research in countries around the world has formed a knowledge network co-occurrence map centered on four keywords: green innovation (262 times), performance (142 times), impact (127 times), and eco innovation (114 times), while management (89 times), research and development (68 times), and eco innovation (114 times) have formed a knowledge network co-occurrence graph, research and development (68 times), sustainability (68 times), empirical evidence (68 times), and environmental innovation (57 times) have also become hot topics around these four core keywords, and it can be seen from the analysis of high-frequency keywords that related research mainly focuses on the research of green innovation development, sustainability development, green innovation management, and green innovation influence factors.

From the above analysis, it can be found that other countries except China pay more attention to topics such as the benefits brought by the implementation of green innovation to the overall social and economic development of their enterprises. Furthermore, combined with the title, abstract, and content analysis of 525 sample literatures, the academic research on green innovation in other countries or regions is divided into the following aspects: (1) The impact of green innovation on the sustainable competitive advantage of organizations. For example, Tu Y and Wu W used structural equation modeling based on a sample of 235 Chinese manufacturers, tested the proposed theoretical model through empirical analysis, and found that green innovation has a positive impact on the competitive advantage of firms [12]. (2) Research on the influencing factors of green innovation. For example, Feng L et al. examined the influence of two dimensions of environmental orientation on two types of green innovation, and the moderating role of political ties. The findings showed that both internal and external environmental orientations can promote green product and green process innovation, and political ties positively moderated the positive impact of internal environmental orientations on green innovation while negatively moderating the positive impact of external environmental orientations [13]. (3) Research on enhancing green innovation performance. Wang M et al. constructed an economic performance transmission model for upgrading green technology innovation and found that both green process innovation and green product innovation can have significant promoting effects on firms’ economic performance, and firms can enhance their economic performance through two mediating paths of improving environmental performance and market competitiveness [14]. (4) Research on the relationship between carbon emissions, energy consumption, and green innovation. This has been a hot topic of research in recent years and is a frontier trend for future research. For example, Du K et al. studied the impact of green technology innovation on carbon dioxide emissions through panel data from 71 economies from 1996 to 2012. The study found that green technology innovation has a single threshold effect on CO₂ emissions for economies with different income levels. Specifically, green technology innovation significantly reduced CO2 emissions only when the income level was above the threshold [15].

2.3. The Relationship between Digitalization and Green Innovation

Many scholars have started to pay attention to and study the impact of digitalization on enterprise innovation. The empirical studies of Li R et al. and Radicic D et al. found that the improvement of digital technology level can promote the optimal allocation of enterprise resources and factors, reduce the input cost of innovation activities, promote enterprise technology research and development, promote management mode change, improve enterprise innovation capacity, absorption capacity and adaptability, improve enterprise innovation performance, and so on [16,17]. There are also studies that analyze the impact of digitization on green innovation. Luo et al. found that in China, the development of the digital economy can improve green innovation levels in indirect ways, such as by boosting the degree of economic openness, optimizing the industrial structure, and expanding the market potential [18]. Mubarak et al. investigated 217 manufacturing enterprises in Malaysia and found that Industry 4.0 positively impacts open innovation, which leads to green innovation behavior that further expands green innovation performance [19]. El-Kassar and Singh investigated the Middle East and North Africa (MENA) and Gulf Cooperation Countries (GCC) and found that digitalization can promote green innovation of enterprises, thus improving economic and environmental performance [20]. Fei, J. divides green innovation into three categories: green technology innovation, green system innovation, and green business model innovation [21]. Zhang Q et al. found that enterprise digitalization can promote enterprise green technology innovation through efficient information sharing and knowledge integration [22]. Li found that enterprise digitalization not only directly promotes green innovation but also promotes employee sharing and enterprise absorption by enhancing human capital, thus promoting green technology innovation [23].

As far as its mechanism is concerned, digital transformation mainly stimulates green innovation of enterprises through the following channels: From the outside of enterprises, digitalization can promote enterprises to increase government subsidies and reduce regulatory pressure, thus promoting green innovation; Internally, digitalization can enhance enterprises’ absorptive capacity [24], learning capacity [25], internal control capacity [26], innovative resource elements [27,28], and other channels to promote green innovation.

This paper reviews the relevant literature on digital development, green innovation, and the impact of digitalization on green innovation in countries around the world and finds that many scholars have conducted extensive research on the impact of digitalization and green innovation on enterprise performance and economic development from different perspectives, but few studies have analyzed the impact of digitalization on green innovation. As for the green innovation effect of digitalization, most scholars only conduct macro-level studies, mainly focusing on various regions, cities, or economic regions, and generally believe that digitalization can reduce innovation costs, improve resource allocation efficiency, and increase green innovation output. Existing studies on the impact of digitalization on enterprise green innovation are mainly theoretical studies and case studies, and few studies verify the specific impact of digitalization on enterprise green innovation from an empirical perspective. Moreover, the existing studies only carry out empirical analysis from the perspective of green technology, and enterprises need not only green technology but also green products and services to realize green innovation. Only from the level of green technology innovation can we study the impact of digitalization on enterprise green innovation. It is inevitable to miss important information from the sample.

3. Theoretical Analysis and Hypothesis

In the era of rapid development of digital technology, manufacturing enterprises can rely on the new generation of digital technology to achieve green transformation, upgrading, and high-quality sustainable development. In this paper, in order to investigate the specific influence mechanism of digitalization on green innovation of manufacturing enterprises, based on a large number of studies on digital development and enterprise green innovation at home and abroad, we introduce mediating variables and construct a conceptual model of the influence of digitalization on green innovation of manufacturing enterprises, as shown in Figure 2.

3.1. Digitalization and Green Innovation in Manufacturing Enterprises

Under the wave of rapid development of the digital economy, digitalization has brought significant empowering effects to manufacturing enterprises, driving changes in technology, products, and business models, broadening feasible space and development paths for better enterprise innovation activities [29], and playing the core function of achieving a balance between economic development and environmental pollution and driving innovation on the basis of sustainable development. The gradual improvement of digital infrastructure level obviously promotes the integration and diffusion of technological innovation, the increase in innovation frequency, the integration of innovation resource elements, the sharing of knowledge and information, the optimization and transformation of processing and manufacturing links, the rapid increase in R&D and design effectiveness and response rate [30], the reconstruction of energy consumption systems, the promotion of green and intelligent manufacturing of enterprises, and the promotion of “innovation-driven” measures. The new generation of information technology applications can be used to promote the efficient combination of “innovation-driven” and “green development” through digital technology [31]. The application of next-generation information technology can promote joint innovation and independent innovation of enterprises through the integration of digital technology, optimize the allocation of green innovation resources, reduce energy consumption, and thus promote the improvement of green innovation levels of enterprises [19].

Relying on digital technology, it can improve the information sharing level of enterprises’ application requirements, production processes, and other information in the innovation network, and the results of green innovation are rapidly shared to create new products and new models [22]. Enterprise digitalization can improve the integration of resource elements, reconfigure and layout the innovation network, improve input–output efficiency and environmental monitoring capability, and become a key link for enterprises to realize green manufacturing, which breaks through the range of traditional innovation intervals and promotes the improvement of green innovation level [32]. Therefore, this paper proposes the following research hypothesis:

Hypothesis 1. 

Digitalization promotes green innovation in manufacturing enterprises.

3.2. Mediating Effects of Digitalization for Green Innovation in Manufacturing Enterprises

During the 14th Five-Year Plan period, the national digitalization process accelerated comprehensively, and digitalization is an effective path to enhance the competitiveness of the manufacturing value chain in the new era. The integration of digital technology changes the spatial layout and value distribution of each link in the value chain, promotes more convenient information transfer in R&D, design, manufacturing, sales, and service, reduces information asymmetry, expands digital network connectivity, reduces the difficulty of connecting the value chain [33], and promotes the upgrade of value chain process connection. Digital technology embedded in each link can reduce the search cost of information, verification cost, and transportation cost of information goods, significantly reduce the transaction cost of enterprises [34], and promote the upgrade of cost reduction in each link of the value chain through a cost-saving effect. Digital technology applications can change the way enterprises embed themselves in each link of the value chain, help enterprises carry out digital transformation for new value creation, provide whole-process closed-loop overall solutions to outstanding problems, form closed-loop value creation, and promote the upgrade of value creation in each process of the value chain. Digitization can provide the necessary data elements for each link of the value chain, accelerate the absorption of digital technology in each link, facilitate the diffusion of innovative knowledge in the field, promote the optimal allocation of manufacturing innovation resources [35], enhance the efficiency of resource allocation, and promote the optimization and upgrading of the configuration of each link of the value chain.

From the perspective of different stages of value chain upgrading [36], digitalization can promote process upgrading, product upgrading, function upgrading, and chain upgrading, and gradually realize the climb of the value chain. Digitization can effectively promote the integration and connection of all links of the value chain in the green innovation process, reduce the innovation cost reduction of the green innovation value chain, effectively integrate green innovation network resource elements, develop green low-carbon technologies and services, and promote enterprises to make full use of digital resources to organize and develop green products and form green innovation value creation. Digital development is conducive to enhancing the digital, intelligent, green, and collaborative capabilities of each link of the value chain, forming a new type of intelligent, interconnected, agile, and green value chain that can better promote green innovation activities. Therefore, this paper puts forward the following hypothesis:

Hypothesis 2. 

Digitalization promotes green innovation in manufacturing enterprises via value chain upgrading.

In the digital era, the servitization of manufacturing is a key choice for manufacturing industries to gain competitiveness, and the essence is to shift the center of the value chain from manufacturing to value-added services to achieve value-added manufacturing benefits [37]. In the process of digital development of enterprises, emerging digital technologies are applied to all aspects of R&D and design, manufacturing, and marketing services, which can realize the sharing and exchange of data and information among the internal networks of enterprises [38], and the use of technology and service innovation can help enterprises make timely feedback to customer needs and promote the innovation of enterprise business models [39], enhance the environmental resilience, organizational management ability, maintain customer relationship and the ability to create customer value [40], enhance the service level of manufacturing industry, and then promote the extension of manufacturing industry to low-energy and high-value industries and the optimization and upgrading of industrial structure [41], reduce the resource consumption of enterprises and improve the possibility of development and innovation.

The development of digitization, on the one hand, promotes the optimization and upgrading of industrial structures by transforming traditional industries. Digital technology has given birth to new digital models and new business models, realized the deep integration of digital technology and traditional industries, updated the key technologies of traditional industries, optimized the traditional business processes, improved the industrial system, enhanced production efficiency, and promoted the transformation of traditional industries [42]. Digitalization accelerates the speed of information transfer within enterprises through digital technology applications, expands access to knowledge, promotes the transformation of traditional manufacturing industries to intelligent production [43], provides kinetic energy for industrial structure optimization, and further empowers the development of green innovation in enterprises. On the other hand, digitization promotes the formation of new industries to drive the optimization and upgrading of industrial structures. Through industrial digitization, i.e., the deep integration of data elements in various traditional industries, which promotes digitalization and intelligent transformation and upgrading [44], emerging technologies have also given rise to new industrial models and production methods and promoted new industrial development fields [45], such as green products and services, including intelligent logistics, sharing economy, network finance, intelligent manufacturing, etc. The green development of traditional industries will enhance market competitiveness and shift the manufacturing value chain from the low-end level to the middle and high-end level, which is more likely to promote the optimization and upgrading of industrial structure, create a complete digital green industrial chain, and then promote the research and development of green technologies and products of enterprises and improve the level of green innovation. Therefore, this paper puts forward the following hypothesis:

Hypothesis 3. 

Digitalization promotes green innovation in manufacturing enterprises via industrial structure optimization.

In the era of the digital economy, digital technologies such as cloud computing and artificial intelligence have reduced the cost of enterprises carrying out innovation activities. Enterprises can analyze the needs of consumer groups with big data and carry out targeted innovation, thus gaining cost advantages [37]. In terms of human capital, the deep penetration and application of digital technology will increase the demand for high-quality human capital, attract an influx of highly skilled personnel, continuously optimize the human capital structure, and promote the improvement of enterprise technological innovation. The application of digital technologies promotes faster information flow within enterprises, faster diffusion of innovation results, and more transparency in all aspects, which is conducive to technological research and development [30]. At the same time, in inter-enterprise R&D collaboration, the application of digital technology breaks the barriers of information flow between enterprises, reduces the time lag of information transmission, promotes rapid and efficient integration of innovation resources such as knowledge and technology, makes inter-enterprise cooperation and connection closer, enhances inter-enterprise R&D collaboration, and promotes the improvement of enterprise technology innovation level. The development of digitalization breaks through the limitations of time and space in carrying out innovation activities, which can encourage more innovation subjects to participate together [37]. Technological innovation is the core driver of innovation activities in enterprises, especially green innovation. With the national goal of “double carbon”, in the competitive market environment, green technology innovation is advantageous in saving energy and reducing costs, which is conducive to manufacturing enterprises gaining competitive advantages in the market and can enhance their willingness to embrace green innovation.

The application of digital technology is conducive to promoting the optimal allocation of innovative technology resources [46], facilitating the integration of digital technology into various links, improving the technology integration capability of enterprises, promoting better integration of green innovation resource elements, driving enterprises to carry out green technology innovation based on existing technologies, amplifying the interaction relationship between old and new technologies, and promoting the generation of new green technologies [47]. The R&D innovation of green low-carbon technologies is the key to solving the problem of green development in the manufacturing industry, and the R&D of green technologies further promotes enterprises to improve green innovation. Therefore, the following hypothesis is proposed in this paper:

Hypothesis 4. 

Digitalization promotes green innovation in manufacturing enterprises via technological innovation.

4. Data and Methods

4.1. Model Construction

According to the theoretical mechanism analysis in this paper, in order to verify the above research hypothesis and to empirically analyze the direct mechanism of the effect of digitalization on green innovation in manufacturing, this paper directly draws on (Fengzheng Wang et al., 2021) the model construction approach of digitalization affecting green technological innovation in enterprises [47], setting up the following basic econometric model:

$$\begin{array}{c}{\mathrm{GI}}_{\mathrm{i},\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{DIG}}_{\mathrm{i},\mathrm{t}}+{\displaystyle \sum }_{\mathrm{j}}{\mathsf{\beta }}_{\mathrm{j}}{ \mathrm{Controls}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}} \end{array}$$

(1)

In Model (1), the explanatory variable ${\mathrm{GI}}_{\mathrm{i},\mathrm{t}}$ are $\mathrm{i}$ firm’s green innovation level in $\mathrm{t}$ year, the green innovation level, the core explanatory variable ${\mathrm{DIG}}_{\mathrm{i},\mathrm{t}}$ is the digitization level of the region where the manufacturing firm is located, and the parameter ${\mathsf{\beta }}_1$ portrays ${\mathrm{DIG}}_{\mathrm{i},\mathrm{t}}$ on the explanatory variable ${\mathrm{GI}}_{\mathrm{i},\mathrm{t}}$ the effect of ${\mathrm{Controls}}_{\mathrm{i},\mathrm{t}}$ denotes the firm-level control variables that may affect the green innovation of manufacturing firms, and ${\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$ denotes the random disturbance term. To control for the effects of macro and industry factors on the explanatory variables, this paper also includes region, industry, and year fixed effects in the model.

In order to test the indirect mechanism of action between digitalization and green innovation in manufacturing, this paper draws on the mediating effect model testing mechanism of Song D.Y. et al. to analyze and set up the following model by stepwise method [48]:

$$\begin{array}{c}{ \mathrm{MED}}_{\mathrm{i},\mathrm{t}}=\alpha +{\mathsf{\beta }}_0{\mathrm{DIG}}_{\mathrm{i},\mathrm{t}}+{\displaystyle \sum }_{\mathrm{j}}{\mathsf{\beta }}_{\mathrm{j}}{ \mathrm{Controls}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}} \end{array}$$

(2)

$$\begin{array}{c}{ \mathrm{GI}}_{\mathrm{i},\mathrm{t}}=\alpha +{\mathsf{\beta }}_0{\mathrm{MED}}_{\mathrm{i},\mathrm{t}}+{\displaystyle \sum }_{\mathrm{j}}{\mathsf{\beta }}_{\mathrm{j}}{ \mathrm{Controls}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}\end{array}$$

(3)

$$\begin{array}{c}{ \mathrm{GI}}_{\mathrm{i},\mathrm{t}}=\alpha +{\mathsf{\beta }}_0{\mathrm{DIG}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_1{\mathrm{MED}}_{\mathrm{i},\mathrm{t}}+{\displaystyle \sum }_{\mathrm{j}}{\mathsf{\beta }}_{\mathrm{j}}{ \mathrm{Controls}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}} \end{array}$$

(4)

Among them, MED contains three mediating variables: value chain upgrading, industrial structure optimization, and technological innovation, and the rest of the variables are consistent with Model (1).

4.2. Variable Design

4.2.1. Explained Variables

The level of green innovation (GI) in manufacturing companies. Green innovation is an innovation that balances economic and environmental benefits with product innovation, technological innovation, conceptual innovation, and institutional innovation [49]. The explanatory variable of this paper is the level of green innovation of manufacturing enterprises, and at present, there are two main methods to measure the level of green innovation: one is to use the number of green patents to measure it; the other is to use DEA and other methods to measure the efficiency of green innovation. The single indicator of green patent output to measure the green innovation level of manufacturing enterprises is not comprehensive enough, so this paper adopts the green innovation efficiency of manufacturing enterprises to measure the green innovation level. Green innovation efficiency needs to have both green and innovation attributes, emphasizing both the reduction of pollution in the whole process and the level of transformation of results generated by unit resource input. Therefore, this paper further highlights the technical inputs and outputs of green innovation based on the study of Cao Ling et al. (2022) in order to reflect its connotation characteristics [50] and constructs a green innovation efficiency index system for manufacturing enterprises that includes both non-desired and desired outputs, as shown in

Table 1. Evaluation index system of green innovation efficiency of manufacturing enterprises.

Indicators	Tier 1 Indicators	Secondary Indicators	Tertiary Indicators	Unit
Green innovation efficiency of manufacturing companies	Inputs	Capital Investment	R&D capital stock New product development expenses	Million yuan Million yuan
		Labor input	R&D personnel full time equivalent	People/year
		Technical input	Technology introduction and renovation expenses	Million yuan
		Energy input	Total energy consumption	Ton
	Desired output	Technology Research and Development Stage	Number of Green Patent Applications	Pieces
		Results transformation stage	New product sales revenue	Million yuan
	Non-desired outputs	Environmental pollution	Total industrial wastewater discharge Industrial sulfur dioxide emissions General industrial solid waste	Million tons Billion standard cubic meters Million tons

below. A non-radial, constant payoff to scale (CRS) super-efficient SBM-DEA model is used to measure the green innovation efficiency of manufacturing enterprises, and the final comprehensive level of green innovation of enterprises is obtained.

4.2.2. Explanatory Variables

The core explanatory variable in this paper is digitalization. There is no consensus on the connotation and measurement of digitalization. Wang, FengZheng et al. classified digitalization into digital foundation, digital talent, and digital application [47]. The China Digital Economy Development Index published by Tencent Research Institute points out that the digital economy includes two major parts: digital industrialization and industrial digitization. Digital industrialization is also called the digital economy foundation. Industrial digitization is the increase in output and efficiency brought by the used sector. Therefore, considering that digitization mainly includes digital infrastructure, digital talent pool, and digital technology application, referring to the studies of Han Lu et al. and Fan Shengyue et al., a digital level indicator system containing nine three-level indicators was constructed based on the scientific and comprehensive nature of the indicator system and the availability of data [46,51]. As shown in

Table 2. Digitalization level index system.

Tier 1 Indicators	Secondary Indicators	Tertiary Indicators
Digital infrastructure	Basic level of communication	Fiber optic cable density
		Cell phone base station density
	Internet base level	Internet access port density
		Cell phone penetration rate
Digital talent pool	Human resource Level	Percentage of information technology employees
	Cultural and educational level	Local financial expenditure on education per capita (yuan/person)
Digital technology applications	Enterprise digital applications	Number of websites per 100 companies
		Number of computers used by enterprises as a percentage
	Enterprise digital trading	Percentage of e-commerce enterprises

below, the comprehensive score of the digital level is calculated according to the entropy weight method to determine the index weights.

4.2.3. Intermediate Variables

(1)

Value chain upgrading (Vcu). This paper mainly analyzes the connotation of the value chain and the key to value chain upgrading for manufacturing enterprises based on the concept of the value chain and smile curve theory. The upgrading level of the manufacturing value chain mainly includes processing and manufacturing capability, R&D and design capability, and branding and marketing capability. Based on Wang Hongqing and Hao Wenwen’s research and considering the availability of data, this paper constructs an evaluation index system for manufacturing value chain upgrading [52], which contains six specific indicators. Specifically, Table 3 below, also according to the entropy weight method, determines the weight evaluation of the comprehensive level of getting the value chain upgraded.

(2)

Industrial structure optimization (Strug). With the development of industrialization, the trend of industrial structure optimization in the middle and late development of manufacturing enterprises will be dominated by high-end technology industries. Therefore, this paper directly draws on the method of Fu Yuanhai et al. [53] to measure the level of industrial structure optimization in the manufacturing industry by the proportion of the total output value of the high-end technology manufacturing industry to the total output value of the mid-range technology manufacturing industry in order to highlight the change trend in the high-end technology industry.

(3)

Technological innovation (Tec). Technological innovation is the key to the development of green innovation and is one of the important influencing factors in improving green and sustainable development. Existing studies mention two measures of technological innovation: the number of patents granted and the total number of patent applications. Patent applications in China go through a long process before they are approved, and the total number of patent applications can reflect a company’s willingness and motivation to engage in technological innovation in a timely and accurate manner compared to the number of patents granted. In this paper, the technological innovation index is measured by the number of patent applications of listed companies.

Table 3. Evaluation index system of manufacturing value chain upgrading.

Evaluation Objectives	Indicator Type	Indicators	Indicator Meaning
Upgrading the value chain of manufacturing companies	Processing and manufacturing capabilities	Cost Margin	Total profit/total cost
		Energy output capacity	Total energy consumption/operating revenue
	R&D and design capabilities	Intangible assets ratio	Intangible assets/total assets
		New product output rate	New product sales revenue/operating revenue
	Branding and marketing capabilities	Net sales margin	Net profit/sales revenue
		Profitability	Net income/average balance of shareholders’ equity

Table 3. Evaluation index system of manufacturing value chain upgrading.

Evaluation Objectives	Indicator Type	Indicators	Indicator Meaning
Upgrading the value chain of manufacturing companies	Processing and manufacturing capabilities	Cost Margin	Total profit/total cost
		Energy output capacity	Total energy consumption/operating revenue
	R&D and design capabilities	Intangible assets ratio	Intangible assets/total assets
		New product output rate	New product sales revenue/operating revenue
	Branding and marketing capabilities	Net sales margin	Net profit/sales revenue
		Profitability	Net income/average balance of shareholders’ equity

4.2.4. Control Variables

Based on the existing literature, a series of control variables that may affect the green innovation of enterprises are included [32,34]. The control variables are firm size (Size), age at IPO (Age), whether or not a state-owned enterprise (Sta), proportion of fixed assets (Fap), equity concentration (Own), financial leverage (Lev), firm growth capacity (Gro), firm operating capacity (Ope), and firm profitability (ROA). The specific definition of each variable is shown in

Table 4. Variable Definitions.

Variable Type	Variable Name	Variable Symbols	Variable Definition
Explained variables	Green innovation level of manufacturing enterprises	GI	Corporate green innovation efficiency value
Explanatory variables	Integrated level of digitalization	DIG_tot	Company location Integrated level of digitalization
Intermediate variables	Value chain upgrade	Vcu	Value chain upgrade index
	Industrial structure optimization	Strug	Total output value of high-end technology manufacturing industry in the region where the enterprise is located/total output value of mid-range technology manufacturing industry
	Technology innovation	Tec	Ln (number of patent applications + 1)
Control variables	Enterprise size	Size	Natural logarithm of total assets
	Corporate listing age	Age	Ln (current year − year of launch)
	Whether state-owned enterprises	Sta	If the state-owned enterprise is recorded as 1, the other is recorded as 0
	Percentage of fixed assets	Fap	Net fixed assets/total assets
	Shareholding concentration	Own	Shareholding ratio of the top 10 largest shareholders of the company
	Financial leverage	Lev	Total liabilities/owner’s equity
	Business growth capability	Gro	Operating income for the year/operating income for the previous year − 1
	Business operating capacity	Ope	Operating income/average current assets
	Corporate profitability	ROA	Net profit/total assets

.

4.3. Data Sources

In view of the accessibility of data on relevant indicators of digitalization and green innovation, as well as the fact that the green innovation output index needs to lag one period, the sample data interval of the empirical measurement in this paper is 2013–2019. The data sources used in the study include: (1) The data for the measured indicators in the digitalization level evaluation index system are obtained from the 2013–2019 China Science and Technology Statistical Yearbook and China Statistical Yearbook, which are manually organized according to the data. (2) The province-level data used in the green innovation efficiency index system of manufacturing enterprises are all from the 2013–2019 China Science and Technology Statistical Yearbook, China Statistical Yearbook, and China Energy Statistical Yearbook, and due to the serious data deficiency in Qinghai and Tibet, the data of 29 provinces and regions in the country except Qinghai and Tibet from 2013–2019 are finally collected as samples. The data on green patents of listed companies were obtained from the China Research Data Service Platform (CNRDS). (3) Other corporate-level relevant data for the study design are obtained from CSMAR and Wind databases and listed companies’ disclosure reports, and some missing years’ data are completed by the interpolation method. Excluding ST, *ST, and enterprises with incomplete data, the data of 158 listed manufacturing enterprises nationwide from 2013–2019 were finally selected through the latest statistical data mentioned above.

5. Empirical Results and Analysis

5.1. Descriptive Statistics of Variables

The results of descriptive statistics for each variable are calculated and shown in

Table 5. Descriptive statistics of variables.

Variables	Obs	Mean	Std. Dev.	Min	Max
GI	1106	0.647	0.379	0.083	1.686
DIG_tot	1106	0.245	0.129	0.038	0.811
Vcu	1106	0.054	0.068	0.005	0.516
Strug	1106	1.788	1.318	0.167	7.443
Tec	1106	1.865	1.554	0	7.519
Size	1106	22.615	1.285	19.138	26.366
Age	1106	2.481	0.577	0.693	3.332
Sta	1106	0.518	0.500	0.000	1.000
Fap	1106	0.307	0.147	0.002	0.808
Own	1106	55.160	13.505	10.57	88.470
Lev	1106	0.448	0.198	0.016	1.352
Gro	1106	0.170	0.852	−0.913	23.998
Ope	1106	1.501	0.927	0.194	6.709
ROA	1106	0.032	0.063	−0.326	0.466

below. According to the table, it can be seen that the overall level of green innovation in China’s manufacturing enterprises is relatively low, and there is a large gap between enterprises. The mean and standard deviation of regional digitalization-integrated water are 0.245 and 0.129, respectively, which indicates that there is also a certain gap in the digitalization degree of different regions in China. The mean value of the manufacturing value chain upgrading index is relatively small, indicating that the overall level of the manufacturing value chain is reduced. The mean value of regional industrial structure optimization and upgrading water is 1.788, and the standard deviation is 1.318, indicating that there is a large difference in the level of industrial structure optimization and upgrading among regions. The mean value of enterprise technology innovation water is 1.865, and the standard deviation is 1.554, indicating that there are some differences in the technological innovation ability of each enterprise. In addition, the largest standard deviation is the concentration of equity, which indicates that there is a great difference in the shareholding ratio of the top ten shareholders of the enterprises in the sample.

5.2. An Empirical Test of Digitalization Affecting Green Innovation in Manufacturing Enterprises

In order to test the direct effect of digitization level on green innovation of manufacturing enterprises, the paper estimated the effect by the constructed econometric model (1), and the test results are shown in

Table 6. Baseline regression results.

Variables	(1)	(2)	(3)	(4)
DIG_tot	1.456 ***	1.493 ***	1.608 ***	1.678 ***
	(0.076)	(0.080)	(0.106)	(0.111)
Constant	0.289 ***	1.519 ***	0.468 ***	1.355 ***
	(0.021)	(0.202)	(0.087)	(0.214)
Control variables	NO	Yes	NO	Yes
Industry/time /regional effect	NO	NO	Yes	Yes
N	1106	1106	1106	1106
R-squared	0.248	0.286	0.432	0.449

Note: (1) *** denote significance levels of 0.01. (2) Standard errors of coefficients are in parentheses.

, which adopts a progressive regression treatment, such as column (1) in the table, which indicates the univariate regression without adding industry/time/region fixed effects and control variables, and column (2) in the table, which indicates the regression with relevant control variables added on top of column (1). The second, as in column (3) of the table, indicates a univariate regression with industry/time/region fixed effects, while column (4) of the table indicates a regression with relevant control variables in addition to column (3). The results show that the coefficients of the digitally integrated level are all positive and significant at the 1% level, and the test results are very robust. The results show that digitalization plays a certain role in promoting green innovation in manufacturing enterprises, and Hypothesis 2 was verified. Digital technologies such as big data and cloud computing break the traditional way of integrating innovation resources, change the innovation model, and promote the integration of innovation factors through the advantages of low diffusion costs and high diffusion speeds. Relying on digital technologies, manufacturing enterprises can build innovation networks to achieve efficient information sharing and obtain certain technology spillover effects, thus enabling efficient allocation of their innovation resources and providing assurance for green innovation.

5.3. Endogeneity Test

(1)

Controlling for possible endogeneity problems caused by omitted variables. In this paper, the fixed-effects model is used to revalidate Hypothesis 1. From the results, it is seen that the regression coefficient of the econometric model is still significantly positive after controlling for individual enterprise fixed effects, and Hypothesis 1 is further tested—that the comprehensive level of enterprise digitalization helps to promote green innovation in manufacturing enterprises.

(2)

Controlling for possible endogeneity caused by bidirectional causality. On the one hand, digitalization can enhance the information sharing and knowledge integration of enterprises to promote green innovation activities; on the other hand, enterprises need more technical support for green innovation, which drives the development process of digital technology. In

Table 7. Regression results of instrumental variables.

Variables	(1)	(2)
	DIG_tot	GI
DIG_iv	0.041 ***
	(8.44)
DIG_tot		3.836 ***
		(7.57)
Constant	−0.388 ***	1.709 ***
	(−6.13)	(6.63)
Control variables	Yes	Yes
Industry/time/region effect	Yes	Yes
Kleibergen–Paap rk LM statistics	80.451 ***
Cragg–Donald Wald F statistic	71.283 [16.38]
Hausman test	p = 0.000
N	1106	1106
R-squared	0.664	0.253

Note: (1) *** denote significance levels of 0.01. (2) Standard errors of coefficients are in parentheses.

, p < 0.05 of the Hausman test rejects the original hypothesis, indicating that the explanatory variable DIG_tot is an endogenous explanatory variable. In order to avoid endogeneity caused by the two-way causality between green innovation of manufacturing enterprises and the digitalization level of enterprises, this paper draws on Zhao Tao et al. [54] and uses the interaction term between the number of telephone sets per 10,000 people in each provincial area at the end of 1984 and the number of Internet broadband access subscribers (time-dependent) in the sample corresponding to the previous year as the instrumental variable of the core explanatory variable. Two conditions of the instrumental variable are satisfied: correlation and exogeneity. The logic lies in the following: on the one hand, digital development relies on the development of information technology, and regions with a better level of telecommunications infrastructure historically tend to be those with faster digital development now, satisfying to a certain extent the relevance requirement of the instrumental variable; on the other hand, with the emergence of new information technology that replaces traditional communication tools, it is almost impossible to influence the development of green innovation in modern manufacturing enterprises, satisfying the exogeneity. Considering that this paper is panel data, the panel instrumental variables are constructed by giving certain time trends to the cross-sectional data by means of interaction terms.

In this paper, the 2SLS model is used to reduce the effect of endogeneity on the model results, and the test results are shown in Table 7. Column (1) in the table shows the test for the instrumental variable (DIG_iv), and the results show that the regression results for the instrumental variable (DIG_iv) in the first stage are significantly positive at the 1% level, indicating that the explanatory variables are significantly correlated with the instrumental variables. Column (2) in the table shows the results of the second stage, and it can be seen that the regression coefficient of the digitally integrated level (DIG_tot) is still significantly positive, which is consistent with the estimated results of the baseline model, and it passes the LM test for non-identifiable instrumental variables and the F test for weak instrumental variables, respectively. In summary, the promotion effect of digitalization on green innovation in manufacturing firms remains robust after the reverse causality of explanatory variables is controlled.

5.4. Robustness Test

To ensure the robustness of the estimation results of the econometric model, the following three robustness tests are used in this paper:

(1)

Substitution of explanatory variable indicators. In this paper, the regression results of the model were further tested by using the traditional DEA-SBM model to measure the green innovation efficiency of manufacturing enterprises and replacing the variables according to the measured efficiency value (GRDEA) as a proxy variable for the explanatory variables. The regression results are presented in column (1) of Table 8 below, and the results show that the estimated coefficients are significantly positive, indicating that after replacing the new explanatory variables, the level of digitization is still significantly and positively correlated with the level of green innovation of manufacturing enterprises, and the obtained regression results are consistent with the previous benchmark results, indicating that the research findings of this paper are robust.

(2)

Replacement of core explanatory variable indicators. In this paper, we use the digital transformation degree indicators of listed companies, i.e., the total frequency of subdivision indicators of artificial intelligence technology, blockchain technology, cloud computing technology, big data technology, and digital technology applications appearing in the annual reports of manufacturing listed companies (DIG_trs), as the replacement variables of the digital integrated level of enterprises, according to the statistics of the Guotaian database, to further test the digital integrated level on the manufacturing enterprises’ The regression results are shown in column (2) of Table 8 below, and the results indicate that after replacing the new core explanatory variables, digitalization still has a robust promotion effect on the green innovation level of manufacturing enterprises.

(3)

Heterogeneity analysis. The development of green innovation in manufacturing enterprises may vary according to their geographical location and the industry to which they belong, and their enterprise digitization levels may also vary, which may lead to a differential impact of digitization levels on green innovation in manufacturing enterprises. Therefore, this paper analyzes the heterogeneity of digitalization on green innovation of manufacturing enterprises from the industry and the region where the enterprises are located.

Table 8. Robustness test results for replacing the core explanatory and explanatory variables.

Variables	(1)	(2)
	GRDEA	GI
DIG_tot	1.343 ***
	(0.091)
DIG_trs		0.045 ***
		(0.012)
Constant	1.194 ***	1.130 ***
	(0.175)	(0.235)
Control variables	Yes	Yes
Industry/time/region effect	Yes	Yes
N	1106	1106
R-squared	0.450	0.340

Note: (1) *** denote significance levels of 0.01. (2) Standard errors of coefficients are in parentheses.

Table 8. Robustness test results for replacing the core explanatory and explanatory variables.

Variables	(1)	(2)
	GRDEA	GI
DIG_tot	1.343 ***
	(0.091)
DIG_trs		0.045 ***
		(0.012)
Constant	1.194 ***	1.130 ***
	(0.175)	(0.235)
Control variables	Yes	Yes
Industry/time/region effect	Yes	Yes
N	1106	1106
R-squared	0.450	0.340

Note: (1) *** denote significance levels of 0.01. (2) Standard errors of coefficients are in parentheses.

• Industry heterogeneity.

Drawing on the guidelines of relevant studies on the division of labor-intensive and technology-intensive degrees of manufacturing industries, the 28 types of manufacturing industries are divided into labor-intensive, capital-intensive, and technology-intensive manufacturing industries [55]. The heterogeneity regression results are obtained as shown in columns (1)–(3) in Table 9 below; it can be seen that the regression coefficients of digitalization level on the green innovation level of manufacturing enterprises in all three industries are significantly The regression coefficients are positive, and the promotion effect of digitalization level on the green innovation of labor-intensive manufacturing enterprises is more obvious than that of capital- and technology-intensive manufacturing enterprises. This indicates that labor-intensive manufacturing enterprises currently pay more attention to digital transformation and digital technology application due to labor costs and new competitive advantages, which leads to a more significant impact of digitalization on green innovation in manufacturing enterprises.

2.

Regional heterogeneity.

According to the provinces where the enterprises are located, the regions where the enterprises are located are divided into eastern, central, and western regions. The heterogeneity regression results are obtained, as shown in columns (4)–(6) in Table 9 below, and it can be seen from the results that the impact of digitalization level on green innovation of manufacturing enterprises in the western region is not significant, probably because the digital infrastructure in the western region is weaker and it is more difficult for enterprises to integrate digital technology, which leads to a low degree of digital transformation of enterprises and has no significant impact. As for the eastern region and the central region, the effect of digitalization on the green innovation of manufacturing enterprises in the central region is greater, probably because the technological level and green innovation level of enterprises in the eastern region have been higher, with less room for progress, and the impact of digitalization technology has entered the stage of diminishing marginal benefits and has not had a greater impact; while the technological and economic levels in the central region are relatively backward, and the state has strengthened in recent years investment in green innovation resources for its manufacturing industry, coupled with the integration of digital technology, has led to an increase in resource allocation efficiency and promoted the improvement of green innovation level, even higher than that of the eastern region.

Table 9. Heterogeneity test results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Labor-Intensive	Capital Intensive	Technology Intensive	Eastern Region	Central Region	Western Region
DIG_tot	1.931 ***	1.551 ***	1.526 ***	1.806 ***	4.301 ***	0.695
	(0.554)	(0.218)	(0.140)	(0.112)	(0.721)	(0.848)
Constant	−0.318	2.116 ***	0.528 **	1.529 ***	1.591 ***	−0.852 *
	(0.815)	(0.395)	(0.264)	(0.280)	(0.526)	(0.435)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes
Industry/time/region effect	Yes	Yes	Yes	Yes	Yes	Yes
N	133	334	639	595	259	252
R-squared	0.645	0.406	0.501	0.543	0.390	0.362

Note: (1) ***, **, * denote significance levels of 0.01, 0.05, and 0.1, respectively. (2) Standard errors of coefficients are in parentheses.

Table 9. Heterogeneity test results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Labor-Intensive	Capital Intensive	Technology Intensive	Eastern Region	Central Region	Western Region
DIG_tot	1.931 ***	1.551 ***	1.526 ***	1.806 ***	4.301 ***	0.695
	(0.554)	(0.218)	(0.140)	(0.112)	(0.721)	(0.848)
Constant	−0.318	2.116 ***	0.528 **	1.529 ***	1.591 ***	−0.852 *
	(0.815)	(0.395)	(0.264)	(0.280)	(0.526)	(0.435)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes
Industry/time/region effect	Yes	Yes	Yes	Yes	Yes	Yes
N	133	334	639	595	259	252
R-squared	0.645	0.406	0.501	0.543	0.390	0.362

Note: (1) ***, **, * denote significance levels of 0.01, 0.05, and 0.1, respectively. (2) Standard errors of coefficients are in parentheses.

5.5. A Test of the Mediation Effect of Digitalization on Value Chain Upgrade

The value chain of manufacturing enterprises includes value links such as R&D and design, manufacturing, sales and service, etc. The emergence and integration of emerging digital technologies have served as a new engine and source of dynamic energy for the transformation and upgrading of the manufacturing value chain. The digital integration of the manufacturing value chain can drive the greening and synergization of the stages of element integration, technology development, and result transformation and enhance the green innovation level of manufacturing enterprises. In view of this, this paper adopts the mediating effect model that digitalization promotes green innovation of manufacturing enterprises by influencing the upgrading of the manufacturing value chain already constructed in the previous paper. The indicators of manufacturing value chain upgrading identified in the previous paper are selected as mediating variables to test whether digitalization promotes green innovation in manufacturing enterprises through promoting manufacturing value chain upgrading.

The results of the regression test are shown in

Table 10. Test results of mediating effects of digitalization for value chain upgrading.

Variables	(1)	(2)	(3)
	GI	Vcu	GI
DIG_tot	1.678 ***	0.184 ***	1.573 ***
	(0.111)	(0.022)	(0.114)
Vcu			0.567 ***
			(0.152)
Constant	1.364 ***	−0.115 ***	1.429 ***
	(0.215)	(0.043)	(0.214)
Control variables	Yes	Yes	Yes
Industry/time/ regional effects	Yes	Yes	Yes
N	1106	1106	1106
R-squared	0.449	0.306	0.456

Note: (1) *** denote significance levels of 0.01. (2) Standard errors of coefficients are in parentheses.

below. Column (1) in the table is the first step test, and the estimated coefficient of digitization integrated level (DIG_tot) on green innovation (GI) of manufacturing companies is positive and significant at the 1% level. Column (2) in the table is the second test step, and the results show that the coefficient of DIG_tot is significantly positive at the 1% level, indicating that digitalization effectively promotes the upgrading of the manufacturing value chain. The estimated coefficient of manufacturing value chain upgrading (Vcu) is significantly positive at the 1% level. The third column shows that the impact of digitalization on green innovation of manufacturing enterprises is significantly positively correlated at the 1% level under the mediating role of value chain upgrading, and the impact of both decreases from the original 1.678 to 1.573, indicating that value chain upgrading plays a partly mediating role. According to the results of the bootstrap mediating effect test, the direct effect of digitalization on green innovation in manufacturing enterprises is significant, and the indirect effect of enterprise value chain upgrading is also significant. It can be seen that digitalization can promote green innovation in manufacturing enterprises through the intermediary channel of value chain upgrading, which verifies Hypothesis 2. Digitization makes all kinds of resources in the value chain come together and helps enterprises improve cooperation and innovation through high-quality integration of resources and technologies. Furthermore, digitization reduces the connection speed and cost of each link in the value chain, which can effectively realize the green innovation path.

5.6. A Test of the Mediation Effect of Digitalization on Industrial Structure Optimization

Digitalization development can promote the upgrading and optimization of manufacturing industry structure, better promote digital industrialization and industrial digitization, enhance the allocation efficiency of production factors, and promote the development of traditional industries in the direction of automation, intelligence, and greening. In view of this, this paper selects the industrial structure optimization (Strug) index identified in the previous paper as a mediating variable through a mediating effect model to test whether digitalization promotes green innovation of manufacturing enterprises by promoting industrial structure optimization of manufacturing industries.

The test results are obtained in

Table 11. Results of the mediating effect test for digital empowerment of industrial structure optimization.

Variables	(1)	(2)	(3)
	GI	Strug	GI
DIG_tot	1.678 ***	10.235 ***	0.531 ***
	(0.111)	(0.249)	(0.173)
Strug			0.112 ***
			(0.013)
Constant	1.364 ***	−0.895 *	1.464 ***
	(0.215)	(0.482)	(0.209)
Control variables	Yes	Yes	Yes
Industry/time/ regional effects	Yes	Yes	Yes
N	1106	1106	1106
R-squared	0.449	0.772	0.483

Note: (1) *** and * denote significance levels of 0.01 and 0.1, respectively. (2) Standard errors of coefficients are in parentheses.

below, and column (1) in the table is the first step test, which has been completed in the previous model test. Column (2) in the table is the second test step, and the results show that the coefficient of DIG_tot is significantly positive at the 1% level, which indicates that digitalization effectively promotes the optimization of manufacturing industry structure. The coefficient of the estimated manufacturing industry structure optimization (Strug) is significantly positive at the 1% level, and the coefficient of the effect of digitization on the level of green innovation in the manufacturing industry is still significantly positive after adding the mediating variables, which decreases compared with the regression coefficient of the baseline regression model, indicating the existence of partial mediating effects. According to the results of the bootstrap mediation effect test, the direct impact effect is significant, and the indirect effect of industrial structure optimization is also significant. It can be seen that digitalization may facilitate green innovation of manufacturing enterprises via the mechanism of industrial structure optimization, which verifies Hypothesis 3. The integration of digital technology and traditional industries promotes the innovation of key technologies in industries, optimizes traditional business processes, and promotes the green transformation of traditional industries; digital development promotes the formation of new industries, such as the field of green products and services, and promotes green innovation.

5.7. A Test of the Mediation Effect of Digitalization on Technological Innovation

The manufacturing industry is able to rely on digital technology, conduct big data analysis, deeply integrate low-carbon and energy-saving technologies, carry out green core technology research, and promote green technology breakthrough innovation. In view of this, the same mediating effect model is used to select the technological innovation (Tec) index identified in the previous section as a mediating variable to test whether digitalization promotes green innovation in manufacturing enterprises by enhancing technological innovation in manufacturing.

The test results are obtained in

Table 12. Results of the test for mediating effects of digitization to enhance technological innovation.

Variables	(1)	(2)	(3)
	GI	Grtc	GI
DIG_tot	1.678 ***	1.236 **	1.645 ***
	(0.111)	(0.483)	(0.111)
Tec			0.026 ***
			(0.012)
cons	1.364 ***	−6.801 ***	1.543 ***
	(0.215)	(0.935)	(0.219)
Control variables	Yes	Yes	Yes
Industry/time/ regional effects	Yes	Yes	Yes
N	1106.000	1106.000	1106.000
R-squared	0.449	0.382	0.456

Note: (1) *** and ** denote significance levels of 0.01 and 0.05, respectively. (2) Standard errors of coefficients are in parentheses.

below, and column (1) in the table is the first step of the test, which was completed in the previous model test. Column (2) in the table is the second step of the test, and the results show that the coefficient of DIG_tot is significantly positive at the 1% level, indicating that digitization has contributed to the increase in the level of technological innovation in manufacturing. Table 3 is the third step of the test; the estimated coefficient of manufacturing technology innovation (Tec) is significantly positive at the 1% level. After adding the mediating variables, the regression coefficient of digitalization on the level of green innovation in the manufacturing industry is still significantly positive, which decreases compared with the regression coefficient of the benchmark model, indicating the existence of a partial mediating effect. According to the results of the bootstrap mediation effect test, the direct effect is significant, and the indirect effect of enterprise technology innovation is also significant. It can be seen that digitalization can promote green innovation of manufacturing enterprises through the mediating channel of enhancing technological innovation in manufacturing, which verifies Hypothesis 4. The development of digitalization breaks through the limitations of time and space when carrying out innovation activities, promotes the flow of innovation resources and the diffusion of innovation results, acquires cost advantages, and promotes the generation of new green technologies.

6. Conclusions and Policy Implications

6.1. Conclusions

This paper empirically examines the impact effect and mechanism of digitization on green innovation of manufacturing enterprises at the micro level using selected data from 158 listed manufacturing enterprises across China from 2013–2019. The empirical findings are as follows: (1) Digitalization has effectively promoted green innovation in manufacturing enterprises. (2) There is significant heterogeneity in how digitalization promotes green innovation in manufacturing enterprises, and the promotion effect of digitalization on green innovation in manufacturing enterprises in labor-intensive industries and central regions is more prominent compared with capital- and technology-intensive industries and enterprises in eastern regions. (3) Digitalization affects green innovation in manufacturing enterprises through three intermediary channels: first, digitalization promotes green innovation in manufacturing enterprises via value chain upgrading; second, digitalization promotes green innovation in manufacturing enterprises via industrial structure optimization; and third, digitalization promotes green innovation in manufacturing enterprises via technological innovation.

6.2. Policy Implications

(1)

Promote the digital transformation of enterprises and cultivate green innovation momentum. Local governments should formulate policies to encourage enterprises to carry out digital transformation, establish digital service management platforms, accelerate the construction of regional digital infrastructure in soft and hard environments, formulate talent strategies to introduce digital talents, enhance the possibility of enterprises for digital technology application, and promote the development of green innovation of manufacturing enterprises with digital policy support; governments should also consider the heterogeneity of digitalization on green innovation of manufacturing enterprises. The government should also consider the heterogeneous impact of digitalization on green innovation of manufacturing enterprises, make policy planning according to local conditions, focus on promoting the integration of digital technology for labor-intensive industries and enterprises in the central region, and jointly build an industrial alliance to promote digital transformation so as to achieve the purpose of industrial cooperation of “1 + 1 > 2”.

(2)

The digital empowerment of upgrading the enterprise value chain to create a green innovation engine. For each link of R&D, design, manufacturing, marketing, and service, intelligent modules, including sensors, processors, and memory, are built into each process to collect data, build a dynamic database of the enterprise, develop an intelligent analysis and decision-making system of the enterprise, effectively use the advantages of data elements to realize real-time monitoring and intelligent control of each link, improve the digitalization and intelligence level of each link, and comprehensively reduce the enterprise. The company will use the advantages of data elements to realize real-time monitoring and intelligent control of each link, improve the level of digitalization and intelligence of each link, comprehensively reduce the energy consumption of production, accurately implement the resource allocation and energy-saving and emission reduction management of each link, and digitally empower the green production and services of enterprises.

(3)

Digital transformation of traditional industrial structures and exploration of the green innovation path. Relying on digital technologies such as cloud platforms and industrial big data, we can transform the manufacturing industry chain and optimize the industrial structure so that the manufacturing industry can move from the low end to the high end of the value chain and extend to R&D and value-added services centered on product manufacturing. The use of digital technology allows SMEs to actively integrate into the industrial chain dominated by large enterprises and gain competitive advantages. Taking industrial digitization as a grip and a green and low-carbon orientation, we actively promote the flow of production factors to new industries that conserve resources and optimize the environment and create a digital green industry chain with international influence and competitiveness.

(4)

Enhance the green technology innovation capability and shape the green innovation advantage. Based on digital technologies such as cloud computing and artificial intelligence, clarify the key points of digitally enhancing the technology level of green innovation in manufacturing enterprises, deeply integrate new energy technologies such as carbon emission reduction and energy efficiency improvement technologies, form a technology integration innovation network, and control the cost of green innovation with integrated development. Encourage small and medium-sized enterprises to rely on industry leaders and research institutes to build a green technology innovation consortium, carry out green low-carbon key core technology research, and promote green technology to empower green innovation and the efficient development of the manufacturing industry.

6.3. Research Limitations and Future Directions

The research object of this paper is Chinese manufacturing enterprises. Due to the completeness and availability of data, there are many missing values when collecting samples, resulting in a small capacity of selected samples. The samples are evenly distributed geographically in order to ensure representativeness, but there will still be some deviations. In the future, other comprehensive evaluation methods can be used to evaluate the level of green innovation of enterprises; for example, text mining methods can be used to conduct text analysis of green innovation-related word frequencies in annual reports of listed companies to collect as complete enterprise data as possible. In terms of studying the mechanism of the role of digitization in influencing green innovation, research methods can be improved, such as using system dynamics methods to explore the evolutionary paths of the three mediating mechanisms.