A Combined IO-DEMATEL Analysis for Evaluating Sustainable Effects of the Sharing Related Industries Development

Abstract

Emerging economies suffer more severe pressure to coordinate sustainable production and consumption, and new industry input–output (IO) solutions are urgent. An emerging service supply and consumption mode, the sharing economy (SE) penetrates various industries and rewrites the technological–economic links between sectors via integrating scattered and idle resources. The sustainable effects of such industrial linkages changes triggered by its specific activities remain unclear. The study aims to clarify the mechanism of SE in coordinating the relationship between production and consumption among industries and demonstrate its particular role in cleaner production and sustainable consumption to reveal its multistage and systematic impact on industrial development through input–output linkages. Grounded on the IO and complex systems theory, we establish an IO-DEMATEL industrial development evaluation model embedded in the IO table. The comparative analysis of IO cross-sectional data in 2007 and 2018 deduces the dynamic evolution of industrial interaction and structure under the penetration of SE, as well as its contributions to the national economy in China. The result reveals that sharing activities raise the industry prominence moderately and promote the cleanliness and resource-conservation of industrial connections. However, the prominence of the five sectors regarding input, output, or comprehensive influence currently ranks in the lower-middle levels among all industries. Industrial structure optimization has only been observed on the supply side. Our findings complement a deepened empirical evidence of SE and its sustainability, which is also of practical significance for policymakers to identify the critical industries that conduct cleaner production and sustainable consumption.

1. Introduction

As sustainable development (SD) was explained as a broad principle, its concrete implementation was broken down into 17 goals [1,2]. For any country, striving to change production and consumption patterns is the primary way to achieve these goals [3,4]. Existing production and consumption patterns are no longer adequate for the developing countries with large populations and rising disposable incomes. Especially in China, the increasing pressure on balancing expanding domestic demand and cleaner production is severe to achieve high-quality development [5]. In this context, as an emerging service supply and consumption model, the sharing economy (SE) is admired as a potential path to sustainability owing to its contribution to resource-saving production and sustainable consumption [6,7,8,9]. The advantage of “low participation cost” attached to its core concept of the separation of ownership and use rights has brought massive corporate and individual participants [10,11]. The SE has penetrated diverse industries of the economy through various organizational forms and business models and exerts far-reaching influence. Supported by Internet technology [12], the SE has gained universal advocacy for its ability to facilitate dynamic and efficient supply–demand matching, lead to high-quality supply, and promote the supply-side structural reform in China. Multiple reports on government work have stressed supporting it as a crucial means to enhance the economy’s resilience, optimize the industrial structure, and construct a modern industrial system [13]. Merely over a decade, SE’s gross merchandise volume (GMV) rose to 3.225% of GDP [14] and grew into an essential part of the national economy.

However, assessments of its sustainable impact are far from adequate and comprehensive [15,16]. Some studies suggest there is apparent win–win cooperation between the sharing-related and other industries, especially those manufacturing sectors with a bulk of idle capacity or surplus supply [17,18]. There are also works verifying that the implementations of sharing activities inevitably lead to cutthroat competition between them and original activities, engendering a squeeze on the market share of the traditional industry and offsetting the positive effect throughout the system [19,20]. There is an unmet need to sort out the transmission mechanism and demonstrate the multiorder influence of SE from the perspective of input and output. Additionally, restricted by methodology and data acquisition, the existing works at home and abroad have been more explored at the microlevel [21,22,23], and the macroscopical contributions are dominated by a qualitative discussion [24,25,26] since there is a lack of statistical data [27]. Few empirical studies have verified SE’s comprehensive impact on sustainability [24]. The research status consisting of massive decentralized, micro, and subjective discussions goes against reaching an in-depth consensus on the chain reaction triggered by the implementation of sharing activities. As sharing activities exist in multiple sectors [28] and bring both positive and negative effects [29], it is urgent to identify the ultimate benefits and losses from a macro and systemic perspective.

In those regards, based on national IO statistics, this study is committed to evaluating the effects of sharing-related industries on promoting cleaner production, sustainable consumption, and upgrading the industrial structure to supplement the lack of empirical studies on the macro level. Deeply, this study aims to systematically comb the mechanism of SE on the relationship between production and consumption from the perspective of supply–demand linkage, demonstrating its complex influence on industrial interaction to deepen scientific consensus.

To narrow the gap, we will address the following specific research questions (RQs):

RQ1.

How to evaluate the impact of sharing activities on the production of relevant upstream and downstream industries?

RQ2.

What benefits does the implementation of sharing activities bring to other departments and the industrial structure?

RQ3.

How prominent are sharing activities in the economic system, and how should they superiorly develop?

The main contribution of this study is to evaluate the sustainable benefits of sharing-related industries via developing a combinatorial approach coupling the respective advantages of the input–output (IO) method and the Decision Making Trial and Evaluation Laboratory (DEMATEL) approach. Based on the comparative analysis of two IO table profiles, this integrated approach is conducive to systematically capturing the positive changes in industry connections and the dynamic evolution of industry interactions and drawing more reliable and robust conclusions.

Our research is organized as follows: the second section illustrates the relationship between the SE and sustainable production and consumption theoretically and reviews previous studies; the third section clarifies the improvements of the combined IO-DEMATEL method and enumerates the related calculation indexes; the fourth section introduces the data and sources; the fifth section briefly presents the results; the sixth section carries out profound discussion, and put forward corresponding policy suggestions; the last section summarizes the conclusions and shortcomings.

2. Sharing Economy, Sustainable Production, and Consumption

2.1. Theoretical Mechanism

2.1.1. Supply Side: Improve the Quality of Supply

As production and supply are the fundamental guarantees of the national economy and the primary driving force for the smooth operation of the economic system, only high-quality supply can ease the conflicts between production and consumption and promote SD. The open-accessing of idle stock resources and production capacity on sharing platforms breaks free from ownership shackles to trade, enriches the supplied form, and broadens the supply boundary [30,31]. Correspondingly, it can facilitate the transaction in a larger radius and shorter time, expanding the trading scale and strengthening supply capacity [32]. Furthermore, timely release of demand information can bridge the loss of industry efficiency caused by the miscommunication and mismatch between supply and demand [33]. It can also conduce to the suppliers’ continuous innovation from the beginning of the production process to efficiently respond to the constantly escalating requirements via flexible product design, refined production and service procedures, and elastic organizational form [34]. In short, the SE promotes intensive, clean, and efficient production through the gradual revolution in production and supply, creates a more efficient supply system, and optimizes supply quality [35].

2.1.2. Demand Side: Stimulate Market Demand

As the final demand, consumption is the core pulling power for economic growth and serves as the engine. Benefiting from the multidimensional innovations of Internet technology, organization, and business model, the SE complies with the trend of consumption upgrading through more product categories, higher service quality, and lower consumption costs to meet the diversified, flexible, and interactive consumer demand [10,32,36]. Additionally, it also breaks the situation of information asymmetry between supplier and demander. As the unimpeded exchange of information contributes to exploring new market opportunities, the SE synchronously facilitates the response to the exposed demand and stimulates new potential demand [23,37,38]. In a nutshell, using information tools as its core, the SE has evolved consuming behavior, improved demand response efficiency, stimulated new orders, and contributed to constructing a sustainable consumption system [39].

2.1.3. Market Entities: Increase Interaction

Market entities, the essential participants in development, including enterprises and individuals, are creators of social wealth. In the economic system, they are the implementer of the technical–economic links between inputs and outputs among industries and the leader of technological progress and business model innovation. The access mechanism of the SE based on the use rights makes it inclusive and open to attract the extensive participation of various market subjects. For individuals, it may be new employment opportunities, income streams, or more accessible ways of spending [10,11,40]. In terms of enterprises, it can also represent a cost-saving advantage, a fresh market opportunity, or a new contender for market share [41]. When the market allocates resources, the SE can entirely trigger the vitality of market subjects by encouraging all kinds of sharing activities such as demand, technology, production material, and service resources [36]. It can improve resource allocation efficiency through active interaction among market players and ultimately contribute to economic restructuring and high-quality development.

As the highlighted impacts in the subtitles of this section, on the supply side, the SE optimizes supply quality by promoting resource conservation and clean production. Meanwhile, on the demand side, it can fully release the market demand via upgrading consuming behaviors and efficient demand response. In addition, it is also conducive to build an efficient matching and responding pathway between supply and demand. In this sense, the SE’s theoretical mechanism to encourage a high-quality balance between cleaner production and more sustainable consumption is constructed and outlined in Figure 1.

2.2. Literature Review

2.2.1. The Debates on Sustainable Effects of the SE

The SE has been highlighted worldwide for its sustainable potential in practice. It has permeated many economic sectors with various organizations and business models, significantly impacting the industry, other sectors, and their interactions and aroused broad interest among scholars [28]. Under the triple bottom line analysis framework [42], numerous empirical studies address the specific impacts of particular departments or sharing activities from isolated economic, environmental, or social perspectives. For instance, examine the ecological benefits [22,43] of sharing mobility in reducing haze [44], carbon dioxide [45], and greenhouse gas [46]; evaluate the direct, indirect, and induced carbon footprint of Airbnb in Sydney [47]; or investigate the economic and social impact of the sharing accommodation on tourism, hospitality, and the residents [48,49,50,51]. These independent contributions provided remarkably in-depth conclusions and implications. Unfortunately, they do not constitute a harmonious and systematic consensus on the impact of the SE; instead, they triggered debates about its sustainability performance [16,28,52].

Roughly, there are two groups. Supporters hold that SE has apparent economic gain [33,53], resource-saving, and environmental protection effects [54,55]. Regarding economic benefits, they claim that usage-based consumption leads to significant changes in the distribution of income and expenditure by providing more flexible working and new income opportunities [56,57,58]. It also makes goods and services available at a lower cost to those who cannot afford them, increasing their utility [20]. Similarly, environmental impacts are considered obvious and directly related to resource efficiency, making the utmost of idle resources and products is conducive to slowing down resource consumption [35] and reducing pollutant emissions [59,60]. In terms of social welfare, the SE is often promoted as participatory democracy, a better quality of life, and a fairer and more open society [7,61]. Still, opponents insist it is less optimistic than promised and will trigger negative externalities of sharing activities in their implementations. A growing body of research demonstrates and confirms the inevitable negative environmental and social impacts of additional consumption and rebound effects deviating from SDGs [62], such as extra carbon emissions, increased inequality in income distribution, and skill discrimination [49,50,63].

Furthermore, the lack of systematic consideration of the industry diversity and the heterogeneity of development backgrounds [64] has led to a definition dilemma despite being covered under “sharing” [11]. Thus, the assessments across different industries and countries have yielded mixed results [43,56]. For instance, the conclusions deriving from sharing mobility cannot generalize to sharing accommodation [55], nor do the suggestions for the developed countries lack adaptability and guidance significance to the developing [43]. In this context, the sustainable impact of SE is controversial and even refutes each other [28], seriously hindering the progress of scientific research.

2.2.2. The Challenges of Methodology and Data Acquisition

Regarding the methodology, the assessment schemes for its sustainable impact mainly consist of bottom-up and top-down structures, distinguished by the data types they adopt [27]. Specifically, life cycle assessment (LCA) is a mature and widely used in bottom-up modeling [65]. Precisely, it is a data-intensive and time-consuming evaluation method requiring highly technical, temporal, and geographically representative data [66,67]. The ambition to map the full life-cycle details of the whole system faces incredible difficulties in data collection and system rules characterization. In this sense, LCA is more suitable for microscopic, well-defined studies of material flows and precise cause-and-effect calculations [68] than for assessing the cross-industry and interaction effects throughout the entire system.

Second, the top-down evaluation approach includes IO analysis [47], IO analysis with environmental extended accounts (EE-IO) [69], and the general equilibrium model (CGE). The establishment of production and consumption functions in econometric models represented by CGE heavily relies on assumptions about current and future behavior, expanding the uncertainty and systematic error. Accordingly, the sustainability assessments of the SE based on CGE are currently in their stage of theoretical deduction and numerical simulation [36], and there is still a lack of empirical tests with accurate data from the objective world. Additionally, the research framework grounded on the IO theory defines the linkages between sectors via the industrial aggregative statistics in national economic accounts [70,71]. Due to the complete and detailed description of the correlation between various departments, the IO analysis is generally cited in identifying key sectors and industrial structures in the economy [47,72]. It has also been applied to model how changes in inputs or outputs in some sectors trigger corresponding changes in the economic system and other sectors to assess systematic ripple effects and overall impacts [73]. Extending with environmental accounts of each industry, the EE-IO could evaluate the policy effects on resource consumption and pollution emission under current technological and economic conditions [27,69]. Similarly, taking economy as a complex system, the DEMATEL, a classical approach to identifying critical elements in a vast and complicated system [74], is developing as an emerging systematic method to evaluate industry prominence and linkages.

In conclusion, it is urgent to supplement more contributions of systematically and comprehensively demonstrating the costs and benefits of sustainability to resolve disputes and deepen consensus. Considering the diverse patterns of sharing activity and their intricate influences [28], an overall impacts study from a macro perspective can get out of the trap of mutually exclusive conclusions of different industries and avoid the disputes driven by boundary-truncated overdetailed discussions. Moreover, the influence of sharing activities must be transmitted and externalized through the IO linkages of products and services of various departments in the economic system, making the IO theory appropriate for our research. As a result, coupling with the national IO table, this study employs the IO-DEMATEL approach to map the complicated industrial relationships and prominence. We present the dynamic evolution among sectors deriving from the penetration of SE via visualizing the comparison of the results to discuss its sustainable impact on production and consumption.

3. Methods

3.1. IO Method and Correlation Indicator

3.1.1. A Brief Introduction to the IO Method

Wassily Leontief originally developed the core concepts of compiling the input–output table (IOT) and its analysis methods. His contribution lies in taking the techno-economic connection between sector couples as the link to explore the industrial correlation and structure across the macroeconomic system [70,73]. The whole national economic IOT in value type consists of three quadrants as intermediate input/use, final use, and added value as graphical in

Table 1. China’s input–output table in 2018 (at buyer’s prices).

Unit: Ten Thousand Yuan (RMB)
	U	TIU	TFU			IM	GO
I			TC	GCF	EX
TII		$x_{ij}$ (i)	(ii)
TVA	VA001	(iii)
	VA002
	VA003
	VA004
TI

Annotation: Refer to Abbreviations for a complete list of abbreviations.

. The three parts comprehensively and systematically reflect the economic relations concluded by the input–output relations of product and service between industries regarding total amount and structure. In the IOT, the consuming relationship is quantified via translating the products and services flow into the corresponding monetary value. Annual data reveal the interdependencies of production and consumption in different sectors of single or multiple economies over one year.

The interdependent and inter-restricted quantitative relationship among departments follow inherent principles of balance, as described:

Supposing the Gross Output of each sector (Total Intermediate Use + Total Final Use) plus Imports equals Total Input (Total Intermediate Input + Total Value Added), shorthand as TIU + TFU + IM = TII + TVA. Then, the relationship between the Intermediate use table ($X$) and Direct consumption coefficient matrix ($A$) is:

$$\begin{gathered} X={\left(x_{ij}\right)}_{n\times n} \\ A={\left(a_{ij}\right)}_{n\times n},a_{ij}=x_{ij}/{\displaystyle \sum }_j\left(x_{ij}\right)\left(i,j=1,2,3,\dots ,n\right) \end{gathered}$$

$x_{ij }$represents the intermediate input of industry $i$ consumed by the output of industry $j$; $a_{ij }$represents the direct consumption coefficient.

The IO method often verifies the industrial correlation effects by calculating the Industrial Correlation Coefficient and the Induced production of final demand. In assessing industrial association, the essence of the Influence Coefficient (IC) is to compare the total demand ripple effects of industry $j$ with the average requirement generated by each sector. The higher the ratio is, the more significant the pulling impact of sector $j$ is. Similarly, the Response Coefficient (RC) contrasts the production-pushing effect of an industry with the average level of all sectors; the larger ratio is, the more notable contribution of sector $i$ is from the supply side. However, the IO analysis could not directly derive the combined strength of push–pull effects by simply adding IC and RC, as the two use different divisors in their calculation [75].

3.1.2. Adopted Indicator

We employ the Production-Induced Measurement of final demand to examine the evolution of the linkage between the sharing-related industries and their upstream and downstream sectors (RQ1). $K_i$ indicates each domestic sector’s output induced by increasing unit investment and consumption in department $i$.

$$\begin{gathered} K_i={[I-(I-\widehat{M})A]}^{-1}\times [(I-\widehat{M})]S_i \\ \widehat{M}=diag\left(m_i\right),m_i=I{M}_i/\left(TI{U}_i+T{C}_i+GC{F}_i\right) \end{gathered}$$

(1)

$S_i$ denotes a column vector whose element in row $i$ is one, and the rest are zero. Assume that $m_i$ is the import proportional coefficient of industry $i$, and $\widehat{M}$ indicates the import coefficient matrix formed via the diagonalization of $m_i$.

3.2. Improved DEMATEL Approach and Correlation Indicators

3.2.1. A Brief Introduction to the DEMATEL Approach

The classical DEMATEL approach is adept at effectively describing the direct and indirect relations between the element couples in complex systems [74]. Its core principle is to address complex problems via fuzzy thinking: (1) divide the complex system into several elements $S\left\{S_1,S_2,S_3,\dots ,S_n\right\}$; (2) define the relationship between elements by pairwise comparison; (3) construct one $n\ast n$ matrix loadding the quantitative correlation to establish a visualized structural model; (4) evaluate the element prominence in the whole system and demonstrate the overall and structural relationship [76]. Notably, the structural model is little more than a relatively stable judgment method via quantifying subjective information and fuzzy relations through matrix operation, differing from mathematical models that pursue accurate quantitative ties [77]. In this regard, the results deriving from subjective data suffer limitations in information load and explanatory power and further restrict the application [78].

3.2.2. The Improved DEMATEL Approach Embedding IOT

In this paper, we develop an improved DEMATEL approach backed by the IOT to quantify the importance of sharing related sectors and their roles in supply and demand (RQ3) and illustrate the cascading effects and dynamic evolution of industry relationships and structures (RQ2).

Specifically, the original relationship matrix in DEMATEL is consistent with the information carried by matrix $X$ in the IOT, namely the quantitative correlation between element pairs. The basic flow table $X={\left(x_{ij}\right)}_{n\times n}$ records the mutual relation between industries pairs; horizontally, $x_{ij}$ displays a supply relationship—the number of products or services produced by $i$ provides for consumption, representing the industry $i$’s direct impact on all industries; vertically, $x_{ij}$ also indicates a demand relationship—the required goods/services quantity provided by each department during the production of $j$, representing the amount that industry $j$ is affected by all industries. The joint on supply and demand constitutes a statistically complete and accurate national economic system. In this sense, the objective data inserted in DEMATEL contributes to overcoming the subjective limitation from respondents scoring elements according to their knowledge reserves. The Direct consumption coefficient matrix $A$ displays the intensity of direct and affected effects.

Additionally, the IO theory is conducive to addressing the unsolved problems in an understandable, accurate, and objective manner of the original DEMATEL applications, such as the theoretical basis of self-dependence of elements, the scoring criterion of strength [79,80], and the issue on defining the negative influence between factor pairs. Under the IO environment, it is comprehensible that industry $i$ has the supply–demand interaction as self-sufficiency and generates a flow of products and services with itself, i.e., $a_{ii}\ne 0$. Besides, from the production inputs perspective, it is also self-evident that the technical–economic relation between industry $i$ and $j$ has quantitative and directional differences, namely $a_{ij}\ne a_{ji}$. Furthermore, identifying vital factors plays a crucial role in understanding complex systems, grasping the primary contradiction, and guiding practice. Researchers have devoted themselves to developing algorithms for identifying key sectors [81,82] and ranking their importance [83,84,85], and carrying out many experiments breaking through the idea of linearity [86,87,88] and weighting [89,90]. It is still indispensable to orient specific decision-making scenarios when comprehensively weighing the management significance of calculated results [91]. The vital industries approved in this study are those conducive to system optimization via optimizing resource allocation and promoting the sustainable transformation of production and consumption.

3.2.3. Correlation Indicators

The four steps to solve the relevant indicators are as follows:

(I) 

Define the direct influence matrix$\mathit{D}$

$$D=sA, d_{ij}=s{a}_{ij}, 0<s\le Sup, i,j=1,2,3,\dots ,n$$

(2)

$d_i={\displaystyle \sum }_{j=1}^n{d}_{ij}$ represents the sum of direct influence intensity of industry $i$ on all industries;$d_j={\displaystyle \sum }_{i=1}^n{d}_{ij}$ represents the sum of the intensity of industry $j$ affected by all industries.

(II) 

Define the total influence matrix$\mathit{T}$

$$T={\displaystyle \sum }_{i=1}^{\propto }{D}^i=D{\left(I-D\right)}^{-1}=\left(t_{ij}\right)$$

(3)

$t_{ij}$ indicates the total influence of industry $j$ on $i$. The multiorder and nonlinear relationships among industries are undeniable in the complex system. Assuming industry $i$ influences industry $j$ through an intermediate sector, the second-order effect is marked as $d_{ij}^{\left(2\right)}={\displaystyle \sum }_{k=1}^N{d}_{ik}{d}_{kj}$, i.e., $D^2=(d_{ij}^{\left(2\right)})$; by analogy, the m-order influence is denoted as$D^m$. The total influence matrix $T$ is the sum of ripple effects, i.e., $T=D+D^2+D^3\dots +D^i$.

(III) 

Measure the industry’s inputs- and outputs-oriented influence strength

$$T_r\left(i\right)={\displaystyle \sum }_{j=1}^n{t}_{ij},T_r=\left(T_r\left(1\right),T_r\left(2\right),\dots ,T_r(n\right){)}^T$$

(4)

$$T_c\left(j\right)={\displaystyle \sum }_{i=1}^n{t}_{ij}, T_c=\left(T_c\left(1\right),T_c\left(2\right),\dots ,T_c(n\right))$$

(5)

$T_r\left(i\right)$displays the sum influence of industry $i$ on all sectors; it reflects the impact of industry $i$ on supply. The larger the value is, the greater influencing power is based on output. $T_c\left(i\right)$ shows the sum influence of all industries on industry $i$, indicating the distribution of the inputs from other industries during $i$’s production under a particular technological environment. Sectors with high values will significantly impact the economy by pulling up the output of different sectors, yet they are accompanied by massive resources and energy consumption. In short, $T_r\left(i\right)$ and $T_c\left(i\right)$ are defined separately as output-based and input-based influence. From the perspective of SE, $T_r\left(i\right)$ also embodies the saving effect of industry $i$ on the supply of various industries by mobilizing idle resources and improving the repeated utilization of resources. In this connection, focusing on prominent affected industries helps capture critical sectors that drive the transformation of production and consumption.

(IV) 

Identify the outstanding industries

To measure the industry prominence degree and clarify the role in supply and demand, we introduce (1) Centrality: $M_i$ carries the combined influencing power of the industry based on input and output, revealing its position and importance in the overall economic system; (2) Causality: $R_i$ is the net value of influence in two directions, illustrating the influence difference of industry itself in input and output.

$$\{\begin{array}{c}{M}_i=T_r\left(i\right)+T_c\left(i\right)\\ R_i=T_r\left(i\right)-T_c\left(i\right)\end{array}$$

(6)

The improved DEMATEL approach can break through the dual dilemma caused by multiple indexes comparison and subjective data. Grounded on the IOT, the approach sets the comprehensive impact of input and output on a more objective and standardized scale. First, the larger the calculated value of $M_i$ is, accordingly, sector $i$ is more prominent, acting as the main driving factor or new economic growth point. Second, $R_i>0$ means that the output-oriented strength of the industry is more significant than that of the input-oriented one, inferring its demander and requiring a large amount of input from other sectors; conversely, $R_i<0$ indicates that it is a supplier. Considering our RQs, we implement the dual identification for highlighting industries from both static and dynamic ways to extract the sustainable impact of the penetration of sharing activities.

4. Materials and Data

China faces profound changes in the economic aggregate, per capita disposable income, and population structure, providing unique conditions for the emergence and exploding of SE. For instance, the Chinese prefer saving, resulting in accumulating and idling massive products and resources as incomes rise, including houses, vehicles, etc., which have completed ownership transactions. The large population has formed a huge domestic demand and abundant labor resources, especially the labor supply capacity of different time slots. As a result, the SE is penetrating the current industries with various formats and modes, affecting them and rewriting the industrial connections. According to the reports released by the Sharing Economy Research Center (SERC) of the State Information Center, the market structure of the SE in China is illustrated in Figure 2. Precisely, the statistics in those reports, named GMV, cannot be directly connected to the IOT since the latter must be strictly measured at current producer prices according to the expenditure method. Hence, the specific activities contained in the sharing fields mentioned in Figure 2 need to be reorganized according to standard industrial classification (SIC).

Since 1987, China’s National Bureau of Statistics (NBS) has compiled a national IOT every five years. NBS compiled a table in both 2008 and 2017, owing to the trial implementation of the new Industrial classification for national economic activities since 1 October 2017. The $IOT{ }^{2018}$ is the latest available edition. Compared with 2007, referring to GB/T 4754—2017, the$IO{T}^{2018}$ has indeed covered many specific sharing activities in the compilation and balancing process. Concurrently, existing studies generally regard 2008 as the birth year of the SE. Since then, the booming development of Uber, Zipcar, Airbnb, Didi, OFO, and other sharing enterprises indicates that, supported by Internet technology, “sharing” is coming into our lives in a new mode and gradually exerting a profound influence on production and consumption activities.

To deduce the dynamic evolution of industrial interaction and structure under the penetration of the SE, we select two economic profiles at discrete time points for comparative analysis, i.e., $IOT{ }^{2007}$ and $IOT{ }^{2018}$. We assume that the$IOT{ }^{2007}$ undisturbed by the SE, while the $IOT{ }^{2018}$ has embedded the emerging changes engendering from its penetration. Furthermore, we screen out the sharing-related industries in the sense of input–output linkages by comparing their respective reference SIC. Only those industries that add, revise, or highlight specific activities closely related to the known areas in Figure 2 are included in the discussion, while those which do not have corresponding revisions in their departmental explanation are temporarily excluded.

Table 2. Revision related to sharing activities.

Original Sector	Changes	Code
Urban Public Transport & Highway Passenger Transport	Add and emphasize public bicycle services, tourist passenger transport	27
Accommodation	Add homestay services, campground services, other accommodation Include short-term rental apartments, timeshare accommodation, and other new forms	30
Food and Beverage Services	New catering delivery and take-out delivery services	31
Renting and Leasing	Subdivide the daily necessities rental, including fixed stalls operating bicycles, furniture & household appliances, linen & textile, clothing & footwear, manual equipment rental services, etc.	35
Services to Households	Revise to include employing domestic workers and self-employment activities of domestic households	39

summarizes the evidence and results of this study for defining sharing-related sectors according to the revised interpretation regarding the industrial categorization of specific activities in $IOT{ }^{2018}$.

In summary, under the dual guidance of the SE practices and the new adjustment of SIC, this study combines 135 commodities in 2007 and 153 commodities in 2018 into 43 industries, of which five sectors are known to contain penetration of sharing activities. Two new $43\ast 43$ flow matrices are obtained for subsequent calculation and comparative research, i.e., $X^{2007}$ and $X^{2018}$ (see sector codes in Appendix A).

5. Results

Supported by MATLAB R2020a, this study uses the integrated approach detailed in Section 3.1 and Section 3.2 to solve the relevant indicators and visualize the analytic solutions.

5.1. Calculation of Upstream and Downstream Industry Connections

Given the matrix $A^{2007}$ and $A^{2018}$, $K_{27},K_{30},K_{31},K_{35},K_{39}$ can be solved according to Equation (1), namely the production amount of each department induced by the final demand (investment and consumption) of one unit of the sharing-related industry.

In a mathematical sense, the higher ${\mathit{K}}_i$ value is, the greater the intensity of induction is. Accordingly, observing the raised point in Figure 3 can help locate the upstream and downstream industries tightly related to the sector. For instance, Food and Beverage Services (Sector 31) closely associates with the upstream Farming, Forestry, Animal Production and Fishery (Sector 1), Manufacture of Food and Tobacco (Sector 3), Chemical industry (Sector 9), and also forms a close development community with the downstream Wholesale Trade and Retail Trade (Sector 26), Transport via Road (Sector 27), etc. It is also self-evident that each industry has the most conspicuous production inducement effect on itself, as the industry development is a self-enhancement movement derived from an endogenous power. Relatively speaking, the self-growth power of Accommodation (Sector 30) is particularly insufficient (far lower than 1), nor does A unit of input of Sector 31 yield the same amount of output.

Using $K_i{}^{2007}$ as a reference, we rank the absolute value of changes in $K_i{}^{2018}$. to capture the significant impact of sharing activities on their upstream and downstream industries in

Table 3. The intensity ranking of production-induced effects of the five sharing-related industries.

Ranking	Sector 27	Sector 30	Sector 31	Sector 35	Sector 39
1	8 (−)	30 (−)	26 (+)	33 (+)	34 (+)
2	33 (+)	22 (−)	31 (−)	36 (+)	9 (−)
3	2 (−)	34 (+)	1 (−)	34 (+)	8 (−)
4	36 (+)	36 (+)	3 (−)	8 (−)	36 (+)
5	26 (+)	39 (−)	27 (+)	19 (−)	2 (−)
6	15 (+)	2 (−)	22 (−)	2 (−)	14 (−)
7	23 (+)	8 (−)	36 (+)	11 (−)	22 (−)
8	22 (−)	26 (+)	9 (−)	22 (−)	11 (−)
9	14 (−)	9 (−)	34 (+)	18 (−)	4 (−)
10	11 (−)	18 (−)	2 (−)	9 (−)	7 (−)

The primary industry contains Sector 1; the secondary industry consists of Sectors 2–25; the tertiary industry includes Sectors 26–43.

. On the one hand, the five sectors induce a dramatic decline in production in their related manufacturing sectors. Taking Sector 27 as an example, the negative eliciting relationships between it and Sector 8/2/22 confirm that the implementation of sharing activities can significantly reduce the industry’s input dependence on its upstream manufacturing industries. On the other hand, the five sectors positively elicit the production of the service sectors, especially in Sector 33/26/36. The two changes mentioned above validate that sharing activities can enrich supply modes (e.g., the production driving effect of Sector 30 on Real Estate (Sector 34)) and release the domestic market’s vitality (e.g., Sector 27 on the Manufacture of Motor Vehicle (Sector 15)).

5.2. Calculation and Results on Industry’s Influence Power

First, the matrix $A$ is normalized according to Formula (2). In this step, regarding the selection of scale factors, based on matrix theory, when the spectral radius $\rho \left(D\right)<1$, the series $F={\displaystyle \sum }_{i=1}^{\infty }{D}^i$ converge to $D{\left(I-D\right)}^{-1}$, $\rho \left(D\right)$ is less than any matrix norm of $D$, so $s$ shall be $0<s\le Sup,$ where $Sup=\frac{1}{ma{x}_{1\le i\le n}{{\displaystyle \sum }}_{j=1}^n\left|a_{ij}\right|} or Sup=\frac{1}{ma{x}_{1\le j\le n}{{\displaystyle \sum }}_{i=1}^n\left|a_{ij}\right|}$. This study selects $max\left(Sup\right),$ namely the maximum value of the sum of columns. To be more specific: $Su{p}_r=\frac{1}{2.4293}<Su{p}_c=\frac{1}{0.8347}$, so $s_{2007}=1.1980$; $Su{p}_r=\frac{1}{2.2238}<Su{p}_c=\frac{1}{0.8441}$, so $s_{2018}=1.1847$. After normalization, we generate the direct relationship matrix $D$, recording the actual production technology connection among sector couples. Running Formulae (3)–(5) step by step could finally solve $T_r\left(i\right),T_c\left(j\right)$.

In addition, due to data collection limitations caused by too many sectors, we failed to complete the processing of a comparable price to these temporal data before the calculation. To address this concern, we further normalize the calculated value, called the contribution rate of the industry in the current year, to eliminate price interference. The contribution rate of the output, input, and combined influence of each department are plotted in the upper, middle, and lower parts of Figure 4, respectively.

The lines “data 1”, “data 2”, and “data 3”, respectively, plot the contribution rate for 2007, 2018, and the average level in 2007.

5.2.1. Results of Influencing Power Based on Outputs

The department with high $T_r\left(i\right)$ contribution rate has absolute control over the demander sector, and any fluctuation of their output will significantly affect the latter’s production. The measured values of two years concurrently display that Sectors 2/9/11/19/22, which occupy a prominent position on the supply side, are the foundation for the others. The five service industries concerned by this study rank in the middle and lower contribution level to the economy. Sectors 27/31 act as a more prominent influence based on output than the remaining three.

Compared to 2007, the contribution rate of those manufacturing industries at the supply level (Sectors 2/8/9/11/14/22) suffers a sharp decline, while the service sectors increase significantly, particularly Sectors 26/32/33/34/36, etc. Different sharing activities have different degrees of output influence, and the sharing-related industries differ in their supply-side adjustment potential. The five sectors affected by sharing activities do not show a uniform upward or downward trend, where Sector 27 leads the higher contribution increase, while Sector 39 regresses. Even so, the sum of the output-based influencing power among them maintains a slight increase (from 0.0389 to 0.0494).

5.2.2. Results of Influencing Power Based on Inputs

The result of $T_c\left(i\right)$ has dual meanings. On the one hand, it is related to the ability of a sector to promote economic development by driving production in other sectors. In this sense, the departments at the forefront of contribution, such as Sector 5/9/15/19, display a more dominant ability to drive economic growth by generating demand for goods and services from the others. Those low-value sectors, such as 21/34/40/1/26, are more independent and less reliant on supplies from the others. Similarly, the five industries also rank in the middle and lower levels of the influence based on input; relatively speaking, Sector 35 currently presents the most outstanding potential for production-pulling and resource savings.

On the other hand, it also indicates an industry’s input demand to other sectors, known as product or services consumption capacity. The secondary industries with large product consumption cardinality, such as Sectors 15/11, offer more potential to drive resource conservation and cleaner production as their input-based influences decline. Those service sectors with a low product consumption but strong growth of $T_c\left(i\right)$, such as Sectors 19/33/37, appear to have more substantial power for sustainable consumption and industrial restructuring. In the meantime, seriously affected by the significant decrease in Sectors 35/39, the aggressive demand-oriented influence of the five could not reverse to increase.

5.2.3. Results of Influencing Power Combined Inputs and Outputs

Given the previous results and the Formula (6), $M_i,R_i$ can be solved to identify the outstanding industries with integrated input–output impact and distinguish their roles in the economic system.

Globally, most secondary sectors show a declining trend of combined impact contribution, while most tertiary sectors rise. Specifically, as measured, Sectors 2/9/22/19 are the central forces of economic construction, especially the Chemical industry (Sector 9) and Mining (Sector 2) have been almost absolutely prominent in the past development. Regarding the sharing-related sectors, the three Sectors 27/30/31 with positive growth of comprehensive influence, theoretically speaking, the supporting behaviors on them have better efficiency than the remaining two at the present stage. Still, the two Sectors 35/39 ultimately fail to improve the overall influence due to their significant contribution decline as a demander. Finally, influenced by the more remarkable growth of the supply side, the combined influencing power of the five still shows an upward trend, contributing a larger share to the national economic construction.

5.3. Calculation and Results of Industrial Structure

This study further summarizes the distribution of the first two kinds of influencing power in the industrial structure in Figure 4. As observed, the results in 2007 and 2018 have some commonalities. First, the output-based influence contribution of the primary industry is higher than average across all sectors and higher than the input-based one. Second, most subsectors in the secondary industry have stronger input-based and output-based influencing power than the average in the current year (except Sectors 21/24); inversely, most tertiary sectors are lower than the average of corresponding indicators. In general, compared with input-based influence, the distribution of output-based contribution is more dispersed, and there are more top-notch prominent industries.

In terms of the aggregative output-based contribution proportion of the three industries, China has achieved preliminary achievements after decades of industrial restructuring efforts. As shown graphically in Figure 5, although China is still relying on the production of the secondary industry to drive the economy, the impetus is gradually weakening (the contribution rate decreased from 0.7897 to 0.6559). Simultaneously, the driving power of the tertiary industry in the production of all sectors is still weak but getting more substantial (the contribution rate enlarged from 0.1648 to 0.2944). Relatively, such positive change in industrial structure is hard to detect from the input-based dimension.

6. Discussions and Policy Implications

6.1. Discussions

Generally speaking, the degree of development of a country is positively correlated with the prominent contribution of its tertiary industry. However, China, affected by its international division and development stage, takes the secondary industry as the pillar of economic development for a long time.

Concerning RQ2, our calculations reveal that the introduction of the SE and related efforts have led to some exciting changes during this prolonged period dominated by secondary industry. First, the manufacturing industries have declined noticeably in influencing power with respect to both input and output orientation, implying significant technological progress or effective resource allocation (business model) in China, such as the SE. Their emergence and applications release the first-order sustainable influence of the stock resources in the economic and environmental dimensions, meeting the higher development requirements with less resource consumption and new product production [45]. Second, the service sectors gradually play a more critical role in both supply and demand and contribute more to economic efficiency and quality. The changes in the pattern of production and consumption are quietly taking place and proving that China’s industrial restructuring has achieved some results [25]. Even so, our findings also reveal that it is relatively difficult to find effective ways and forceful levers to adjust the economic structure from the demand side under China’s national conditions. In contrast, the supply-side reform is more likely to bring us remarkable results.

Regarding RQ1, the weakening of inducement force of sharing activities to their upstream manufacturing activities reveals their potential to decouple consumption growth from resource consumption and improve supply quality through optimizing idle resources and allocation. Enhancing the inducing intensity with their downstream service industries also appears to have the potential to stimulate demand and release the vitality of market entities. In conclusion, the emergence of sharing activities can boost the growth of the service industry, promote cleaner production in the manufacturing industry, and contribute to industrial structure upgrading.

However, as for RQ3, we also find that all three kinds of influencing power of them contribute a lower than average share throughout the system, so the sustainable potential has not been entirely released. Notably, sharing activities for revaluing the stock and idle resources also inevitably affect the production of re-expansion [23]. Their penetration into traditional sectors may trigger inappropriate competition between sharing and traditional activities, which counteracts the initial positive effects and leads to the loss of industry efficiency [48]. In summary, the SE does have significant and positive sustainability in integrating production and consumption, taking place at a cost.

6.2. Identification of Prominent Industries and Corresponding Policy Suggestions

From the IO perspective, the interaction between industry couples can briefly classify into the production-impulse influence as a supplier (the quadrants 1,2 in Figure 5) or the demand-pulling effect as a demander (the quadrants 3,4 in Figure 5). Relatively, it is much more complicated to assess outstanding departments in the economic system. Given that the accurate identification of the role and prominence of each industry is the prerequisite for formulating an appropriate industrial development strategy, we implement a more in-depth interpretation of the results and put forward our suggestions. The prominent sectors and their effects are abstracted in

Table 4. The evaluation results of industry prominence degree.

	Prominent Role	Sorting Formula	Prominent Sectors
Static (displayed in 2018)	Leading sector	$\max (T_r\left(i\right)+T_c\left(j\right))/M_i$	9/2
	Supplier-oriented	$\max \left(T_r\left(i\right)\right)$	9/2/26/19/22/33/12/1/36/3
	Demander-oriented	$\max \left(T_c\left(i\right)\right)$	19/17/18/15/5/16/14/20/4/6
Dynamic (compared with 2007)	Promoting cleaner production and sustainable consumption	$\max \left|(T_r\left(i\right)-T_c\left(j\right))\right|$
		$\left(1\right) \max \left(-∆T_c\left(i\right)\right)$	11/2/22
		$\left(2\right) \max \left(∆T_r\left(i\right)\right)$	33/34/36
		$\left(3\right) \max (∆T_c\left(i\right)$)	38/33/32/31
	Promoting industrial structure upgrading	$\min \left|(T_r\left(i\right)-T_c\left(j\right)\right|$ & $\max \left(M_i\right)$	14/3/8

, while more detailed role distribution information for all industries in 2018 can be found in Figure 6.

6.2.1. Static Identification and Suggestions

We derive the overall static outcomes of industrial prominence using data in 2018 since it is the closest profile to the current situation. In general, the ideal form of industrial development is that an industry can exert a powerful influence on both inputs and outputs synchronously; that is, those with high $M_i$value are regarded as the current key and driving industries. Owing to their dominant position, their investment actions and development policies can yield the most efficient return throughout the system. In this regard, giving priority to the Chemical industry and Mining is still a firm and unshakable strategy in China now and in the future. Unfortunately, as calculated, limited by the production relationship, few industries can concurrently combine their prominent degree as suppliers and demanders. Most departments gravitate toward being influential in unidirectional, making the appropriate policy solutions must be firmly grounded in their roles. For instance, emphasizing the development of supplier-oriented sectors, such as computers and nonferrous metals, is conducive to maintaining the stability of economic construction from a supply perspective. Promoting those industries with an outstanding contribution on the demand side, such as computers, and manufacturing of the transportation in water, road, air, etc., will help boost domestic demand, invigorate the market, and inject a steady flow of endogenous power into economic growth.

6.2.2. Dynamic Identification and Suggestions

Dynamically tracking the evolution of the influencing power helps to screen out and target those industries having the potential to facilitate cleaner production and sustainable consumption. Specifically, the cleaner manufacturing activities refer to those which reduce the resource and product consumption demand on the others in their production process via technological progress or optimal allocation of resources, i.e., $\max \left(-∆T_c\left(i\right)\right)$. The sustainable service industries include (1) those which have a more notable impact as a supplier, i.e., $\max \left(∆T_r\left(i\right)\right)$. and (2) those which have a growing role as a demander in driving production, i.e., $\max \left(∆T_c\left(i\right)\right)$. Pinpointing the three categories of sectors is a prerequisite for effective and integrated policies for sustainable transformation. Respectively, for manufacturing sectors such as ferrous metals, mining, electricity, steam, etc., the critical issue is to improve internal efficiency via continuously strengthening R&D investment and establishing the appropriate resource allocation model. Only the efforts to address this concern relieve the current oversupply pressure, reconcile the imbalance between supply and demand, and accelerate the transition to cleaner production. Similarly, for those service sectors with low resource consumption but growing influence, innovation in various dimensions is the main path to enhance the efficiency and sustainability of interindustry contacts, for instance, technological advances and model optimization. In terms of services whose output-oriented influence is rising, such as wholesale and retail, it is crucial to maintain an open attitude to support business and profit model innovation and improve supply capacity and quality, while for those pure consumer sectors, such as the finance, software, and information technology services, catering, etc., increasing the investment and expanding their scale can give full play to the potential to boost domestic demand.

In addition, given the disparate implementation effects of supply-side and demand-side reforms in China, there is still an urgency to precisely diagnose the core sectors conducive to upgrading industrial structure. An efficient industrial system should present a state of mutual integration and interactive development of manufacturing and service, known as greater servitization in manufacturing and greater mechanization and automation in service. That is the sector whose absolute value of $R_i$ is closer to 0 and, coupling with a higher level of $M_i$, i.e., min$\left|(T_r\left(i\right)-T_c\left(j\right))\right| \& \max \left(M_i\right)$. Pinpointing these industrial sectors with similar roles in production and consumption conduces to locate the acting points for structure upgrading, such as general- and special-purpose machinery, refined petroleum, and food. Finding breakthroughs in those industries could gradually guide the upgrading of the industrial structure via integrating the relationship between production and consumption, accelerating the integration of the manufacturing and service industry.

6.2.3. Identification and Suggestions for the Sharing Related Industries

In promoting cleaner production and sustainable consumption, it is worth noting that five sharing-related industries play different roles but contribute a sizeable positive impact overall. Analogously, concerning entirely unleashing their potential for sustainability, it is also essential to formulate policies supporting and regulating their development based on their prominent roles. As marked in red in Figure 6, Sector 27 has completely transformed into a supply sector from a demander, revealing its potential for promoting supply-side reform by avoiding unnecessary production and saving resources. Accordingly, the follow-up policy options should engage in improving their service supply capacity and quality via enriching supplied forms, such as thoroughly exploiting idle private cars and traveling seats. Still, the remaining four departments play the role of demanders, such that their effects in expanding domestic demand and promoting consumption upgrades will be fully released. By carrying out innovation in organizational forms and business models, they could adapt to the escalation of the market, encourage sustainable consumption, and cater to the supply–demand balance at a higher level. In the meantime, timely and effectively monitoring and regulating the counteraction of positive effects triggered by unfair competition in the accommodation and catering could improve their internal efficiency and sustainability.

7. Conclusions and Future Research Directions

7.1. Conclusions

In developing countries, notably China, the scenarios of enormous consumption demand, excess capacity, and a mismatch between supply and demand more easily coexist. The emergence and vigorous development of the SE do provide an essential pathway for reforming the supply and consumption sides and upgrading the industrial structure. As verified in a comparative analysis, the tertiary sector has shown great potential in supply-side reform. Although China still heavily relies on the secondary industry for development at present, the adjustment of industrial structure has achieved initial results. According to the changes in the production-induced relationship between the sharing-related industries and their upstream and downstream industries, the SE does have the potential to optimize production technology linkages and promote cleaner production and sustainable consumption. Regarding production and supply, sharing activities can decouple consumption growth from resource consumption by improving the efficiency of resource allocation and promoting cleaner production in primary manufacturing departments (Sectors 8/2/22/11). Concerning consumption, it can also induce the symbiosis and development of emerging services modes (Sectors 33/26/36), stimulate demand, and release the potential of the consumer market. According to the DEMATEL analysis, the prominence of the five sectors in terms of input, output, or comprehensive influence is lower than the average level. It implies that scholars may place too high expectations on it due to complicated reasons in the past. More profoundly, their backward influencing powers reveal the sustainable potential urgent been further released to play an outstanding role in economic construction.

7.2. Future Research Directions

As a burgeoning business form and model, the development of the SE inevitably goes hand in hand with its governance. Unfortunately, although the Chinese government has introduced adaptive and supportive policies to safeguard and promote growth, various sharing-related sectors still suffer problems frequently since 2018. First, in addition to the sharing areas such as online knowledge, skills, and medical care that remain resilient, other fields requiring closed transaction loops through offline activities have seen a decline in transaction scale and enterprise expectations due to the ravages of COVID-19. Second, the application of Internet and mobile information technology has accelerated its development yet posed the challenge of data security governance. Third, sharing activities were based on using rights mobilize idle resources to broaden the supply boundaries, yet they unavoidably suppressed consumer demand and prevented the re-expansion of production, triggering destructive competition within the industry and partly disconnection between production and consumption among sectors. Fourth, according to the calculation of the export-induced amount of a unit of final demand, it is evident that the SE is not conducive to promoting export, mainly attributed to the heterogeneity in consumption culture and stock supply between the Chinese market and the others. The homegrown sharing activities in China may face new challenges in global promotion. In those regards, it is urgent to establish an effective mechanism and policy system, including legislation, benefit distribution, product and service quality, and taxation, to promote its long-term and healthy development.

This study complements the macro empirical analysis on evaluating the sustainable impact of SE. The main contribution lies in introducing the IO-DEMATEL approach to verify the sustainable benefits of sharing-related industries and investigate the dynamic evolution of industrial interaction and structure under SE penetration. Innovatively embedding IO data into the causality and salience identification of system elements, this work expands and deepens the application of input–output and complex system theory. The combined research scheme breaks through the limitation that a single IO analysis cannot evaluate the comprehensive pulling–pushing effect and makes up for the restriction of subjective data existing in the DEMATEL approach. The results have practical significance for accurately identifying key industries contributing to sustainable transformation. Furthermore, the findings provide an essential reference for Chinese policymakers to formulate appropriate policies for a healthy SE development, orderly upgrading of the industrial structure, and ongoing promotion of clean production and sustainable consumption.

However, limitations still exist in our work. First, the latest IO data is merely updated to 2018, making our research lose the observation and evaluation of the performance of the SE over a more extended period. Second, the SE is an inclusive collection involving participation in various industries; the industry data in the current IO table is highly aggregated and lacks a detailed description of sharing activities; the market share of sharing activities in their related industries needs further clarity. Hence, more scale data of segmented sectors is required to decompose and rebalance the IOT to assess accurately in the future.