Comprehensive Evaluation of China’s Input–Output Sector Status Based on the Entropy Weight-Social Network Analysis Method

Abstract

In order to understand the source of China’s global influence, the key sectors and important connections in China’s economic and trade networks should be identified. Based on China’s 2002–2018 input–output data, this study uses the entropy weight method to determine weights, and then combines the social network analysis method in order to construct a comprehensive index system for industry status evaluation. Research results indicate that the high-risk financial industry occupies the top position among all industries. Additionally, resource-consuming industries represented by the coal mining industry and highly polluting industries still occupy an important position in China’s economy. Machinery, electronics manufacturing, and other real industries show increasing value-added products, steadily improving technology intensity, and continue to climb the industrial value chain. The conclusions of this study provide a comprehensive and detailed industrial understanding of the formulation of comprehensive and systematic national economic strategies as well as targeted industrial policies to achieve sustainable development.

1. Introduction

The global economic system is a large network composed of national subnets. As the world’s second largest economy and the largest trading nation, a slight change in China will have a huge impact on the interconnected global economy and trade [1]. An economy is composed of interdependent departments that rely on the exchange of products and services. The industrial structure, as a concentrated expression of the level of economic development, reflects various complex proportional relationships between different industrial sectors in accordance with certain economic and technological connections, as well as the impact of each industry on the entire industrial network [2,3,4]. Understanding the economic structure (i.e., how the industrial sectors interact with each other) is crucial to determine how the economic system functions [5,6]. Therefore, identifying the key sectors and important connections in China’s economic and trade networks is crucial for also understanding the source of China’s global influence and long term sustainable development.

To deepen the application of sustainable development theory in economic and trade, “Agenda 21” has a comprehensive elaboration on sustainable development capacity building: “A country’s sustainable development capacity depends to a large extent on the capacity of its people and institutions under its ecological and geographical conditions, specifically, the capacity building includes the development and enhancement of a country’s human, scientific, technological, organizational, institutional, and resource capabilities”. The basic goal of capacity building is to improve the ability to evaluate and select policies and development models. “The application of sustainable development theory in economic and trade is to realize the sustainable development of trade, that is, to make full use of various economic resources, to ensure the lasting economic trade capacity, and to maintain the sustainable use capacity and function of resource regeneration and utilization”. China has also put forward the goal of a carbon peak by 2030 and carbon neutralization by 2060, namely, the “double carbon goal,” to support sustainable development. This paper aims to better promote the sustainable development of trade and optimize the industrial structure by identifying key sectors in the trade network. Development and how to achieve sustainable economic development has become a hot topic in the fields of ecology and economics, making sustainability coordinated with economy, ecological environment, and society to meet the needs of current and future generations of people.

The remainder of this paper is organized as follows. Section 2 describes the literature review. Section 3 introduces the research methods and data sources, and describes the research process. Section 4 reports the evaluation results of the industrial status from the cross section and time series, and then discusses the reasons and economic significance of the industrial status change based on economic reality and the policy situation. Finally, Section 5 draws the research conclusions.

2. Literature Review

The input–output table is an effective tool for supporting the analysis of the structural relationship between national or regional economic sectors. It traces the direct and indirect supply–demand relationship between various departments in the entire economic system, and comprehensively describes the complex national economic system and the economic structure at the departmental level [7]. It is widely used in agriculture [8,9], industrial economic transformation [10], heavy industry [11], construction industries [12], etc. The social network analysis (SNA) method is a theory and technique for measuring the structure of complex systems. The SNA method can characterize the internal structure of the industry by measuring the input–output flow between industrial departments, revealing the complex relationships between departments [13], and identifying key departments at the center of power and influence. The economic system based on the input–output table can be regarded as a network, wherein nodes represent industrial sectors and the edges connecting nodes represent transactions between sectors [5]. Each department in the network acts as a producer and consumer simultaneously. On the one hand, it produces and distributes inputs provided to other departments; on the other hand, it consumes inputs from other departments to complete its own transformation process [14]. In order to better understand the relationship between inter-departmental dependence and its impact on the entire economic system, a large number of documents have been combined with input–output tables and SNA methods to investigate the industrial structure, mainly in three areas.

First, the overall structure of the industrial network and its evolutionary characteristics, whether at the global level or that of a single country or region, has been fully studied [15,16,17]. This type of research mainly focuses on the topological structural characteristics of the network, and it often uses SNA indicators, such as network scale, density, degree distribution, connectivity, reciprocity, core-peripheral structure, and cohesive subgroups, to describe the overall performance of the network. An important feature of this subset of the literature is that it focuses on the analysis of mathematical and statistical characteristics of the input–output network, without paying too much attention to the meaning of the policy, or that it involves policy, but at a superficial level.

The second subset of the literature focuses on the roles and functions of the industrial sector. In this literature, some scholars typically have a subjective bias. For example, to determine whether logistics and transportation sectors have become the center of the U.S. economy over time, Lyengar et al. [18] use the input–output table data of the U.S. Bureau of Economic Analysis for more than two decades in order to conduct SNA. Ma et al. [19], Li et al. [20], and Liu et al. [21] intend to study the internal evolution of China’s construction industry, evolution of the coal industry chain, and network relationship between the power industry and economic growth, respectively, as well as to provide a reference for the formulation of industrial policies. Among China’s 139 sector input–output networks in 2012, wholesale, retail, and agricultural products have the greatest impact on total output fluctuations through network connections, and can be used as key industries [22]. However, in the 42 sector input–output networks in 2017, the three major industries, namely construction, public management, and social security and transportation equipment, showed clear advantages and achieved a higher status than other industries [23]. It can be found that because various scholars may use data from different years, different statistical calibers, different evaluation indicators, or diverse conclusions are often obtained.

The third subset of the literature is the application of ecological network analysis to input–output data with environmental expansion, by analyzing the interaction between the economy and the environment through the flow of energy, resources, and emissions. In order to understand how the economic structure affects the country’s environmental conditions [5], this type of research is often related to energy flow and carbon emissions. According to the input–output relationship between industrial sectors, energy flow and carbon emission networks can be constructed indirectly, and the source of energy demand and carbon emissions can be determined by analyzing the flow of various substances in the network. Departments have provided evidence to improve global environmental issues [24,25,26,27,28,29,30,31,32,33]. Because research topics closely follow leading real-world issues in today’s world, research results in this field are quite rich. However, an important flaw is that the data used for empirical explanations often lag in timeliness.

These results have laid the foundation for this paper, which has also made the following expansions. Firstly, in terms of data, this paper uses China’s latest input–output table data (2002–2018), which, currently, has not yet appeared in the literature. It is believed that the analysis results based on the latest data can provide new insights regarding China’s economic reality. Secondly, in terms of industrial sector selection, this study uses the input–output data of sub-industrial sectors that have not been merged for calculation and evaluation, which overcomes the limitations that the previous literature had using the input–output table and the social network method to analyze the industrial structure. This makes it easy to miss the original input–output information between industries. In terms of research methods, this paper solves the problem that the previous literature had drawing different conclusions based on various evaluation indicators. The entropy weight method is used to determine the weight, and various measurement indicators of the SNA method are combined in order to construct a set of comprehensive evaluation indicator systems. It provides effective ideas for future research and the development of new methods for individual status evaluation. Finally, in terms of the analytical framework, cross-sectional analysis and time-series analysis were combined. On the one hand, a horizontal comparison of the status of industry sectors is carried out under three strength relationship networks. On the other hand, vertical changes in the status of specific sectors are discussed under the industry category. This provides a comprehensive and detailed industrial cognition for the formulation of comprehensive and systematic national economic strategies and targeted industrial policies.

3. Materials and Methods

3.1. Research Ideas

3.1.1. Input–Output Network Construction

Industrial networks with different intensity relationships can be constructed based on the different amounts of input or output between industrial sectors. In this study, based on the basic input–output flow scale, the input–output or output over CNY 100 million, CNY 1 billion, and CNY 10 billion is regarded as the standard of the relationship between the two industrial sectors. The constructed input–output networks were labeled as weak, relatively strong, and strong networks, respectively. UCINET software was used to draw the network topology diagram. The network topology diagram of 2018 is shown in Figure 1 and Figure 2 (see Table S1 for the names of industrial departments corresponding to node numbers in the figure, and File S1 for all network diagrams).

See Table S1 for the industrial sector name corresponding to the node serial number in the figure. The more nodes in the figure are connected, the higher the status of the country in the network. The more central the location is, the greater the impact on the trade network.

3.1.2. Index System Construction

In the social network analysis, nine metrics were selected from three aspects of centrality, structural hole, and influence in order to comprehensively evaluate the status of subdivided industry sectors;

Table 1. Industrial status evaluation index system.

Total Goal	Secondary Indicators	Three-Level Indicators	Efficacy	Indicator Meaning
Industry position	Centrality	Relative degree center degree	positive	Number of other nodes directly related to a node (standardized)
		Relative center middle degree	positive	The degree to which a node controls resources
		Relatively close to the center degree	positive	The reciprocal of the sum of the shortest distance between a node and all other points, the degree to which a node is not controlled by others
		Eigenvector centrality	positive	Weighted centrality based on the centrality of adjacent nodes (the centrality of a node is a linear function of the centrality of other nodes related to it)
	Structural holes	Efficient scale	positive	Non-redundant factors in networks
		Efficiency	positive	Ratio of effective size to actual size
		Limitation	negative	The strength of the node being bound by a single relationship
	Influence	Influence degree	positive	The influence of a node on other nodes
		Cause degree	positive	The difference between the degree of influence and the degree of being influenced: greater than 0 is the cause factor, and less than 0 is the result factor

Source: authors’ computation based on the primary survey.

presents the specific indicators.

The input–output table data used in this study were obtained from the National Bureau of Statistics of China (https://data.stats.gov.cn/index.htm, accessed on 16 March 2022). The National Bureau of Statistics of China has provided national input–output tables for eight years: 2002, 2005, 2007, 2010, 2012, 2015, 2017, and 2018. Based on the research purpose of this study and the timeliness and consistency of data, the basic tables of input–output in 2002, 2007, 2012, 2017, and 2018 were selected as the data basis for network construction. The number of industrial sectors each year was 122 in 2002, 135 in 2007, 139 in 2012, 149 in 2017, and 153 in 2018. As the comprehensive status index (CSI) calculated in this study is a relative value (i.e., the relative position of each industry sector in all industry sectors in the year), the inconsistency of the number of industry sectors in each year has little influence on the evaluation results.

3.2. EW-SNA Method

3.2.1. Social Network Analysis

SNA is a set of theories and methods used to analyze the structure, properties, and attributes of a network of relationships between social actors [34]. Extensive connections exist between industrial departments. Resources such as labor, capital, technology, and intermediate inputs flow between departments so that a tight network system is formed between various industrial departments, and each industrial department becomes a node in the industrial network. Therefore, the industrial relationship matrix based on the input–output table and the SNA method can be used to examine the overall situation of the industrial structure as well as the correlation structure of the economic network within the industry. It also aims to describe the interrelationship and mutual influence between industrial sectors, and to analyze status changes in different sectors in the industrial structure.

• Strong and weak relationships theory. The theory of strong and weak ties was proposed by the American sociologist Granovetter [35]. Strong relationships refer to the strong homogeneity of an individual’s social network. In contrast, weak relationships are characterized by strong personal social network heterogeneity.
• Quantification of power: centrality. From the perspective of social networks, people have power because they have relationships with other people and can influence others. Social network scholars conduct quantitative research on power from the perspective of “relationships” and provide a variety of quantitative indicators about power, such as centrality. Centrality includes a variety of indicators. The most commonly used ones are degree, intermediate, proximity, and eigenvector centrality.
• Structural holes and middlemen. A structural hole is a concept proposed by American sociologist Burt [36]. It refers to the gap in the social network. This implies that some individuals in the social network are directly connected with some individuals, but not directly connected with other individuals; that is, a direct or intermittent relationship does not exist. From the network perspective, it seems that a cave appears in the network structure. The intermediary can realize value by connecting structural holes and bridges, thus having information advantages and control advantages.
• Decision-making trial and evaluation laboratory. Dematel (decision inspection and evaluation methods) effectively combines the graph and matrix theories. Through the logical relationship and direct influence matrix between the various elements in the system, the degree of influence of each element on other elements and the degree to which they are affected can be calculated, thereby calculating the cause degree and centrality of each element and determining the causal relationship between elements and the position of each element in the system.

3.2.2. Entropy Weight Method

The SNA method contains a number of indicators, and each indicator is suitable for investigating the status change in the industry. It would be one-sided to consider only one when evaluating the status of the node sector in the industrial network, because these measures contain the importance, reputation, influence, and other aspects of the node. Therefore, a comprehensive method for evaluating the industrial status is to combine all types of measurement indexes of SNA, and then to construct a comprehensive evaluation index system. In this process, the weight of each metric index is difficult to determine, but the entropy weight method can solve this problem.

Entropy is a measure of uncertainty in information theory. The greater the amount of information, the lower the uncertainty, and the smaller the entropy; the smaller the amount of information, the greater the uncertainty, and the greater the entropy. According to the characteristics of entropy, we can determine the randomness and disorder degree of an event by calculating the entropy value, as well as the dispersion degree of an index by using the entropy value. The greater the degree of dispersion of an index, the greater the influence of the index on the comprehensive evaluation. Therefore, according to the degree of variation of each indicator, the weight of each indicator can be calculated using the information entropy tool to provide a basis for the comprehensive evaluation of multiple indicators. The entropy method is an objective weighting method that uses the amount of information provided by the entropy value of each index to determine the weight of the index. The entropy weighting method works as follows:

(1)

The entropy weighting method can avoid the interference of human factors in the weighting of each evaluation index, making the evaluation results more realistic. It overcomes the problem of the weighting process of indicators being influenced by human factors in the current evaluation method.

(2)

By calculating the entropy value of each indicator, we can measure its amount of information, so as to ensure that the established indicator can reflect most of the original information.

This study adopted the entropy method to weigh the basic evaluation indicators in the SNA method, and comprehensively measured the status of various sub-industry sectors in China from 2002 to 2018. The specific steps are as follows:

Step 1: Standardize each index in the evaluation system.

$$\mathrm{Positive}\mathrm{index}:X_{ij}=\frac{x_{ij}-\min \left\{x_{1j},\dots ,x_{nj}\right\}}{\max \left\{x_{1j},\dots ,x_{nj}\right\}-\min \left\{x_{1j},\dots ,x_{nj}\right\}}$$

(1)

$$\mathrm{Negative}\mathrm{index}:X_{ij}=\frac{\max \left\{x_{1j},\dots ,x_{nj}\right\}-x_{ij}}{\max \left\{x_{1j},\dots ,x_{nj}\right\}-\min \left\{x_{1j},\dots ,x_{nj}\right\}}$$

(2)

where i represents the i-th industry, the maximum value is n, j represents the total number of indicators, and the maximum value is m.

Step 2: Calculate the proportion of the i-th industry in the index under the j-th index.

$$P_{ij}=\frac{X_{ij}}{{\displaystyle \sum _{i=1}^n{X}_{ij}}}$$

(3)

Step 3: Calculate the entropy value of the j-th index.

$$e_j=-k{\displaystyle \sum _{i=1}^n{P}_{ij}\ln P_{ij}}$$

(4)

Among them, $k=1/\ln (n)>0$ satisfy $e_j>0$.

Step 4: Calculate the information entropy redundancy.

$$d_j=1-e_j$$

(5)

Step 5: Calculate the weight value of each indicator.

$$w_j=\frac{d_j}{{\displaystyle \sum _{j=1}^m{d}_j}}$$

(6)

Step 6: Calculate the composite position index (CSI) for each industry sector.

$$P{S}_i={\displaystyle \sum _{j=1}^m{w}_j{X}_{ij}}$$

(7)

Evidently, the larger the value of $P{S}_i$, the higher the status of the industry sector.

The index weights calculated according to the above steps are presented in

Table 2. Weight of each indicator.

Total Goal	Secondary Indicators	Weight	Three-Level Indicators	Weight
Industry position	Centrality	65.61%	Relative degree center degree	7.91%
			Relative center middle degree	48.04%
			Relatively close to the center degree	3.95%
			Eigenvector centrality	5.72%
	Structural holes	13.74%	Efficient scale	12.32%
			Efficiency	1.22%
			Limitation	0.20%
	Influence	20.64%	Influence degree	18.60%
			Cause degree	2.04%

Source: Authors’ computation based on the primary survey.

.

4. Results

4.1. Position Ranking of Subdivided Industry Sectors under the Input–Output Network of Three Relationship Strengths

Table S2 reports the ranking of the comprehensive status index in the input–output network of the three relationship strengths in 2018. In Table S2, in the input–output network of different relationship strengths, a large, but incomplete, overlap occurs in the ranking of the status of sub-industry sectors, which can be divided into six situations. These are in the (1) top 20 (or the bottom 20) in the weak relationship network, (2) top 20 (or bottom 20) of relatively strong relationship networks, (3) top 20 (or bottom 20) of strong relationship networks, (4) top 20 (or bottom 20) in weak relationship and relatively strong relationship networks, (5) top 20 (or bottom 20) in relatively strong relationship and strong relationship networks, and (6) top 20 (or bottom 20) of the three relationship networks simultaneously. The specific division is presented in

Table 3. Classification of the top 20 sub-industrial sectors in the input–output network with three relationship strengths in 2018.

Classification	Industrial Sector	Quantity
Only in the top 20 in weak relationship networks	Internet and Related Services, Insurance, Telecommunication, Water Cargo Transportation and Auxiliary Activities, Railway Passenger Transportation	5
Only in the top 20 in relatively strong relationship networks	Paper Making and Paper Products, Information Technology Services, Technology Promotion and Application Services	3
Only in the top 20 in strong relationship networks	Agriculture, Steel Rolling Products, Residential Building Construction, Electronic Components, Basic Chemical Raw Material, Non-ferrous Metals and Their Alloys, Auto Parts and Accessories, Transmission and Distribution and Control Equipment, Synthetic Material	9
Ranks top 20 in weak relationship and relatively strong relationship networks	Multimodal Transport and Transportation Agency, Accommodation, Other Services, Postal, Textile Clothing and Apparel, Handling and Warehousing	6
Ranks top 20 in relatively strong relationship and strong relationship networks	Plastic Products, Special Chemical Products and Explosives, Pyrotechnics and Fireworks Products	2
Ranked in the top 20 of all three relationships	Business Services, Monetary Finance and Other Financial Services, Road Freight Transportation and Auxiliary Activities, Retail, Wholesale, Catering, Electricity, Heat Production and Supply, Petroleum and Nuclear Fuel Processing, Metal Products	9

Source: Authors’ computation based on the primary survey.

and

Table 4. Classification of the bottom 20 sub-industrial sectors in the input–output network with three relationship strengths in 2018.

Classification	Industrial Sector	Quantity
Only in the bottom 20 in weak relationship networks	Ferrous Metal Mining and Dressing, Fertilizer, Waste Resources and Waste Materials Recycling and Processing Products, Steel	4
Only in the bottom 20 in relatively strong relationship networks	/	0
Only in the bottom 20 in strong relationship networks	Ecological Protection and Environmental Governance, Refined Tea, Radio, Television, Film and Video Recording Production, Metal Products, Machinery and Equipment Repair Services, Cultural and Office Machinery, Audio Visual Equipment, Capital Market Services, Radio and Television Equipment and Radar and Ancillary Equipment, Medical Instruments and Equipment	9
Ranks bottom 20 in weak relationship and relatively strong relationship networks	Software Services, Dairy Products, Convenience Food, Wool Textile and Dyeing and Finishing Processing, Condiments, Fermented Products, Pesticide, Feed Processing, Aquatic Products Processing, Hemp Textile, Silk Textile and Fine Processing	9
Ranks bottom 20 in relatively strong relationship and strong relationship networks	Social Security, Water Conservancy Management, Culture and Art, Press and Publishing	4
Ranked in the bottom 20 of all three relationships	Sports, Social Work, Radio, Television and Satellite Transmission Services, Sugar Manufacturing, Mining Ancillary Services and Other Mining Products, Knitting or Crochet Weaving and Its Products, Water Passenger Transportation	7

Source: Authors’ computation based on the primary survey.

.

In Table 3, departments ranked in the top 20 in all three relationship networks have extensive and in-depth connections with other departments, and their input or consumption is also very large. Therefore, these sectors can be considered the most important sectors in the input–output network. Most of them are tertiary industries, including business services, monetary finance and other financial services, road freight transportation and auxiliary activities, and other productive services, as well as retail, wholesale, catering, and other life services. In addition, the secondary industries, such as electricity, heat production and supply, petroleum and nuclear fuel processing, and metal products, ranked high. These three sectors are all energy-intensive industries which are specified in the statistical Report on National Economic and Social Development 2010. This indicates that energy-intensive and high-polluting industries still play a crucial role in China’s national economy in 2018.

In Table 4, we also focus on the sectors that rank in the bottom 20 of the three relational networks. These sectors only have input–output relationships with few other sectors, and the amount of input and consumption remains quite small. The centrality of some departments in the strong relationship network is 0. This implies that no direct input or consumption with any department exists, which mainly includes sports, social work, radio and television, satellite transmission services, water passenger transportation, and other life services. The service objects of these sectors are relatively specific, mainly residents rather than enterprises. Thus, the service scope is relatively narrow, and the volume is not large. These include sugar manufacturing, knitting, crochet weaving and its products, and others in the manufacturing industry which provide products only to a few specific industries.

4.2. Relationship between the Comprehensive Status Ranking of the Sub-Industry Sector and the Sub-Item Ranking

The above analysis examines the comprehensive status ranking of China’s sub-industry sectors under the input–output network of three relationship strengths. The following will further determine the source of their status by analyzing the relationship between the comprehensive status ranking and the three sub-item rankings. To highlight the key points, this paper mainly analyzes the average top 20 and bottom 20 industrial sectors in the input–output network of the three relationship intensities, as well as representative industrial sectors that are beneficial to resources, ecology, and environmental protection. This includes waste resources and waste materials, recycling and processing products, water conservancy management, ecological protection, and environmental governance.

As shown in Figure 2, Figure 3, Figure 4 and Figure 5, the comprehensive status ranking of China’s subdivided industrial sectors is highly consistent with the ranking of centrality, structural holes, and influence. Additionally, industrial sectors with higher status rank high in terms of centrality, structural holes, and influence. However, industrial sectors with lower status rank quite low in terms of centrality, structural holes, and influence. This indicates that, in the actual input–output network, the status of industrial sectors is comprehensively determined based on many aspects, including direct connection between industries (centrality), indirect connection (structural hole), and direction of the connection (influence). However, the three are interlinked under normal circumstances. That is, industrial sectors with direct input–output relationships with most other sectors have stronger control over resources and information, and more easily influence other sectors.

Firstly, for the most important industrial sector in the input–output network (i.e., the industry sector ranked top 20 in all the three networks), producer services, especially road cargo transport and transport auxiliary activities, ranked higher than the other indicators in terms of influence. Road freight transport and transport support activities rank 2nd and 1st, respectively, in the weak and relatively strong relation networks, although their overall position ranks 8th and 6th (Figure 2a and Figure 3a). It can be found that the importance of services is largely due to the large amount of material or product (both tangible and intangible) input they make to a large number of other industrial sectors, thus bringing strong spillover effects.

Secondly, for the bottom 20 industrial sectors, a common phenomenon is that the influence ranking is relatively high relative to other indicators. This is because the bottom 20 industrial sectors are mostly service industries, such as sports, social work, radio and television, satellite transmission services, and water passenger transportation. The service industry itself is an industry that provides support to other industries. Therefore, the influence delivered is much greater than the influence received, and its cause degree coefficient will be relatively large. However, because they are part of the public service industry and the scope of their services is relatively narrow, their overall influence remains relatively small.

Finally, for those representative industrial sectors that are conducive to resources, ecology, and environmental protection, although their ranking of influence is slightly higher than that of other indicators (Figure 5), the impact of these industrial sectors is far from reaching the “dual carbon” goal and meeting the need for sustainable development. The degrees of influence and being affected were relatively low, especially for the latter. This shows that other industries have quite little investment in eco-environmental protection industries. Due to insufficient resources and funds, these industries have developed poorly and, therefore, cannot provide a role to their positive environmental externalities.

4.3. Temporal Changes of the Status of China’s Segmented Industry Sectors from 2002 to 2018

The following is a further study of the temporal changes in the status of China’s segmented industry sectors from 2002 to 2018. To simplify the analysis and avoid bias and extremes in the conclusion, the input–output network of the medium strength relationship (CNY 1 billion) is mainly considered here.

4.3.1. Changes in the Ranking of the Status of the Sub-Industry Sectors of the Primary Industry

According to the “Classification of National Economic Industries” (GB/T 4754-2017), the primary industry mainly includes four sub-sectors: agriculture, forestry, animal husbandry, and fishery.

According to Figure 6, the vertical coordinates represent the composite status index and ranking, respectively, and the composite status index represents a quantitative score of industrial status. An increase in the chart is an increase in the score, which is an increase in industrial status. In general, the status of the primary industry showed a sharp downward trend. This downward trend was not due to the poor performance of a certain indicator, because the rankings of the centrality, structural hole, and influence of the primary industry have dropped significantly. In fact, in the past 10 years, the direct connection between the primary industry and other industrial sectors has increased (as observed in the numerical change in degree centrality). Although this enhancement is an inevitable trend of industrial development brought about by national economic and technological development, all industries can and tend to strengthen connections with other industries. However, compared with the rapid development of other industries (e.g., service industries) during the same period, the primary industry was affected by the process of China’s industrialization and urbanization. Further, the outflow of labor, capital, and other factors led to insufficient development momentum, so the industrial status declined relative to others.

In the primary industry, compared with the other three subsectors, the comprehensive status index and ranking of agriculture have been far ahead, and the ranking decline is relatively small. However, recent studies [37] have found that agricultural production (including planting and animal husbandry) and land use change are also major sources of greenhouse gas emissions, which account for approximately 25% of the total global greenhouse gas emissions. Agriculture may not seem to account for a large proportion of emissions compared to fossil fuels. However, the fact is that emissions from agricultural activities are inevitable, especially from grains and beef due to their dominant positions in people’s diets. These bring about the largest total and the largest unit intensity of agricultural emissions, respectively. Therefore, the achievement of current international climate goals may depend on significant reductions in agricultural emissions, which will require new demands for improved agricultural production technology and policies.

4.3.2. Changes in the Status Ranking of Subdivided Industrial Sectors of the Secondary Industry

According to the National Economic Industry Classification (GB/T 4754-2017), the secondary industry can be divided into four categories: mining, manufacturing, supply, and construction. As the construction industry had only one industrial sector in 2002 and 2007, it was subdivided into multiple industrial sectors in subsequent years, and the annual data are not comparable. This research mainly analyzes the subdivided industrial sectors of the mining, manufacturing, and supply industries.

• Mining industry

Overall, the industrial status of the five major industrial segments of the mining industry declined from 2002 to 2018. This also reflects the gradual decline in the status of resource-consuming industries with the sustainable transformation of China’s economic development in the past decade. Among them, the industrial status of the oil and natural gas exploitation industry fluctuates significantly. Previously, the industrial status was relatively high, but it has decreased greatly in recent years, which is closely related to having less resource reserves and less advanced exploitation technology.

Most notable in Figure 7 is the coal mining industry remained relatively high in the rankings. Despite a slight decline in the rankings, this industry still ranked high in 2018. This implies that China’s energy use structure has not changed much in more than a decade, and has been dominated by coal. However, the average power efficiency of coal in the country remains quite low, and there is much room for improvement if more advanced technology is used, which would also be useful for developing renewable energy sources.

2.

Manufacturing

China’s manufacturing is mainly divided into three categories: the textile industry, the resource processing industry, and machinery and electronics manufacturing. For the convenience of analysis, only a few representative industry sectors were selected for each type.

Figure 8 shows the large status gap between the subsectors of the textile industry. While some sectors rank quite high (e.g., paper and paper-related products), some sectors rank quite low (e.g., grain mill products). The changes in the status of the different industrial sectors vary. While the status of some industrial sectors has been declining (e.g., cotton, chemical fiber textiles, and printing and dyeing finishing products), the status of individual industrial sectors has risen sharply (e.g., tobacco products, culture education, sports, and entertainment products).

Papermaking and paper-related products are the industry sectors with the highest ranking and the most stable status in the textile industry, and have been stable in the top 20 for more than 10 years. Recent studies [38] have shown that some provinces in southwestern China (e.g., Yunnan, Guizhou, and Guangxi provinces) and northeastern China (e.g., Heilongjiang and Jilin provinces) have established a rapid afforestation model. In the past 10–15 years, the provincial forest area has increased by 40,000–440,000 hectares each year. The large-scale expansion of these fast-growing plantations has brought about the vigorous development of China’s timber exports and paper products industry.

The declining status of grain grinding products, cotton, chemical fiber textiles, and printing and dyeing finishing products, as well as the rising status of tobacco products, culture education, sports, and entertainment products, reflect the changes in demand structure accompanied by economic growth and the improvement of residents’ living standards, and the demand structure determines the industrial structure. In recent years, residents’ demand has increasingly shifted from simple clothing and food to spiritual culture, which is bound to subvert the status of the corresponding industry.

As shown in Figure 9, in the resource processing industry, only waste resources and waste material recycling and processing products have quite a low status, while other industrial sectors have a relatively high status, especially metal products, refined petroleum and nuclear fuel-processed products, and plastic products, which have stabilized at about 10 each year (at quite a high level). Therefore, regardless of the cross-sectional view or time series change, the resource processing industry with high pollution and high energy consumption has always occupied a high position in China’s industrial relationship network, while the resource recycling industry with high utilization and low emission status is quite low and continuously declining.

This reflects the current reality of China’s resource demands and utilization. On the one hand, the pursuit of economic growth goals has resulted in a rigid increase in the demand for traditional resources and energy, and the resource processing industry has developed steadily. On the other hand, China’s energy utilization efficiency is not high, energy consumption per unit of GDP and water consumption are still significantly higher than the world average, and resource security is facing greater pressure.

Generally, the industrial status of machinery and electronics manufacturing is mostly at the middle level, and the range of fluctuations over time is not significant. In particular, the status of electronic components, instruments, and meters is relatively high; that of railway transportation, urban rail transportation equipment and metal processing machinery is relatively low; and the industrial status of household appliances has shown a significant decline in the past 10 years. The changes in the status of the machinery and electronic manufacturing sub-industry sectors, as shown in Figure 10, reflect the continuous increase in the added value of China’s manufacturing products, the steady increase in technology intensity, and the continuous rise of the industrial value chain.

As shown in Figure 11, the status of electricity and heat production and supply is quite high, and there has been little change in more than 10 years. However, from the perspective of secondary indicators, relative to the centrality and structural hole indicators, the ranking of its influence indicators fluctuates widely. This is because, from 2002 to 2018, the influence degree and the affected degree of power and heat production and supply have been quite high. This results in the cause degree coefficient, as the difference between the two is becoming increasingly small (ranking from 1 to 6). Consequently, the overall influence decreased slightly (ranking from 2 to 5). This further demonstrates that, while the influence of the entire industrial chain of power and heat production and supply plays a driving role in downstream industries, the change in demand also affects the development of upstream industries.

In contrast, the status of water production and supply and gas production and supply has been quite low. This is mainly because the supply objects of water and gas are mainly residents rather than enterprises. Thus, they have relatively few links with other industries, and the centrality and structural hole indicators are relatively low. However, their influence is not low, which is manifested in a high degree of influence, a low degree of influence, and a high degree of use. The production and supply of water and gas, as a supply industry, provide resources to other industries to a far greater extent than other industries that invest in them.

4.3.3. Changes in the Ranking of the Status of the Tertiary Industry Segmented Industry Sectors

According to the Industrial Classification of the National Economy (GB/T 4754-2017) and the statistical classification of producer services and living services of the National Bureau of Statistics (2019), the tertiary industry can be divided into producer services, living services, and public services.

• Productive service industry

The productive service industry refers to the service industry that provides guaranteed services for maintaining the continuity of the industrial production process, promoting industrial technological progress, industrial upgrading, and improving production efficiency. It mainly includes productive support services, transportation, the storage and postal industry, information transmission, the software and information technology service industry, the financial industry, the leasing and business service industry, and the scientific research and technical service industry.

According to Figure 12, in general, except for the producer support services serving agriculture, forestry, animal husbandry and fishery, which have a narrow service scope and few connections with other sectors, the status of other producer services is high.

Of these, finance is by far the highest-ranked of all industrial sectors, with a combined position that was No. 3 in 2002, but was raised to No. 1 in the following year, 2007, and has remained there since. Business services is the industry sector that is currently second only to finance in terms of position, having improved its overall position from No. 8 in 2002 to No. 2 in 2017 and stabilized at that position in 2018. A similarly high-status industrial sector is road freight transport and transport support activities, whose overall position ranking has fluctuated around 5th place for more than a decade. This indicates that modern service industries, represented by finance, business services and transportation, have become the main sectors in the current development of China’s service industry.

2.

Living service industry

The living service industry refers to the service activities that meet the final consumption needs of residents, including wholesale and retail, accommodation and catering, real estate, residential services, repairs, other service industries, and other industrial sectors.

From the overall view in Figure 13, the status of the life service industry is generally higher, but most of these exhibit a slight downward trend. Among them, the wholesale and retail industries dropped from rank 1 in 2002 to rank 5 and 4 in 2018. When the ranking changes of the financial and business service industries are compared, this reflects the optimization and upgrading of the internal industrial structure of China’s service industry—the transformation from a labor-intensive traditional service industry (wholesale and retail) to a knowledge-intensive modern service industry (financial and business services).

However, with the deepening of industrial integration, mutual promotion between the transformation of the traditional service industry to the modern service industry and that of the modern service industry to the traditional service industry will occur. The modern service industry transforms and upgrades the traditional service industry through its remarkable knowledge innovation characteristics and its high technology and management level. Thus, the traditional service industry retains and continues its original composition and content, and the service mode is more advanced, the service content is richer, and the service quality is higher. The accommodation and catering industries (shown in Figure 13) and the postal industry (shown in Figure 12) are typical traditional service industries that have been transformed and promoted. Therefore, their industrial status has not fallen, but improved.

In addition, the status of the real estate industry has encountered a process of initial rise and subsequent fall, reflecting the transformation of the Chinese real estate market from the age of settlement to the age of speculation to government regulation. At present, with the release of various government control policies, the real estate industry has gradually returned from investment and financial attributes to residential attributes. This industry no longer requires excessive capital investment from other industries. It seems that industrial connections have decreased and industrial status has declined. Nevertheless, in the long run, the order of the real estate market has improved significantly, with reduced industrial risks and increased stability.

3.

Public service industry

The public service industry refers to the service activities performed by government departments, as well as the ability of state-owned enterprises and institutions and related intermediary agencies to perform statutory duties, provide assistance, or handle related affairs according to the requirements of citizens, legal persons, or other organizations. This includes water conservancy, environment and public facilities management, education, health and social work, culture, sports and entertainment, public management, social security and social organizations, and international organizations and other industrial sectors.

According to Figure 14 it can be seen that unlike that of producer and living service industries, the status of the public service industry is generally lower and continues to decline. This is closely related to the characteristics of the public service industry. Its non-profit, non-exclusive, and non-competitive characteristics make the main public service providers into mainly government departments and state-owned enterprises and institutions, and the service targets into mostly residents, rather than enterprises. Therefore, quite a few connections exist between the public service industry and other profit-making industries and enterprises. Moreover, from the perspective of the degree of influence, degree of being affected, and degree of cause, the causal degree of these industries is basically negative. That is, the influence generated by the public service industry is generally lower than the influence received, which is not in line with the functional objectives of the industry to manage society and serve the public.

Among these, the public services that are mostly concerned with and most needed by the majority of urban and rural residents are education, health, and social security. These services must be provided in order to establish a social safety net and to protect the basic rights of survival and development of all members of society. Specifically, the Chinese government has always attached great importance to these services. In fact, education has always been highly influenced, ranking among the top 5 almost every year; that is, from the perspective of the level of commitment of resources, investment in the education sector is obtained quite adequately. However, it is not proportional to the output to input, the impact force is quite small, and it is far from being able to play its fundamental and leading role in economic development and national rejuvenation. Further, the problem with the social security industry is that the resources invested and the impact are quite insufficient. This is mainly because the content of the state’s direct management in China’s social security management system is quite high, and the participation of all parties in society, including enterprises, is not high, and the socialization component is insufficient.

In addition, the status of environmental protection industries is particularly low. As shown in Figure 14, water conservancy management and environmental management industries are ranked at the bottom. These industries aim to protect natural resources, the natural and social environment, and Earth’s ecology and biology, which is particularly important for sustainable development. However, they have not received enough attention, and other industries have invested quite little in them. Thus, environmental protection industries that do not receive sufficient resources and funds have developed slowly, thereby failing to produce significant positive environmental impacts.

4.4. Robustness Check

In order to further ensure the robustness of the study findings, the eigenvector centrality tests were analyzed using raw continuous data for the input–output sectors (see Table S3 for eigenvector centrality for all sectors), and the results were as follows:

As can be seen in

Table 5. Eigenvector centrality of major sectors.

Sectors	2002	2007	2012	2017	2018
Coal Mining and Washing Products	13.432	12.705	12.548	12.185	12.093
Petroleum and Nuclear Fuel Processing Industry	13.432	12.757	12.548	12.239	12.144
Agricultural Products	12.05	11.449	11	10.602	10.348
Forest Products	12.243	12.232	12.194	11.945	11.862
Livestock Products	11.458	12.033	11.212	10.106	9.806
Financial Industry	13.358	12.757	12.548	12.239	12.144
Electronic Equipment Manufacturing	13.432	12.307	12.133	12.164	12.072
Education	13.358	12.757	12.548	12.239	12.144
Healthcare	13.358	12.757	12.337	12.17	12.077
Electricity, Heat Production and Supply Industry	13.358	12.757	12.548	12.239	12.144
Metal Products	13.432	12.757	12.548	12.239	12.144
Plastic products	13.432	12.705	12.548	12.239	12.144
Water Production and Supply Industry	13.358	12.757	12.548	12.239	12.144
Railroad Transportation and Urban Rail Transit Equipment	13.358	12.757	11.464	10.492	10.338
Social Security and Social Welfare Industry	11.229	12.667	12.158	12.239	12.144
Environmental Resources and Public Facilities Management Industry	13.358	12.757	12.247	12.084	11.994
Wholesale and Retail Trade Industry	13.358	12.757	12.548	12.239	12.144

, the financial sector tops the list of industries. Resource-consuming industries, represented by coal mining, as well as high-energy-consuming and high-polluting industries, such as electricity and heat production and supply, refined petroleum and nuclear fuel processing products, metal products, and plastic products, still occupy an important position in the Chinese economy. Environmental resources and public facilities management, social security and social welfare, and railroad transportation and urban rail transportation equipment have a lower position in the Chinese economy. Agricultural products, forest products, and livestock products have the lowest position in the Chinese economy.

5. Conclusions

Based on China’s latest input–output table data, combined with SNA indicators, the entropy weight method was used to construct a comprehensive set of evaluation index systems of industry positions. The establishment of this indicator system is scientific and reasonable, and this proves that the status of the industrial sector is comprehensively determined by multiple aspects, including direct connection between industries (centrality), indirect connections (structural holes), direction of the connection (influence), and breadth and depth of contact (weak network, strong network). Among them, the influence index often shows a trend that is not completely consistent with other indexes. This is because the contribution of the degree of influence and the degree of being affected by the comprehensive status index cannot be measured only from a mathematics and statistics perspective. In fact, the influence index can only clarify the direction of the influence, but it cannot determine the quality of the influence. It should be combined with the economic reality in order to make a comprehensive judgment.

According to the existing literature, the industrial structure is closely related to economic efficiency, economic stability, carbon emissions, and sustainable development. The following are the results of the comprehensive evaluation of the status of China’s sub-industry sectors from 2002 to 2018:

• From the economic efficiency perspective, on the one hand, changes in the status of the subsectors of the machinery and electronics manufacturing industries reflect the continuous increase in the added value of China’s manufacturing products, as well as the increase in technology intensity and the continuous rise of the industrial value chain. On the other hand, the high and stable status of the coal mining industry indicates that low-efficiency fossil energy still accounts for the main proportion of the country’s energy utilization structure. The consumption of fossil energy is damaging the environment, threatening people’s safety and health. Facing the new constraints brought by carbon reduction and carbon neutrality, it is necessary to further improve traditional energy efficiency, optimize the energy structure, promote energy transformation, and deepen the energy revolution so that human beings can enjoy a better and healthier life.
• From the economic stability perspective, although the chaos in the real estate industry (known as the “bubble”) has been improved to a certain extent, it currently occupies the top position in the high-risk and non-entity industry—the financial industry. Hence, this is worthy of caution. In the future, guidance, regulation, and supervision of the financial and insurance industry should be strengthened, and attention should be paid to the diversified and balanced development of the modern service industry [39] in order to improve the overall stability of the industrial network. Additionally, attention should be paid to the performance of education investment, and the use of education funds should be improved. In terms of social security, while the country actively assumes more responsibilities, a social security management system based on socialized management should be established.
• From the perspective of environmental impact, high energy consumption, and high-polluting industries (such as electricity and heat production and supply), refined petroleum and nuclear fuel, metal, and plastic products still occupy an important position in China’s economy. Meanwhile, industries that are beneficial to ecological environmental protection and sustainable development (e.g., waste resources and waste material recycling and processing, water conservancy management, ecological protection, and environmental governance) have not been paid attention to, lack sufficient funds, and cannot effectively play their role in positive environmental changes. Considering China’s current important national strategy, the “dual carbon goals,” future industrial support policies should be focused on eco-environmental protection industries.

In the past, a large number of documents have been combined with input–output tables and SNA methods in order to investigate the industrial structure. The overall structure of the industrial network and its evolutionary characteristics, whether at the global level or that of a single country or region, has been fully studied [15,16,17]. This type of research focuses on the analysis of mathematical and statistical characteristics of the input–output network, without paying too much attention to the meaning of the policy. Some studies focus on the role of industrial sectors, aiming to find the key industries in the input–output network and thus provide the basis for targeted policy proposals. In the literature, some scholars have a subjective bias; they only study the power and status of one or a few sectors because they are more interested in those sectors [18,19,20,21,22,23]. These studies lay the foundation for this paper, which differs in the following ways: first, this paper uses the latest input–output table data in China, which has not been studied so far. Second, in terms of the selection of industrial sectors, this paper uses unconsolidated input–output data of subdivided industrial sectors for calculation and evaluation. This overcomes the limitation of the previous literature, that the original input–output information among industries is easily missed when using input–output tables and social network methods to analyze industrial structure. By combining various metrics of the social network analysis method through the entropy weighting method in order to determine the weights, a comprehensive evaluation index system is constructed, which provides an effective idea for future research and development of new methods used for individual status evaluation.

The final possible limitation of this study is that the input–output table provided by the National Bureau of Statistics of China has inconsistencies in the number of industrial sectors each year; the names of industry sectors are also different. Consequently, some important industrial sectors (such as the construction industry), whose internal structure changes greatly each year, cannot perform the time-series analysis of their status. We also regarded subdivided industry sectors with similar names in different years as the same sector. This treatment is more subjective, and may cause inaccurate analysis results. However, in general, we are most concerned about the industrial sector every year; the number of specific sectors and department names is relatively stable, and, therefore, has little effect on the key conclusions.