Does the Digital Economy Improve the Innovation Efficiency of the Manufacturing Industry? Evidence in Provincial Data from China

Abstract

This paper selects panel data on 30 Chinese provinces from 2011 to 2020 and uses an econometric approach to empirically examine the impact of development in the digital economy on the innovation efficiency of China’s manufacturing industry. The results of this study show that during the period under investigation, the innovation efficiency of China’s manufacturing industry is in a stage of continuous improvement, and the development in the digital economy significantly promotes the level of manufacturing innovation efficiency, becoming an important driving force for the improvement in manufacturing innovation efficiency in the digital era. Moreover, the digital economy not only directly promotes manufacturing innovation efficiency but also has an indirect impact on such efficiency with industrial structure upgrading and technology spillover. With upgrading industrial structure and technology spillover, the digital economy has an enhanced effect on the innovation efficiency of the manufacturing industry. Using an analysis with the spatial Durbin model, it is found that the digital economy has a positive impact on the innovation efficiency of the manufacturing industry with spatial spillover characteristics.

1. Introduction

The manufacturing sector is the “ballast stone” of a large country’s economy, a significant component of the real economy, and the fastest-growing sector in terms of productivity in the national economy. This sector is crucial for driving economic growth and improving employment quality. Since the reform and opening up of the country over 40 years ago, China’s economy has achieved high growth rates, and the expansion of the manufacturing sector and the increase in its share of gross domestic product (GDP) have become indispensable. Since 2012, China’s manufacturing sector has increased its global share from 22.5% to nearly 30%, ranking first in the world in terms of volume for 11 consecutive years and accounting for nearly 30% of the global manufacturing sector; thus, this sector is considered the cornerstone of China’s long-term economic growth. The term “innovation-driven” means that economic growth relies primarily on innovation in science and technology to increase the output rate of production factors with technological change, thus achieving an intensive growth approach that most rationally and effectively promotes sustainable economic and social development. Manufacturing is the fastest-growing sector in the national economy in terms of productivity and is also the most active area of innovation. Innovation is the key to the high-quality development of the manufacturing industry. In 2020, the expenditure on R&D in China’s regulated industrial enterprises reached RMB 1527.13 billion, accounting for 62.60% of the national expenditure on R&D. In that same year, the number of patent applications was 1,243,927, accounting for 23.95% of all national patent applications. China’s manufacturing value added as a proportion of GDP experienced a process of rising then falling during 2004–2019, reaching a peak of 32.45% in 2006, and showing a faster decline after 2012, eventually falling to 27.17% in 2019. The scale of innovation input and output has increased significantly, yet the proportion of the manufacturing industry on this scale is decreasing year by year, with innovation output corresponding to innovation input being the embodiment of innovation capability; innovation output is also the key to restricting the innovation development of the manufacturing industry.

Since 2015, countries worldwide have introduced policies related to the digital economy, considering its development as the key to achieving prosperity, maintaining competitiveness, and making deployments to attain world leadership and comprehensively promoting digital transformation. Countries worldwide have accelerated the layout of new infrastructure, with new information infrastructure represented by 5G, artificial intelligence, and the Internet of Things. The industrial Internet and satellite Internet have gradually become the new driving force behind global economic growth. Digital industrialisation and industrial digitisation go hand in hand, with manufacturing becoming the main battleground of the digital economy. Innovation is an important force promoting economic development and social progress, and sustainable development is a necessary condition for the development of a country and region. Innovation and sustainable development are intrinsically linked, and while continuously promoting innovation, we should consider the direction and goals of sustainable development to achieve the harmonious development of economic, social, and environmental aspects.

In recent years, a proportion of the manufacturing industry has been experiencing a premature and rapid decline, the innovation capacity in key areas has become insufficient, and the supply chain of the industrial chain has become constrained by other factors that have not yet fundamentally changed. Therefore, it is necessary to accelerate the transformation and upgrading of traditional manufacturing industries, improve the innovation capacity of manufacturing industries, and promote the application of digital technology in manufacturing industries. The transformation and upgrading of the manufacturing industry are fundamental to the construction of a new “double cycle” model for China’s economy. The basis for quality development is a higher-level and more competitive manufacturing industry. Against the backdrop of current trade frictions and the outbreak of the new pneumonia epidemic, the global supply chain has been affected, placing higher demands on China’s manufacturing sector. Given the background of a booming digital economy, how efficient is China’s manufacturing industry in terms of innovation? Can the development of the digital economy contribute to the efficiency of innovation in China’s manufacturing sector? Through which channels does the digital economy affect the efficiency of innovation in China’s manufacturing industry? Considering the significant differences in economic development and other conditions across regions in China, what are the spatial and temporal differences in the impact of the digital economy on manufacturing innovation efficiency in different regions at different stages of development? By exploring the above questions, we can deepen our understanding of the relationship between the digital economy and the innovation efficiency of China’s manufacturing industry and provide specific ideas for promoting the development of the digital economy and innovation development in the manufacturing industry.

The major contributions of this paper are summarised as follows: (1) A more comprehensive construction of China’s provincial-level digital economy development-level indicator system is used to measure the development of the digital economy in five dimensions, including digital infrastructure, digital industrial development, industrial digital development, digital development environment, and digital inclusive finance. A superefficient slacks-based measure (SBM) data envelopment analysis (DEA) model is used to measure the manufacturing innovation efficiency of each province from input and output perspectives. On the output side, the output process of high-tech industries is divided into two stages: the output of knowledge and technology and the conversion of that knowledge and technology into new products, providing market value. (2) The empirical analysis in this study finds that the digital economy not only promotes manufacturing innovation efficiency directly but also has an indirect impact on manufacturing innovation efficiency with industrial structure upgrading and technology spillover. With upgrading industrial structure and technology spillover, the digital economy has an enhanced effect on the innovation efficiency of China’s manufacturing industry. Using an analysis with the spatial Durbin model, it is found that the digital economy has a positive impact on manufacturing innovation efficiency with spatial spillover characteristics. This not only enriches the empirical evidence on the economics of evolution and innovation in the digital economy but also provides a solid empirical basis on which the government can better promote innovation development.

The remainder of this paper is organised as follows. Section 2 presents a literature review. Section 3 presents the theoretical analysis and research hypotheses. Section 4 presents the research methodology, including the data structure, research model, and variables. Section 5 presents the data and results from the empirical analysis, demonstrating the measurement results on the digital economy and discussing the impact of the digital economy on the efficiency of manufacturing innovation and spatial spillover effects. Section 6 concludes this paper and proposes suggestions for future research.

2. Literature Review

2.1. Digital Economy

The concept of the digital economy can be traced back to the mid-1990s. At that time, the US government promoted the development of information technology and the Internet, and the US economy continued to grow at an average annual rate of 4% in GDP for more than a decade due to the positive interaction between economic policy and technological innovation [1]. In this context, Tapscott (1996) [2] formally introduced the concept of the “digital economy” and clearly defined the knowledge-driven, virtualised, disintermediated, and globalised characteristics of the digital economy. The emergence of the digital economy attracted the attention of international organisations, national government statistical agencies, and relevant scholars to the statistics and measurement of the size of the digital economy. The research in this area can be divided into four main categories [3]: first, methodological research on national accounts [4]; second, research on value-added measurements [5]; third, research on relevant indices [6,7,8,9,10]; and fourth, research on the construction of satellite accounts [11]. The Organisation for Economic Co-operation and Development (OECD) selected 38 indicators with which to measure the level of development of the digital economy covering four areas: investing in smart infrastructure, empowering society, unleashing innovation, and achieving growth and employment [6]. Liu et al. measured two dimensions: digital foundation and digital application [7]. Yang et al. constructed a digital economy development-level indicator system from the perspective of the digital industry [8]. Jiao et al. integrated previous studies and added digital innovation-related indicators to enrich the index dimension [9]. However, a prominent problem is the very large differences in the results using various measurements, which make understanding the scale of the digital economy not only unclear but also more confusing, and thus, such measurements are not conducive to having an accurate grasp of the digital economy.

2.2. Innovation Efficiency of the Manufacturing Industry

Innovation efficiency is defined in different ways by academics. Some domestic scholars believe that innovation efficiency is an absolute concept and that its connotation is externalised to the innovation input spent per unit of innovation outcome or the innovation outcome obtained per unit of technological innovation input based on the environment of innovation resource allocation [12]. Some foreign scholars argue that innovation efficiency is a relative concept, focusing on the optimal allocation of innovation factors between inputs and outputs. The second type of research is the measurement of regional innovation efficiency. These measurement methods are broadly divided into parametric and nonparametric techniques. Nonparametric estimation is performed using methods such as data envelopment analysis (DEA) [13,14] and factor analysis [15], and parametric estimation is performed using methods such as stochastic frontier methods [16,17,18,19]. The third type of study is the regional innovation classification study. Based on the distinction between “innovation capability” and “innovation efficiency”, some scholars have empirically examined the spatial spillover effects of the development of the digital economy and the quality improvement in innovation capability [20]. Some scholars have noticed that there are differences in the outputs obtained at different stages of innovation and have thus classified regional innovation efficiency into business innovation efficiency and technological innovation efficiency [21]. In this paper, efficiency is defined as technical efficiency, i.e., the ability to maximise output for a given input or minimise input for a given output, and innovation efficiency is the technical efficiency of innovation.

2.3. Digital Economy and Innovation Efficiency

The most relevant research on the innovation effects of the digital economy is focused on the innovation effects of the Internet. There are different views on the role of the Internet in innovation. On the one hand, some scholars have shown that the Internet has a significant positive impact on innovation, both at the international and domestic regional levels as well as at the micro-firm level. In both high- and low-income countries, technological advances in the Internet have a significant technical efficiency-enhancing effect [22]; moreover, the diffusion of information and knowledge using the Internet facilitates enterprise innovation [23], with more efficient firms and those with greater learning and absorptive capacity benefiting more from Internet innovation spillover effects [24]. The Internet makes a significant contribution to regional innovation efficiency in China, with some lagging effect on part of this positive spillover effect [25]. On the other hand, some scholars have argued that the Internet is not conducive to innovation, arguing that overcrowded Internet investment has a “crowding out” effect on other resource inputs, which is detrimental to enterprise innovation [26]. Moreover, such scholars argue that the community effect of Internet development has contributed to the formation of monopolies in the Internet industry, weakening the initiative and motivation of enterprises to engage in competitive innovation activities [27]. The contribution of Internet resources to regional innovation capacity is characterised by a “high-level trap”, i.e., the more crowded the input of Internet resources is, the weaker its contribution to regional innovation [28].

At the macro-level, the literature considers mainly the factors influencing innovation efficiency in terms of government behaviour, economic environment, foreign investment, and internet construction [29] in addition to other influencing factors such as manufacturing innovation efficiency in terms of technology introduction cost, industrial outwards orientation, property rights structure, and enterprise-scale [30,31,32]. Most studies on the impact of the Internet on regional innovation efficiency are closely related to studies on the digital economy, while few studies examine the impact of the development of the digital economy on innovation efficiency in the real economy and its mechanism of action. At the micro-level, some scholars explored the relationship between digital development and enterprise innovation, and the empirical results affirm the positive impact of digital technology development on enterprise innovation [33,34].

3. Theoretical Analysis and Research Hypotheses

3.1. Theoretical Analysis by Which the Development of the Digital Economy Affects the Efficiency of Manufacturing Innovation in China

Compared with the traditional economy, the digital economy denotes the industrialisation and marketisation of the information technology revolution. It has undergone comprehensive changes in terms of production factors, production relations, and productivity [1]. The manufacturing industry, with its intensive factor inputs, complex production processes, and significant scale effects, is a carrier of innovation activities, promoting the industrialisation of innovation results and fostering a positive innovation cycle. Therefore, this industry is an important factor in establishing and maintaining an innovative economy in a country [30]. Schumpeter’s innovation theory suggests that innovation is congruent with recombining production factors. Digital technology innovation, such as information and communication technologies (ICTs), can effectively integrate factors such as labour, raw materials, and equipment into the production process and promote the digitisation of enterprises’ production processes and procedures, thereby reducing energy consumption in the production process and promoting technological innovation in enterprise production. The direct impact of the digital economy on the innovation efficiency of China’s manufacturing industry can be seen in terms of factors of production, and thus, digital data resources have become a key factor of production. Data, as a new production factor, not only improve the quality and efficiency of use for other factors of production, such as labour and capital, but data also change the way economic activities are organised, accelerate the reorganisation of resources, improve total factor productivity, and inject new momentum into modern economic development [3]. In terms of production relations, network infrastructure constitutes a new type. The emergence of the digital economy has constructed a new type of production relationship with modern information networks, digital infrastructure, and digital platforms. The original information segregation and information monopoly have been broken by digital transformation, prompting more transparent cooperation between people and between sectors. In terms of productivity, digital technology has become an important productive force. Digital technology is the technical basis for the development of the digital economy and has “positive externalities”, with exponential growth effects, generating very-high-level development momentum. The enhancement of innovation capabilities can bring automated and intelligent production equipment to the traditional manufacturing industry, effectively reducing resource losses and related costs in the production process, thus improving the production efficiency, capital-to-output ratio, and profitability of this industry [32].

The digital platform derived from the digital economy has a connectivity function that strengthens the interaction between enterprises and consumers. This interaction not only reduces the business risks caused by the mismatch between supply and demand but also creates a new situation in which consumers’ diversified needs drive enterprises to generate product innovation, avoiding ineffective production and resource waste, and ultimately promoting enterprise product innovation. Digital technology can improve the production efficiency of enterprises, solve the imbalance between supply and demand in the market, and promote the transformation and upgrading of enterprises, thus promoting the upgrading of the industrial chain. On the one hand, the momentum of the transformation and upgrading of the manufacturing industry originates from expanding the boundary of the industrial chain organisation in terms of division of labour, reduction in transaction costs, transfer of value distribution and demand change reversal under digital technology [14], On the other hand, the digital economy promotes the transformation and upgrading of the manufacturing industry from four aspects—data-driven, innovation-driven, demand-driven, and supply driven—and emphasises the need to guide the integration and development of the manufacturing industry using the Internet and new technologies [35]. At the same time, the digital economy can expand the scope of economic applicability, significantly reduce production costs, and maximise the effect of economies of scale. Larger enterprises tend to perform better at the same level of digitalisation as that of smaller enterprises. In addition, digital transformation promotes the servitisation of manufacturing and the development of modern manufacturing services and thus has a positive impact on strengthening the manufacturing industry’s independent innovation capacity.

The digital operation of enterprises in the production, management, and sales interaction processes creates a large amount of data but also gradually becomes an important source of profit and core competitiveness [36]. The enterprise digital process formed in the database for statistical analysis and trend fitting can accurately predict future market demand and innovation trends, reduce uncertainty in research and development, and avoid blind innovation. The digital development of the manufacturing industry is a strategic choice for a conceptual change in the traditional manufacturing industry and a necessary path for transformation and upgrading. The digital economy for manufacturing transformation and upgrading refers to the process of traditional industries using digital technology to reshape their business and improve quality and efficiency. Such industries specifically rely on open and shared data resources to accurately match diverse product supply and heterogeneous user demand, promote the transformation of manufacturing from large-scale production to personalised customisation, and realise the precision of manufacturing services and the digitalisation of manufacturing processes [14]. The target path relies mainly on the deep integration of digital technology, the traditional manufacturing industry, and the driving role of technological innovation, data elements, precise demand and effective supply.

Hypothesis 1.

The digital economy has contributed significantly to improving innovation efficiency in China’s manufacturing sector.

3.2. Transmission Path for the Development of the Digital Economy Affecting the Innovation Efficiency of China’s Manufacturing Industry

The digital economy is not only a growth point for economic development but also a support point for upgrading the industrial structure [37]. The ability to innovate independently is a top priority for long-term industrial restructuring, and an improvement in innovation capacity stems not only from the continuous increase in R&D investment but also from the improvement in innovation efficiency. The digital economy influences upgrading the industrial structure by transforming traditional industries, giving rise to new industries, and reshaping the structure of supply and demand. Benefiting from the high-synergy, low-cost, and high-penetration characteristics of the digital economy, the development of the digital economy has a positive effect on upgrading the industrial structure [38,39]. First, the digital economy is conducive to deepening the interconnectedness of industries. The widespread application of digital technology has broken down the resource and technological barriers between regions and improved the relatively independent industrial dynamics of traditional industries in the past, and the interconnectedness of industries has increased with the penetration of digital devices and technologies such as networks, communications, and computers. Second, the digital economy has revolutionised the development model and evolutionary path of traditional industries. Relying on ICT, the digital economy has changed the traditional transaction model and improved matching efficiency by reducing transaction costs. The application of new technologies has given rise to new products and models, brought about new levels of supply and demand, and promoted innovation in new industries and the transformation of traditional industries. Third, the digital economy promotes the development of industrial integration. The versatility and permeability of digital technology have enabled the digital economy to expand its scope of action from the Internet and information industry to the consumer and production sectors and to deepen its integration with the real economy in the process of cross-industry integration.

Under the rapid development of the digital economy, information and data are key factors of production, with which new business models are constantly being created and digital infrastructure is constantly being improved. All of these factors can effectively promote knowledge and technology spillovers within and between regions, thereby improving the efficiency of industrial innovation in the region. On the one hand, the development of the digital economy can effectively reduce the cost of technology spillovers. “Moore’s law” shows that technological progress, especially the accelerated renewal and iteration of digital technologies, continues to lead to a decline in the price of digital technologies [40]. Unlike traditional production, innovations in digital technology can often be produced on a large scale, and due to their replicability, non-depletion, and low marginal cost, these innovations can rapidly become part of all aspects of manufacturing innovation, promoting the efficiency of innovation and product quality in the region [41]. When knowledge, information, and technology are diffused and shared within the region, the manufacturing industry gains good innovation-driven and learning environments. Such production services are technology-intensive industries that can quickly absorb the most cutting-edge new technologies, methods, and processes and apply them to other intermediaries and manufacturing enterprises in the region. This process is conducive to manufacturing enterprises’ involvement in the high-end aspects of the value chain and the enhancement of manufacturing technology. Due to their geographical proximity and economic interconnectedness, enterprises with different characteristics influence each other and generate knowledge spillover, developing together in terms of accumulating a large amount of knowledge and technological capital, reducing innovation costs, and improving innovation efficiency.

Hypothesis 2a.

The digital economy indirectly contributes to innovation efficiency in China’s manufacturing sector by upgrading the industrial structure.

Hypothesis 2b.

The digital economy indirectly contributes to manufacturing innovation efficiency in China through technology spillovers.

3.3. Spatial Spillover Effects

Given the continuous development of new economic geography theory, it is believed that economic development is closely related to neighbouring regions, i.e., spatial externalities exist. There is a close relationship between innovation subjects in each region in innovation activities, with such subjects being influenced by other innovation subjects in the adjacent space. In the process of manufacturing innovation, the innovation ability of innovative enterprises is not only determined by their own strength but also influenced by the interactive effects between innovative enterprise subjects within the region or between regions, especially under the development of the digital economy.

According to regional innovation theory, innovative knowledge and technology are highly specific to geographical locations, and economic exchanges and mutual imitation are more likely to occur between regions that are geographically close or that have similar technological standards, resulting in innovation agglomeration [42]. On the one hand, there are significant differences in resource endowment and economic development among regions, as well as different levels of talent structure and capital development. Hence, innovation factors, such as innovation talent and capital, flow between neighbouring regions due to physical distance, time, and other objective factors, and the process of factor flow can generate a certain degree of knowledge spillover between regions, thus creating a spatial spillover effect on the efficiency of manufacturing innovation in surrounding regions. On the other hand, the digital economy has greatly facilitated the transfer of information between regions with the “Metcalfe’s law” effect and the “network effect” of the Internet, compressing the space–time distance between regions, making the spatial clustering and diffusion of innovation factors simpler and more flexible, and thus enhancing the depth and flexibility of economic exchanges between regions. The spatial clustering and diffusion of innovation factors become easier and more flexible, thus enhancing the depth and breadth of economic exchanges between regions [43] and expanding the spatial spillover effects with respect to the innovation efficiency of the manufacturing industry in surrounding regions. Due to the large geographical differences across different regions in terms of talent structure and innovation resources, such as capital and the level of development of the digital economy, the spatial spillover performance of manufacturing innovation efficiency in surrounding regions may vary across regions. At the same time, China has had different development goals at different stages of economic development, and the digital economy has developed rapidly along with the emergence of breakthrough mobile network technologies such as 3G and 4G. Thus, the spillover effect of the digital economy on manufacturing innovation efficiency at different time stages may also have different characteristics. Therefore, the following research hypothesis is proposed:

Hypothesis 3.

The digital economy affects manufacturing innovation efficiency in neighbouring regions through spatial spillovers, and there are spatial and temporal differences in such spatial spillovers.

4. Methodology

4.1. Model Construction

4.1.1. Baseline Regression Model

Based on the theoretical analysis, to study the direct impact mechanism of the digital economy on the efficiency of regional manufacturing innovation, a general panel econometric model is constructed as follows:

$$\ln t{e}_{it}={\alpha }_0+{\alpha }_1\ln di{g}_{it}+{\alpha }_2{X}_{it}+{\delta }_t+{\lambda }_i+{\epsilon }_{it}$$

(1)

In Equation (1), i denotes the province, t denotes time, $\ln t{e}_{it}$ denotes the regional manufacturing innovation efficiency level, $\ln di{g}_{it}$ denotes the regional digital economy development level, X_it denotes a set of control variables, ${\alpha }_0$ denotes the model intercept term, ${\alpha }_1$ and ${\alpha }_2$ denote the regression coefficients of the corresponding variables, δ_t denotes time fixed effects (FE), ${\lambda }_i$ denotes individual FE, and ${\epsilon }_{it}$ denotes the random disturbance term.

Due to a possible endogeneity problem in Equation (1), it is further estimated using a generalized method of moments (GMM) model, which is able to solve the endogeneity problem caused by the one-period difference in the lagged innovation efficiency and consists of two methods: systematic GMM (SYS-GMM) and first-order-difference GMM (DIF-GMM) estimation. Compared to DIF-GMM estimation, SYS-GMM estimation, which uses a level-valued lag term as the instrumental variable for the difference variable, uses both the level and difference equations in the estimation process. Therefore, the relatively more robust SYS-GMM method is finally chosen, and the reasonableness of the model setting is judged using Hansen’s test and an AR (2) test, with the following model setting:

$$\ln t{e}_{it}={\alpha }_0+{\alpha }_1\ln t{e}_{it-1}+{\alpha }_2\ln di{g}_{it}+{\alpha }_3{X}_{it}+{\epsilon }_{it}$$

(2)

4.1.2. Multiple Mediating Effects Model

To test the existence of industrial structure and technology spillover transmission channels, the SYS-GMM stepwise regression method is used to test the specific model settings as follows:

$$\ln t{e}_{it}={\alpha }_0+{\alpha }_1\ln t{e}_{it-1}+{\alpha }_2\ln di{g}_{it}+{\alpha }_3{X}_{it}+{\epsilon }_{it}$$

(3)

$$\ln Me{d}_{it}={\beta }_0+{\beta }_1\ln Me{d}_{it-1}+{\beta }_2\ln di{g}_{it}+{\beta }_3{X}_{it}+{\epsilon }_{it}$$

(4)

$$\ln t{e}_{it}={\gamma }_0+{\gamma }_1\ln t{e}_{it-1}+{\gamma }_2\ln di{g}_{it}+{\gamma }_3\ln Me{d}_{it}+{\gamma }_4{X}_{it}+{\epsilon }_{it}$$

(5)

In the above equation, $\ln Me{d}_{it}$ represents the mediating variable; ${\alpha }_0$, ${\beta }_0,$ and ${\gamma }_0$ represent the model intercept terms; and ${\alpha }_1, {\alpha }_2$, ${\alpha }_3$, ${\beta }_1$, ${\beta }_2$, ${\beta }_3$, ${\gamma }_1$, ${\gamma }_2, {\gamma }_3$, and ${\gamma }_4$ represent the regression coefficients of the corresponding variables, where ${\alpha }_2$ represents the total effect of the digital economy on manufacturing innovation efficiency, ${\beta }_2$ represents the effect of the digital economy on the mediating variable, ${\gamma }_2$ represents the direct effect of the digital economy on manufacturing innovation efficiency after accounting for the mediating effect, ${\gamma }_3$ represents the effect of the mediating variable on manufacturing innovation efficiency, and the remaining indicators have the same meaning as above. The existence of the mediating effect is determined mainly by the significance of ${\alpha }_2$, ${\beta }_2$, and ${\gamma }_3$. When all three values pass the significance test, a mediating effect is shown to exist.

4.2. Metrics

4.2.1. Explained Variables

Innovation efficiency (te). In this paper, efficiency is defined as technical efficiency, i.e., the ability to maximise output for a given input or minimise input for a given output [44], and innovation efficiency is the technical efficiency of innovation [45].

This paper uses the superefficient SBM-DEA model to measure the efficiency of manufacturing innovation in each province. On the input side, inputs to R&D innovation activities usually include both personnel and financial inputs. The full-time equivalent of R&D personnel is chosen to measure personnel input, i.e., full- and part-time staff add up to the total amount of workload after discounting it on a full-time basis. The internal expenditure of R&D funds is used to measure the financial input of R&D activities, which includes the direct costs of research projects and indirect costs, such as management and service costs. Therefore, this paper selects the internal expenditure stock of R&D funds to measure capital input and uses the perpetual inventory method to convert the internal expenditure of R&D funds into capital stock by referring to Xie Ziyuan [46], the equation of which is as follows:

$$k_{it}=\theta i\left(t-1\right)+\left(1-\delta \right)k_i\left(t-1\right)$$

(6)

where $\delta$ represents the depreciation rate of funding, which is set at 15% according to the practice of most scholars; $\theta (t-1)$ represents the funding input after discounting in $t-1; k_i(t-1)$ represents the internal expenditure stock of R&D funding in year t−1; and the initial value of the internal expenditure stock of R&D funding is $K_{i0}={\theta }_{i0}/(g+\delta )$, where g represents the average annual growth rate of internal expenditure of R&D funding. The R&D funding index is constructed by Li et al. [47]: R&D funding index = fixed asset investment. The R&D expenditure index is constructed using the method of Li et al. [47]: R&D expenditure index = fixed asset investment index * 46% + consumer price index * 54%.

On the output side, the output process of high-tech industries is divided into two stages. The first stage is the output of knowledge and technology, with patents being a common indicator of research results at this stage, and the number of patents applied for is used in this paper to characterise the level of innovation output. The second stage is the conversion of knowledge and technology into new products or providing market value, and the benefits of the industrialisation of R&D results are usually measured using the sales revenue of new products.

4.2.2. Explanatory Variables

The level of development of the digital economy (dig). This paper constructs a digital economy development-level indicator system with five dimensions, including digital infrastructure, digital industry development, digital economy application, digital development environment, and digital financial inclusion (see

Table 1. Table of indicators for the level of development of the digital economy.

	Tier 1 Indicators	Secondary Indicators
Level of development of the digital economy	Digital infrastructure	Total telecommunications services per capita
		Internet penetration rate
		Mobile phone penetration rate
	Digital industry development	ICT industry output as a percentage of GDP
		ICT industry fixed asset investment as a proportion of total fixed-asset investment
		ICT industry employment as a proportion of total employment
	Digital development of industry	Industrial value added
		Expenditure on technical transformation of industrial enterprises above a designated size
	Digital development environment	GDP per capita
		R&D investment as a percentage of GDP
		Proportion of people with tertiary education or above
	Digital financial inclusion	China Digital Inclusive Finance Index

). Among them, digital infrastructure directly reflects the existing level of the hardware conditions in the digital economy; digital industry development reflects the degree of digitalisation of related industries; digital economy application reflects the degree of penetration of digital technology into existing industries; the digital development environment reflects the degree of existing social support for the development of the digital economy; and digital financial inclusion is an important reflection of the development of the digital economy. The measurement of digital financial inclusion uses the China Digital Inclusive Finance Index, jointly compiled by the Digital Finance Research Centre of Peking University and Ant Financial Services Group. The index measures three main aspects: the breadth of digital financial coverage, depth of use, and degree of digitalisation. To eliminate the influence of scale, this paper standardises the data and applies the global principal component method to calculate the level of digital economic development in each province in China from 2011 to 2020.

4.3. Intermediate Variables

Level of industrial structure upgrading (upg). Industrial structure upgrading refers to the evolution of the industrial structure from a lower form to a higher form, from the dominance of the primary industry to that of the secondary and tertiary industries. As of 2020, China’s primary, secondary, and tertiary industries accounted for 8%, 38%, and 54% of GDP, respectively, and the tertiary industry accounted for half of the national economy. This paper draws on the methodology of Ke Jun [48] to measure the level of industrial structure upgrading in each province using the industrial structure hierarchy coefficient and the following formula:

$$\mathrm{TN}={\displaystyle \sum _{i=1}^3{q}_i\times i}$$

(7)

where $q_i$ is the share of the output value of industry $i$. The higher the industrial structure level coefficient is, the higher the level of industrial structure upgrading. At the same time, the industrial structure level coefficient focuses on a comparison of the level of industrial structure development at different times and in different provinces rather than reflecting its absolute level.

Technology spillover (sp). Inter-regional knowledge, stock gaps, and spatial distance are important factors affecting a region’s access to technology spillover. Drawing on Caniels (2000) [49], this paper uses a knowledge spillover model to measure the level of interregional technology spillover by measuring the technology spillover received by the region.

4.4. Control Variables

(1) Quality of personnel in the industry (lq). In terms of the internal environment for innovation, maintaining the smooth running of innovation activities and the efficiency of innovation depends mainly on the quality of R&D personnel, which is expressed by the number of personnel in the science and technology institutions/average number of employees. (2) Innovation climate (ia). In terms of the external environment for innovation, an increasing number of science and technology institutions are being created, which leads to competition and cooperation, thus creating a strong atmosphere for innovation, as expressed by the number of science and technology institutions/number of enterprises in this paper. (3) Policy support (gii). The government’s behaviour reflects its investment preferences to a certain extent, and financial expenditure on science and technology has an impact on innovation efficiency to varying degrees. (4) Market structure (ms). Schumpeter’s theory suggests that industries with a high level of market concentration can help motivate enterprises to conduct R&D activities, and the market structure can reflect the degree of competition in the market. This structure is represented by the number of industrial enterprises above the scale selected in this paper.

4.5. Data Sources

Using 30 provinces in China from 2011 to 2020, the data are obtained mainly from the China Statistical Yearbook, China Science and Technology Statistical Yearbook, China Industrial Statistical Yearbook, China Electronic Information Industry Statistical Yearbook, China Internet Development Statistical Report, provincial statistical yearbooks, CEE statistical database and Wind database, taking into account that data from Tibet and Hong Kong, Macao, and Taiwan are missing. The missing data from Tibet and Hong Kong, Macao, and Taiwan are excluded, and the missing values for the indicators are linearly interpolated. The results for the descriptive statistics for all variables are shown in

Table 2. Descriptive statistics.

Type of Indicator	Variable	Variable Symbol	Sample Size	Average	Standard Deviation	Minimum Value	Maximum Value
Explained variable	Innovation efficiency	te	300	0.503	0.332	0.078	3.212
Explanatory variable	Level of development of the digital economy	dig	300	0.606	0.042	0.556	0.739
Intermediate variable	Level of industrial structure upgrading	uis	300	0.012	0.011	0.001	0.091
	Technology spillover	cs	300	0.012	0.024	0.001	0.161
Innovative environment	Quality of industry personnel	lq	300	0.534	0.529	0.000	2.585
	Innovative atmosphere	ia	300	0.523	0.042	0.312	0.973
	Policy support	gii	300	0.102	0.112	0.271	0.584
	Market structure	ms	300	1.222	1.367	0.056	6.436

.

5. Empirical Study and Results

5.1. Baseline Regression Results

Table 3. Table of baseline regression results.

	OLS		FE		GMM
	(1)	(2)	(3)	(4)	(5)	(6)
L.lnte					0.052 ***	0.062 ***
					(0.049)	(0.004)
lndig	0.282 ***	0.276 ***	0.253 ***	0.183 ***	0.419 ***	0.443 ***
	(0.026)	(0.022)	(0.021)	(0.031)	(0.012)	(0.020)
lnlq		0.022		−0.091 ***		0.142 ***
		(0.037)		(0.024)		(0.005)
lnia		0.003		−0.042 *		−0.076 ***
		(0.041)		(0.024)		(0.007)
lngii		−0.092 ***		0.022 *		−0.015 ***
		(0.011)		(0.013)		(0.004)
lnms		0.064 ***		0.031 **		−0.052 ***
		(0.016)		(0.015)		(0.007)
_cons	−0.296 ***	−0.356 ***	−0.358 ***	−0.240 ***
	(0.052)	(0.080)	(0.092)	(0.073)
AR(1)					−0.91	−1.08
					[0.351]	[0.276]
AR(2)					−1.56	−1.04
					[0.115]	[0.292]
Hansen					29.62	28.43
					[0.582]	[0.438]
Hausman			0.17	42.52
			[0.669]	[0.000]
R-sq	0.256	0.403	0.256	0.195
N	300	300	300	300	240	240

Note: L.lnte denotes the first-order lagged term of the explanatory variable—manufacturing innovation efficiency; *, **, and *** denote significance at 10%, 5%, and 1%, respectively; and values in parentheses and brackets are standard errors and p-values, respectively.

reports the estimation results for the ordinary least squares (OLS), FE, and SYS-GMM models both with and without the control variables. As seen from the table, the coefficient for the impact of the digital economy on manufacturing innovation efficiency is significantly positive at the 1% level for all models with and without control variables, indicating that the digital economy makes a significant contribution to manufacturing innovation efficiency. The SYS-GMM model has the advantage of solving the endogeneity problem caused by one-period lagged manufacturing innovation efficiency, and the results show that the standard error in the regression coefficients of the SYS-GMM model is smaller than those in the OLS and FE models, indicating that the SYS-GMM regression results show better statistical characteristics. The Hansen and AR(2) test results both pass the validity tests for the instrumental variables, indicating that no second-order autocorrelation exists in the residual terms of the SYS-GMM model after differencing; therefore, the SYS-GMM model, which accounts for the endogeneity problem and has more robust regression results, is finally selected. According to the regression results from the SYS-GMM model, the regression coefficient for the effect of the digital economy on manufacturing innovation efficiency is 0.443 and significant at the 1% level, indicating that the functions of the digital economy, such as intertemporal dissemination and data creation, significantly contribute to the improvement in regional manufacturing innovation efficiency. Therefore, Hypothesis 1 is verified.

For the control variables, the regression coefficient for industry personnel quality is significantly positive at the 1% level, indicating a significant contribution to the innovation efficiency of the manufacturing industry. The regression coefficient for the innovation climate is significantly negative at the 1% level, which is perhaps because the current level of infrastructure construction in China is relatively lagging and cannot effectively improve innovation efficiency [50]. Another possibility is that the growth rate of manufacturing innovation efficiency in provinces with better infrastructure is lower than that in provinces with poorer infrastructure [51]. The regression coefficient for policy support is significantly negative at the 1% level, which is perhaps due to the low level of innovation efficiency in China’s manufacturing industry during the sample period and the fact that enterprises’ environmental management, according to government regulations, is only at the level of emission control. This generates costs that crowd out capital investment in manufacturing technology R&D, thus inhibiting improvement in manufacturing innovation efficiency. The regression coefficient for market structure is significantly negative at the 1% level, which is perhaps because the capital inflow to the manufacturing industry in the domestic market has already become insufficient and foreign direct investment has further worsened the shortage of domestic manufacturing input capital, thus inhibiting improvement in manufacturing innovation efficiency.

5.2. Analysis of the Transmission Mechanism Underlying the Impact of the Digital Economy on the Efficiency of Manufacturing Innovation

The direct effect of the digital economy on regional manufacturing innovation efficiency is tested above, and a multiple mediating effects model based on the SYS-GMM model is further constructed to examine the transmission mechanism underlying the impact of the digital economy on manufacturing innovation efficiency in terms of promoting industrial structure upgrading and technology spillover channels. The second and fourth columns in

Table 4. Table listing the estimated results for the transmission mechanism of the digital economy on innovation efficiency in manufacturing.

	Total Effect	Industrial Structure		Technology Spillover
	lnte	lnuis	lnte	lncs	lnte
L.lnte	0.063 ***		0.214 ***		0.005
	(0.006)		(0.040)		(0.016)
L.lnuis		0.016
		(0.032)
L.lncs				−0.135 ***
				(0.012)
lndig	0.444 ***	0.964 ***	0.158 ***	−0.482 ***	0.422 ***
	(0.025)	(0.074)	(0.019)	(0.064)	(0.020)
lnuis			0.072 ***
			(0.020)
lncs					−0.042 ***
					(0.002)
lnlq	0.144 ***	0.180 ***	−0.168 ***	−0.995 ***	0.176 ***
	(0.008)	(0.031)	(0.012)	(0.121)	(0.011)
lnia	−0.071 ***	0.080 ***	−0.042 ***	−0.303 ***	−0.088 ***
	(0.005)	(0.014)	(0.013)	(0.034)	(0.006)
lngii	−0.014 **	−0.028 ***	0.115 ***	0.137 ***	0.021 **
	(0.004)	(0.007)	(0.032)	(0.044)	(0.010)
lnms	−0.002	−0.022	0.022 ***	−0.651 ***	0.007
	(0.007)	(0.040)	(0.005)	(0.052)	(0.008)
AR (1)	−1.08	−2.63	−2.97	−4.42	−0.93
	[0.277]	[0.009]	[0.003]	[0.000]	[0.356]
AR (2)	−1.04	0.65	−1.28	1.52	−0.43
	[0.298]	[0.508]	[0.196]	[0.131]	[0.660]
Hansen	28.46	27.04	27.14	29.54	26.92
	[0.439]	[0.513]	[0.512]	[0.129]	[0.939]
N	240	240	240	240	240

Note: L.lnte, L.lnuis, and L.lncs denote the first-order lagged terms of the natural logarithm of manufacturing innovation efficiency, industrial structure, and technology spillover, respectively. **, and *** denote significance at 5%, and 1%, respectively. Values in parentheses and square brackets are standard errors and p-values, respectively.

report the regression results for the effect of the digital economy on industrial structure upgrading and technology spillover, respectively. Their regression coefficients are both significant at least at the 5% level, indicating that the digital economy significantly promotes industrial structure upgrading and affects technology spillover. The third and fifth columns in Table 4 report the joint impact of the digital economy and mediating variables on manufacturing innovation efficiency. The regression coefficients for both mediating variables are significant at least at the 10% level, indicating that the digital economy can significantly enhance the level of regional manufacturing innovation efficiency using the channels of industrial structure upgrading and technology spillover. Moreover, the coefficients for the impact of the digital economy on manufacturing innovation efficiency are 0.158 and 0.422, respectively, after adding mediating variables. These values are significantly lower compared to the total effect—0.444, indicating that the mediating effect of the digital economy with industrial structure upgrading and technology spillover variables is only a partial mediating effect. Therefore, Hypothesis 2 is verified.

On the basis of its existence, the proportion of the mediating effect of the digital economy with industrial structure and technology spillover variables on the innovation efficiency of the manufacturing industry is further calculated using the following formula: “regression coefficient of explanatory variables on mediating variables × regression coefficient of mediating variables on explanatory variables/total regression coefficient of explanatory variables on explanatory variables”. This calculation is based on the approach of Wen and Ye (2014) [52]. The results are shown in

Table 5. Decomposition of the contribution of the digital economy to the transmission mechanism of innovation efficiency in manufacturing.

Intermediate Variables	Digital Economy → Mediating Variables	Intermediate Variables → Manufacturing Innovation Efficiency	Intermediary Effect Volume	Share of Total Effect
lnuis	0.964	0.072	0.069	15.632%
lncs	−0.482	−0.042	0.020	4.559%

. The mediating effects of the digital economy on manufacturing innovation efficiency with industrial structure and technology spillover variables are 0.069 and 0.020, respectively. Furthermore, in terms of the degree of indirect effects, among the total effects of the digital economy on manufacturing innovation efficiency, the relative contribution share of promoting upgrading in industrial institutions is 15.632%, and the inhibition of technology spillover explains 4.559% of the causal relationship between the digital economy and manufacturing innovation efficiency. This finding suggests that upgrading the industrial structure is an important channel with which the digital economy contributes to the efficiency of manufacturing innovation, but the mediating effect of technology spillover should not be overlooked.

5.3. Spatial Spillover Effects

Considering the possible spatial effects of the digital economy and manufacturing innovation efficiency, this paper uses the Moran index to analyse the spatial characteristics of the samples of the digital economy and manufacturing innovation efficiency, the results of which are shown in

Table 6. Table listing the spatial characteristics of innovation efficiency in the digital economy and manufacturing.

Year	te		dig
	Moran’s I	Z Value	Moran’s I	Z Value
2011	0.3542 ***	3.3819	0.2732 ***	2.6543
2012	0.3558 ***	3.4268	0.3058 ***	3.2963
2013	0.3676 ***	3.5642	0.3264 ***	3.0475
2014	0.3613 ***	3.5442	0.2527 ***	2.4336
2015	0.3823 ***	3.7473	0.3153 ***	2.9054
2016	0.3865 ***	3.7625	0.2924 ***	2.7843
2017	0.4023 ***	3.8635	0.2448 ***	2.3664
2018	0.4176 ***	3.9937	0.2567 ***	2.4523
2019	0.4351 ***	4.1158	0.2353 **	2.2675
2020	0.4533 ***	4.2313	0.2152 *	2.1446

Note: *, **, and *** denote significance at 10%, 5%, and 1%, respectively.

. According to the table, both the digital economy and manufacturing innovation efficiency show significant spatial aggregation characteristics. First, there is a strong spatial aggregation in the manufacturing innovation efficiency index, which is significant at the 1% level, and the aggregation effect gradually increases over time. Based on the increasingly close economic and social ties between regions, the spatial spillover effect of manufacturing innovation efficiency gradually decreases due to the increasingly close exchange of physical, human, and social capital across regions. This situation shows an increasing marginal effect. Second, the digital economy index has significant spatial aggregation but shows a decaying feature over time, indicating that the spillover effect of the digital economy has a decreasing marginal effect. When the difference in development between regions is large, less developed regions may exhibit a strong catch-up effect under the influence of the digital economy, which is a rapid bridging of the development imbalance, manifesting as a rapid development speed in the early stage. However, as the difference between regions gradually decreases, the development of the digital economy returns to its endogenous technological growth momentum, and the spatial spillover effect gradually diminishes. All the above analyses demonstrate the existence of spatial interaction effects in our model.

6. Conclusions and Recommendations

The development of the digital economy has provided important opportunities and an impetus to enhance the innovation efficiency of China’s manufacturing industry. This paper explains the transmission mechanism of the digital economy to promote the innovation efficiency of China’s manufacturing industry in terms of both industrial structure upgrading and technology spillover. Using provincial panel data from the period 2011 to 2020, this paper separately measures the level of development of the digital economy and manufacturing innovation efficiency to empirically test the overall role of the digital economy in promoting manufacturing innovation efficiency, the indirect mechanism, and the spatial spillover effects, the main findings of which are presented below.

(1) The digital economy significantly contributes to the improvement in manufacturing innovation efficiency in China, and the results remain significant after taking robustness and endogeneity into account. This finding is similar to the conclusion that the digital economy plays a significant role in promoting enterprise innovation [53], indicating that the digital economy has become a strong engine driving the development of manufacturing innovation in China. (2) The digital economy can indirectly improve manufacturing innovation efficiency with mediating effects such as promoting industrial structure upgrading and technology spillover, indicating that the positive impact of the digital economy on regional manufacturing innovation efficiency does not exist in isolation but relies on multiple channels such as promoting industrial structure upgrading and technology spillover to induce an effect. (3) Innovation efficiency is also influenced by the innovation environment. From the perspective of the internal environment, the quality of industry personnel has a positive effect on the innovation efficiency of the manufacturing industry, while the external environment, such as policy support, innovation atmosphere and market structure, also has a positive impact. (4) The development of the digital economy has a spatial spillover effect on the improvement in manufacturing innovation efficiency. The vigorous development of the digital economy in a region is not only conducive to the achievement of manufacturing innovation efficiency in that region but also has a positive impact on the improvement in manufacturing innovation efficiency in neighbouring regions.

The following points of inspiration are offered in relation to the above findings:

First, we comprehensively promote the development of the digital economy and accelerate its integration with the real economy. The digital economy has become a powerful driving force and an important way in which to promote the efficiency of manufacturing innovation. At present, there is still substantial space for improvement in the development level of the digital economy, and the difference between the development level of the digital economy and that of the real economy in different regions is fully considered to improve regional coordination and the overall level of real economic development. Second, the scale of the manufacturing industry should be maintained, and the industrial structure should be optimised. The digital economy affects the innovation efficiency of the manufacturing industry by promoting upgrades in the industrial structure, and it is necessary to encourage the development of “R&D and manufacturing linkage” industries, give full play to the “large and complete” advantages of China’s manufacturing industry to form a driving force for R&D, and achieve linkage R&D and innovation breakthroughs. With the continuous development of the digital economy, the phenomenon of the “digital divide” has gradually been highlighted between regions; thus, the digital economy empowers the regional layout of the manufacturing industry. It is not only related to the coordinated development of regions but also the key to the adjustment of the manufacturing structure, gradually optimising the industrial and economic structure and giving full play to the role of the digital economy in boosting the regional real economy. Third, an innovative environment should be built, and the level of manufacturing innovation should be improved across the board. The government should continue to deepen the reform of the science and technology system, establish a reasonable market competition institutional environment, increase support for digital technology innovation and business innovation, strengthen the construction of the legal system as it relates to digital technology, create a good institutional ecology of digital technology, and provide full protection for digital innovation. There are various limitations with respect to this study. First, based on the availability of data, this paper studies the digital economy and manufacturing innovation efficiency at the provincial level, but future studies could further analyse it from a more micro-level, such as the levels of cities and enterprises. Second, the development of the digital economy and the real economy is dynamic, and the relationship between the two can be studied from different angles in follow-up research to further consider the impact mechanism, which needs further exploration.