Ecological Efficiency Evaluation, Spatial Difference, and Trend Analysis of Logistics Industry and Manufacturing Industry Linkage in the Northeast Old Industrial Base

Abstract

The ecological efficiency of industrial linkage considering unexpected output is an important indicator to measure the coordinated development of industrial linkage, resources, and the environment. It is an important basis for realizing the sustainable development of industry linkage. Taking the composite index of carbon emissions of the logistics industry and pollution emissions of the manufacturing industry as the unexpected output, we used the unexpected SBM model to evaluate the ecological efficiency of industrial linkage between the logistics industry and the manufacturing industry in Northeast China from 2011 to 2019, and used the spatial autocorrelation analysis method to analyze the spatial differences in industrial linkage efficiency. The results show that (1) considering the unexpected output, in Northeast China, the ecological efficiency cannot reach a high level of linkage development stage. (2) The results of the spatial correlation show that there are spatial differences between H-H agglomeration and L-L agglomeration in the linkage ecological efficiency of the two industries, and the spatial agglomeration attribute is relatively stable. (3) The analysis results of spatial agglomeration characteristics show that the spatial agglomeration of the two industries has a spatial evolution process from the southern coastal area to the central region. (4) Spatial trend analysis shows that in Northeast China, the western region is slightly higher than the eastern region, while the southern region is higher than the northern region. (5) From the development trend of linkage ecological efficiency, the linkage ecological efficiency of the study area will be improved in the future, but in the short term, the linkage ecological development level is not high and may still be at the primary linkage level.

1. Introduction

In China’s total social logistics, industrial goods logistics account for more than 90% of the total logistics, and more than 90% of the whole process from raw material collection, production, and manufacturing to consumption is in the logistics link. With the expansion of the economic scale of industrial clusters, environmental pollution and carbon emissions are also increasing. The imbalance between economy and environment has restricted the high-quality development of China’s economy. Exploring the change rule of industrial ecological efficiency is conducive for the government to formulate reasonable environmental regulation policies, promote the improvement of environmental quality and high-quality industrial development, and achieve the “win-win” goal of sustainable economic and ecological development.

Facing increasingly fierce international and domestic market competition, if the manufacturing industry and logistics industry want to seek comparative advantages in the long term and continue to grow in a competitive environment, they must be able to adapt to the new green and sustainable development environment [1]. Green development has become an important coordinating factor for manufacturing and logistics industries to eliminate development bottlenecks [2], an urgent requirement for improving supply chain performance [3], and an important part of the transportation power strategy [4].

Since the 18th National Congress of the Communist Party of China proposed that ecological civilization construction be an important part of promoting the “five in one” overall layout and the coordinated promotion of the “four comprehensive” strategic layout, the importance and necessity of ecological civilization construction have been further highlighted. Ecological efficiency is considered an important indicator for the comprehensive evaluation of environmental performance, which can directly and accurately reflect the dynamic synergy between economic benefit output and resources and environment input. Its core is “low input, more output, less emissions”, which is consistent with the requirements of ecological civilization construction. Therefore, improving urban ecological efficiency is a key way of ecological civilization construction [5].

Sustainable development is to meet the needs of contemporary people without compromising the ability of future generations to meet their needs. Based on this development model, some scholars believe that integrating the concept of sustainable development into the core business functions of the supply chain can enable business organizations to obtain a competitive market position in the global environment. In view of the positive role of sustainability in the supply chain, sustainable supply chain management has gradually attracted the interest of many researchers. Sustainable supply chain management (SSCM) was born from the process of integrating the concept of sustainable development into supply chain management. At first, it mainly focused on environmental aspects, as well as traditional economic aspects. Later, it expanded to social aspects, especially after Elkington introduced the concept of the triple bottom line [6].

Due to the increasing pressure of development on the environment, sustainability is an unavoidable topic that must be addressed in the development of all industries in today’s society. Similarly, the sustainable development of linkage efficiency is also more and more closely related to the environment. Freight transport and logistics industries contribute to increasing environmental and health problems [7]. As for manufacturing enterprises, scholars have found that environmentally sustainable manufacturing practices may be positively correlated with the results of competition [8]. More attention should be paid to the environmental factors when researching the linkage efficiency of the logistics industry and the manufacturing industry. The continuous expansion of the industrial agglomeration scale of the two industries has brought about the rise in the costs of production factors and environmental pollution, making environmental problems prominent [9].

Previous studies have shown that supply chain entities involved in manufacturing and logistics can obtain more benefits by performing green contracts [10], and theoretically, the deepening of the integration of “two industries” will help to increase economic output and reduce carbon emissions [11]. Therefore, studying the ecological efficiency of the logistics industry and the manufacturing industry has important theoretical and practical significance for promoting the high-quality development of these industries, promoting the sustainable green growth of the regional economy, realizing the transformation and upgrading of industrialization, and realizing the development of China’s low-carbon economy [12,13,14].

Northeast China is the birthplace of China’s manufacturing industry. It has a complete manufacturing system, and its manufacturing strength has a prominent position in the country. The rapid development of the manufacturing industry has led to many environmental problems, including excessive energy consumption, excessive resource consumption, excessive discharge of industrial waste, and deterioration of the ecological environment [15]. The rapid development of the logistics industry has also brought huge negative externalities to society, and the problems of energy consumption and pollutant emissions caused by its economic activities have become increasingly prominent [16]. Studying the ecological efficiency of the logistics industry and the manufacturing industry has important practical significance for promoting the transformation and upgrading of the manufacturing industry in Northeast China, realizing the high-quality development of low-carbon and green regional logistics, promoting regional economic development, and realizing the goal of “made in China 2025” [17]. According to the relevant policy documents issued by the state, the state attaches great importance to the revitalization of the old industrial base in the Northeast, and strives to promote the industry in the region to the middle and high-end levels, making remarkable achievements in the quality of economic development and economic benefits, and achieving a high level of sustainable supply chain capacity.

Based on the above background, in order to establish the ecological efficiency evaluation index system of the logistics industry and the manufacturing industry, we performed the ecological efficiency evaluation, spatial difference, and trend analysis of Northeast China according to the index system. We aimed to solve three problems.

The first is the index selection of the sustainable supply chain performance evaluation index system. The second is to ensure the rationality of the sustainable supply chain performance evaluation index system. The third is to promote the continuous improvement of sustainable supply chain performance.

The contribution of this research can be summarized as follows:

(1)

Based on the four dimensions of economy, society, ecology, and resources, we selected and studied the performance evaluation indicators of sustainable supply chains.

The existing literature on the performance evaluation of sustainable supply chains is mostly based on the triple bottom line theory. We further divided the environmental dimension into ecological and resource dimensions; that is, we studied the performance evaluation of sustainable supply chains based on four dimensions. In addition, the existing research on the ecological dimension is insufficient, so this paper also further enriches the indicators under the ecological dimension.

(2)

The performance evaluation index system of sustainable supply chains is constructed. On the basis of the literature research and empirical research, the performance evaluation index system of sustainable supply chains is preliminarily constructed. It enriches the existing research and points out the future research direction for the development of supply chain sustainability evaluation tools.

(3)

The super-efficiency SBM-DEA model with unexpected output is used to evaluate the sustainable supply chain, and the unexpected output is introduced into the performance indicators of the sustainable supply chain to make the performance evaluation results more realistic. At the same time, this DEA model based on non-radial and non-angle is used for sustainable supply chain performance evaluation, which also makes up for the shortcomings of traditional CCR and BCC models for sustainable supply chain performance evaluation.

The main structure of this paper is as follows.Section 2 reviews the latest literature closely related to this study. Section 3 mainly introduces the five models used. Section 4 introduces the selection of indicators and the acquisition of unexpected data. Section 5 contains the empirical analysis. Section 6 summarizes the conclusions of this paper.

2. Literature Review

2.1. Related Research on Logistics and Manufacturing

First, the current research on low-carbon logistics at the macro level focuses on industrial agglomeration [18], environmental regulation [19], technological heterogeneity [20], and low-carbon port evaluation [21]. At the micro level, it focuses on supply chain sustainability [22], supply chain transportation strategy design [23], supply chain decision making and coordination [24], supply chain carbon policy [25], supply chain integration [26], low-carbon storage design [27], corporate environmental responsibility [28], tripartite low-carbon game [29], low-carbon human resource evaluation [30], sustainable distribution [31], and low-carbon multimodal transport path optimization [32].

Secondly, the current research on low-carbon manufacturing at the macro level focuses on carbon locking in manufacturing [33], green assessment of high-tech manufacturing [34], and green development of manufacturing [35]. At the micro level, it focuses on low-carbon workshop collaborative optimization [36], multi-party low-carbon evolutionary game [37], and low-carbon manufacturing workshop facility layout [38].

Finally, research on sustainable supply chains focuses on the sustainable supply chain system [39], digital sustainability [40], digital twins of production logistics [41], Industry 4.0 [42], blockchain technology [43], green supply chain management and strategic alliance [44], the intermediary role of supply chain strategy [45], sustainable supply chain management under big data [46], digital transformation [47], and key obstacles to sustainable supply chains [48].

To summarize, the sustainable supply chain has become a key area in recent research on the development of low-carbon industries closely related to this study.

2.2. Related Research on Spatial Econometric Methods

The research on spatial econometric methods in the logistics industry includes the spatial spillover effect of logistics agglomeration on economic growth in prefecture-level cities of Anhui Province [49], research on the spatial spillover effect of logistics industry agglomeration on regional economic growth in the prefecture-level cities of the Chengdu Economic Zone [50], poverty reduction effect measurement of logistics industry development in contiguous poverty-stricken areas [51], the economic spillover effect and spatial heterogeneity of inter-provincial logistics industry agglomeration [52], and spatial econometric analysis of the impact of regional logistics on regional economy [53].

To summarize, the current application of spatial econometric methods in the logistics industry is increasingly focused on the research of city-level scale.

The research on spatial econometric methods in the manufacturing industry explores the impact of industrial synergy on regional economic growth [54], the economic benefit analysis of industry agglomeration in Yangtze River Delta [55], and the transfer of producer services and manufacturing industries [56].

To sum up, the current application of spatial econometric methods in the manufacturing industry has increasingly focused on the research of manufacturing and producer services.

Therefore, starting from the current research focus and trends, we adopted the spatial econometric method to focus on the relevant research of logistics and manufacturing at the city level.

2.3. Related Research on Index Selection

In terms of indicator selection, the indicators of the manufacturing industry focus on market size [57], operation level [58], human capital level [59], infrastructure investment level [60], etc. The indicators of the logistics industry focus on market size [61], infrastructure level [62], human capital level [63], and the external economic environment [64].

To sum up, the indicator selection does not involve the unexpected output of the industry.

3. Materials and Methods

3.1. Unexpected SBM Model

The DEA model is a commonly used efficiency evaluation method. The greatest disadvantage of the traditional DEA model is that the output index of the model is generally the expected output index of economic and social categories, but it is not applicable to the unexpected output index, does not take into account the relaxation of input and output, and there may be efficiency measurement deviation in the selection of the radial and angle. To solve these problems, Tone put forward a kind of non-radial measurement DEA model, an evaluating decision-making unit (DMU) method with the slacks-based measure (SBM). The SBM model adds the slack variable into the objective function, so the economic interpretation of the model is to maximize the actual profit [65]. At the same time, Tone proposed a super SBM model for evaluating SBM-effective DMUs, making up for the inability to calculate the efficiency values of all DMUs [66].

In the super SBM evaluation, the SBM model needs to evaluate the DMUs, and the effective DMUs for the SBM are evaluated by the ultra-efficient SBM model. In this work, we introduce unexpected outputs into the super-efficient SBM, which results in an improved super-efficient SBM model that takes unexpected outputs into account. DMUs are uncountable numbers, and each DMU consists of input m, an expected output r1, and some unexpected output r2. The forms of vector are denoted as $\mathrm{x}\in {\mathrm{R}}^{\mathrm{m}}$, ${\mathrm{y}}^d\in {\mathrm{R}}^{\mathrm{r}1}$, ${\mathrm{y}}^u\in {\mathrm{R}}^{\mathrm{r}2}$, where $X$, ${\mathrm{y}}^d$, and ${\mathrm{y}}^u$ are matrices, and $X=[x_1,\dots ,x_n]\in R^{m*n}$, $Y^d=[Y_1^d,\dots ,Y_n^d]\in R^{r_1*n}$,$Y^u=[Y_1^u,\dots ,Y_n^u]\in R^{r_2*n}$. The SBM model is expressed as follows:

$$\min \rho =\frac{1-(1/m){\displaystyle \sum _{i=1}^m(w_i^{-}/x_{ik})}}{1+1/(r_1+r_2)({\displaystyle \sum _{s=1}^{r_1}(w_s^d/y_{sk}^d}+{\displaystyle \sum _{q=1}^{r_2}(w_q^u/y_{qk}^u})}$$

(1)

Subject to

$$\begin{gathered} x_{ik}={\displaystyle \sum _{j=1}^n{x}_{ij}}{\lambda }_j+w_i^{-}i=1,\dots ,m \\ {y}_{sk}^d={\displaystyle \sum _{j=1}^n{y}_{sj}^d}{\lambda }_j-w_s^{d}s=1,\dots ,r_1 \\ {y}_{qk}^u={\displaystyle \sum _{j=1}^n{y}_{qj}^u}{\lambda }_j+w_q^{u}q=1,\dots ,r_2 \\ {\lambda }_j>0j=1,\dots ,n \\ {w}_i^{-}\ge 0i=1,\dots ,m \\ {w}_s^d\ge 0s=1,\dots ,r_1 \\ {w}_q^u\ge 0q=1,\dots ,r_2 \end{gathered}$$

According to the above model, the efficiency of the manufacturing and logistics industries in the three northeast provinces from 2011 to 2019 is calculated. The following coupling and coordination method is used to express the linkage efficiency, and LD (manufacturing and logistics linkage efficiency) is used to express it.

The manufacturing industry and logistics industries are regarded as a composite system. The manufacturing industry and the logistics industry are the subsystems of the composite system. The coordinated development process of the two industries is determined by the internal order parameters of the composite system. Assuming that the order parameter of the subsystem is Ui (i = 1, 2), we calculate the order degree U (U_ij) with u (U_ij) ∈ [0, 1] according to Equation (3). The order degree U (Ui) of the subsystem is calculated according to Equation (4), and the greater the value of U (Ui) ∈ [0, 1], the higher the development level of the corresponding subsystem.

$$\mathrm{U} ({\mathrm{u}}_{\mathrm{ij}})=\left\{{}_{({\alpha }_{ij}-{\mathrm{x}}_{ij})/({\alpha }_{ij}-{\beta }_{ij})}^{({\mathrm{x}}_{\mathrm{ij}}-{\beta }_{\mathrm{ij}})/({\alpha }_{ij}-{\beta }_{ij})}\right.$$

(2)

$$\mathrm{U} ({\mathrm{u}}_{\mathrm{i}})={\displaystyle {\sum }_{\mathrm{j}=1}^{\mathrm{n}}{\omega }_{\mathrm{ij}}}\mathrm{U}({\mathrm{u}}_{\mathrm{ij}})$$

(3)

From the above calculation, assuming that the order degree of the manufacturing subsystem is U₁ and the order degree of the logistics subsystem is U₂, the coupling degree C of the manufacturing industry and the logistics industry is calculated according to the coupling coefficient model in physics, and C ∈ [0, 1], as shown in Equation (5).

$$\mathrm{C}=\sqrt{U_1\times U_2}/(U_1+U_2)$$

(4)

The closer C is to 1, the higher the coupling degree is.

The coupling degree can reflect the degree of correlation between manufacturing and logistics, but it cannot reflect the degree of coordinated development between manufacturing and logistics. This paper continues to build the coordination degree model between manufacturing and logistics to further measure the relationship between the development of the two industries. It is assumed that the coordination degree of the composite system composed of the manufacturing industry and the logistics industry is LD, the comprehensive coordination index of the two subsystems is t, and a and b are undetermined coefficients, which are the subsystem weights of the manufacturing industry and the logistics industry in the composite system, respectively. Referring to relevant literature [67,68], assuming that the logistics industry and the manufacturing industry are equally important, the undetermined coefficients are determined to be 0.5, that is, a = b = 0.5, which indicates that the logistics industry and the manufacturing industry are equally important. The calculation formula is shown in Equations (5) and (6).

$$\mathrm{T}=\mathrm{a}{U}_1+\mathrm{b}{U}_2$$

(5)

$$\mathrm{LD}=\sqrt{\mathrm{C}\times \mathrm{T}}$$

(6)

The closer LD is to 1, the higher the level of coordinated development.

3.2. Spatial Correlation Analysis Model

According to the first law of geography, the linkage ecological efficiency of cities in the study area may have spatial correlation, which exists in a certain type of spatial diffusion model. In this paper, the global and local Moran’s index (Moran’s I) are used for spatial correlation test. The values of global and local Moran’s index are in the range [−1, 1]. If the global Moran index is greater than 0, there is a positive spatial correlation between linkage ecological efficiency, that is, cities with high (low) levels are adjacent to cities with high (low) levels, and if it is less than 0, it is the opposite. The meaning of the local Moran index is similar to that of the global Moran index. If it is greater than 0, it means that cities with high (low) ecological efficiency are surrounded by cities with high (low) ecological efficiency, and vice versa if it is less than 0.

In the exploratory spatial data analysis, the spatial weight should reflect the close degree of the connection between the adjacent regional transport infrastructure network and the regional economic, social, and environmental factors. The existing spatial weight matrix setting methods mainly include 0–1 spatial weight, a distance weight matrix, and a transport network element weight matrix. This paper takes the regional carbon emissions as the unexpected output and the spatial correlation of ecological efficiency as the starting point to analyze its spatial relationship. Due to the diffusion effect of carbon dioxide gas and the characteristics of ecological self-absorption, the most commonly used spatial weight is 0–1; that is, it is assumed that the attributes of geospatial units are only related to adjacent spatial units. Therefore, in this study, the 0–1 weight matrix is used for the spatial relationship analysis. In this paper, 1 and 0 are used to represent the adjacency relationship between city I and city J in spatial attributes [69]. For the calculation method and interpretation of the global and local Moran index, please refer to the works of Chen Qiang [70].

3.3. Spatial Clustering Effect Model

The spatial clustering analysis method is conducive to the deep mining of spatial data clustering characteristics and to better identifying the degree of clustering. The distance between objects in space can be represented by location distance, while the similarity degree of features can be reflected by attribute distance. Set the rectangular coordinates of area center Pi as (x_i, y_i) The corresponding r attribute vectors are (a_i1, a_i2, a_i3, ⋯, a_in). Then, the position distance D_P and attribute distance Da between points P_I and P_J can be expressed as follows:

$${\mathrm{D}}_{\mathrm{p}}=\sqrt{({\mathrm{x}}_{\mathrm{i}}{-\mathrm{x}}_{\mathrm{j}}{)}^2+({\mathrm{y}}_{\mathrm{i}}{-\mathrm{y}}_{\mathrm{j}}{)}^2}$$

(7)

$${\mathrm{D}}_{\mathrm{a}}=\sqrt{{\displaystyle \sum {}_{\mathrm{k}=1}^{\mathrm{n}}}({\mathrm{a}}_{\mathrm{ik}}{-\mathrm{a}}_{\mathrm{jk}}{)}^2}$$

(8)

The spatial distance DS is as follows:

$${\mathrm{D}}_{\mathrm{s}}={\omega }_{\mathrm{p}}\sqrt{({\mathrm{x}}_{\mathrm{i}}{-\mathrm{x}}_{\mathrm{j}}{)}^2+ ({\mathrm{y}}_{\mathrm{i}}{-\mathrm{y}}_{\mathrm{j}}{)}^2}+{\omega }_{\mathrm{a}}\sqrt{{\displaystyle \sum {}_{\mathrm{k}=1}^{\mathrm{n}}}({\mathrm{a}}_{\mathrm{ik}}{-\mathrm{a}}_{\mathrm{jk}}{)}^2}, ({\omega }_{\mathrm{p}}+{\omega }_{\mathrm{a}}=1)$$

(9)

Given that some regions are close to each other and have certain similar characteristics, we used the K-means clustering method to reasonably divide the clustering types in the region: high-value agglomeration, secondary high-value agglomeration, medium concentration, secondary low-value agglomeration, and low-value agglomeration.

3.4. Spatial Trend Development Model

Trend surface analysis is a commonly used spatial statistical method in geographic information science. It is a mathematical method to simulate the spatial distribution and change trend of geographical elements by using mathematical surface. It is often used to simulate the spatial distribution of geographical elements. In this study, the ArcGIS 10.2 tool is used to construct the trend surface analysis chart in order to analyze the systematic change in logistics and manufacturing linkage efficiency in the whole study area.

3.5. Trend Prediction Model

The generalized regression neural network (GRNN) was proposed in 1991 according to the changes in neural networks. It is a kind of commonly used nonlinear parametric regression model. GRNN’s learning ability is powerful and accurate, so it can effectively build prediction models [71,72,73].

$${\mathrm{R}}_{\mathrm{i}}(X)=\exp \left[-{(x-x_i)}^T(x-x_i)/2{\sigma }^2\right],i=1,2,\dots ,n$$

(10)

$${\mathrm{S}}_D={\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{R}}_{\mathrm{i}}}(X),{\mathrm{S}}_{Nj}={\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{y}}_{ij}{\mathrm{R}}_{\mathrm{i}}}(X),j=1,2,\dots ,k y_j=S_{Nj}/S_D$$

(11)

4. Indicators and Unexpected Data

4.1. Selection of Indicators

We referred to the research results of logistics industry and manufacturing industry efficiency [74,75] and constructed the index system as shown in

Table 1. Index system of manufacturing industry and logistics industry linkage ecological efficiency evaluation.

Evaluation System	Pointer Type	Name of Index	Unit
Manufacturing system	Input indicators	Number of employees	10,000 people
		Total assets	100 million
	Output indicators	Industrial added value	100 million
		Main business income	100 million
	Unexpected output	Industrial waste water Industrial waste gas Industrial solid waste Carbon emissions from manufacture	Tons Billion standard cubic meters Tons Tons
Logistics system	Input indicators	Number of employees	10,000 people
		Fixed capital investment	100 million
		Energy consumption	10,000 tons of standard coal
	Output indicators	Freight volume	Tons
		GDP of transportation industry	100 million
		Cargo turnover	Million ton-km
	Unexpected output	Carbon emissions from transportation	Tons

.

4.2. Calculation Method of Carbon Emission in Logistics Industry

We adopted the basic methods provided by the 2006 IPCC Guidelines for National Greenhouse Gas Inventories. This analysis takes 21 kinds of energy as the statistical object. We calculated the carbon emission (D) of the logistics industry and manufacture industry [76], as shown in Formula (12).

$$D={\displaystyle \sum A_i\times }{B}_i\times C_i\times N_i\times \frac{44}{12}$$

(12)

5. Empirical Analysis

5.1. Evaluate the Ecological Efficiency

The original data are from China Statistical Yearbook, China Environmental Statistical Yearbook, and the provincial statistical yearbooks of Liaoning, Jilin, and Heilongjiang in from 2010 to 2020. In the statistical process, according to the industry classification standard of the national economy, the logistics industry belongs to the freight and storage part of transportation and warehousing, post, and telecommunications in the tertiary industry, and the logistics industry data are counted according to this standard. The output value of China’s manufacturing industry accounts for more than 90% of the total industrial output value and plays a dominant role in the industrial structure. Due to the availability of data in the statistics, some data of the manufacturing industry are counted with industrial data.

As shown in

Table 2. Ecological efficiency of manufacturing and logistics industries in 2011–2019.

Year	LE	ME	LD
2011	0.561	0.591	0.476
2012	0.573	0.528	0.451
2013	0.583	0.508	0.451
2014	0.584	0.606	0.495
2015	0.577	0.508	0.443
2016	0.567	0.503	0.451
2017	0.579	0.512	0.456
2018	0.585	0.523	0.460
2019	0.596	0.535	0.465

, first of all, from 2011 to 2019, the ecological efficiency of the manufacturing industry in Northeast China generally maintained stable development in fluctuations. The phased analysis found that from 2011 to 2013, this number increased slightly, which may be due to the positive stimulus of the national policy to revitalize the northeast at that time. From 2013 to 2015, the ecological efficiency was relatively stable and slightly decreased, which may be related to the slowdown of economic growth in Northeast China, which is consistent with the relevant research conclusions [77]. From 2015 to 2019, ecological efficiency increased steadily, which is closely related to the national policy of paying more attention to ecological and environmental protection during the 14th Five Year Plan period. Secondly, the same trend is also reflected in the ecological efficiency of the logistics industry. This shows that the ecological efficiency of the two industries is closely related to local economic development and environmental policies, which is consistent with the relevant research conclusions [78]. Finally, from 2011 to 2019, on the whole, the linkage ecological efficiency between the two industries maintained a stable development, indicating that the industrial development and environmental policies in Northeast China have better implemented the national policies, paying more attention to environmental protection and emission reduction. However, the overall linkage efficiency is not high, and there is still room for improvement.

5.2. The Result of Spatial Correlation of the Ecological Efficiency

5.2.1. Global Moran’s Index

The results of Moran’s I value in the study area are calculated by the ArcGIS10.2 software (see

Table 3. Moran’s index of the linkage ecological efficiency.

Year	Moran’s Index	Z Score	p-Value	Agglomeration
2011	0.4245	5.09106	0	Yes
2016	0.5384	6.3324	0	Yes
2019	0.5049	5.9013	0	Yes

). The results show that the annual Z value is positive and the p value is less than α (0.05), indicating significant spatial autocorrelation. Therefore, it can be explained that the regions with similar (high or low) development levels of logistics and manufacturing in Northeast China from 2011 to 2019 are spatially concentrated (2011 and 2016 are the beginning years of the 12th Five Year Plan and the 13th Five Year Plan, respectively. Due to the availability of data, only 2019 data can be obtained at the latest, so these three years are selected as the research time points).

5.2.2. Local Moran’s Index

According to Lisa’s calculation results, each city in the study area is divided into four quadrants according to different agglomeration characteristics (as shown in

Table 4. The Lisa analysis of the linkage ecological efficiency.

Features	2011	2016	2019
High–High	Shenyang, Dalian	Shenyang, Dalian	Shenyang, Changchun
High–Low	Harbin	Harbin	Harbin, Daqing
Low–High	Dandong, Siping	Dandong	Siping, Liaoyuan
Low–Low	Hegang, Jiamusi, etc.	Hegang	Shuangyashan, Qitaihe, etc.

).

The cities in the HH cluster type have the characteristics of high linkage efficiency and high efficiency of surrounding cities. The result showed that these cities are concentrated in the southern part of the study area, and most of them are cities with concentrated economic populations in the region. The local industrial development itself is relatively developed, and it is also the key planning and development area of the region, so it is easy to form an agglomeration area with efficient development.

The cities in the LH cluster type have the characteristics of low linkage efficiency and high efficiency of surrounding cities. The result showed these cities are concentrated in the central part of the study area, and most of them are distributed around the cities with concentrated economic populations in this area. Most of the local industrial development is the supplement and support of the surrounding high-efficiency cities, but there is still a certain gap compared with the high-efficiency cities.

The cities in the LL cluster type have the characteristics of low linkage efficiency and low efficiency of surrounding cities. The result showed that these cities are concentrated in the northern part of the study area, most of which are border cities. The local industrial development is relatively singular, and it is not the key planning and development area of the region, so it is easy to form marginal areas.

The cities in the HL cluster type have the characteristics of high linkage efficiency and low efficiency of surrounding cities. The result showed that these cities are concentrated in the central and northern parts. There are too many inefficient cities around, which shows that Harbin and Daqing, as the core cities of the Hachang urban agglomeration, have limited driving force in the linkage efficiency of the two major industries in the surrounding cities, which is also related to the decline in the overall economic development environment in Northeast China.

Based on the above analysis, it is obvious that the spatial agglomeration attribute of linkage efficiency of the logistics industry and the manufacturing industry in Northeast China is relatively stable, and regardless of agglomeration type, the impact on surrounding cities is not strong. Therefore, it is found that the spatial diffusion effect in this region is not obvious. Shenyang, Dalian, and Changchun can produce positive spillover effects on the surrounding cities in HH agglomeration-type cities, forming a development trend of multiple agglomeration areas. Siping and Liaoyuan are the typical cities in the LH type of agglomeration. These cities are located in the transition zone between the central and southern Liaoning urban agglomeration and the Harbin Changcheng city group, surrounded by these developed cities. The typical cities in the LL agglomeration type are Shuangyashan and Qitaihe. These cities are located in the border area of the region, and close to the border, with small populations. Historically, the industrial structure is singular and the industrial foundation is weak, forming a low concentration area. Harbin is a typical city of the HL agglomeration type, with good accessibility of transportation infrastructure and a good manufacturing foundation, which produces spatial spillover effects on the surrounding area.

5.3. The Result of Spatial Agglomeration Effect of the Ecological Efficiency

Figure 1 shows the spatial agglomeration distribution of logistics industry efficiency in Northeast China in 2011, 2016, and 2019. The low-value areas of logistics industry development are light blue, and the high-value areas are dark blue. From the perspective of the spatial evolution process, the spatial agglomeration of logistics industry efficiency of cities in Northeast China changes from high-efficiency value concentration in southern coastal cities to secondary high-efficiency value concentration in the Central South. From the perspective of time evolution, the distribution of the logistics industry in Northeast China can be divided into two stages: from 2011 to 2016, the development of the logistics industry was mainly concentrated in Liaoning Province, in which Dalian, Jinzhou, Yingkou, Dandong, Shenyang, Jinxi, and other areas were the high value-gathering areas; Yichun, Heihe, Jiamusi, Jixi, Hegang, and other areas were the low value-gathering areas of the logistics industry. From 2016 to 2019, the logistics industry developed rapidly, with the addition of Changchun, Siping, and other high value-gathering areas. The spatial autocorrelation of Dalian, Shenyang, and other areas was not significant, while the low value-gathering areas decreased.

Figure 2 shows the spatial agglomeration distribution of manufacturing efficiency in Northeast China in 2011, 2016, and 2019. The low-value region of manufacturing development is light blue, and the high-value region is dark blue. Based on the analysis results of the spatial agglomeration of the manufacturing industry in Northeast China from 2011 to 2019, from the perspective of the spatial evolution process, the spatial agglomeration of the manufacturing industry efficiency in Northeast China has evolved from the sub-high-efficiency value concentration of southern coastal cities and small regions in the middle to the high-efficiency value concentration in the middle and sub-high-efficiency value concentration in the middle. From the perspective of time evolution, the agglomeration and distribution of the manufacturing industry in Northeast China can be divided into two stages. From 2011 to 2016, the high-value areas of manufacturing development were mainly concentrated in Liaoning Province, among which Siping, Jinzhou, Yingkou, Dandong, Shenyang, and other areas were the high-value areas of manufacturing development; Yichun, Hegang, Heihe, Jixi, Jiamusi, Shuangyashan, and Tonghua were the low-value areas of manufacturing development. From 2016 to 2019, the high-value areas of manufacturing development are mainly concentrated in Jilin Province and Heilongjiang Province, of which Harbin, Changchun, Siping, Jilin, and Dandong are the high-value areas of manufacturing development; Yichun, Hegang, Heihe, Jixi, Jiamusi, and Shuangyashan are the low-value areas of manufacturing development.

Figure 3 shows the spatial clustering distribution of the two industries’ linkage efficiency of cities in Northeast China in 2011, 2016, and 2019. Among them, the low-value area of the two industries’ linkage development is light blue, and the high-value area is dark blue. From the perspective of the spatial evolution process, the spatial agglomeration of the two industries’ linkage efficiency of each city in Northeast China changes from high value and sub-high-efficiency value of southern coastal cities to high value of central cities and sub-high-efficiency value of southern and central cities. From the perspective of time evolution, it can be divided into two stages: from 2011 to 2016, the high-value areas of two industries’ linkage development are mainly concentrated in Liaoning Province, among which Jinzhou, Yingkou, Panjin, Dalian, and other southern coastal areas are the high-value areas; the low-value areas of the two industries’ linkage development are mainly concentrated in Heilongjiang Province, Yichun, Heihe, Hegang Jiamusi, Jixi, Shuangyashan, and other areas. From 2011 to 2016, the high-value areas of the two industries’ linkage development are mainly concentrated in Jilin Province and Heilongjiang Province, of which Harbin, Changchun, Jilin, and other areas are the high value-gathering areas of the two industries’ linkage development; the low-value areas of the two industries’ linkage development are mainly concentrated in Liaoning Province, namely Jinzhou, Yingkou, Panjin, Dalian, and other southern coastal areas. From 2016 to 2019, the spatial aggregation of the two industries showed a spatial evolution process from the southern coastal area to the central region.

5.4. Empirical Analysis of Spatial Trend Development of the Ecological Efficiency

In this study, ArcGIS10.2 software was used to analyze the trend development of efficiency (see Figure 4, Figure 5 and Figure 6). In the three-dimensional space, the east–west (green lines) and north–south (blue lines) orthogonal planes are used to explain and compare the relevant problems.

The figures show that the distribution of the logistics industry in Northeast China in 2011 is in the east–west direction, higher in the west than in the east, U-shaped in the north–south direction, low in the middle, and high on both sides. In 2016, the overall distribution of the logistics industry in Northeast China showed a downward trend in the east–west direction, and a U-shaped distribution in the north–south direction, low in the middle and high on both sides, and higher in the south than in the north, which is consistent with the trend of the logistics industry in 2011. In 2019, the overall distribution of the logistics industry was basically a straight line in the east–west direction, the distribution in the north–south direction was flat, and the southern part was slightly higher than the northern part.

As shown in Figure 4, in the east–west direction, the development of the logistics industry in Northeast China is higher than that in the west, and that in the south is higher than in the north. The result of trend surface analysis shows that the development of the logistics industry in Northeast China is relatively high in the middle and south of Northeast China, which is basically consistent with the results of spatial autocorrelation analysis and cluster analysis.

As shown in Figure 5, the overall distribution of the manufacturing industry in Northeast China in 2011 is inverted U-shaped in the east–west direction, high in the middle and low on both sides, and the distribution in the north–south direction is lower than in the southern region. In 2016, the overall distribution of the manufacturing industry in Northeast China presents a slightly inverted U-shape in the east–west direction, and the distribution in the north–south direction is lower than in the south, which is consistent with the trend of the manufacturing industry in 2011. The results of the overall layout of the manufacturing industry in 2019 are basically consistent with those in 2011.

As shown in Figure 6, in 2011, the linkage ecological efficiency in Northeast China increased linearly in the east–west direction, and the eastern region was lower than the western region in the east–west direction; it decreased linearly in the north–south direction, and the north–south distribution was lower than the southern region in 2011. The distribution trend of the linkage ecological efficiency in Northeast China in 2016 is consistent with that in 2011. In 2019, the overall layout of the manufacturing industry showed an inverted U-shape in both east–west and north–south directions.

To summarize, this phenomenon can be observed by studying the spatial distribution characteristics of linkage efficiency in Northeast China: the western region is slightly higher than the eastern region, while the southern region is higher than the northern region. The efficiency trend surface analysis results are basically consistent. The main reason is that the eastern and northern parts of Northeast China belong to the Changbai Mountain Area and the Daxinganling Area, respectively. Due to the restrictions of traffic conditions, the industrial development is relatively backward and the industrial structure is singular.

5.5. The Results of Trend Prediction of Ecological Efficiency

The calculation process of trend prediction was programmed by MATLAB 2018. From the calculation results (

Table 5. Network training error test of the ecological efficiency from 2011 to 2019.

Region	Three Northeast Provinces
Index	LE			ME			LD
Result	SV	AV	TE	SV	AV	TE	SV	AV	TE
2011	0.591	0.586	0.005	0.561	0.554	0.007	0.476	0.479	−0.003
2012	0.528	0.522	0.006	0.573	0.572	0.001	0.451	0.455	−0.005
2013	0.508	0.503	0.005	0.583	0.588	0.005	0.451	0.453	−0.003
2014	0.606	0.599	0.007	0.584	0.582	0.002	0.495	0.497	−0.002
2015	0.508	0.498	0.010	0.577	0.572	0.005	0.443	0.441	0.002
2016	0.503	0.501	0.002	0.567	0.561	0.006	0.451	0.446	0.005
2017	0.512	0.509	0.003	0.579	0.577	0.002	0.456	0.458	−0.002
2018	0.523	0.517	0.006	0.585	0.582	0.003	0.460	0.461	−0.001
2019	0.535	0.531	0.004	0.596	0.592	0.004	0.465	0.467	−0.002

SV, AV, and TE are sample values, analog values, and test errors, respectively.

), it can be seen that when the diffusion value is 0.4, the prediction effect of the network is the best.

The simulation results in

Table 6. Development trend of the ecological efficiency from 2020 to 2025.

Region	Three Northeast Provinces
Year	LE	ME	LD
2020	0.492	0.580	0.466
2021	0.496	0.585	0.471
2022	0.507	0.584	0.496
2023	0.513	0.588	0.511
2024	0.526	0.583	0.525
2025	0.534	0.587	0.531

show that the linkage efficiency values of three provinces (LE, ME, LD) will continue to rise after 2019, but the overall linkage development level is not high and is still at the primary linkage level.

In order to improve the accuracy of prediction, the K-fold cross-validation method was used for data cross-validation, the GRNN network was trained, and the best SPREAD was found by cycle. The calculation process was programmed and completed with MATLAB 2018, and the calculated value of the coupling coordination degree of the two industries from 2011 to 2019 was used as the test sample for simulation. From the calculation results, it can be found that when the SPREAD value of the network is 0.4, the prediction effect of the network is the best, the simulation value is relatively consistent with the actual value, and the model is highly feasible (as shown in Table 5).

6. Conclusions

6.1. Theoretical Significance

In this study, we considered the unexpected output of the industry, used the unexpected SBM model to evaluate the ecological efficiency, used the spatial autocorrelation analysis method to analyze the spatial differences of industrial linkage efficiency, and used Moran’s model to analyze the spatial differences of industrial linkage efficiency. We employed Lisa analysis, K-means spatial clustering, and neural networks to analyze and compare the spatial and temporal differentiation characteristics, spatial diffusion effects, aggregation effects, and development trends of urban logistics industry efficiency, manufacturing industry efficiency, and two-industry linkage efficiency in Northeast China. We quantitatively analyzed the spatial differences in structure and function between the whole and local areas of the three provinces in Northeast China. The main results can be summarized as follows:

(1)

The calculation results of industrial linkage ecological efficiency in the study area from 2011 to 2019 show that on the whole, the efficiency of the manufacturing industry has maintained a stable development in fluctuations; the efficiency of the logistics industry has improved slightly in fluctuation. Considering the unexpected output, the study area cannot achieve a higher linkage development level.

(2)

From the perspective of spatial correlation characteristics, there are spatial differences between high–high agglomeration and low–low agglomeration in the industrial linkage ecological efficiency of the study area from 2011 to 2019, and the attribute of spatial agglomeration is relatively stable. High-efficiency areas are mainly concentrated in the southern coastal areas and inland provincial capital cities, while low-efficiency areas are mainly concentrated in the northern border areas.

(3)

From the perspective of spatial agglomeration characteristics, the spatial agglomeration of the two industries shows a spatial evolution process from the southern coastal area to the central region. The pattern distribution of spatial agglomeration type may be related to the economic development level of each region. The technology level of the logistics industry and the manufacturing industry is relatively high in areas with a higher economic level, the technology level is relatively backward in areas with a low economic level, and the efficiency level of the two industries’ linkage is also relatively low.

(4)

From the perspective of spatial trend development, in Northeast China, the western region is slightly higher than the eastern region, while the southern region is higher than the northern region. The efficiency trend surface analysis results are basically consistent. The main reason is that the eastern and northern parts of Northeast China belong to Changbai Mountain Area and the Daxinganling Area, respectively. Due to the restriction of traffic conditions, the industrial development is relatively backward and the industrial structure is singular.

(5)

From the development trend of linkage ecological efficiency, the linkage ecological efficiency of the study area will be improved in the future, but in the short term, the linkage ecological development level is not high and may still be at the primary linkage level.

Considering the five conclusions, we put forward corresponding policy recommendations.

(1)

According to the analysis results of the ecological efficiency of industrial linkage, we should strengthen the promotion of green and sustainable supply chains in Northeast China. In the long run, the establishment of a green supply chain can help enterprises reduce costs and improve long-term benefits. At the same time, the green practice activities of the supply chain can also help enterprises improve their corporate image, fulfill their social responsibilities, and enhance their long-term benefits.

(2)

According to the analysis results of spatial correlation, we should strengthen the driving effect of industries in high ecological efficiency cluster areas, and take the green supply chain as the link to strengthen the promotion of green logistics in the border areas of northeast China, so as to reduce the low ecological efficiency cluster areas.

(3)

According to the analysis results of the characteristics of spatial agglomeration, in order to achieve a balanced development between different regions, we can guide the cross-regional flow of human resources, technology research and development, and management experience related to the logistics and manufacturing industries, and pay attention to the spatial spillover effect of highly concentrated cities, so as to improve the linkage efficiency between the dynamic logistics industry and the manufacturing industry and achieve more economic benefits.

(4)

According to the analysis results of spatial trend development, we should give full play to the industrial advantages of various regions in Northeast China, strive to promote the development of featured agriculture and tourism in the surrounding border areas, change the status quo of the singular local industrial structure, develop more green and environmentally friendly multimodal transport to the surrounding areas, and drive the industrial cost reduction and efficiency increase in border areas.

(5)

We should focus on improving the linkage level of ecological efficiency. In the process of linkage between the manufacturing industry and the logistics industry, the coordination of the linkage mechanism is related to the enthusiasm of both parties to participate in linkage and the stability of linkage, which is an important guarantee to improve the linkage level.

6.2. Implications for Sustainability

Advanced and high-end manufacturing is the symbol of economic revitalization in Northeast China, and also the basis for regional competitiveness and innovation. As an important strategic industry supporting the development of the national economy, the logistics industry plays an important role in the improvement and upgrading of the economy in Northeast China. In addition, the logistics industry and the manufacturing industry have a two-way interaction relationship of coordinated development and mutual promotion. Therefore, the development of the manufacturing industry and the logistics industry has not only become one of the key points to promote the coordinated development of Northeast China, but also has important practical significance for changing the mode of economic growth in Northeast China and optimizing the industrial structure.

The study has the following implications for sustainable development. First, at the macro level, in the face of the policy requirements for an environmentally friendly country, it is a complicated process for logistics and manufacturing to achieve sustainable development. Thus, relevant decision makers from the perspective of diversification must collect more comprehensive information, deepen the understanding of sustainable decision-making problems, and carry out scientific and high-quality decision making for sustainable industry development. Secondly, at the micro level, the sustainable development and the success of interactive integration between logistics and manufacturing depend on the effective integration of logistics and manufacturing systems to achieve win-win cooperation. Information fusion is an effective way to realize system fusion. By strengthening the information exchange and cognitive sharing within the two systems, different viewpoints can be integrated into the comprehensive thinking of decision-making problems at the enterprise level, so as to avoid the failure of decision making due to the lack of an environmental perspective.

6.3. Limitations and Future Research

The limitations of this paper are as follows. (1) The measurement method of ecological efficiency needs to be improved. Although the unexpected output SBM model effectively considers the problem of variable relaxation and unexpected output, there are several research units with equal efficiency that cannot be compared microcosmically. In future research, it will be necessary to consider adopting the super-efficiency SBM model, which can effectively avoid multiple cases of the same efficiency, and more scientifically and comprehensively measure the development level of ecological efficiency. In recent years, the modified gravity model has been widely used in the research of space networks, but there are still limitations in the incomplete consideration of relevant factors. (2) The spatial connection mechanism of ecological efficiency is complex. We only examined the spatial connection network of urban ecological efficiency in Northeast China from the perspective of urban gravitational connection, and thus cannot accurately measure the interaction between cities’ ecological efficiency. In the discussion of the formation mechanism of the spatial correlation network of urban ecological efficiency, only indicators are selected from the socio-economic environment, and physical geographical factors and other possible influencing factors are not included in the study. Therefore, future research should focus on comprehensively discussing the spatial interaction and linkage mechanism of ecological efficiency.