Digital Economy and the Sustainable Development of China’s Manufacturing Industry: From the Perspective of Industry Performance and Green Development

Abstract

Digital transformation is increasingly crucial to the upgrading and sustainable development of China’s manufacturing industry with the rapid development of the digital economy. To study the impact of the digital economy on the sustainable development of the manufacturing industry, this study analyzed the theoretical basis of the digital economy’s impact on the promotion of the sustainable development of the manufacturing industry. Then, based on the panel data of manufacturing sectors in 2002, 2005, 2007, 2010, 2012, 2015, and 2017, empirical tests and mechanism analysis were conducted by means of the two-way fixed effect model and the mediating effect model. The results were as follows: (1) Digital services can significantly improve the industrial performance of the manufacturing industry, while the effect of digital products is nonsignificant; (2) Mechanism analysis revealed that digital services can promote the industrial performance of the manufacturing industry through the intermediary mechanisms of reducing production costs rather than transaction costs; (3) Digital services can also reduce carbon emissions and promote the green development of the manufacturing industry through the intermediary mechanisms of innovation. In conclusion, digital services can promote the sustainable development of China’s manufacturing industry. This paper provides evidence for the integration of the manufacturing industry and the digital economy. Furthermore, it has important implications for the formulation of digital economy policies and the sustainable development of the manufacturing industry.

1. Introduction

The digital economy refers to a broad range of economic activities that includes using digitized information and knowledge as the key factor of production, modern information networks as an important activity space, and the effective use of information and communication technology (ICT) as an important driver of productivity growth and economic structural optimization (since there is no widely accepted definition of the digital economy in academia, this paper adopts the definition in “G20 Digital Economy Development and Cooperation Initiative”). The digital economy can be divided into two parts: digital industrialization and industrial digitization. Digital industrialization is equivalent to the traditional ICT industry, including digital products and digital services. As a result of the digital transformation of other industrial sectors, the added value of the digital economy is generated in traditional industries, and this portion represents the industrial digitization part [1]. In recent years, the digital economy has grown rapidly in China. According to the latest “White Paper on China’s Digital Economy Development (2022)”, released by the China Academy of Information and Communication Technology, the share of China’s digital economy in GDP increased year by year from 2005 to 2021, from 14.2% to 39.8%. As a general-purpose technology (GPT), digital technology permeates all aspects of economic activities, and the digital economy has become a key driving force for China’s economic development and industrial upgrading [2].

With the disappearance of the demographic dividend, China’s manufacturing industry is in a critical period of transformation and upgrading to achieve sustainable development [3]. The Chinese government regards the digital transformation of the manufacturing industry as the key to achieving industrial upgrading and high-quality sustainable development [4,5], and it successively issued a number of major plans and policy guidelines regarding the digital economy in recent years. The Report to the 20th National Congress of the Communist Party of China (CPC) pointed out: “We will accelerate the development of the digital economy, further integrate it with the real economy, and build internationally competitive digital industry clusters.” To promote the digital transformation of the manufacturing industry, China has issued a series of strategies and policy measures, such as the Action Plan on Promoting Big Data Development, Guidelines for the “Internet Plus” Action Plan, Guidelines for Deepening the Integrated Development of Manufacturing Industry and Internet, and the Industrial Internet Development Action Plan (2018–2020). Therefore, it is of practical significance to study the impact of the digital economy on the sustainable development of the manufacturing industry. How to effectively release the dividends of the digital economy in the sustainable development of the manufacturing industry has become a topic of concern for the government and society.

Several studies have discussed the socio-economic impact of the digital economy. The existing research mainly focuses on the following three aspects. First, on the macro level, the digital economy facilitates technological innovation [6,7] and high-quality economic development featuring green development, innovative development, and coordinated development by optimizing resource allocation [8]. In addition, the digital economy significantly promotes the upgrading of the technological progress and human capital of Chinese cities, thus promoting the upgrading of the industrial structure. The digital economy can also have an indirect impact on the upgrading of industrial structures by affecting the innovation level of the region [9,10,11]. Heterogeneity analysis reveals that digital finance plays a more significant role in cities with developed economies, low levels of financialization, and small urban–rural income gaps [12]. Second, on the meso level, the digital economy has an impact on the development of the manufacturing industry. Big data [13] and other digital technologies can promote the digital transformation of the manufacturing industry [2] and improve its total factor productivity [14] by reshaping the production function of firms [15] and enhancing the innovation of firms [16,17]. Third, on the micro level, the digital economy can improve the total factor productivity of manufacturers and thus their performance by improving the innovation capability of the firms [18], optimizing the production process [19], and reducing the cost of information interaction [20]. To sum up, there is extensive research on the impact of the digital economy at the macro level and the enterprise level, which provides a theoretical basis and empirical support for this study. However, there are few studies that examine the role of the digital economy at the industrial level, and no research has used empirical analysis to respond to the relationship between the digital economy and the sustainable development of manufacturing.

The digital economy has significantly changed the forms of production and organization. We therefore ask whether the digital economy can promote the sustainable development of the manufacturing industry. Specifically, can the digital economy benefit the industry performance and green development of manufacturing sectors? If so, what are the mechanisms of this? These questions, which have attracted widespread attention from industry and academia, constitute the research propositions of this paper. This paper first clarifies the theoretical mechanism of the digital economy promoting the sustainable development of manufacturing sectors, and proposes the research hypotheses of this study; second, we construct an index reflecting the integration level of the manufacturing industry and the digital economy; and third, we empirically test the impact and mechanism of the digital economy on the industry performance and green development of manufacturing sectors by using the fixed effect model. The empirical tests show that the integration of the manufacturing industry with the digital service industry has significantly improved the performance of the manufacturing industry, but the integration of the manufacturing industry with the digital manufacturing industry has had no significant impact on its performance. After changing the measurement index of the industry performance, the first conclusion still stands. Further research shows that the integration of the manufacturing industry and the digital service industry improves industry performance by reducing manufacturing costs, which means manufacturing cost is an important mediating variable for digital technology affecting the industry performance; however, the aforementioned integration does not affect sales and administrative expenses, which means transaction cost is not a mediator variable. In further discussion, we find that the integration of the manufacturing industry and the digital service industry can significantly reduce the emissions of carbon dioxide and promote the green development of the manufacturing industry.

The contributions of this paper are as follows. First, the existing literature pays more attention to the impact of the digital economy on economic transformation and upgrading, and most of the studies do not focus on the impact of the digital economy on the sustainable development of manufacturing sectors. Therefore, this paper aims to fill this gap by analyzing the impact of the digital economy on the sustainable development of manufacturing sectors. Second, most of the existing literature conducts studies from the perspective of countries and provinces, while this paper analyzes the impact of the digital economy on the sustainable development of manufacturing sectors at the industrial level, using the industry performance and emissions of carbon dioxide. Third, we explore the heterogenous impacts of the digital manufacturing industry and the digital service industry on the sustainable development of the traditional manufacturing industry, and further discuss the micro-mechanism. This paper provides guidance for China to accelerate the development of “Digital China”, promote the digital transformation of the manufacturing industry, and build it into a manufacturing powerhouse.

The rest of this paper is arranged as follows: the second part analyzes the influence channels of the digital economy on the manufacturing industry, and proposes hypotheses to be tested; the third part introduces the empirical models and variables; the fourth part empirically tests the hypotheses through econometric models; the fifth part provides a further discussion; and the last part presents the conclusion.

2. Theoretical Analysis and Research Hypotheses

The concept of sustainable development is complex and multidimensional. However, sustainable development should encompass two fundamental aspects: increasing economic value and having a positive impact on the environment. Therefore, the sustainable development of the manufacturing industry in this paper is defined from the perspective of industrial performance and green development.

2.1. Direct Impact of the Digital Economy on the Sustainable Development of Manufacturing Sectors

As a GPT, digital technology affects all aspects of the economy. First, the digital economy has promoted the improvement of production efficiency, which not only improves the efficiency of labor production, but also improves the total factor productivity and the green total factor productivity. Specifically, the digital economy promotes the coordination of all production factors, improves the efficiency of resource allocation, realizes the full utilization of resources, and promotes a reduction in carbon emission intensity. Second, the development of the digital economy has improved the Chinese government’s governance ability, including its environmental governance ability. The improvement of the Chinese government’s governance ability can improve the market environment, thereby guiding capital to flow to more efficient economic activities; the improvement of the government’s environmental governance ability can reduce carbon emission intensity [11]. Third, at the enterprise level, the development of the digital economy has formed an economic environment with economies of scale, economies of scope, and long-tail effects. Digital transformation is conducive to the specialization of enterprises [1]. Based on the above analysis, it is reasonable to believe that the digital economy can promote the sustainable development of the whole economy, including the manufacturing industry. Therefore, we propose the following hypothesis.

Hypothesis 1.

The digital economy can significantly promote the sustainable development of the manufacturing industry, including improving industrial performance and promoting green development by reducing carbon emissions.

2.2. Indirect Impact Mechanism on Industrial Performance: Reduction in Production Costs

Digital technology, such as the Internet, IoT, AI, etc., is applied in all aspects of manufacturing, leading to changes in production modes, reducing manufacturing costs, and improving industry performance. First, there is the penetration effect. As a GPT, digital technology can penetrate all production links, improve the synergy between digital technology and other production factors, and improve industry performance [21,22]. Second, there is the substitution effect. According to Moore’s Law, technological progress is accompanied by a decline in the prices of related technical products and services, thus facilitating the substitution of digital technology for other production factors and enhancing the digitalization level and overall performance of the industry [23,24].

Penetration effect of digital technology. First, with information communication technologies, dematerialization can occur, for example, as new communication technologies optimize production processes, eliminate redundancies and waste, and reduce travel and commuting [25]. This will allow the enterprise to save manufacturing costs [19,26]. For example, digital technology facilitates the numerical control of equipment and the automation of production [27], which effectively reduces the energy consumption of manufacturers [28,29,30]. Secondly, digital products and services, as intermediate inputs, directly improve production efficiency by complementing other production factors. The integration of digital technology and the manufacturing industry creates human–machine collaboration and synergy and frees more workers from tedious and repetitive work, thus engaging them in more creative work. For example, recent computerization has been substituted for low-skilled workers in performing routine tasks while complementing the abstract, creative, problem-solving, and coordination tasks performed by highly educated workers [31]. Digital technology could use online data to perform dynamic optimization to improve the system operation efficiency and quality of the manufacturing system while reducing the cost [19,32].

The substitution effect of digital technology. Due to the price drop caused by technological progress, the hardware and software system architectures of the digital economy can replace other capital and affect the input structure of factors [22,33]. For example, industrial robots can replace the labor force for automatic production, improve the efficiency of factor allocation, and improve productivity. Taking automobile manufacturing as an example, with the wide application of digital technology, especially intelligent technology, the production efficiency of the automobile industry has been continuously improved, and automobile manufacturing and finished products are increasingly digitalized.

In addition, digital technology is applied in business operations, which brings about changes in production, marketing, and financial management. The application of digital technology is not limited to production and processing links, but rather has been extended to the whole life cycle [34,35], and the whole manufacturing system including upstream and downstream chains [28]; it improves the daily operation and management of firms [36].

First, in production management, digital technology helps manufacturers monitor the production process closely, adopt refined management, realize the rational allocation of production factors [37], and reduce the management cost. For example, digital technology can realize the real-time monitoring of machine operation and provide remote maintenance, fault prediction, performance optimization, and other services, thus improving management efficiency. The development of computer and Internet technology has greatly accelerated the flow of information within and between firms, and promoted the standardization and flattening of management. Secondly, in marketing management, the integration of manufacturing and digital technology has led to “platform firms”, which improve information transmission efficiency and reduce sales expenses. For example, the use of ICT by the business sector may lead to a reduction in trade costs and the expansion of trade relations between the countries [38,39]. Digital technologies enhance firms’ ability to obtain upstream and downstream feedback, to better meet customer needs, to reduce information asymmetries, and to significantly diminish the risk of creating lemon markets [40]. Thirdly, in terms of financial management, data-based Internet supply chain finance can effectively facilitate enterprise financing channels and reduce financing costs for manufacturing SMEs by replacing traditional mortgage-backed financing with data transaction pledges. Finally, digital technology can also enhance the innovation capability of firms from the perspective of information collection and resource integration [41].

Therefore, digital technology, as a GPT, participates in all aspects of enterprise management and profoundly changes the production and organization modes of manufacturing enterprises in China. It can effectively reduce the manufacturing, management, consumption, and financial costs of enterprises, improve the operational efficiency, and enhance overall industry performance and promote sustainable development. Its influence pathway is shown in Figure 1. Therefore, the following research hypothesis is proposed.

Hypothesis 2.

The digital economy improves the industrial performance of the manufacturing industry by reducing production costs.

2.3. Indirect Impact Mechanism on Industrial Performance: Reduction in Transaction Costs

The digital economy not only affects industry performance by changing the internal organization and production mode of firms, but it also affects the competition and cooperation dynamics between firms, the supply–demand relationship between firms and customers, and the market and industrial structure.

The digital economy has changed the supply–demand relationship between firms and customers. First, based on big data analysis technology, the digital economic platform improves the matching efficiency of information; reshapes the matching mode of supply and demand information in the terminal of the industrial chain and stimulates new consumption potential [20]; and expands the market scale. Second, digital technology enables firms to achieve customization. Digital technology such as the Internet and big data provide an online communication platform for producers and buyers, which reduces their information cost and enables producers to quickly obtain the personalized needs of customers [42]. Meanwhile, digital technologies reduce the production conversion cost and make the production process more flexible, which makes it possible to meet the diversified and personalized needs of customers [43,44]. Under the digital economy, the production mode of the manufacturing industry has changed from traditional product-centered mass production to customer-centered mass customization, effectively increasing the added value of products.

The digital economy has changed the competition and cooperation dynamics between firms and has reconstructed industrial organization. First of all, the utilization of big data by traditional manufacturers reduces the cost of information collection, analysis, and application, and therefore reduces the constraints of information asymmetry on the flow of factors. The Internet and big data generate a large number of quality signals [32,34,45], and the factors of production flow to the field where user value is concentrated according to quality signals. Therefore, in the digital economy, user value has become the core index to allocate production factors.

Secondly, digital technology promotes the industrial division of labor and production coordination in the manufacturing industry. Digital technology helps firms to outsource production so as to focus on their core business with comparative advantages; it also helps to reduce manufacturing costs [46] and promote the expansion and deepening of the division of labor [47]. In addition, digital technology enables firms to form networked collaborations. The production links of the traditional manufacturing industry are complex-serial-link-based, with links separated from each other and often adopting different technical standards, such that it is difficult for them to interact with each other. With the popularization of the industrial Internet, networked collaborative manufacturing makes it possible for different market actors to collaborate closely [48]; transforms serial work into parallel work, forming a manufacturing network; and realizes the large-scale integration and optimization of production factors and resources.

Thirdly, the digital economy facilitates the upgrading of the manufacturing industry. The digital transformation of traditional industries will promote cross-border integration of industries, forming a digital ecology; the digital economy can allow entrepreneurs to more conveniently obtain innovative resources, greatly improving innovation efficiency. Therefore, it can upgrade industrial technology [9,11]. Furthermore, digital transformation reconstructs the competition pattern of industrial organizations, strengthens the competition mechanism, and helps to improve the efficiency of resource utilization; it also promotes the fair distribution of income and the continuous optimization of industrial organizations [10,49]. The enhancement of firm productivity and market competitiveness promotes industrial transformation and upgrading, and the performance and sustainable development of the industry. Based on the above analysis, this paper puts forward a logical framework for the role of the digital economy in the market behavior of China’s manufacturing industry, as shown in Figure 2. Therefore, the following research hypothesis is proposed.

Hypothesis 3.

The digital economy improves the industrial performance of the manufacturing industry by reducing transaction costs.

2.4. Indirect Impact Mechanism on Green Development: Improvement of the Innovation Level

For manufacturing enterprises, the digital economy is one of the important driving forces of innovation [7], including green innovation. Firstly, the digital economy provides chances for enterprises to upgrade traditional production lines or create new production lines using innovation, which can promote the green development and a reduction in the carbon emissions of manufacturing enterprises [50]. Secondly, the development of the digital economy can provide more financial support for the innovation activities of manufacturing enterprises by alleviating financing constraints. The development of the digital economy can alleviate the financing constraints of manufacturing enterprises by easing the information asymmetry and reducing the principal-agent cost [51]. Finally, the development of the digital economy accelerates the spillover of knowledge among different industries through network externalities [52], and further enhances the green innovation level of different industries. Therefore, the hypothesis is proposed as follows.

Hypothesis 4.

The digital economy promotes the green development of the manufacturing industry by improving the innovation level.

According to the above analysis, the digital economy can not only directly improve the industrial performance of the manufacturing industry by improving productivity and the market environment, but can also indirectly improve it by improving production and organization modes, thus reducing manufacturing and transaction costs. In addition, the basic industries of the digital economy include not only the manufacturing sector, such as communication equipment, computers, and other electronic equipment manufacturing; but they also include the service sector, such as information transmission, software, and information technology, making the impacts of these two sectors on the performance of the manufacturing industry heterogeneous. Therefore, this paper performs a separate analysis in our empirical study. Similarly, the direct promotional effects of the digital economy on the green development of the manufacturing industry, and the indirect impact mechanism in terms of innovation, can be analyzed separately. Accordingly, this paper puts forward conceptual models, shown in Figure 3, to study the impact of the digital economy on the performance and green development of the manufacturing industry.

3. Research Design

3.1. Model Specification

According to the above theoretical analysis, to test the impact of the digital economy on the sustainable development of the manufacturing industry, this paper adopts ordinary least squares (OLS) with clustering robust standard deviation for a year and individual fixed effects, expressed herein as Equation (1):

$${Sus}_{it}={{\beta }_0+{\beta }_{1}DCC}_{it}+\mathit{C}\mathit{o}\mathit{n}\mathit{t}\mathit{r}\mathit{o}\mathit{l}\mathbf{\ast }{\mathit{\beta }}_2+{\mu }_{it}$$

(1)

where $i$ represents a specific industry; $t$ represents a specific year; ${Sus}_{it}$ is the sustainable development of manufacturing industry i during the period t, including industrial performance and green development; and ${DCC1}_{it}$ is the level of digitalization of the manufacturing industry $i$ during the period $t$—that is, the level of integration of the manufacturing industry and the digital economy. $\mathit{C}\mathit{o}\mathit{n}\mathit{t}\mathit{r}\mathit{o}\mathit{l}$ represents a vector of control variables; ${\mu }_{it}$ is a random perturbation term; and ${\beta }_1$ and ${\mathit{\beta }}_2$ are the coefficients to be estimated.

This paper divides the digital economy into the digital manufacturing industry and the digital service industry in order to test their respective impact on the performance of the manufacturing industry, and it constructs model 2 for empirical testing.

$${Sus}_{it}={{\gamma }_0+{\gamma }_{1}DCC1}_{it}+{{\gamma }_{2}DCC2}_{it}+\mathit{C}\mathit{o}\mathit{n}\mathit{t}\mathit{r}\mathit{o}\mathit{l}\mathbf{\ast }{\mathit{\gamma }}_3+{ϵ}_{it}$$

(2)

where ${DCC1}_{it}$ denotes the integrated development of the manufacturing industry $i$ and the digital manufacturing industry during the period $t$; while ${DCC2}_{it}$ denotes the degree of integration of the manufacturing industry $i$ and the digital service industry during the period $t$. Other variables are defined in the same way as above, so they will not be repeated.

The digital economy may indirectly affect the sustainable development of the manufacturing industry by reducing production costs and transaction costs, or by promoting the innovation level. To explore the indirect impact mechanism of the digital economy on the sustainable development of the manufacturing industry, this paper adopts the mediation effect model to conduct further empirical analysis. The specific model is shown in Formulas (3) and (4).

$$Mediation={{\beta }_{01}+{\beta }_{11}DCC1}_{it}{+{\beta }_{21}DCC2}_{it}+\mathit{C}\mathit{o}\mathit{n}\mathit{t}\mathit{r}\mathit{o}\mathit{l}{\mathbf{\ast }\mathit{\beta }}_{31}+{\mu }_{it}$$

(3)

$${Sus}_{it}={{\beta }_{02}+{\beta }_{12}DCC1}_{it}{+{\beta }_{22}DCC2}_{it}+{\beta }_{32}{Mediation}_{it}+\mathit{C}\mathit{o}\mathit{n}\mathit{t}\mathit{r}\mathit{o}\mathit{l}{\mathbf{\ast }\mathit{\beta }}_{42}+{\mu }_{it}$$

(4)

where ${Mediatior}_{it}$ represents the production cost, the transaction costs, and the innovation level of the manufacturing sector $i$ in the period $t$, respectively; other variables are defined in the same way as above, so they will not be repeated. ${\beta }_{11}$ represents the effect of the digital economy on mediation variables. ${\beta }_{12}$ represents the direct effect of the digital economy on the sustainable development of the manufacturing industry, ${\beta }_{22}$ represents the effect of mediating variables on the sustainable development of the manufacturing industry.

Specifically, as the dependent variable is industrial performance, if the coefficient ${\beta }_1$ in Formula (1) is significantly positive, it indicates that the digital economy can promote the industrial performance of the manufacturing industry. If the coefficient ${\gamma }_1$ or ${\gamma }_2$ in Formula (2) is significantly positive, it indicates that the digital manufacturing industry or digital service industry can enhance the industrial performance of the manufacturing industry. On testing of the mediation effect, if coefficients ${\beta }_{11}$ and ${\beta }_{32}$ hold simultaneously and both are significantly negative, it shows that the integration of the traditional manufacturing industry and the digital manufacturing industry significantly reduces the transaction costs or production costs, thus improving the operating performance of the manufacturers. If coefficients ${\beta }_{21}$ and ${\beta }_{32}$ hold simultaneously and both are significantly negative, it indicates that the integration of the manufacturing and digital service industries significantly reduces transaction costs or production costs, and thus improves the operating performance of manufacturing firms. If the coefficient ${\beta }_{12}$ or ${\beta }_{12}$ is significantly positive, it indicates that the effect is partial mediation; otherwise, it is full mediation. The explanations also apply to the model in which the dependent variable is green development, which is omitted here.

3.2. Data Source and Variables Design

In this paper, the direct input coefficient was used to measure the integrated development of the digital economy and the manufacturing industry [53,54], so we needed to calculate the direct input coefficient matrix based on the input–output table. In China, the basic input–output table is compiled every five years in the years ending in 2 and 7, and the extended input–output table is updated every five years in the years ending in 0 and 5. Given the availability of data, this paper used the input–output tables of 2002, 2005, 2007, 2010, 2012, 2015 and 2017 for analysis (the input–output table was also compiled in 2018 and 2020, but considering the availability of other data and the impact of the COVID-19 epidemic, this paper did not add the data for 2018 and 2020). The data on industry performance, transaction cost and production cost of manufacturing industry were obtained from the China Statistical Yearbook and the China Industrial Statistical Yearbook. The industrial innovation level is obtained from Kou and Liu [55].

In the input–output table, the digital economy mainly includes “communication equipment, computer and other electronic equipment manufacturing” and “information transmission, software, and information technology”; the former falls into the category of the digital manufacturing industry and the latter falls into the category of the digital service industry. In addition, this paper eliminated the digital manufacturing industry (i.e., communication equipment, computer, and other electronic equipment industries) from the traditional manufacturing sector. After elimination, the manufacturing industry includes 15 sectors: food and tobacco; textiles; textiles for clothing, shoes, hats, leather, down and its products; wood products and furniture; paper making, printing, cultural, educational, and sporting goods; petroleum, cooking products and nuclear fuel processed products; chemical products; nonmetallic mineral products; metal smelting and rolling products; metal products; general and special equipment; transportation equipment; electrical machinery and equipment; instrument and meter; and other manufactured products.

Based on the research purpose and data availability, this paper designed dependent variables, core independent variables, control variables, and mediator variables. See

Table 1. Explanations of variables.

Variable	Name	Formula	Source
Dependent variables
PROFIT1	Return on total assets (ROTA)	Net profit/total assets	China Statistical Yearbook
PROFIT2	Return on capital (ROC)	Added value/net fixed assets	China Industrial Statistical Yearbook
CO2_Asset	Carbon dioxide emissions per unit asset	Carbon dioxide emissions/total assets	China Industrial Statistical Yearbook
CO2_Market	Carbon dioxide emissions per unit market value	Carbon dioxide emissions/total market value	China Industrial Statistical Yearbook
Independent variables
DCC	Direct input coefficient of manufacturing industry to the digital economy	-	China Input–Output Table
DCC1	Direct input coefficient of manufacturing industry to the digital manufacturing industry		China Input–Output Table
DCC2	Direct input coefficient of the manufacturing industry to the digital service industry	-	China Input–Output Table
Control variables
PRI	Ownership structure	The proportion of paid-in capital of private firms	China Industrial Statistical Yearbook
SCALE	Market concentration	The gross output value of industry i/number of firms	China Industrial Statistical Yearbook
FOR	Market openness	The proportion of total foreign assets	China Industrial Statistical Yearbook
Mediator variables
COGS	Main business cost	-	China Industrial Statistical Yearbook
OE	Admin expense	-	China Industrial Statistical Yearbook
SE	Sales expense	-	China Industrial Statistical Yearbook
INNO	Innovation index	-	FIND Report on City and Industrial Innovation in China

for their specific meanings, calculation methods, and data sources.

Dependent variables. Industrial performance refers to the economic benefits achieved by firms in an industry in terms of price, cost, profit, output, quality, and technological progress under a certain market structure and market behavior. Industry performance is often measured by net profit and productivity [56]. Drawing on the practices of Sun and Wang [57], this paper first used Return on total assets (ROTA) to measure industry performance. ROTA is the ratio of net profit to total assets. In the analysis of robustness, this paper further measured industry performance from the perspective of production efficiency by using the ROC index, which is the ratio of industrial-added value to net fixed assets. To measure the green development of the manufacturing industry, we adopted the carbon dioxide emissions per unit asset as the dependent variable, and replaced it with carbon dioxide emissions per unit market value in the robustness test.

Core independent variables. The patent coefficient, the Herfindahl index, and the direct input coefficient are commonly used methods to measure the degree of industry convergence. Both methods use the correlation coefficient of the number of patents between different industries to measure the degree of industrial convergence. However, these methods may have some problems. First, it is difficult to determine the exact number of patents of various industries; second, the patent dimension may not fully capture industrial convergence, and the role of digital technology in the manufacturing industry does not directly affect the number of patents and other indicators. By comparison, the direct input coefficient reflects the interdependence between the digital economy and the manufacturing industry more comprehensively; so this paper used the direct input coefficient as a measure of the integrated development of the digital economy and manufacturing industry. The direct input coefficient (${DCC}_{sm}$) represents the value of the products or services of sector s that are directly consumed by sector m to produce a unit of output. It is calculated as the ratio of the value of products or services of sector s directly consumed by sector m to the total input of industry sector m:

$${DCC}_{sm}=\frac{x_{sm}}{x_m}\dots i,m=\mathrm{1,2},\dots ,n$$

(5)

Control variables. In the study of industry performance, the “Structure–Conduct–Performance” (SCP) framework proposed by Bain [58] is the most relevant one. The SCP framework holds that industrial structure determines firm conduct, which in turn determines industry performance. Based on the SCP framework, and referring to the existing research on industry performance, this paper selected market structure, ownership structure, and market openness as control variables. Specifically, this paper chose the average firm size as an index to reflect market concentration. The larger the average size, the more concentrated the market is, and vice versa. It is calculated as follows: the average firm size in industry i = gross output value of industry $i$/the total number of firms in industry i. The market ownership structure reflects the ownership composition in the industry. This paper selected the proportion of paid-in capital of private firms to the total paid-in capital of the industry as the proxy variable of market ownership structure, and it is calculated as follows: ownership structure of industry $i$ = paid-in capital of private firms in industry $i$/total paid-in capital in industry i. Market openness reflects the openness of the industry. This paper selected the share of foreign-invested assets (plus those invested by Hong Kong-, Macao-, and Taiwan-based firms) of the total assets of the industry as the measurement index of market openness. It is calculated as follows: market openness of industry $i$ = foreign-invested assets in industry i/total assets in industry $i$.

Mediator variables. The ratios of the administrative expense and sales expense of various industries are used as proxy variables for transaction cost; and the ratio of the main business cost of various industries is used as a proxy variable for manufacturing cost. To explore the mediation effect of innovation on the relationship between the digital economy and green development, the industrial innovation index is involved as a mediation variable; it was obtained from Kou and Liu [55] and was logarithmized in this study. Specifically, the ratio of administrative expense of the industry $i$ = the administrative expense of the industry $i$/main business income of industry $i$; the ratio of sales expense of the industry $i$ = the sales expense of the industry $i$/main business income of industry i; and the ratio of main business cost of the industry $i$ = main business cost of industry i/main business income of industry $i$.

3.3. Descriptive Statistics

The descriptive statistics are shown in

Table 2. Summary statistics.

Variable	Observations	Mean	Std	Min	Max
PROFIT1	112	0.074	0.029	−0.018	0.141
PROFIT2	112	0.266	0.126	−0.043	0.536
CO2_Asset	112	1.408	3.303	0.007	18.400
CO2_Market	112	1.000	1.536	0.004	7.839
DCC	112	0.056	0.125	0.001	0.538
DCC1	112	0.051	0.124	0.000	0.524
DCC2	112	0.005	0.005	0.000	0.022
SCALE	112	1.857	2.361	0.196	15.785
FOR	112	0.296	0.134	0.072	0.726
PRI	112	0.223	0.116	0.024	0.514
COGS	112	0.849	0.034	0.736	0.950
OE	112	0.044	0.013	0.021	0.085
SE	112	0.029	0.011	0.008	0.063
INNO	96	53.004	107.559	0.150	560.430

. It can be seen that the mean value of profit of the manufacturing industry in China from 2002 to 2017 is 0.074, and the minimum and maximum are −0.018 and 0.141, respectively. The mean value of carbon dioxide emissions per unit asset is 1.408, and the minimum and maximum are 0.007 and 18.400, respectively. The mean values of our primary variables of interest, DCC, DCC1, and DCC2, are 0.056, 0.051, and 0.005, respectively. The integration of the digital economy and the manufacturing industry mainly occurs in the digital manufacturing industry. It is noted that the observations of the industrial innovation index are 96, because the observations in 2017 are missing. The descriptive statistics of other variables are shown below.

4. Empirical Results and Analysis

Based on the models and variables mentioned above, this paper first analyzes the basic regression results and then the robustness of the models by using different indicators of industry performance. Finally, by using the results of the mediation effect model, this paper analyzes the internal mechanism for the digital economy to affect the performance of the manufacturing industry. Using the same method, we also analyze the impact of the digital economy on the green development of the manufacturing industry.

4.1. Basic Regression Results

4.1.1. Direct Impact of the Digital Economy on Industrial Performance

Based on the panel data of 15 manufacturing sectors in China for 7 years, the empirical results obtained in this paper are shown in

Table 3. Effects of the digital economy on the industrial performance of the manufacturing industry.

Independent Variables	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
	PROFIT1	PROFIT1	PROFIT1	PROFIT1	PROFIT1
DCC1	-	−0.173 (0.149)	−0.172 (0.152)	−0.124 (0.172)	0.034 (0.145)
DCC2	-	0.775 ** (0.328)	0.789 ** (0.317)	0.981 *** (0.319)	1.176 *** (0.361)
DCC	0.011 (0.163)	-	-	-	-
SCALE	0.003 (0.002)	-	0.000 (0.002)	0.000 (0.002)	0.003 * (0.002)
FOR	−0.027 (0.048)	-	-	−0.039 (0.072)	−0.045 (0.047)
PRI	0.216 *** (0.055)	-	-	-	0.218 *** (0.054)
CONS	0.023 * (0.012)	0.034 *** (0.006)	0.034 *** (0.006)	0.042 ** (0.018)	0.016 (0.011)
Industry fixed	Yes	Yes	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes	Yes	Yes
R2	0.776	0.732	0.733	0.736	0.788
N	105	105	105	105	105

Note: *** p < 0.01, ** p < 0.05, * p < 0.1, the values in brackets are robust standard errors.

. According to the results of the Hausman test and the F test, we find that the fixed effect model should be used; and to eliminate the heteroscedasticity problem that may exist in the model, this paper uses clustering robust standard error for estimation. Specifically, the core independent variable in Column (1) is the direct input coefficient of the manufacturing industry to the digital economy, and the core independent variables in Columns (2)–(6) are the direct input coefficient of the manufacturing industry to the digital manufacturing industry and digital service industry. In Columns (3)–(5), we add control variables such as market structure, ownership structure, and market openness, successively.

In terms of the impact of the digital economy on the performance of the manufacturing industry, the regression coefficient is 0.011. The result is a positive number, but it fails to pass the significance test, indicating that the integration of the digital economy and the traditional manufacturing industry has no obvious facilitation effect on the performance of the manufacturing industry, which is inconsistent with the previous estimation. To get to the bottom of the reason for this, the digital economy is divided into the digital manufacturing industry and the digital service industry; and their impact on the performance of manufacturing industry is investigated, respectively. The results show that the direct input coefficient of the manufacturing industry to the digital manufacturing industry does not pass the significance test, which shows that the integration of the digital manufacturing industry and the traditional manufacturing industry has no obvious facilitation effect on the performance of the manufacturing industry. The regression coefficient of the direct input coefficient of the manufacturing industry to the digital service industry is significantly positive, which shows that the integration of the two industries has obviously improved the performance of the manufacturing industry. Hypothesis 1 of this paper is verified.

The integration of the digital manufacturing industry and the digital service industry with the traditional manufacturing industry shows a heterogeneous impact on the performance of the manufacturing industry. Regarding the heterogeneity, this paper holds that a possible reason is that the digital manufacturing industry and the digital service industry play different roles in the traditional manufacturing industry. For example, the digital manufacturing products are mainly involved in the production process as parts and components; however, the large-scale digital transformation in recent years, especially in the applications of big data, the industrial Internet, and cloud computing, has effectively improved the management and production processes of firms, and the resulting technological progress has largely affected the production processes of firms in the form of intermediate services. Therefore, the technological progress brought about by the digital transformation of the manufacturing industry is mainly caused by the integrated development of the digital service industry and the traditional manufacturing industry.

4.1.2. Direct Impact of the Digital Economy on Green Development

Industrial performance reflects the sustainable development of manufacturing from the perspective of production [59]. With increasingly widespread attention paid to green development, the level of green development can also be considered as one of the targets of the sustainable development of the manufacturing industry. The digital economy may improve green innovation ability [60], reduce carbon emission intensity [61], and promote the green development of the manufacturing industry. Therefore, the industrial performance in Equations (1) and (2) is replaced by carbon dioxide emissions, and empirical analysis is carried out to explore the impact of the digital economy on the green development of the manufacturing industry. It should be noted that to eliminate the impact of industry size, the dependent variable here is carbon dioxide emissions per unit asset, which is calculated by carbon dioxide emissions divided by total assets.

The regression results are shown in

Table 4. Effects of the digital economy on the green development of the manufacturing industry.

Independent Variables	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
	CO2_Asset	CO2_Asset	CO2_Asset	CO2_Asset	CO2_Asset
DCC	4.755	-	-	-	-
	(0.886)	-	-	-	-
DCC1	-	4.341	6.346	3.888	3.299
	-	(0.569)	(1.023)	(0.741)	(0.709)
DCC2	-	−91.17	−60.21 *	−70.05 *	−70.78 *
	-	(−1.444)	(−1.825)	(−2.043)	(−2.086)
SCALE	−0.667 ***	-	−0.631 ***	−0.659 ***	−0.669 ***
	(−14.26)	-	(−9.156)	(−0.998)	(−16.07)
FOR	0.873	-	-	2.020	2.039
	(0.455)	-	-	(1.105)	(1.120)
PRI	−0.709	-	-		−0.812
	(−0.230)	-	-		(−0.300)
CONS	1.421 ***	2.663 ***	1.914 ***	1.869 ***	1.739 ***
	(3.063)	(4.499)	(6.697)	(6.697)	(3.754)
Industry fixed	Yes	Yes	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes	Yes	Yes
R2	0.757	0.302	0.782	0.787	0.788
N	105	105	105	105	105

Note: *** p < 0.01, * p < 0.1, and the values in brackets are robust standard errors.

. The results in Column (1) show that the overall integration degree of the digital economy and manufacturing industry has no significant impact on the green development of the manufacturing industry. The results in Columns (2)–(4) show that, from the perspective of sub-items, the integration degree of the digital manufacturing industry and the manufacturing industry has no significant impact on the green development of the manufacturing industry, but the integration degree of the digital service industry and the manufacturing industry has a significant negative impact on the carbon dioxide emissions of the manufacturing industry. That is, the penetration of the digital service industry, to some extent, has reduced the carbon dioxide emissions of the manufacturing industry and promoted the green and sustainable development of the industry. Overall, Hypothesis 1 is verified—that is, the integration of the digital service industry and the manufacturing sector significantly promotes the sustainable development of the manufacturing industry, improves its operating performance, and promotes green development.

4.2. Robustness Test

The test results in Table 3 and Table 4 show that the development of the digital economy can effectively promote the sustainable development of the manufacturing industry. However, the robustness of the basic regression results has not been tested. This paper further verifies the robustness of the basic regression results by changing the core dependent and independent variables.

4.2.1. Robustness Test on the Direct Effect of the Digital Economy on Industrial Performance

To verify the robustness and sensitivity of the estimated results of the benchmark model, we first estimate models in Columns (1)–(5) by using different indicators of industry performance. ROTA measures industry performance from the perspective of net profit, and industry performance is also often measured from the perspective of productivity. For this reason, we use ROC as the measurement index, and the estimated results are shown in

Table 5. Robustness test 1 on the direct effect of the digital economy on the industrial performance of the manufacturing industry.

Independent Variables	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
	PROFIT2	PROFIT2	PROFIT2	PROFIT2	PROFIT2
DCC1	-	−3.331	−3.453	−4.727	−3.408
	-	(2.889)	(2.821)	(3.223)	(2.149)
DCC2	-	14.109 **	12.239 **	7.135	8.765 *
	-	(5.511)	(4.331)	(4.128)	(4.825)
DCC	−3.647	-	-	-	-
	(2.114)	-	-	-	-
SCALE	0.044 **	-	0.038 ***	0.023 *	0.045 **
	(0.016)	-	(0.007)	(0.013)	(0.016)
FOR	1.196 **	-	-	1.048 *	1.005 *
	(0.440)	-	-	(0.517)	(0.474)
PRI	1.802	-	-	-	1.819
	(1.401)	-	-	-	(1.378)
CONS	8.887 ***	9.256 ***	9.246 ***	9.030 ***	8.813 ***
	(0.182)	(0.103)	(0.088)	(0.123)	(0.165)
Industry fixed	Yes	Yes	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes	Yes	Yes
R2	0.357	0.270	0.303	0.330	0.373
N	105	105	105	105	105

Note: *** p < 0.01, ** p < 0.05, * p < 0.1, the values in brackets are robust standard errors. The coefficient of DCC2 in model (4) is not significant, but the p value is 0.106.

.

The results in Column (1) of Table 5 show that the coefficient of the digital economy fails to pass the significance test, indicating that the integration of the digital economy and the traditional manufacturing industry has no obvious facilitation effect on the performance of the manufacturing industry. The results in Columns (2)–(4) of Table 5 show that the coefficient of the digital manufacturing industry does not pass the significance test, and the coefficient of the digital service industry is significantly positive. The basic regression results in Table 3 do not change due to the change in the measurement of dependent variables, and the basic conclusions are robust.

In addition, considering that the integrated development of the digital economy and manufacturing industry may have an incremental effect on the performance of the manufacturing industry, and that the technological progress brought out by the digital economy has a lagging effect on industry performance, this paper uses the incremental value of the integration of the digital economy and the manufacturing industry as an independent variable and the increment of industry performance as a dependent variable to test the robustness of the results. The results are shown in

Table 6. Robustness test 2 on the direct effect of the digital economy on the industrial performance of manufacturing industry.

Independent Variables	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
	PROFIT2	PROFIT2	PROFIT2	PROFIT2	PROFIT2
DDCC1	-	−0.041	−0.039	−0.050	−0.012
	-	(0.095)	(0.092)	(0.085)	(0.084)
DDCC2	-	0.986 **	0.984 *	1.076 **	1.299 **
	-	(0.307)	(0.396)	(0.370)	(0.342)
DDCC	−0.005	-	-	-	-
	(0.084)	-	-	-	-
SCALE	0.003	-	0.001	0.001	0.003
	(0.002)	-	(0.002)	(0.002)	(0.002)
FOR	0.035	-	-	0.026 **	0.041 *
	(0.023)	-	-	(0.010)	(0.017)
PRI	0.052	-	-	-	0.057 *
	(0.027)	-	-	-	(0.025)
CONS	−0.021	0.009	0.007	-	−0.023 *
	(0.014)	(0.000)	(0.004)	-	(0.011)
Industry fixed	Yes	Yes	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes	Yes	Yes
R2	0.095	0.028	0.052	0.072	0.140
N	90	90	90	90	90

Note: ** p < 0.05, * p < 0.1, and the standard error in brackets are robust. The dependent variable is the added value of the profit rate of each industry compared with the previous period, and the core independent variable is the incremental value of the integrated development of the digital economy and the manufacturing industry.

.

The results in Column (1) of Table 6 show that the coefficient of the digital economy still fails to pass the significance test, indicating that the integration of the digital economy and the traditional manufacturing industry has no obvious facilitation effect on the performance of the manufacturing industry. The results in Columns (2)–(4) of Table 6 show that the coefficient of the digital manufacturing industry does not pass the significance test, and the coefficient of the digital service industry is significantly positive. The basic regression results in Table 3 do not change due to the change in the measurement of independent variables, and the basic conclusions are robust.

To sum up, the digital economy promotes industrial performance, and the conclusions do not change due to the change in the core independent and dependent variables.

4.2.2. Robustness Test on the Direct Effect of Digital Economy on Green Development

To test the robustness of the direct impact of the digital economy on carbon dioxide emissions, the dependent variables in Table 4 are replaced by the carbon dioxide emissions per unit market value, which is calculated as the carbon dioxide emissions divided by the total market value. The regression results are shown in

Table 7. Robustness test 1 on the direct effect of the digital economy on the green development of manufacturing industry.

Independent Variables	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
	CO2_Market	CO2_Market	CO2_Market	CO2_Market	CO2_Market
DCC1		2.073	2.013	3.910	2.977
		(0.598)	(0.579)	(0.919)	(0.620)
DCC2		−25.13 *	−26.06 *	−18.47 *	−19.62 **
		(−1.977)	(−1.986)	(−2.085)	(−2.287)
DCC	3.421				-
	(0.671)
SCALE	0.0268		0.0190	0.0412	0.0261
	(0.750)		(0.949)	(1.380)	(0.737)
FOR	−1.885			−1.560	−1.529
	(−1.289)			(−1.270)	(−1.126)
PRI	−1.255				−1.287
	(−0.749)				(−0.758)
CONS	1.064 ***	1.310 ***	1.332 ***	1.367 ***	1.161 ***
	(3.617)	(5.970)	(5.729)	(5.740)	(3.516)
Industry fixed	Yes	Yes	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes	Yes	Yes
R2	0.223	0.217	0.218	0.229	0.233
N	105	105	105	105	105

Note: *** p < 0.01, ** p < 0.05, * p < 0.1, the values in brackets are robust standard errors.

. It is revealed that after replacing the dependent variables with the carbon dioxide emissions per unit market value, the results still hold.

Similarly, we also use the increment of carbon dioxide emissions as a dependent variable to further test the robustness of the results. The results are shown in

Table 8. Robustness test 2 on the direct effect of the digital economy on the green development of manufacturing industry.

Independent Variable	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
	$∆$CO2_Market	$∆$CO2_Market	$∆$CO2_Market	$∆$CO2_Market	$∆$CO2_Market
DCC1		1.213	1.247	1.822	3.771
		(0.242)	(0.249)	(0.298)	(0.584)
DCC2		−85.60 **	−85.26 **	−83.45 **	−81.14 **
		(−2.921)	(−2.889)	(−2.925)	(−2.893)
DCC	5.155				-
	(0.749)
SCALE	0.0253		−0.0167	−0.00799	0.0171
	(0.676)		(−0.487)	(−0.211)	(0.457)
FOR	−1.772			−0.602	−0.850
	(−1.208)			(−0.416)	(−0.598)
PRI	2.693				2.581
	(1.156)				(1.271)
CONS	0.692	0.812	0.795	0.833	1.027
	(1.133)	(1.529)	(1.489)	(1.364)	(1.535)
Industry fixed	Yes	Yes	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes	Yes	Yes
R2	0.188	0.235	0.236	0.236	0.241
N	90	90	90	90	90

Note: ** p < 0.05, the values in brackets are robust standard errors.

. It is revealed that the regression results still hold if the incremental effects of carbon dioxide emissions are involved—that is, the integration of the digital service industry and the manufacturing sector significantly promotes a reduction in carbon dioxide emissions and the green development of the manufacturing industry.

Above all, Hypothesis 1 is firmly verified—that is, the integration of the digital service industry and the manufacturing sector significantly promotes industrial performance and the reduction in carbon dioxide emissions, thus contributing to the sustainable development of the manufacturing industry.

5. Mechanism Analysis

5.1. Mechanism Analysis on Industrial Performance: Reduction in Production Costs and Transaction Costs

The above basic regression and robustness test results show that the integration of the digital service industry and the traditional manufacturing industry can significantly improve the performance of the manufacturing industry; but how the digital economy plays this role is still worth further exploring. Therefore, this paper will verify the internal mechanism of the digital service industry enhancing the performance of the manufacturing industry. As mentioned above in the theoretical analysis, the digital service industry can not only directly improve the performance of the manufacturing industry, but indirectly improve it by reducing transaction and manufacturing costs. Therefore, this paper chooses the administrative expense ratio and sales expense ratio as proxy variables of the transaction cost, and the prime cost ratio as a proxy variable of production cost; and tests the effectiveness of the transmission mechanism with the mediation effect model. Specific regression results are shown in

Table 9. The mediation effect of selling expense between the digital service industry and the industrial performance of the manufacturing industry.

Variables	Model (1)	Model (2)	Model (3)
	PROFIT1	SER	PROFIT1
DCC2	1.176 *** (0.361)	0.131 (0.162)	1.242 ** (0.417)
Mediator variable (SER)	-	-	−0.501 (0.553)
Control variable	Yes	Yes	Yes
Industry fixed	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes
R2	0.788	0.682	0.791
N	105	105	105

Note: *** p < 0.01, ** p < 0.05, and the values in brackets are robust standard errors.

,

Table 10. The mediation effect of the administrative expense ratio between the digital service industry and the industrial performance of the manufacturing industry.

Variables	Model (1)	Model (2)	Model (3)
	PROFIT1	AER	PROFIT1
DCC2	1.176 *** (0.361)	0.112 (0.195)	1.198 *** (0.360)
Intermediate variable (AER)	-	-	−0.197 (0.356)
Control variable	Yes	Yes	Yes
Industry fixed	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes
R2	0.788	0.847	0.788
N	105	105	105

Note: *** p < 0.01, and the values in brackets are robust standard errors.

and

Table 11. The mediation effect of prime costs between the digital service industry and the industrial performance of the manufacturing industry.

Variables	Model (1)	Model (2)	Model (3)
	PROFIT1	PCR	PROFIT1
DCC2	1.176 *** (0.361)	−0.013 *** (0.001)	0.986 ** (0.394)
Mediator variable (PCR)	-	-	−0.468 *** (0.143)
Control variable	Yes	Yes	Yes
Industry fixed	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes
R2	0.788	0.550	0.846
N	105	105	105

Note: *** p < 0.01, ** p < 0.05, and the values in brackets are robust standard errors.

.

Table 9 reports the estimation results of the mediation effect model using the selling expense ratio (SE) as the mediation variable. The results in Column (2) of Table 9 show that the coefficient of the digital economy on the selling expense ratio fails to pass the significance test, indicating that the digital economy cannot effectively reduce the selling expense ratio. The results in Column (3) of Table 9 show that the coefficient of the selling expense ratio on industry performance does not pass the significance test, indicating that the selling expense ratio cannot effectively promote selling industry performance. The results in Table 9 indicate that the digital economy cannot promote the sustainable development of the manufacturing industry by reducing the selling expense ratio.

Table 10 reports the estimation results of the mediation effect model using the administrative expense ratio (AE) as the mediation variable. The results in Column (2) of Table 10 show that the coefficient of the digital economy on the administrative expense ratio fails to pass the significance test, indicating that the digital economy cannot effectively reduce the administrative expense ratio. The results in Column (3) of Table 10 show that the coefficient of the administrative expense ratio on industry performance does not pass the significance test, indicating that the administrative expense ratio cannot effectively promote selling industry performance. The results in Table 10 indicate that the digital economy cannot promote the sustainable development of the manufacturing industry by reducing the administrative expense ratio.

The results in Table 9 and Table 10 show that the digital service industry has no significant impact on either the administrative expense ratio or the sales expense ratio. Adding the two ratios to the benchmark model, respectively, has no significant impact on the performance of the manufacturing industry either, so transaction cost is not a mediating variable between the digital service industry and the performance of the manufacturing industry. Hypothesis 3 of this paper is not verified.

Table 11 reports the estimation results of the mediation effect model using the prime cost ratio (PC) as the mediation variable. The results in Column (2) of Table 11 show that the coefficient of the digital economy on the prime cost ratio (−0.013) is significantly negative, indicating that the digital economy can effectively reduce the prime cost ratio. It is evident from Column (3) that the coefficient for the relationship between the prime cost ratio and industry performance is significantly negative at the 1% level, at −0.468. When the prime cost ratio is added to the benchmark model, the regression coefficient of the direct input coefficient of the manufacturing industry to the digital service industry is still significant, at a 95% confidence level, indicating that production cost has a mediating effect between the digital service industry and manufacturing industry performance to some extent. Hypothesis 2 of this paper is verified.

5.2. Mechanism Analysis about Green Development: Promotion of Innovation

To test Hypothesis 4, the mediation model is constructed, in which the mediation variable is the innovation level of manufacturing industries measured by the innovation index of every industry. The innovation index was obtained from Kou and Liu [55], and it ranges from 2001 to 2016 and contains 80 industries.

The results from mechanism analysis are shown in

Table 12. The mediation effect of innovation between the digital service industry and the green development of the manufacturing industry.

Variables	Model (1)	Model (2)	Model (3)
	CO2_Asset	INNOINNO	CO2_Asset
DCC2	−84.08 *	4.022 *	1.373
	(−2.111)	(2.126)	(0.408)
Mediator variable (INNO)	-	-	1.396 **
	-	-	(2.329)
CONS	2.155 ***	1.264 ***	−0.614
	(4.235)	(10.11)	(−0.907)
Control variable	Yes	Yes	Yes
Industry fixed	Yes	Yes	Yes
Time fixed	Yes	Yes	Yes
R2	0.790	0.984	0.794
N	75	75	75

Note: *** p < 0.01, ** p < 0.05, * p < 0.1, and the values in brackets are robust standard errors.

. In Column (2), the dependent variable and independent variable are the innovation index and the integration of the digital service industry and the manufacturing sector, respectively. The results indicate that the digital economy can significantly increase the innovation level of the manufacturing industry. The results in Column (3) prove that innovation has a significant full mediation effect in the relationship between the digital economy and carbon dioxide emissions, meaning that Hypothesis 4 holds.

6. Discussion and Conclusions

6.1. Conclusions

Digital transformation is the key for China to achieve industrial structure upgrading and the sustainable development of the manufacturing industry. Based on the role of the integration of the digital economy and the traditional manufacturing industry in the development of the manufacturing industry, and the internal mechanism of this role, this study focused on the analysis of the impact of the digital economy on the sustainable development of the manufacturing industry. Specifically, based on the input–output tables, this study used the direct input coefficient of the manufacturing industry to the digital economy to measure the integration level of the digital economy and the manufacturing industry, and tested the impact of the digital manufacturing industry and the digital service industry on industry performance and the green development of manufacturing sectors, respectively.

Our main research conclusions are as follows. First, digital services can significantly improve the industrial performance of the manufacturing industry, while digital products have no significant impact on the industrial performance of the manufacturing industry. Second, digital services can promote the industrial performance of the manufacturing industry through the intermediary mechanisms of reducing production costs rather than transaction costs. Finally, digital services can also reduce carbon emissions and promote the green development of the manufacturing industry by increasing the innovation level. In conclusion, digital services can promote the sustainable development of China’s manufacturing industry.

6.2. Marginal Contribution to Related Research

The present study also contributes to the existing literature. First, most of the existing papers conduct studies from a regional perspective, such as within provinces [6,7] or cities [9,10,11,12]; only a few papers conduct empirical research at the industry level [54]. This paper analyzes the impact of the digital economy on the sustainable development of manufacturing sectors at the industry level. Second, this paper uses the direct input coefficient to measure the degree of integration between traditional manufacturing sectors and the digital economy, providing a method to measure the degree of digitization in manufacturing sectors. Most of the existing papers use the entropy method to construct a measurement index of the digital economy, which lacks clear economic connotations [11,62]. Third, differing from the existing studies, this paper divides the digital economy into digital products and digital services, and analyzes the impact of each on the sustainable development of the manufacturing industry, providing new ideas for digital economy-related research. The existing studies largely ignore the heterogeneity of the impact of digital products and digital services on manufacturing.

6.3. Implications

The findings of this paper are relevant for the digital transformation of the manufacturing industry: (1) The government should speed up the construction of digital infrastructure and actively promote the digital service industry, such as cloud computing service platforms and the industrial Internet. As found in this paper, the application of products of the digital economy has no significant impact on the performance of the traditional manufacturing industry; what really matters is the services of the digital economy; (2) Firms should give full play to digital technology and extend its application to all areas, such as organization, management, production and sales. As found in this study, China has not yet realized the pathway of reducing the transaction cost and thus enhancing the performance of the manufacturing industry by leveraging the digital service industry. Manufacturing firms should make full use of digital technology in management, operation, and production processes.

This study analyzed the influence of the digital economy on the sustainable development of the manufacturing industry. It complements the empirical evidence on this research question. However, there are still some limitations to this in the research work. First, in this study, the sustainable development of the manufacturing industry is defined from the perspective of industrial performance and green development, but the sustainable development of the manufacturing industry also includes other connotations, such as lower income disparities, which should be considered in future research. In addition, whether the research of this paper can be extended to the sustainable development of manufacturing in the international scope is yet to be elucidated.