Environmental Decentralization, Resource Endowment and Urban Industrial Transformation and Upgrading: A Comparison of Resource-Based and Non-Resource-Based Cities in China

Abstract

Based on panel data of resource-based and non-resource-based cities in China from 2011 to 2021, we empirically investigate the effects of environmental decentralization and resource endowment on urban industrial transformation and upgrading using a fixed-effects model and a threshold model. It is found that environmental decentralization promotes industrial transformation and upgrading in both types of city in China. However, the combined effect of resource endowment and environmental decentralization inhibits industrial transformation and upgrading in resource-based cities, while promoting it in non-resource-based cities. In addition, the influence of environmental decentralization and resource endowment on industrial transformation and upgrading varies depends on the degree of implementation. Therefore, it is imperative to continuously improve the environmental decentralization management system, scientifically utilize resource endowment and promote industrial transformation and upgrading based on the unique characteristics of resource-based and non-resource-based cities. Lastly, we should focus on the transformation and upgrading of traditional resource-based industries and strengthening the green innovation of new industries, aiming to achieve a win–win situation for ecological environmental protection and economic development.

1. Introduction

Resource-based cities are a special type of industrial city that emerge from local natural resources, and they account for about 42.7% of all types of city in China in 2021 [1]. In the early stages of urban development, more resource-based cities turn the advantages of natural resources into productivity and achieve a rapid development of their industrial economy by virtue of their unique resource endowment. However, as resources are exploited with excessive intensity, some resource-based cities face problems, such as resource depletion, ecological environment degradation, economic structure imbalance and industrial upgrading obstruction, falling into the “resource curse” economic dilemma [2,3,4]. Environmental decentralization is a special type of environmental management system that aims to achieve an optimal distribution of environmental management power among different levels of government. Its goal is to improve the efficiency of environmental management, improve the quality of the regional ecological environment and promote the industrial transformation and upgrading of resource-based cities by stimulating the “innovation compensation effect” of enterprises. This can lead to a win–win situation between ecological protection and economic development, and help to solve the dilemma of sustainable development in resource-based cities [5,6].

China’s resources are rich and widely distributed, and include coal, petroleum, iron ore, copper ore and non-metallic mineral resources (such as lead, limestone and granite). Cities such as Bijie City in Guizhou Province, Huainan City in Anhui Province, Huzhou City in Zhejiang Province, Jingdezhen City in Jiangxi Province, Xuzhou City in Jiangsu Province, Zibo City in Shandong Province and Shaoguan City in Guangdong Province are typical resource-based cities, and these resource-based cities are important strategic bases of energy resources for China’s industrial economic construction and high-quality development. At the same time, China ascribes great importance to ecological construction and actively participates in global environmental governance practices, having introduced many environmental policies and regulations, such as the Reform Program of the Ecological and Environmental Damage Compensation System, the Environmental Impact Assessment Law, the Cleaner Production Promotion Law and the Rules for the Preparation and Revision of National Ecological and Environmental Standards. These environmental policies and regulations strictly limit the emission of pollutants in accordance with environmental protection goals and national environmental quality standards. The government imposes penalties for environmental pollution and ecological damage violations, regulates through environmental taxation, and strengthens environmental monitoring and evaluation. In addition, they actively promote the development of a green economy and encourage enterprises to take measures to reduce emissions. In this way, the energy structures have been adjusted and production methods have been improved. The results have been effective. In 2011, China emitted 22.18 million tons of sulfur dioxide, 23.09 billion tons of industrial wastewater and 56.8% of industrial solid waste. By 2021, China’s sulfur dioxide emissions will have decreased by 87.61% compared to 2011, industrial wastewater emissions will have decreased by 15.74% compared to 2011 and the comprehensive utilization rate of industrial solid waste will have increased by 16% compared to 2011. The average annual growth rate of China’s green industry output since 2011 is about 10%, and exceeded CNY 8 trillion in 2021. The share of clean energy consumption increased from 14.5% in 2011 to 25.5% in 2021. The proportion of coal consumption decreased from 68.5% in 2011 to 56.0% in 2021. Energy consumption per unit of GDP decreased by 26.4% from 2011 to 2021. Based on the comparison of resource-based and non-resource-based cities, this paper explores the mechanisms of the influence of environmental decentralization on industrial transformation and upgrading, as well as the combined effect of resource endowment and environmental decentralization on these cities. The heterogeneity of the effects of different degrees of environmental decentralization and resource endowment on industrial transformation and upgrading was also analyzed in order to explore the effect of environmental decentralization on the industrial transformation and upgrading of resource-based cities and to encourage it to develop in the correct direction. This provides an innovative approach to the dilemma of industrial transformation and upgrading in resource-based cities.

2. Literature Review

2.1. The Impact of Environmental Decentralization on Industrial Transformation and Upgrading

The impact of environmental decentralization on industrial transformation and upgrading has been widely studied in recent years. Some scholars believe that the implementation of environmental decentralization is not conducive to industrial transformation and upgrading. Yang et al. [7] argued that environmental decentralization is the division of environmental affairs based on decentralization, and the greater the local government’s power to manage environmental affairs, the more likely it is to cause a low performance in environmental governance, which in turn inhibits industrial transformation and upgrading [8,9]. Gao et al. [10] argued that the decentralization of power tends to lead to a vicious circle of “pollution for growth” among local governments, and this vicious circle will inhibit industrial transformation and upgrading [11]. Michael et al. [12] pointed out that sacrificing the environment for short-term economic growth will not only lead to the deterioration of urban environmental quality, but also hinder industrial transformation and upgrading [13]. Simon et al. [14] found that the implementation of moderate environmental decentralization can reduce environmental pollution and improve the efficiency of resource utilization, thus contributing to the promotion of industrial transformation and upgrading [15,16]. Abhishek et al. [17] pointed out that technological innovation is an important driving force for urban industrial development, and the implementation of moderate environmental decentralization can force enterprises to innovate technology and thus promote industrial transformation and upgrading [18,19].

2.2. The Impact of Environmental Decentralization on Resource Endowment

Jurgelevicius et al. [20] found that the overdependence of industrial economic development on resource endowments can lead to environmental problems and thus fall into the “resource curse” trap [21]. Syed et al. [22] argued that the “resource curse” is long standing and has a close correlation with environmental sustainability [23]. David et al. [24] pointed out that environmental decentralization can help local governments to implement environmental regulations in a reasonable way, thus reducing the dependence of industrial development on resources and helping to reduce the “resource curse” [25,26]. James et al. [27] suggested that resource dependence can be broken by strengthening technological innovation, and it also contributes to improving environmental quality [28,29]. Yuldashev et al. [30] argued that the implementation of environmental decentralization is conducive to increasing foreign direct investment, thus introducing clean production technology and advanced management experience from developed countries, which helps to accelerate green transformation [31].

2.3. The Impact of Resource Endowment on Industrial Transformation and Upgrading

Jack et al. [32] found that natural resources are “divine food” and an indispensable element for industrial transformation and upgrading, and cities with abundant natural resources can provide a material basis for industrial transformation and upgrading by virtue of their resource endowment [33]. Iman et al. [34] argued that resource-based industries receive limited financial support from local governments, while clean industries receive more financial support. Therefore, production factors flow to clean industries, thus promoting industrial transformation and upgrading [35]. However, some scholars believe that resource endowment inhibits industrial transformation and upgrading. Fatima et al. [36] suggested that the development of resource-based industries tends to form a “path dependence” on resource endowment, which inhibits industrial transformation and upgrading. However, the path lock-in of resource dependence can be broken by continuously strengthening technological innovation. Peng et al. [37] argued that local governments in cities with abundant natural resources prefer to rely on them to develop resource-based industries, and the over-exploitation of natural resources is detrimental to promoting industrial transformation and upgrading [38,39]. Anser et al. [40] argued that the excessive reliance on natural resource endowments for the development of resource-based industries is likely to lead to environmental pollution problems, which is detrimental to the promotion of industrial transformation and upgrading, and that the implementation of appropriate environmental decentralization is conducive to breaking this problem [41].

In summary, existing studies have mainly focused on environmental decentralization and environmental pollution, resource-based industrial development and the “resource curse”. However, there is a lack of research that comprehensively explores the relationship between environmental decentralization, resource endowment, and industrial transformation and upgrading. Additionally, there is a dearth of comparative studies between resource-based and non-resource-based cities. Based on panel data from the two types of city in China from 2011 to 2021, this paper explains the mechanisms of the influence of environmental decentralization on industrial transformation and upgrading, as well as the combined effect of resource endowment and environmental decentralization on these cities. It empirically investigates the effects using fixed-effects models and threshold models. The heterogeneity of the effects of different degrees of environmental decentralization and resource endowment on industrial transformation and upgrading in resource-based and non-resource-based cities is also explored.

3. Theoretical Mechanism Analysis and Research Hypothesis

Environmental decentralization is based on the division of power and responsibility between central and local governments in environmental management matters, seeking the optimal allocation of environmental management power between government levels. Compared with the central government, local governments have a better understanding of the pollution levels caused by polluting enterprises and residents’ choice preferences in their jurisdictions. Therefore, they can implement environmental regulations according to the local conditions, which will improve the efficiency of environmental management. It will promote industrial transformation and upgrading by both increasing investment in environmental management and encouraging enterprises to embrace technological innovation. However, there are two views on the implementation of environmental decentralization. One is the “compliance cost theory”, based on the traditional school of neoclassical economics, which analyzes the implementation of environmental management policies at the static level that increase the cost of pollution treatment for enterprises and thus inhibit industrial transformation and upgrading [42,43,44]. In other words, with the increase in environmental decentralization, strict environmental regulations by the government will lead to the internalization of external costs. It will increase the production burden of enterprises and cut their profits, and will thus reduce their investment in production and technological innovation. Consequently, it is not conducive to the promotion of industrial transformation and upgrading. Another view is the dynamic level of “innovation compensation” [45,46]. As the degree of environmental decentralization increases, strict environmental regulations imposed by the government will increase the production costs of enterprises. However, through the improvement of production technology and equipment and processes, the production efficiency and clean production level of enterprises will be significantly improved, and the marginal cost of pollution control will decrease. In the long run, the innovation compensation benefits obtained by enterprises will considerably exceed the cost of following environmental regulations, and the enterprises will be transformed and upgraded. The transformation and upgrading of enterprises will lead to the transformation and upgrading of the overall industry, and then promote the industrial transformation and upgrading of resource-based and non-resource-based cities, as shown in Figure 1. Based on this, Hypothesis 1 is proposed:

Hypothesis 1.

Environmental decentralization promotes industrial transformation and upgrading in both resource-based and non-resource-based cities.

Natural resources are an important material basis for economic construction and social development. Since the exploitation of natural resources can release a considerable “resource dividend” in the short term [47,48], resource-based cities with rich natural resources will focus on resource-based industries in the early stage of urban development. However, considering that resource-based industries are the leading industries in this case, the industrial structure of resource-based cities will gradually become unbalanced. Furthermore, the industrial development of resource-based cities will lead to “path lock-in” of excessive reliance on natural resources [49,50,51], which seriously restricts the industrial development of cities. Due to the pressure of economic growth, resource-based cities will be forced to choose to relax environmental regulations and lower the entry threshold for enterprises, which can attract polluting enterprises with high energy consumption and high emissions. This is because they can create a high economic output in the short term. However, this will cause a series of serious environmental pollution and resource wastage problems, which will lead to an increase in the cost of pollution control in resource-based industries. It will also trigger a vicious circle of a “race to the bottom” among local governments, thus inhibiting the industrial transformation and upgrading of resource-based cities. In this way, resource-based cities can fall into the “resource curse” trap. Since resource-based industries occupy a large proportion in the industrial structure of resource-based cities, the development space of non-resource-based industries is crowded out and suppressed, resulting in an imbalance in the industrial structure of resource-based cities. Furthermore, the development of resource-based industries is overly reliable on resources, which leads to a serious reduction in resource allocation efficiency and also hinders the improvement of labor productivity, industrial competitiveness and the added value of industrial technology, thus making economic efficiency and profit seriously affected. Additionally, the development of a resource-based economy is prone to considerable waste and loss of resources due to the over-intensive exploitation and unreasonable use of resources. Resource-based industries are characterized by “high pollution, high energy consumption and high emissions”, causing serious problems of pollution and damage to the environment, thus easily creating the vicious circle of a “resource–environment–economy” system, as shown in Figure 2. Based on this, Hypothesis 2 is proposed:

Hypothesis 2.

The combined effect of resource endowment and environmental decentralization inhibits industrial transformation and upgrading in resource-based cities.

Compared with resource-based cities, the leading industries in non-resource-based cities are not resource-based, so they do not have an excessive reliance on natural resources, which can improve the efficiency of resource allocation and economic development. At the same time, the clean production capacity, technological innovation level and the ecological environment of non-resource-based cities are better than those of resource-based cities. Therefore, the implementation of moderate environmental decentralization in non-resource-based cities is more conducive to stimulating the vitality of the green technological innovation of market players. Therefore, they can obtain more innovation compensation, continuously improve the efficiency of resource utilization and clean production level, and reduce the marginal pollution control cost, thus maximizing industrial transformation and upgrading, as shown in Figure 3. Based on this, Hypothesis 3 is proposed:

Hypothesis 3.

The combined effect of resource endowment and environmental decentralization promotes industrial transformation and upgrading in non-resource-based cities.

4. Construction of the Measurement Model and Selection of Indicators

4.1. Econometric Model Construction

For the purpose of empirical analysis, the following econometric model was constructed:

$${\mathrm{ITU}}_{it}={\alpha }_0+{\alpha }_{1}E{D}_{it}+{\alpha }_{2}R{E}_{it}+\beta X_{it}+{\mu }_i+v_t+{\epsilon }_{it}$$

(1)

where i represents each resource-based and non-resource-based city; t represents the year; the explained variable ITU_itis the industrial transformation and upgrading index; the explanatory variable ED_it is the environmental decentralization degree; RE_it is the resource endowment; X_it is a set of control variables; ${\epsilon }_{it}$ is a random disturbance term satisfying the independent homogeneous distribution with finite variance; and ${\mu }_{\mathrm{i}}$ and ${\nu }_{\mathrm{t}}$ are individual fixed effects and time fixed effects, respectively.

4.2. Selection of Indicators

4.2.1. Explained Variables

Industrial transformation and upgrading (ITU): Industrial transformation and upgrading is concerned with the coordination of two dimensions: industrial structure rationalization and industrial structure advancement. Industrial structure rationalization reflects both the degree of inter-industry coordination and the degree of effective resource utilization. Industrial structure advancement implies the evolution of an industrial structure from primary to secondary and then to tertiary industries, which is based on industrial structure rationalization. In this paper, we constructed an RIS index to measure the rationalization of industrial structures, drawing on the approach of Satoshi [52]. We used the ratio of the output value of a tertiary industry to that of a secondary industry to measure the advancement in industrial structures, and used the entropy value method to synthesize the RIS index and the ratio of the output value of a tertiary industry to that of a secondary industry to derive the index of industrial transformation and upgrading. The formula of the RIS index is:

$$RI{S}_{it}={\displaystyle \sum _{k=1}^n(\frac{Y_{kt}}{Y})}\ln (\frac{Y_{kt}}{L_{kt}}/\frac{Y}{L})={\displaystyle \sum _{k=1}^n(\frac{Y_{kt}}{Y})}\ln (\frac{Y_{kt}}{Y}/\frac{L_{kt}}{L})$$

(2)

where k denotes the kth industry; n is the number of industrial sectors; Y_kt and L_kt denote the output value and employment of industry i in year t in two types of city; and Y_kt/L_kt denotes the productivity level.

4.2.2. Explanatory Variables

(1) Environmental decentralization (ED): In this article, drawing on the approach of Zhang et al. [53], the distribution of employed persons in urban units in the environment and public facilities management industry in each city was used to measure the degree of environmental decentralization. In addition, considering the endogeneity of environmental decentralization and economic development, the economic deflation factor (1 − (GDP_it/GDP_t)) was introduced according to the work of Ran et al. [54]. The environmental decentralization index was calculated as follows:

$$E{D}_{it}=\frac{L{E}_{it}/L{P}_{\mathrm{i}t}}{G{E}_t/G{P}_t}[1-(GD{P}_{it}/GD{P}_{\mathrm{t}})$$

(3)

where ED_it denotes the environmental decentralization index of each resource-based and non-resource-based city; LE_it denotes the number of urban units employed in the environment and public facilities management industry in year t in city i; LP_it denotes the total population at the end of year t in city i; GE_t denotes the number of urban units employed in the environment and public facilities management industry in year t; and GP_t denotes the total population at the end of year t.

(2) Resource endowment (RE): Resource endowment plays a crucial role in industrial transformation and upgrading, and most scholars use resource-based industry-related indicators to measure resource endowment. The larger the proportion of GDP accounted for by the output value of resource-based industries in a city, the more it reflects the dependence of the city’s industrial transformation and upgrading on resource endowment. Therefore, this article used the method of Eskander et al. [55], adopting the proportion of the output value of resource-based industries to GDP to measure resource endowment.

4.2.3. Other Control Variables

(1) Government intervention (GI): Drawing on the work of Yi et al. [56], the share of the total industrial output value of state-owned and state-controlled enterprises in the total industrial output value above the scale was used as a measure. (2) Resident consumption level (HC): Based on the work of Bao et al. [57], the total retail sales of social consumer goods per capita was selected as a proxy variable for resident consumption level, and the price deflator was used to adjust the total retail sales of social consumer goods before calculating the per capita result. (3) Regional openness (RO): Based on the work of Orji et al. [58], the actual use of foreign direct investment in each region was used as a measure. (4) Technological innovation capability (TI): Based on the work of Shi et al. [59], the full-time equivalent of R&D personnel was selected as a proxy variable for technological innovation capability. (5) Social investment intensity (SI): Drawing on the work of He et al. [60], we measured social investment intensity by using the social fixed-asset investment in each region.

4.3. Data Sources

This paper selected 110 resource-based cities and 145 non-resource-based cities from 2011 to 2021 as the study sample based on the list of resource-based and non-resource-based cities in the Sustainable Development Plan for Resource-based Cities in China. Prefecture-level cities with important missing data were excluded. The division of the data into resource-based and non-resource-based cities was conducted to compare the influence of environmental decentralization on the industrial transformation and upgrading of the two types of city, as well as the combined effect of resource endowment and environmental decentralization on these cities. China started to implement environmental decentralization management systems in 2011, such as the River Chief System, Lake Chief System, Forest Chief System, and Field Chief System, after which the statistical caliber of environmental decentralization statistics in the China Environmental Statistics Yearbook changed. The indicators selected in this paper were analyzed for the years 2011–2021 in order to maintain the continuity of data and also to facilitate regression analysis. The data selected above were obtained from the China Urban Statistical Yearbook, China Environmental Statistical Yearbook, China Industrial Economic Statistical Yearbook and Statistical Yearbook of Scientific and Technological Activities of Industrial Enterprises, and relevant data released by the websites of provincial and municipal government departments. We also used linear interpolation to supplement some missing data. All variables were logarithmized to eliminate the interference of heteroskedasticity and dimensionality problems. The descriptive statistics of the variables is shown in

Table 1. Descriptive statistics of the variables.

City Category	Variable Names	Abbreviations Used	Measurement	Data Sources	Observations	Average	Standard Deviation	Minimum	Maximum
Resource-based cities	Industrial transformation and upgrading	ITU	Three-product method with RIS index-weighted synthesis	China Industrial Economic Statistical Yearbook	1100	−1.610	1.071	−6.907	0.747
	Environmental decentralization	ED	$\begin{array}{l}E{\mathrm{D}}_{\mathrm{it}}=\frac{L{E}_{it}/L{P}_{\mathrm{i}t}}{G{E}_t/G{P}_t}\\ [1-(GD{P}_{it}/GD{P}_{\mathrm{t}})\end{array}$	China Environmental Statistical Yearbook	1100	−0.161	0.615	−3.290	8.036
	Resource endowment	RE	Output value of resource-based industries/GDP	China Urban Statistical Yearbook	1100	0.211	0.549	−3.399	2.871
	Government intervention	GI	Total industrial output value of state-owned and state-controlled enterprises/Total industrial output value above the scale	China Urban Statistical Yearbook	1100	−1.646	0.433	−3.126	−0.193
	Resident consumption level	HC	Total retail sales of social consumer goods per capita	China Urban Statistical Yearbook	1100	1.114	0.595	0.540	1.907
	Regional openness	RO	Actual FDI used	China Urban Statistical Yearbook	1100	−6.801	1.486	−13.20	−1.942
	Technological innovation capability	TI	Full-time equivalent of R&D staff	Statistical Yearbook of Scientific and Technological Activities of Industrial Enterprises	1100	0.024	0.083	0.002	0.095
	Social investment intensity	SI	Social fixed-asset investment by region	China Urban Statistical Yearbook	1100	15.97	0.900	9.284	17.97
Non-resource-based cities	Industrial transformation and upgrading	ITU	Three-product method with RIS index-weighted synthesis	China Industrial Economic Statistical Yearbook	1450	1.525	0.459	−1.281	5.766
	Environmental decentralization	ED	$\begin{array}{l}E{\mathrm{D}}_{\mathrm{it}}=\frac{L{E}_{it}/L{P}_{\mathrm{i}t}}{G{E}_t/G{P}_t}\\ [1-(GD{P}_{it}/GD{P}_{\mathrm{t}})\end{array}$	China Environmental Statistical Yearbook	1450	−0.126	0.799	−2.287	8.970
	Resource endowment	RE	Output value of resource-based industries/GDP	China Urban Statistical Yearbook	1450	0.283	0.493	−2.657	1.452
	Government intervention	GI	Total industrial output value of state-owned and state-controlled enterprises/Total industrial output value above the scale	China Urban Statistical Yearbook	1450	−1.813	0.379	−4.176	0.396
	Resident consumption level	HC	Total retail sales of social consumer goods per capita	China Urban Statistical Yearbook	1450	1.135	0.815	0.542	2.281
	Regional openness	RO	Actual FDI used	China Urban Statistical Yearbook	1450	−6.405	1.228	−13.33	−3.510
	Technological innovation capability	TI	Full-time equivalent of R&D staff	Statistical Yearbook of Scientific and Technological Activities of Industrial Enterprises	1450	0.649	0.287	0.267	0.548
	Social investment intensity	SI	Social fixed-asset investment by region	China Urban Statistical Yearbook	1450	16.43	0.840	14.02	18.65

.

Furthermore, a Variance Inflation Factor (VIF) test was conducted to assess the multicollinearity among the independent variables. The results reveal that the VIF values of all variables were below 10, with the highest VIF values for RE (2.85) in resource-based cities and GI (3.19) in non-resource-based cities. This suggests the independence assumption was satisfied and that there was no severe multicollinearity present. Additionally, Cronbach’s alpha was employed to evaluate the reliability of the indicators. The reliability coefficients of all indicators surpassed 0.800, thus indicating a high reliability. Meanwhile, the KMO values of all variables ranged between 0.6 and 0.7, and the approximate chi-squared value of the Bartlett’s sphericity test was significant at the 0.001 level. As a result, the indicators selected in this article had high validity.

5. Empirical Analysis

5.1. Baseline Regression Results

In order to analyze the influence of environmental decentralization on the industrial transformation and upgrading of resource-based cities and non-resource-based cities, as well as the combined effect of resource endowment and environmental decentralization on these cities, a benchmark regression analysis was first conducted and the results are shown in

Table 2. Baseline regression results.

Variables	Resource-Based Cities		Non-Resource-Based Cities
ED	0.089 *	0.541 ***	0.078 ***	0.886 ***
	(1.90)	(3.19)	(2.83)	(2.72)
RE	−1.365 ***	−1.119 ***	1.164 ***	1.428 ***
	(−3.85)	(−8.41)	(7.35)	(5.25)
ED × RE	-	−0.578 *	-	0.514 ***
	-	(−1.67)	-	(5.74)
GI	−0.458 ***	−0.287 **	0.177 **	2.051 ***
	(−3.63)	(−2.51)	(2.29)	(5.47)
HC	0.034 **	0.142 ***	0.143 ***	0.195 ***
	(2.02)	(3.04)	(3.25)	(3.13)
RO	−0.167 **	0.087	0.057 **	0.254 **
	(−2.34)	(1.58)	(2.20)	(2.06)
TI	0.006 *	0.077 ***	0.079 ***	0.153 **
	(1.62)	(3.58)	(3.72)	(2.13)
SI	0.240 ***	0.243 ***	0.120 ***	0.011 *
	(5.80)	(6.99)	(6.05)	(1.76)
_cons	−2.111 ***	−2.145 ***	−2.249 ***	−0.004 *
	(−2.97)	(−3.60)	(−6.15)	(−1.73)
F-value	17.03	23.31	21.31	18.61
R-squared	0.576	0.706	0.960	0.633

Note: ***, **, and * indicate significance at the 1%, 5% and 10% levels, respectively, and t-statistics for the parameter estimates are in parentheses.

.

According to Table 2, as shown by the Hausman test results, all the p-values are lower than 0.1; therefore, the fixed-effect model was selected for empirical testing. The regression results show that the effect of ED on ITU are significantly positive in both resource-based and non-resource-based cities, indicating that environmental decentralization plays a role in promoting industrial transformation and upgrading in both types of city. The implementation of environmental decentralization can form an intrinsic incentive mechanism to drive industrial transformation and upgrading by imposing resource and environmental constraints on market players. The transformation and upgrading of enterprises will promote the restructuring of the industry and the optimal allocation of resources, which will promote the industrial transformation and upgrading of the two types of city. Therefore, Hypothesis 1 is verified. The regression results of RE on ITU for resource-based cities are significantly negative, while those for non-resource-based cities are significantly positive. According to the regression results of the interaction term between ED and RE, every 1% increase in the interaction term between ED and RE increases the inhibitory effect on industrial transformation and upgrading in resource-based cities by 0.578%, while the promotional effect on industrial transformation and upgrading in non-resource-based cities increases by 0.514%. This is because resource-based cities are prone to the considerable waste and loss of resources due to the excessive and unreasonable exploitation of natural resources. Resource-based industries also have the characteristics of “high pollution, high energy consumption and high emission”, which cause serious pollution and damage to the ecological environment. Their development is overly dependent on resources, forming a “path lock-in” and seriously inhibiting the transformation and upgrading of resource-based industries. Compared with resource-based cities, non-resource-based cities have better environmental conditions and industrial bases. Since these cities have a relatively high level of green technology innovation and intensive resource utilization, they are more conducive to promoting industrial transformation and upgrading.

Further examining the regression results of the control variables, the regression results of GI in resource-based cities are significantly negative, while those in non-resource-based cities are significantly positive. It is due to the path dependence of industrial development in resource-based cities, which distorts the resource allocation mechanism to a certain extent. Government intervention in non-resource-based cities is more focused on changing the economic development mode and boosting industrial transformation and upgrading. The regression results of HC in both resource-based and non-resource-based cities are significantly positive, but the consumption level of residents in resource-based cities is lower than that in non-resource-based cities. The regression results of RO were significantly negative for resource-based cities and significantly positive for non-resource-based cities, which is due to the opening up of resource-based cities to the outside world, causing polluting enterprises to move in, exacerbating regional environmental pollution and hindering industrial transformation and upgrading. The opening up of non-resource-based cities to the outside world will improve regional greening development by introducing advanced technology and management experience, thus promoting industrial transformation and upgrading. From the regression results of TI, the technological innovation capacity of non-resource-based cities is significantly higher than that of resource-based cities. The regression results of SI are significantly positive for both types of city, which is due to the fact that expanding the scale of investment can integrate the required funds for environmental management and technological innovation.

5.2. Robustness Test

To further enhance the robustness of the estimation results of the econometric model, this article used the replacement of explanatory variables, the instrumental variable method and GMMs to test the robustness of the sample data. The results are shown in

Table 3. Robustness test analysis.

Variables	Resource-Based Cities				Non-Resource-Based Cities
	Replacement of Explanatory Variable	Instrumental Variable Method	SYS-GMM	DIF-GMM	Replacement of Explanatory Variable	Instrumental Variable Method	SYS-GMM	DIF-GMM
ED	0.166 *	0.603 **	0.277 ***	0.310 *	0.320 *	0.309 **	0.279 ***	0.210 **
	(1.81)	(2.42)	(2.71)	(1.75)	(1.88)	(2.22)	(3.07)	(2.35)
RE	−1.271 ***	−1.293 ***	−0.567 ***	−0.634 ***	0.605 ***	0.662 ***	1.053 ***	1.304 ***
	(−9.10)	(−4.72)	(−5.02)	(−5.24)	(4.36)	(6.03)	(3.37)	(3.26)
ED × RE	0.198 **	0.581 **	0.313 ***	0.305 **	0.301 **	0.240 *	0.256 ***	0.241 ***
	(2.41)	(2.47)	(2.78)	(2.00)	(2.07)	(1.81)	(3.10)	(3.09)
Add control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
_cons	−3.027 ***	0.325 *	1.112 *	0.723 **	−3.495 ***	−4.024 ***	1.186 *	1.033 **
	(−4.84)	(1.74)	(1.89)	(2.11)	(−3.36)	(−2.90)	(1.88)	(2.36)
F-value	71.81	96.95	92.47	85.52	86.50	69.99	97.94	79.29
R-squared	0.825	0.950	-	-	0.915	0.904	-	-
AR (1) value	-	-	0.003	0.000	-	-	0.008	0.065
AR (2) value	-	-	0.793	0.481	-	-	0.185	0.697
Sargan value	-	-	0.996	0.975	-	-	0.215	0.112

Note: ***, **, and * indicate significant at the 1%, 5% and 10% levels, respectively, and t-statistics for parameter estimates are in parentheses.

.

5.2.1. Replacement of the Explanatory Variables

In this paper, the “industrial structure hierarchy coefficient” proposed by Chang et al. [61] was used as an additional proxy variable to measure industrial transformation and upgrading. The formula for calculating the industrial structure hierarchy coefficient $S_{it}$ is as follows:

$$S_{it}={\displaystyle \sum _{j=1}^n{\beta }_{jt}{q}_{jt}}$$

(4)

where $q_{jt}$ represents the share of the output value of industry j in the GDP of the region; ${\beta }_{jt}$ represents the weights of primary, secondary and tertiary industries; i represents each resource-based and non-resource-based city; and t represents the year.

The regression results for the two types of city after replacing the explained variables did not change significantly from the baseline regression results in the previous section; therefore, the model can be considered to be set up and estimated with strong robustness.

5.2.2. Instrumental Variable Method

To alleviate the endogeneity problem, this article adopted the 2SLS method and selected the average level of environmental decentralization in neighboring regions as the instrumental variable of environmental decentralization in this region [62]. The estimation results of the instrumental variables show that the F-statistics for both the two types of city are above the critical value of 10, indicating that there is no weak instrumental variable problem. In addition, the regression results of the endogeneity test for each variable remain largely consistent with the baseline regression results in the previous section. Therefore, the model can be considered to have strong robustness in its setting and estimation results.

5.2.3. GMM Models

Considering the possible heteroskedasticity of the randomly perturbed terms and other problems, the GMM estimation method was further used to construct a dynamic panel model to test the robustness of the sample data. In this paper, the lagged term of the interaction term between environmental decentralization and resource endowment was used as an instrumental variable and tested using two estimation methods: systematic GMM and differential GMM. According to the estimation results, the AR (1) values of the regression results of both the two types of city are less than 0.1. The AR (2) values are greater than 0.1, and the Sargan values are greater than 0.1, which indicates that the instrumental variables were selected effectively and there was no over-identification problem. Meanwhile, the regression results of each variable of the GMM test remained basically consistent with the regression results of the fixed-effects estimation; therefore, the model setting and estimation results can be considered to have strong robustness.

5.3. Analysis of the Threshold Effect

The influence of environmental decentralization on the industrial transformation and upgrading of resource-based cities and non-resource-based cities, as well as the combined effect of resource endowment and environmental decentralization on these cities, may vary depending on the different degrees of environmental decentralization and resource endowment. Therefore, we used environmental decentralization and resource endowment as the threshold variables and adopted a threshold regression model to analyze the nonlinear influence.

5.3.1. Threshold Effect Analysis with Environmental Decentralization as the Threshold Variable

According to the principle of Hansen’s threshold effect test, the bootstrap self-sampling method was used to repeatedly sample 300 times to test whether there was a threshold effect on the sample data. The test results are shown in

Table 4. Results of the threshold effect test with environmental decentralization as the threshold variable.

City Category	Number of Thresholds	F-Value	p-Value	10% Critical Value Level	5% Critical Value Level	1% Critical Value Level
Resource-based Cities	Single Threshold	15.97 **	0.037	12.317	15.411	23.798
	Double Threshold	0.460	0.997	11.774	14.346	20.710
	Three-fold threshold	3.340	0.603	9.240	10.602	17.449
Non-resource-based cities	Single Threshold	9.310 *	0.087	8.515	10.280	15.165
	Double Threshold	3.040	0.530	10.616	12.934	19.723
	Three-fold threshold	3.100	0.807	17.886	21.404	31.597

Note: ** and * indicate significant at the 5% and 10% levels, respectively.

.

It can be seen that, when environmental decentralization is the threshold variable, the single threshold F-value is significant at the 10% significance level for both resource-based and non-resource-based cities. However, none of the other threshold tests are significant, thus indicating a single threshold effect.

In addition, the threshold values of resource-based and non-resource-based cities were estimated by using the great likelihood ratio estimation method. The threshold value of ${\eta }_1$ was estimated as −3.785, and ${\eta }_2$ was estimated as −0.738. In order to render the estimation results visually, the great likelihood ratio functions of the estimated results of the threshold variable values ${\eta }_1$ and ${\eta }_2$ were also plotted (see Figure 4 and Figure 5).

According to the panel threshold regression results in

Table 5. Regression results of the threshold effects with environmental decentralization as the threshold variable.

Variables	Resource-Based Cities	Variables	Non-Resource-Based Cities
ED-I (ED ≤ −3.785)	0.337 ***	ED-I (ED ≤ −0.738)	0.108 ***
	(4.72)		(3.01)
ED-I (ED > −3.785)	0.104 *	ED-I(ED > −0.738)	−0.471 ***
	(1.88)		(−3.33)
GI	−0.413 **	GI	−1.250 ***
	(−2.33)		(−3.24)
HC	0.318 **	HC	0.625 **
	(2.29)		(2.52)
RO	0.212 ***	RO	0.362 **
	(2.76)		(2.55)
TI	0.139 ***	TI	0.212 *
	(3.11)		(1.68)
SI	0.155 ***	SI	−0.327 ***
	(3.62)		(−3.73)
_cons	−2.612 ***	_cons	5.199 ***
	(−3.16)		(3.03)
F-value	8.51	F-value	12.06
R-squared	0.359	R-squared	0.510

Note: ***, **, and * indicate significant at the 1%, 5% and 10% levels, respectively, and t-statistics for parameter estimates are in parentheses.

, the effect of ED intensity on ITU was different for diverse intervals; therefore, the effect of environmental decentralization on industrial transformation and upgrading of resource-based cities and non-resource-based cities can be considered nonlinear. The regression coefficient of ITU was 0.337 (p < 0.01) when ED was less than or equal to −3.785. When ED was greater than −3.785, the regression coefficient of ITU was 0.104 (p < 0.1), which indicates that, as the degree of environmental decentralization increases, the role of environmental decentralization in promoting industrial transformation and upgrading in resource-based cities diminishes. This is because the increase in environmental decentralization will make companies pay increasingly high environmental regulation. If the increasing marginal cost of environmental regulation is not met by the marginal innovation compensation benefit, the production burden of the firm will increase. Thus, the reverse effect on industrial transformation and upgrading will be weakened. According to the regression results of non-resource-based cities, when ED was less than or equal to −0.738, the regression coefficient of ITU was 0.108 (p < 0.01). The regression coefficient of ITU was −0.471 (p < 0.01) when ED was greater than −0.738, which indicates that environmental decentralization with an exceedingly high intensity will be detrimental to the promotion of the industrial transformation and upgrading of non-resource-based cities. For non-resource-based cities, the reversal effect of environmental decentralization on industrial transformation and upgrading is relatively limited due to the better environmental conditions and industrial base compared to resource-based cities. When the degree of environmental decentralization exceeds a certain threshold range, it will have the side effect of inhibiting industrial transformation and upgrading. From the results of the panel threshold regression of the control variables, the regression coefficients of GI in two types of city were significantly negative; the regression coefficients of HC, RO and TI in the two types of city were significantly positive; and the regression coefficients of SI in resource-based cities were significantly positive. Those in non-resource-based cities were significantly negative. In addition, the F-values and R-squared values of the panel threshold regression model were above the average, indicating that the panel threshold regression model has a better goodness of fit and the regression results are more robust.

5.3.2. Threshold Effect Analysis with Resource Endowment as the Threshold Variable

When resource endowment was the threshold variable, the same bootstrap self-sampling method was used to repeatedly sample 300 times to test whether there was a threshold effect in the sample data. The test results are shown in

Table 6. Results of the threshold effect test with resource endowment as the threshold variable.

City Category	Number of Thresholds	F-Value	p-Value	10% Critical Value Level	5% Critical Value Level	1% Critical Value Level
Resource-based Cities	Single Threshold	13.020 *	0.076	12.774	17.606	26.881
	Double Threshold	3.310	0.643	10.873	14.613	22.162
	Three-fold threshold	4.670	0.567	17.595	21.547	31.860
Non-resource-based cities	Single Threshold	23.530 ***	0.003	14.057	17.209	21.050
	Double Threshold	2.770	0.833	15.230	19.069	25.466
	Three-fold threshold	2.470	0.683	11.397	14.009	21.827

Note: *** and * indicate significant at the 1% and 10% levels, respectively.

.

It can be seen that, when resource endowment is the threshold variable, the single threshold F-value is significant at the 10% significance level for both resource-based and non-resource-based cities. However, none of the other threshold tests are significant, thus indicating a single threshold effect.

In addition, the threshold values of resource-based and non-resource-based cities were estimated using the great likelihood ratio estimation method. The threshold value of ${\eta }_3$ was estimated to be −0.743, and ${\eta }_4$ was estimated to be −0.795. In order to render the estimation results visually, the great likelihood ratio functions of the estimated results of the threshold variable values ${\eta }_3$ and ${\eta }_4$ were also plotted (see Figure 6 and Figure 7).

According to

Table 7. Regression results of the threshold effects with resource endowment as the threshold variable.

Variables	Resource-Based Cities	Variables	Non-Resource Cities
ED-I (RE ≤ −0.743)	0.056 *	ED-I (RE ≤ −0.795)	0.044 *
	(1.91)		(1.79)
ED-I (RE > −0.743)	−1.126 ***	ED-I (RE > −0.795)	0.330 ***
	(−5.02)		(4.50)
GI	−0.294 *	GI	−0.964 ***
	(−1.79)		(−2.64)
HC	0.066 ***	HC	1.192 **
	(4.32)		(2.57)
RO	0.124 *	RO	0.341 ***
	(1.92)		(2.64)
TI	0.091 ***	TI	0.148 ***
	(3.14)		(3.38)
SI	0.205 ***	SI	−0.313 ***
	(4.63)		(−3.94)
_cons	−3.474 ***	_cons	5.289 ***
	(−4.08)		(3.39)
F-value	7.93	F-value	23.38
R-squared	0.343	R-squared	0.586

Note: ***, **, and * indicate significant at the 1%, 5% and 10% levels, respectively, and t-statistics for parameter estimates are in parentheses.

, it can be seen that, when the RE of resource-based cities was less than or equal to −0.743, the regression coefficient of ED on ITU was −0.056 (p < 0.1). Additionally, when RE was greater than −0.743, the regression coefficient of ED on ITU was −1.126 (p < 0.01), which indicates that the inhibitory effect of environmental decentralization on industrial transformation and upgrading in resource-based cities increases with the increase in resource endowment. This is because the industrial development of resource-based cities is highly reliant on resources. However, with the improvement of resource endowment, the industrial development of resource-based cities falls into the “resource curse” trap, which inhibits industrial transformation and upgrading. Thus, Hypothesis 2 is verified. The regression results from non-resource-based cities show that the regression coefficient of ED on ITU was 0.044 (p < 0.1), when RE was less than or equal to −0.795. Additionally, when RE was greater than −0.795, the regression coefficient of ED on ITU was 0.330 (p < 0.01), which indicates that, with the increase in resource endowment, the promotion effect of environmental decentralization on industrial transformation and upgrading in non-resource-based cities continually increases. Unlike resource-based cities, the industrial structure of non-resource-based cities is more balanced, and the improvement of resource endowment can be rapidly combined with various production factors, thus releasing increasingly more resource dividends, and the promotion of industrial transformation and upgrading becomes stronger. Therefore, Hypothesis 3 is verified. From the results of the panel threshold regression of control variables, the regression coefficients of GI for the two types of city were significantly negative; the regression coefficients of HC, RO and TI for the two types of city were significantly positive; and the regression coefficient of SI for resource-based cities was significantly positive, while that of non-resource-based cities was significantly negative. In addition, the F-values and R-squared values of the panel threshold regression model were above the average, indicating that the panel threshold regression model has a better goodness of fit and the regression results have better robustness.

6. Conclusions and Policy Recommendations

It was found that environmental decentralization promotes industrial transformation and upgrading in both resource-based and non-resource-based cities. The combined effect of resource endowment and environmental decentralization inhibits industrial transformation and upgrading in resource-based cities, while it promotes it in non-resource-based cities. Specifically, for every 1% increase in the interaction between resource endowment and environmental decentralization, the inhibitory effect on industrial transformation and upgrading in resource-based cities increases by 0.578%, while the promotional effect on industrial transformation and upgrading in non-resource-based cities increases by 0.514%. In addition, the effects of different degrees of environmental decentralization and resource endowment on industrial transformation and upgrading differ. When environmental decentralization is the threshold variable, it promotes industrial transformation and upgrading in both resource-based and non-resource-based cities. Additionally, when resource endowment is the threshold variable, the inhibitory effect of environmental decentralization on industrial transformation and upgrading in resource-based cities increases with resource endowment, while it promotes industrial transformation and upgrading in non-resource-based cities.

Based on the findings of the study, the following policy recommendations are proposed:

First of all, it is important to continuously optimize and improve the environmental decentralization management system. According to the empirical results, environmental decentralization promotes the industrial transformation and upgrading of both resource-based and non-resource-based cities. Therefore, local governments in both two types of city should implement environmental regulations based on a precise understanding of the behavioral preferences of market players and residents in their jurisdictions, and actively promote green technological innovation while strengthening ecological and environmental governance. They should also actively promote the transformation and upgrading of traditional resource-based industries. New industries should be cultivated as a means to actively promote industrial transformation and upgrading. In addition, the central government should increase the supervision of ecological and environmental governance of local governments, and supervise the autonomous management of environmental affairs to avoid the vicious circle of “race to the bottom” among local governments in environmental regulation. For resource-based cities, the relationship between the central government and local governments should be scientifically defined and reasonably divided in terms of both financial and administrative powers. This can continuously improve administrative efficiency and policy implementation, so that the local government has both the financial resources and corresponding power to improve the quality of the ecological environment. In addition, governments should also scientifically design and formulate assessment objectives, highlight the guiding role of coordinated ecological and economic development in the industrial transformation and upgrading of resource-based cities, and incorporate the performance of ecological environmental protection and industrial development of resource-based cities into the government assessment.

Secondly, we should make scientific use of resource endowment, fully utilize the advantages of resource dividends, and avoid falling into the trap of the “resource curse”. According to the empirical results, the combined effect of resource endowment and environmental decentralization inhibits industrial transformation and upgrading in resource-based cities, while it promotes industrial transformation and upgrading in non-resource-based cities. The cities should embrace the positive role of resource endowment in promoting industrial transformation and upgrading, and scientifically formulate resource exploitation. They should also utilize natural resources in a reasonable and moderate manner, avoid shrinking and depletion of resources due to over-exploitation, and maintain ecological balance while minimizing the negative impact on the ecosystem. They should also continuously improve the efficiency of the comprehensive utilization of resources and continuously reduce the energy consumption of the industrial economy and the degree of reliance on resources. As for resource-based cities, they should continuously improve industrial competitiveness and economic efficiency by strengthening technological innovation and extending the industrial chain. Moreover, they can also promote the synergistic development of related industries and actively develop an outward-oriented economy. In this way, they would avoid the “resource curse” trap and accelerate industrial transformation and upgrading.

Additionally, the specific characteristics of resource-based and non-resource-based cities should be addressed, and their industrial transformation and upgrading should be promoted based on the reverse law of environmental decentralization. According to the empirical results, the combined effect of resource endowment and environmental decentralization on these cities may vary depending on the different degrees of environmental decentralization and resource endowment. Resource-based cities should balance the relationship between comprehensive resource development and utilization, economic growth and ecological environmental protection. They should vigorously promote the transformation and upgrading of traditional resource-based industries, and process resources and ecological restoration in depth. In this way, the production efficiency would be improved and the marginal pollution control costs would be reduced, which can promote the transformation of industrial production methods from careless to intensive. It is also necessary to actively cultivate and develop successive alternative industries, bring together new industries and new dynamics that reflect the qualities of “green” and “innovation”. Non-resource-based cities should focus on the pushback effect of environmental decentralization on industrial transformation and upgrading. They should actively develop new generation information technology, high-end equipment manufacturing, new materials, biomedical, new energy, energy conservation and environmental protection and digital creative. The industrial chain should be extended and the gathering of advantageous elements and resources should be strengthened. Furthermore, they should optimize resource allocation and create new industrial clusters so that the win–win situation of environmental protection and economic development can be continuously promoted.