Spatial and Temporal Variation Characteristics and Driving Mechanisms of Multidimensional Socio-Economic Development Levels in Resource-Based Cities

Abstract

As resources are depleted, resource-based cities face unique challenges in the process of socio-economic development. We constructed a multidimensional socio-economic development level model by adopting Entropy Value Method, Analytical Hierarchy Process, time series weighting method, and Game Theory approach for the data of 10 indicators in 4 dimensions of 115 resource-based cities in China from 2004 to 2019 to explore the spatial and temporal divergence characteristics of multidimensional socio-economic development level and the driving mechanism of its pattern of evolution. The results show that: (1) the overall socio-economic development level of resource-based cities has improved from 2004 to 2019, but the overall level is low. Large differences exist in the spatial distribution of socio-economic development levels between cities with more significant regional spatial aggregation characteristics. (2) Secondary industry, tertiary industry, retail trade goods sales, urban construction land area, and total freight transport have a significant positive impact on socio-economic development; the correlation coefficient between the number of schools and the socio-economic development level index is negative. (3) Retail trade merchandise sales contribute the most to the Gini coefficient, where the percentage of secondary industry and urban construction land area have a higher cumulative contribution to growing cities (55.02%), the percentage of secondary industry has the lowest contribution to regenerating cities (10.94%), and the percentage of tertiary industry has an increasing contribution to declining cities year by year. Based on the above findings, some specific suggestions are provided to provide reference for resource-based city development planning.

1. Introduction

Resource-based cities are a special type of city in which resource exploitation and utilization is the leading industry to support the economic development of the whole city, and they have made great contributions to the overall economic growth of China in the past decades [1,2]. China’s resource-based cities are characterized by a high dependence on resources at the expense of resource consumption, which is likely to cause problems such as increased costs and low utilization of resources in the development process, thus forming a development model with high energy consumption, high input, and high pollution [3]. In addition, the city’s ecological industry is imbalanced and lacks comprehensive urban services. In the long run, if resource-based cities do not restructure their industries and develop non-extractive industries, their economic development space will be limited [4,5].

Influenced by the industrial revolution and the colonial activities of developed countries, early resource-based cities emerged in developed countries in Europe and America. With the decline of resource production and the emergence of new energy sources, resource-based cities were transformed one after another, and the policies and plans of resource-based cities mainly focused on revitalizing the local economy, such as the Ruhr industrial base (coal and steel) in Germany, Sudbury (nickel) in Canada, and Houston (oil) and Pittsburgh (steel) in the United States, etc. China put forward the issue of resource-based city transformation in the 1990s, in which economic transformation is the core of resource-based city transformation [6,7]. In 2013, the Chinese government classified resource-based cities into four types according to their resource security capacity and sustainable development capacity: growth, maturity, decline, and regeneration. The General Office of the State Council of the People’s Republic of China’ report [8] shows that nearly a quarter of resource-based cities have entered the decline stage. In 2020, the fourteenth five-year plan once again puts the real economy in a prominent position, and for the first time proposed to “maintain the basic stability of the proportion of manufacturing industries” and promote the advanced industrial base and modernization of the industrial chain [9]. A scientific understanding of the changing characteristics of the socio-economic development level of resource-based cities is of great significance for the study of the restrictive factors and influencing mechanisms of the economic development of resource-based cities.

There are many studies around resource-based cities, including urban efficiency, sustainable development, the impact of transformation on the economic development of resource-based cities, and the impact of government policies on urban transformation. While domestic studies mainly focus on the unsustainability of resources, foreign studies emphasize more on the single resource economy, the risk of resource industries in economic activities, and the resulting unsustainability, and thus the research on resource-based cities abroad is broader in scope [10,11,12]. Resource-based city transformation research is mainly focused on five dimensions, including economic, social, ecological environment, resources, and innovation of science and technology. The prominent problems of resource-based cities are single resource factor type, uncoordinated industrial structure, insufficient innovation capacity, and lack of integrated urban service management. Compared with ordinary cities, resource-based cities have obvious characteristics such as cyclical urban development, loose spatial structure, and low efficiency utilization of resources [13,14,15,16,17,18]. The current research focuses on the phenomenological description of the overall evolutionary characteristics of a region or a resource-based city or a specific development stage, and the research mostly starts from the perspective of a single city or a single region, while neglecting the comparative analysis between regions but paying insufficient attention to the differences in economic development between cities. The study is therefore not conducive to the identification of problems in the process of urban development in cities. Meanwhile most of the studies explored the degree of influence using data from constructed indicators, and the results are not enough to explain the nature of economic characteristics of specific cities, making it difficult to explore the causes and mechanisms behind socio-economic development [19,20]. Due to the vast territory of China, resource-based cities have significant differences in terms of resource endowment, historical background, economic base, and natural environment, and these differences have caused development differences among cities, which affect the sustainable development of cities. In this paper, long time series data are used in analysis in order to obtain the trajectory of inter-city development, and the introduction of time factor weights can extend the study of the problem from the static field to the dynamic field, which is of great practical significance for achieving the sustainable development of resource-based cities.

The choice of evaluation methods is a key element in building the model. Analytical Hierarchy Process (AHP) is the most commonly used assessment method for resource-based cities, it can be used for qualitative analysis, or both qualitative and quantitative decision analysis. Its core function is to rank and select solutions, but the results may not objectively reflect the weight of the judged solutions [21,22]. The Entropy Value Method (EVM) is highly inclusive of the degree of variation in the selection of indicators and can exclude the influence of subjective factors on the results and obtain a more objective evaluation [23,24]. The gray comprehensive evaluation method can deal with the situation of insufficient data, but the processing results obtained are easily affected by the extreme values of the data. The data envelopment analysis method does not need to consider functional relationships, estimate parameters, or dimensionless processing of data, but it needs to avoid linearity between input variables before use [25,26]. Game Theory (GT) can regard the subjective and objective sides of the game and seek the optimal combination of weights for both sides to reach in consistency [27,28]. Therefore, in this paper we combine AHP, EVM, and GT as an analysis method. Because the long time series data are more conducive to explore the mechanism of urban development, in order to make the indicators more comparable between years, we tried to add time series weights to jointly construct a multidimensional socio-economic development level model.

Based on the above, this paper identifies three research components: (1) constructing a socio-economic development level model with 4 dimensions and 10 indicators for the data of 115 resource-based cities in China; (2) analyzing the multidimensional socio-economic level and multidimensional socio-economic change patterns of each resource-based city for analysis; (3) exploring the determinants of the socio-economic development level affecting resource-based cities in China and providing suggestions for narrowing the development differences among cities.

2. Materials and Methods

2.1. Study Area and Data Source

According to the definition and classification of the National Sustainable Development Plan for Resource-based Cities (2013–2020) issued by the State Council, there are 262 resource-based cities nationwide, including 115 prefecture-level cities. The prefecture-level cities have larger economies and face more challenges in development, and to avoid double counting and to ensure the consistency of the study area level, this paper takes 115 resource-based prefecture-level cities as the research object. The specific locations are shown in Figure 1. In this paper, time series data related to the socio-economic development of 115 resource-based cities from 2004 to 2019 are selected from the China City Statistical Yearbook (Department of Urban Survey, National Bureau of Statistics of China, 2004–2019) [29]. Pre-processing such as data checking and elimination of abnormal data before using the data, part of the missing data by the statistical yearbooks of each city, the statistical yearbooks of the corresponding provinces, and the statistical bulletins on national economic and social development of each city in previous years to supplement the city vector base data were obtained from the data sharing platform of the Institute of Earth System Science (http://www.geodata.cn, accessed on 15 April 2022).

2.2. Research Methodology

2.2.1. Multidimensional Socio-Economic Development Level Model

The multidimensional socio-economic development level is a response to the level of economic development, population level, social living conditions, and the degree of influence of resource use on urban development from a multidimensional scale, so the multidimensional socio-economic development level index can reflect both the essential properties of the region in economic development and the process influenced by social development and resource changes [30,31,32]. According to the principles of scientific, dynamism, importance, comparability, and reliability of data acquisition in the selection of indicators, four dimensions of gross product, social and human development, investment, and consumption level and resource utilization level were included in the multidimensional economic development level model [33,34], and 10 indicators from four dimensions were used to construct the multidimensional socio-economic development model, it is as shown in

Table 1. Multidimensional socio-economic development index system of resource-based cities.

Target Layer	Guideline Layer	Indicator Layer
Socio-economic development index system for resource-based cities	Total production value	Regional production ($x_1$)
		Per capita GDP ($x_2$)
	Social and human development	Total population at the end of the year ($x_3$)
		Natural population growth rate ($x_4$)
		Population density ($x_5$)
	Investment consumption level	Retail sales of social consumer goods ($x_6$)
		Actual amount of foreign capital used ($x_7$)
		Total investment in fixed assets ($x_8$)
	Resource utilization level	Per capita household domestic water consumption ($x_9$)
		Electricity consumption per resident ($x_{10}$)

, and the technical route is shown in Figure 2. To ensure that the indicators are comparable between various units, the indicators of different dimensions should be normalized using extreme difference normalization, the formula is: $X=\frac{(x-x_{min})}{\left(x_{max}-x_{min}\right)},$ where x is each indicator’s value.

For the processed dimensionless indicators, we performed indicator assignment to construct a multidimensional socio-economic development level model: First, we assigned the indicators with AHP and EVM indicators respectively, and combined the results of them with GT, and then performed GT assignment with the time series weights from 2004 to 2019 to obtain the indicator weights of 115 cities from 2004 to 2019. Finally, the indicators and weights were multiplied to obtain the multidimensional socio-economic development level index. Figure 3 shows the indicator weights of Tangshan City, and the specific process is divided into 4 steps.

The first step is EVM. EVM is to judge the amount of information carried by the system itself based on the sample data, and to determine the degree of its influence on the comprehensive evaluation by the size of the information carried by the basic index variables [23,35]. Assuming that there are m indicators and n samples in the assessment, where the contribution of the $i$th indicator $A_i$ under the $j$th sample, $P_{ij}$ can be defined as Equation (1):

$$P_{ij}=\frac{x_{ij}}{{{\displaystyle \sum }}_{i=1}^m{x}_{ij}}$$

(1)

the entropy value $E_j$ of the $j$th indicator can be calculated by Equation (2):

$$E_j=\frac{-{{\displaystyle \sum }}_{i=1}^m{P}_{ij}\ln \left(P_{ij}\right)}{\ln m}$$

(2)

When $E_j$ tends to 1, the contribution of each scheme of the attribute also tends to the same attribute. The size of the difference determines the size of the weight coefficient. The difference coefficient $d_j=1-E_j$ indicates the degree of consistency of the contribution of the programs under the $j$th attribute.

As shown in Equation (3), the attribute weights $W_J$ are determined as:

$$W_J=\frac{1-E_j}{{{\displaystyle \sum }}_{\mathrm{j}=1}^m{d}_j}$$

(3)

The second step is AHP whose basic idea is to construct a comparative judgment matrix by comparing the importance degree between two adjacent indicators in the sequence, calculate the feature vector, each indicator weight factor, etc., and then compare the judgment matrix for consistency testing until passing the test criteria. This paper uses the 1~9 scale method to construct a judgment matrix for each layer [36].

The following are the details of the AHP application process: building judgment matrix A, then normalizing the judgment matrix by columns to produce $a_{ij}$, where $a_{ij=}{a}_{ij}/{\displaystyle \sum }_{k=1}^n{a}_{kj}(i,j=1,2,3\cdots n$); the vector $W_i$ is then normalized using the formula $W_{i=}{w}_i/{\displaystyle \sum }_{k=1}^n{w}_{ki}(i=\left(1,2,3\cdots n\right)$ to obtain $W_i$, and $W_i$ is the AHP weight. Next, the normalized judgment matrix is summed by the row direction to obtain $w_i$. Finally, the consistency index (CI) must be calculated in order to determine whether the matrix is consistent, where $CI=\frac{{\lambda }_{max}-n}{n-1}$ and ${\lambda }_{max}={\displaystyle \sum }_{j=1}^n\frac{{\left(AW\right)}_i}{n{W}_i}$ is the greatest eigenroot. The consistency increases as the CI approaches zero. We require an expert opinion in order to rank each variable in order of importance.

The third step is Game Theory (GT), also known as response theory, which can analyze the logic and laws of competition and the problem of optimal decision making under conditions of conflict and confrontation [37]. The fundamental idea is as follows: assume that there are L ways to weight evaluation indicators, and the associated set of basic weight vectors is $w_k=\left\{w_{k1},w_{k2},\dots ,w_{kj}\right\}$, ($K=1,2,\dots ,L$) and the combination weight coefficients are $\beta =\left\{{\beta }_1,{\beta }_2,\dots ,{\beta }_L\right\}$, if L weight vectors are arbitrarily linearly merged as:

$$W_j={\displaystyle \sum }_{\mathrm{k}=1}^L{\beta }_k{W}_{kj}{}^T$$

(4)

the calculation uses the minimization of the deviation of $W_j$ and $W_{kj}$ as the objective in order to seek consistency and compromise between various weights, and then optimizes the L linear weight combination coefficients ${\beta }_k$ in Equation (5) to provide the optimal weights $W_j$. In accordance with the matrix differential property, the best-optimized linear equation system with the first-order derivative requirement is:

$$\left[\begin{array}{ccc}{W}_1{W}_1{}^T& \cdots & W_1{W}_L{}^T\\ ⋮& \ddots & ⋮\\ W_L{W}_1{}^T& \cdots & W_L{W}_L{}^T\end{array}\right]\left[\begin{array}{c}{\beta }_1\\ ⋮\\ {\beta }_L\end{array}\right]=\left[\begin{array}{c}{W}_1{W}_1{}^T\\ ⋮\\ W_L{W}_L{}^T\end{array}\right]$$

(5)

The optimal combination coefficient ${\beta }_k$ can be obtained by normalizing Equation (5). ${\beta }_k^{*}=\raisebox{1ex}{${\beta }_k$}\!\left/ \!\raisebox{-1ex}{${{\displaystyle \sum }}_{\mathrm{k}=1}^L{\beta }_k$}\right.$ in Equation (6), and the equivalent $j$-indicator game-theoretic combination assignment of the combination weights is:

$$W_j^{*}={\displaystyle \sum }_{\mathrm{k}=1}^L{\beta }_k^{*}{W}_{kj}{}^T$$

(6)

where the value of L is 2, $W_1$ means the EVM weight, and $W_2$ means the AHP weight.

The fourth stage is to assign time series weights. Aiming at the problem of discrepancies in the forecast data of the time series multi-period model, the method of weighting the forecast results of the multi-period model can be used to solve the problem. The time weight method increases the weight of the recent model to more sensitively adapt to time series changes. The calculation is as follows:

$$GD{P}_t=\frac{gd{p}_t}{Gdp}$$

(7)

$$W_t=\frac{GD{P}_t}{\sum GD{P}_t}$$

(8)

where $Gdp$ is the is the overall growth rate of each city from 2004 to 2019; $gd{p}_t$ is the growth rate of the next year compared with the previous year. The share of GDP growth rate is shown as time weight $W_t$.

2.2.2. OLS Linear Regression Model

Ordinary least squares (OLS) mainly finds the best function match of the data by minimizing the sum of squares of the errors, and uses it to estimate parameter. The model can quantify the marginal effects of factors affecting the level of urban socio-economic development [35]. Its expression is:

$$Economic=a_0+a_{1}gdpse{c}_i+a_{2}gdpthir{d}_i+a_{3}locmk{t}_i+a_{4}builtrat{e}_i+a_{5}schoo{l}_i+a_{6}Traffi{c}_i+{\epsilon }_i$$

(9)

where $i$ denotes each city; $a_0$ denotes the constant term; $a_1~a_6$ denotes the marginal impact of each factor, and ${\epsilon }_i$ denotes the error term. Figure 4 shows the results of multiple regression fitting for the socio-economic development level index of 115 cities in 2019.

2.2.3. Gini Coefficient

Gini coefficient is an important indicator for estimating income inequality [38,39], and assuming that n represents the size of something, the indicator y is income, which can be calculated as:

$$Gini=\frac{{{\displaystyle \sum }}_{i=1}^n{{\displaystyle \sum }}_{j=1}^n\left|y_i-y_j\right|}{2n^2\overline{y}}$$

(10)

2.2.4. Shapley Decomposition Model

Shapley value is a mathematical expression proposed by Shapley (1953) to solve the problem of reasonable distribution of internal interests in cooperative game events through a combination of regression models and inequality decompositions and conducts stability analysis by weighting the marginal contribution, which has high feasibility and rationality. In the article, indicators that reflect the development differences between cities are used to decompose the contribution of all variables to the differences in multidimensional socio-economic development levels [40]. The Shapley value decomposition concept can be explained as follows:

$$Pi=\sum W\left(\left|S\right|\right)\left[v\left(S\right)-v\left(\frac{S}{I}\right)\right], \left(i=1,2,3,\cdots ,n\right)$$

(11)

$$W\left(\left|S\right|\right)=\frac{\left[\left(n-\left|S\right|\right)!\left(\left|S\right|-1!\right)\right]}{n!}$$

(12)

where $n$ represents the number of variables; $S$ is the possible set of all variables; $v$ is the corresponding benefit obtained in the cooperation; $v\left(\frac{S}{I}\right)$ denotes the benefit obtained by removing the variable i from the set S; $W$ is the weighting factor; and $Pi$ is the expected contribution of the $i$th member under cooperation.

3. Results

3.1. Spatial and Temporal Divergence Characteristics of Socio-Economic Development Level Index

Figure 5 shows the index calculated by using the multi-dimensional socio-economic development level model, and its range is from 0.00 to 0.67. It can be seen from the results, the overall socio-economic development level of resource-based cities improved from 2004 to 2019. In 2019, most regions reached 0.30 except for the western and northeastern regions. There are 16 cities with an index greater than 0.40, of which the top five performers are Xuzhou, Tangshan, Luoyang, Ordos and Linyi. Cities with lower indices are Jixi, Hegang, Yichun, Heihe, and Longnan, three of which belong to Heilongjiang Province. The index difference is 0.58 between the higher index city (Xuzhou) and the lowest index city (Longnan), indicating that the social and economic development level of the resource-based cities is quite different.

We calculated the statistics of cities of each grade in each year to obtain Figure 6. In 2004, 75% of cities with socio-economic development level index of 0.00–0.30, 5.22% of cities with socio-economic development level of 0.50–0.67, and by 2019 the cities with socio-economic development level index of 0.00–0.30 accounted for 57%, the proportion of cities with the index of 0.50–0.67 is 7.83%. This indicates that the social development level of resource-based cities has improved, but most cities have a low socio-economic development level index, the change is small and cities have greater development potential.

Figure 7 shows the trend analysis of socio-economic development level using linear regression analysis, all trend value of cities are greater than 0.05, which indicates that the socio-economic development level of different cities improved during 2004–2019, among which are Xuzhou, Tangshan, Linyi, Luoyang, Zibo, Nanyang, Jining, Xianyang, Handan, Baoji, Eerduosi, Baotou, Hengyang, Dongying, Taian, Ganzhou, and Jilin. The remaining cities are relatively slow in development. The 14 slowest cities are Huludao, Zhangye, Hechi, Jixi, Lijiang, Wuhai, Fuxin, Jinchang, Heihe, Qitaihe, Shuangyashan, Hegang, Yichun, and Karamay. From Figure 7, it can be seen that the trend changes of socio-economic development level among resource-based cities in 2004 to 2019 has obvious differences in spatial location, and the rapidly developing cities are concentrated in the central and southeastern regions, mainly in Jiangsu Province, Fujian Province, Henan Province, and Hunan Province, and there is a trend of patchwork in spatial distribution. The cities with declining development are basically in Heilongjiang, Jilin, and Liaoning provinces. The top five cities with better development in 2004–2019 have better economic bases themselves and are in the more developed provinces, which means that the gap between the less developed provinces of resource-based cities and the more developed provinces is gradually expanding. From an overall perspective, the level of urban socio-economic development is better after the national implementation of the resource-based sustainable development plan, but the spatial distribution of the socio-economic development level of resource-based cities shows large development differences and a more significant regional spatial aggregation feature.

3.2. Driving Mechanism of Socio-Economic Development Level in Resource-Based Cities

3.2.1. Selection of Influencing Factors

Based on the socio-economic impact of six aspects: industrialization level, development of service industry, urban area change, education level, consumption level, and transportation, the indicators of secondary industry share of GDP, tertiary industry share of GDP, retail trade goods sales, urban construction land area, total number of schools, and total freight transport were selected to measure their contribution to the socio-economic level. Resource-based cities are classified by planning into four categories: growing city, declining city, mature city, and regenerative city (2021). Since these four types of cities differ in terms of their resource profiles and influence mechanisms, this paper explores each of the four types of cities.

3.2.2. Determinant-Based Regression Analysis of Socio-economic Development Level Index

This paper quantitatively analyzes the marginal impact of the development of resource-based cities, and the results are shown in

Table 2. OLS regression analysis model results.

Variable	Regression Coefficient	t Value
Proportion of secondary industry to GDP	0.1713	2.73
Proportion of tertiary industry to GDP	0.0185	0.33
Retail trade commodity sales	0.2247	3.89
Urban construction land area	0.0185	0.37
Total number of schools	−0.0448	−0.97
Cargo volume	0.00088	0.03
Number of cities	115	115

. The regression correlation coefficients of secondary industry, tertiary industry, retail trade goods sales, urban construction land area and total freight transport for the socio-economic development of resource-based cities are positive, indicating that these five indicators play a facilitating role for socio-economic development, and the number of schools is negative for urban development. Figure 8 reflects the trend of regression correlation coefficients of influencing factors. The results show that the regression coefficient of socio-economic development level of mature cities is decreasing with the proportion of secondary industry and increasing with the proportion of tertiary industry, while the index of socio-economic development level of this type is increasing, which indicates that the development of tertiary industry is increasing its influence on the socio-economic level of mature cities. Regenerative cities have been largely freed from resource dependence, and the correlation coefficient between their socio-economic development level index and the share of secondary industry is on a decreasing trend, and an increasing trend with the share of tertiary industry during 2005 to2016. The correlation coefficient between the share of secondary industry and socio-economic development level is higher in growth-oriented cities, which is due to the fact that the economic growth model of growth-oriented cities is dominated by resource investment and maintains the economic structure dominated by secondary industry. The correlation coefficient between the number of schools and the socio-economic development level of growing cities shows a decreasing trend and changes from positive to negative, which indicates that the number of schools plays a negative role in the development of growing cities. The regression correlation between total freight transport and socio-economic development level index in declining and regenerating cities shows an overall increasing trend. The coefficient of secondary industry in declining cities tends to decline first and then increase. Due to the depletion of resources, the development of declining cities is hindered. At present, in the absence of transformation and upgrading, inevitably, fading cities will have issues including slow economic growth and pronounced social strife, but tapping the city’s own advantages and characteristics and developing non-resource-based industries are the driving force for sustainable development of declining cities.

3.2.3. Analysis of the Differences in Socio-Economic Development Levels between Cities

Based on the standard of the Gini coefficient released by the National Bureau of Statistics in 2021, the Gini coefficient lies between 0.30 and 0.40 indicating a more reasonable development, and it can be seen from Figure 9 that the Gini coefficient of the socio-economic development level of resource-based cities from 2004 to 2019 is more reasonable, but it does not reach a more average level, and there are significant differences in socio-economic development levels between cities. If the gap cannot be reduced in the socio-economic development level, it will have a negative impact on sustainable development.

3.2.4. Analysis of Factors Influencing Social Level Differences Based on Shapley Decomposition Model

Shapley decomposition model is used to quantify the contribution of determinants to the Gini coefficient, and the results are shown in Figure 10, with significant differences in the contribution of each factor. Among them, the contribution of trade goods sales and the increase in urban construction land area significantly increase the level of urban socio-economic development, and the cumulative contribution can reach more than 50.00%, so these two factors can explain most of the socio-economic development differences between cities. By calculating the annual average of factor contribution, the contribution of urban construction land area to the Gini coefficient of regenerative cities is 31.88%. Compared with other cities, the share of secondary industry has the lowest contribution of 10.94% to regenerative cities. From the trend of change, the contribution of retail trade goods sales to mature cities is greater than 27.00%, and the interannual change is stable. The contribution of urban construction land to growing cities gradually increased from 2015 to 30.43% in 2019. The contribution of secondary industry to GDP is increasing year by year, reaching 24.59% in 2019, and the cumulative contribution of the two indicators with urban construction land area is 55.02%, which indicates that secondary industry and urban construction promote the development of growing cities. The proportion of tertiary industry to GDP to the socio-economic development of declining cities is gradually increasing, from 5.75% in 2004 to 7.29% in 2019, but the annual average contribution of the secondary industry share is as high as 33.59%, the resources of declining cities are being depleted, and timely acceleration of industrial structure adjustment is the key to promote the development of this type of city.

4. Discussion

In China, cities driven by resources used to be the growth centers of regional economies. Resource-based industries have significantly changed the development path of urban development, making resource-based cities distinct. However, with the bankruptcy of a large number of state-owned enterprises, resource-based cities have declined. With the continuous deepening of reform and opening up, the development trajectory of China’s resource-based cities has clearly differentiated. From the perspective of Chinese resource-based cities a whole, there are obvious differences between resource-based cities and non-resource-based cities in terms of spatial structure. In this study, time series weighting factors are added to determine the index weights to study the use of the constructed multidimensional socio-economic development model based on time series to derive the development index of resource-based cities and quantitatively analyze their urban development evolution characteristics at key time points, in addition to using Shapley analysis to address the contribution of inter-city development level differences. The research shows that the social and economic development level of resource-based cities improved from 2004 to 2019, but only a few have reached the maximum level. The level of economic development in most cities is relatively low. In addition, Sun et al. also noted that from 2000 to 2008, the overall efficiency of resource-based cities in China showed a slight improvement trend [41]. The top 6 cities with faster urban development from 2004 to 2019 are regenerative cities, and the 14 cities with slower development are declining cities accounting for 43%, which indicates that the economic development trends of resource-based cities at different development stages are obviously different. Declining cities are in the primary stage of the development of resource-based cities, there has been some development but at a slow pace.

There are large differences in socio-economic development and spatial distribution among cities, which are consistent with the findings of some researchers [2,19,42]. The regions with relatively slow development show an aggregation trend in spatial distribution, mainly in the western and northeastern regions. Miao et al. [43] noted that the resource-based cities in northern China have a higher level of path dependence than those in the south, and the path dependence of these cities is mainly manifested in the excessive dependence on the mining industry. For resource-based cities, the economy is highly concentrated and dependent on natural resources, which will cause the city to lack buffer space due to the single industrial structure in the face of external shocks, thereby increasing the risk of recession [44,45]. In addition, through the analysis the data, the development of Jilin and Liaoning Province is better than that of Heilongjiang Province. The freight volume of Heilongjiang Province is much smaller than the remaining two provinces. The reason for this is not only related to its own economic base, but also related to the location in the region, where most of the slow economic development is in remote areas [18]. The study shows that the mining economy is in the stage of recession and regeneration development, and when its dominant position gradually declines, the degree of agglomeration of economic activities shows a more obvious positive correlation; therefore, the concentration of economic activities is a factor that can affect the development of two types of cities to a certain extent [2].

Although the types of resource-based cities are different, most cities encounter the same dilemma. However, the four types of cities have different resource-based industrial structures, and their impact mechanisms are also different. Shanxi [46] is a mature city, except for Shuozhou, whose main resource is coal. Before 2011, due to the unreasonable industrial structure, it was in trouble. Resource-based industries account for the vast majority of the secondary industry. The lock-in and crowding-out effect caused by this structure not only hinders the expansion of the manufacturing industry, but also inhibits the development of the primary and tertiary industries. Fan Jie et al. [47] noted that a single production activity is not conducive to the construction of urban innovation system and entrepreneurial environment in the analysis of China’s mining cities. Martin et al. noted that the local industrial structure (diversity level, enterprise scale, and enterprise innovation) and regional vulnerability is closely related to resilience [48,49]. The economic structure of resource-based cities has a relatively high proportion of resource exploitation and related industries. When external market needs change, the vulnerability of these industries may cause the vulnerability of the city’s overall economy, a single industrial structure also results in weaker cities. Through experiments to analyze the marginal impact, the impact of the secondary industry and economic development in Shanxi Province has shown a downward trend since 2008, and the tertiary industry has shown an upward trend, and the social and economic level has increased year by year. By analyzing the Shapley values of each city, the study shows that the higher the socio-economic development level index, the greater the contribution of the tertiary industry to development. For example, the relative contribution rate of the tertiary industry in Tangshan City is 17.39%, but the contribution of the tertiary industry in Lijiang City is only 8.35%. The contribution of retail trade goods sales in the lower socio-economic development level index is far higher than that of cities with a higher index, which indicates that the adjustment of industrial structure has a more significant impact on cities with a better economic foundation. Therefore, the government’s main measures to change the structure of a single industry are to reorganize the economy of resource-based cities and develop extended industries and alternative industries. For example, Houston, a city that used to rely on oil as its main resource, has achieved a successful transformation through the extension of the industrial chain and the development of machinery, cement, electricity, and transportation [50]. Regions such as Lorraine, Pittsburgh, and Ruhr have adopted the industrial substitution model, bringing in processing industries to transform them [50,51,52].

It can be seen that in order to improve the level of socio-economic development, the existing industrial structure and technological innovation of cities should be adjusted in a timely manner [53,54], and the restructuring of resource-based industries can be achieved by extending the industrial chain, cultivating alternative industries, and reforming traditional industries [55,56]. For growing and mature cities, the service industry should be vigorously promoted and encouraged; meanwhile, industrialization should also be promoted, the establishment of urban infrastructure should be accelerated, and local consumption patterns and education systems should be adjusted and changed to meet demand. For declining and regenerating cities, the development of innovative industries can be increased to reduce over-dependence on natural resources and implement industrial upgrading. In addition, in order to achieve sustainable economic growth, scholars have placed more emphasis on the role of technological innovation in promoting the industrial transformation of resource-based cities, and accelerated the development of the manufacturing industry by increasing technological innovation [57,58]. Therefore, the government should establish a sound long-term assistance program for the sustainable development of resource-based cities in slow-developing districts, which can work by introducing new projects and increasing investment in infrastructure, especially transportation infrastructure. Rather than just relying on the development of tertiary industries, it is recommended that new industries combine clean production, comprehensive utilization of resources, ecological design, and sustainable consumption patterns [5]. Finally, the implementation of policies should be adjusted according to the differences in urban development, addressing key local issues according to the development differences of different types of resource-based cities, encouraging energy-saving and environmental protection industries, and supporting the development of high-tech industries. In addition, the results of this study note that the number of schools has a negative impact on the development of growing cities, and the negative impact is greater as the urban social level index increases, so more public spending should be used for cities with lower indices, especially on vocational training, improving the quality and capacity of the local population, and changing consumption patterns.

5. Conclusions

For 115 resource-based cities in China, this study constructs the time series-based multidimensional socio-economic change model, and explores the mechanisms driving the evolution of the spatio-temporal patterns of multidimensional socio-economic development levels from 2004 to 2019, drawing the following conclusions: Firstly, the overall socio-economic development level of resource-based cities improved from 2004 to 2019, with trend values greater than 0.05, but is low overall. The spatial distribution varies widely, and the regional spatial aggregation characteristics are more significant. The fast-growing cities are concentrated in the central and southeastern regions, the slowly developing regions are in the northeast. This indicates that the spatial distribution varies widely, and the regional spatial aggregation characteristics are more significant. Secondly, secondary industry, tertiary industry, retail trade sales, urban construction land area, and overall freight transport are benefits for the socio-economic development of resource-based cities, and the number of schools has a negative impact on the development of growing cities, and its negative impact increases with the increase in the urban social level index. The contribution of retail trade goods sales to the Gini coefficient is the largest, and the contribution of secondary industry to regenerating cities is the lowest (10.94%). The tertiary industry has a positive effect on the socio-economic development level of resource-based cities, and the impact is more significant for cities with a better economic base. This shows that the implementation of policies should be carried out according to the type of urban development. The experience of this paper provides important theoretical and practical implications for exploring the influence mechanism of resource-based city development and for further research. However, the multidimensional socio-economic level model constructed in this paper has different indicator weights for each city according to the research data, which may produce bias in the evaluation process, and future work should select representative cities for empirical testing.