Resource Economy

Relationship between Industrialization, Urbanization and Industrial Ecology in Western China: A Panel Vector Auto-Regression Model Analysis

  • WANG Yajun , *
  • School of Economics and Management, Ningxia University, Yinchuan 750021, China
*WANG Yajun, E-mail:

Received date: 2020-06-14

  Accepted date: 2020-09-15

  Online published: 2021-03-30

Supported by

National Social Science Fundation of China(17XJY020)

National Natural Science Foundation of China(71963028)

Discipline Construction Project for Ningxia Institutions of Higher Education (Discipline of Theoretical Economics)(NXYLXK2017B04)


As the foundation of modern economic development, industry is the engine of industrialized and urbanized development. Industrial ecology is a high-level form of industry that is achieved after it has reached a certain stage, which guides the coordinated industrial development balancing mankind and nature. The implementation of industrial ecology is an important method and effective approach to realize the sustainable development of industrialization and urbanization. In this article, based on the inter-provincial panel data of western China during 2003-2018, the spatial development trends of industrialization, urbanization and industrial ecology are analyzed, and an empirical method is employed to conduct a robustness test based on the Panel Vector Auto-Regression (PVAR) model to determine the long-term interactions among these three aspects. The results show that it is difficult to manifest the short-term causal relationships among industrialization, urbanization and industrial ecology. After lagging for three periods, they present the Granger causality, the industrial ecology and industrialization have promoted urbanization, and the coefficient for the influence of industrial ecology on urbanization is 0.4612. However, industrialization and urbanization have negative impacts on industrial ecology, and with a 1% increase in industrialization or urbanization, the industrial ecology will decline by 0.2261% or 0.2850%, respectively. With the continuation of the lagging period, industrial ecology will have better interpretability than industrialization and urbanization, and industrialization and eco-friendly development have strong self-accumulation development mechanisms, while the self-accumulation mechanism of urbanization is not obvious, and it might even have a decline. By fulfilling the role of the regional leading industry, the state of unbalanced internal development can be improved, so as to realize mutual promotion between industrialization and urbanization. By improving the utilization rates of resources and energy, efforts should be made to implement green production, significantly promote industrial ecology, and boost high-quality development of both the regional economy and society.

Cite this article

WANG Yajun . Relationship between Industrialization, Urbanization and Industrial Ecology in Western China: A Panel Vector Auto-Regression Model Analysis[J]. Journal of Resources and Ecology, 2021 , 12(1) : 68 -79 . DOI: 10.5814/j.issn.1674-764x.2021.01.007

1 Introduction

Industrialization is the power source of economic growth, and urbanization is the spatial carrier of industrial development. Industrialization cannot be separated from urbanization, and urbanization is the manifestation of industrialization. They complement and promote each other. Ecological crisis and environmental damage are inevitable during the development of industrialization and urbanization, and the sustainable development of the economy and society can only be achieved by solving the conflicts between industrialization & urbanization and between resources, energy & environment. As the foundation of modern economic development, industry is the engine of both industrialization and urbanization. Industrial ecology refers to carrying out production activities according to the ecologic and economic rules, which can reflect the unity of economic, social and ecological benefits. It is a high-level form of industry after it has developed to a certain stage, which guides the coordinated industrial development balancing humankind and nature. Implementation of industrial ecology is an important method and a necessary condition for realizing the sustainable development of industrialization and urbanization. The western regions of China have a wide distribution of the ecologically vulnerable human-nature system, which are also the regions with most severe ecological impacts. In these areas, water and soil loss and desertification are severe, geological disasters are frequent, and the development and utilization of ecological resources are semi-extensive. Their industrial development model is of the classic resource- and energy-dependent type, and they are at the stage of industrial transformation and upgrade. Various factors have led to the continuous deterioration of the environment and aggravated the ecological problems. Under the promotion of the Belt and Road initiative and a national strategy of western development in the new era, the relationships among industrialization, urbanization and industrial ecology in the western regions are also becoming increasingly closer, all three aspects can generate certain positive externality to the economic and social development, and if any one aspect is unbalanced, then coordinated development cannot be achieved. Under the guidance of a new round of western development policy, industrialization, urbanization and ecological civilization construction have become the subjects of regional development. The question of whether real problems encountered during the industrialization and urbanization processes of western regions can be effectively solved through industrial ecology has important realistic significance in order to reduce the gap with the coastal developed areas under the new normal, achieve industrial development in high-end field, alleviate and reduce poverty, implement the coordinated development of ecological resources and realize common prosperity.

2 Literature review

Among the published studies on the industrialization in developing countries, Kuznets (1964) believed that industrialization has two basic characteristics, an increased proportion of industry and an increased employment of the industrial labor force during the industrialization process, which can be used to define the concept of industrialization in a relatively accurate way. From the perspective of the transformations of industrial structure and employment structure, the Indian economist Thaker (1986) believed that industrialization is a process in which the proportions of the production and employment of agriculture in the national economy continue to decline, while those of manufacturing and service industries continuously increse. Western economists, including Lewis, Prebish Singer and Hirschman, successively proposed the dual economic model, trade deterioration theory and linkage effect theory. They defined industrialization as a strategic objective, which is the theoretical basis for stipulating related strategy and policy for national economic development and a necessary measure for improving the living standard (Zhang, 1984). Zhang (1991) defined industrialization from the perspective of production factors, i.e., “a continuous process of breakthrough change (or revolution) from low level to high level based on a series of basic productions factors (or the combination of productions factors) in national economy”. In the meantime, he also emphasized that the role of agriculture should not be ignored during the industrialization process, and that agriculture can also carry out its effects separately. The economic growth of an agricultural country can only be realized based on the overall development of agriculture, manufacturing and other sectors (rather than just agriculture and manufacturing), and all these sectors are organic components of industrialization, which should not be separated from each other. Industrialization must be the modernization of agriculture and industry. Furthermore, he also believed that the improvement of the scientific and technological innovation level is a key factor of industrialized development. Balasubramanyam et al. (2000) believed that whether industrialization can be smoothly completed mainly depends on interference and promotion from national or local governments, and that a series of incentive policies should be employed to fully coordinate and deploy various market resources and elements, and promote the development of industrialization.
Chinese researchers mainly study urbanization from the aspects of labor migration, urban development and ecological civilization construction in cities. They focus on the spatial distribution of urbanization, new urbanization and other related issues, which involve various aspects, such as the improvement and rationalization of industrial structure, ecological urbanization and urbanization from the perspective of rural revitalization. Based on the theoretical mechanism and empirical research, a series of countermeasures and suggestions have been proposed to promote and improve urbanization construction. Most of them involve research on the interaction between industrialization and urbanization, and analyze their internal connections and interactive mechanisms. On the other hand, according to the coordination between industrialization and urbanization, estimations can be made, related causes can be explained, and the issues related to urbanization, industrialization and synchronous development of the four modernizations (modernization of agriculture, industry, national defense and science and technology) can be discussed (Guo and Xu, 2017). Furthermore, various factors can be incorporated into the systematic research on the interactions between industrialization and urbanization, such as population agglomeration (Li and Wang, 2018), agricultural modernization (Ye and Chen, 2017), land use and allocation (Mi, 2014), guidance of the government (Xiong and Yu, 2017), regional innovation (Fu and Zhuo, 2018), and by selecting some area or the entirety of China as the research object, specific operational policies and measures can be proposed.
During the industrialization process, related ecological and environment problems are inevitable, which will affect the sustainable development of the economy and society. Based on such a background, most scholars discuss the action mechanism for industrial ecology according to the negative influences of industry on the ecological environment. As the natural ecology, biosphere and industrial ecology have close relationships, they suggest implementing the resource-saving industrial structure, clean production and environmental-friendly development mode and building an efficient industrial ecosystem with low pollution and low consumption, so that the coordinated development of industry and ecology can be achieved. The ecological view of Karl Marx also emphasizes coordinating the harmonic relationship between humans and nature during industrial production activity by humans, and according to the process of “exchange of matter between man and nature”, the human intelligence should be fully employed to transform nature. In recent years, foreign scholars have focused on studying the evolutionary model and driving mechanism of micro or meso enterprise clusters and eco-industrial parks (Marx, 1995). Smith et al. (2015) pointed out that industrial ecology reflects the integration between the eco-environmental system and the social-industrial system, that the life cycle of industry is decided by the reutilization of resources and waste, and that industrial sustainability can be effectively guaranteed by strengthening control of energy mobility and enterprise production supervision. Odum and Elisabeth (2000) built an interactive response process model between urbanization and environmental quality based on system dynamics, and studied their interaction mechanisms and relations. They believed that an effective approach for promoting the coordinated development of urbanization and a low-carbon economy is to build an internal urban mobility system with zero carbon emission, and to establish the industrial ecology mode. Korhonen (2002) discussed industrial development from the macro perspective, and believed that efforts should be made to reduce the consumption of energy and resources, achieve efficient energy and material transformation consistent with the ecological environment, and realize unity between the industrial economy and ecological benefits. Guo et al. (2019) built a comprehensive industrial ecology criteria system. They employed various measurement methods (including the entropy method, coupling coordination analysis, explorative spatial data analysis method and spatial econometrics) to conduct comprehensive research on the ecologicalization level, spatial-temporal variation characteristics and influencing factors in the 17 prefecture-level cities in Shandong Province during 2005-2016. The results showed that industrial ecology presented the spatial gradient differentiation characteristics of decreasing progressively along the east-middle-west direction.
In the above studies, the industrialization, industrial ecology, urbanization, the interaction between industrialization and urbanization and the relationship between urbanization and industrial ecology are all studied from different perspectives, and rich research results and pertinent policy suggestions have been accumulated. However, with the acceleration of industrialization and urbanization, the birth of industrial civilization has changed the scale of the social economy and promoted the progress of human society. In the meantime, the ecological crisis is also exposed within the industry, which threatens the resources and ecological environment and affects the sustainable development of urbanization. Industrialization, urbanization and industrial ecology are all internal subsystems in the social, economic and environmental composite system. During the development of urbanization and industrialization, the production management activities and industrial ecology system are employed to affect their efficiency and state of development. In the industrial ecology system, the internal industry, external resources, environment and other factors have also provided the material basis for industrialization and urbanization, which has restricted the scale of the industrialization and urbanization system to a certain extent. When analyzing the relationships between industrialization and urbanization, most researchers utilize a linear regression model. However, if the variable of industrial ecology is incorporated, the relationships among these three aspects are no longer simple linear relationships, and alternative influences will be reflected with the passage of time. In this paper, the interactions between industrialization and urbanization are placed under the objective of industrial ecology to conduct further in-depth research on the interactions among these three aspects in 12 western provinces. Because these three variables are mutually restricting and affected by the previous state, the endogenous Panel Vector Auto-Regression (PVAR) model is adopted so that the variables are regarded as the lag-period function of all endogenous variables. This methodology can prevent the problem that regression analysis cannot reflect time lag in a timely manner, and in this way, the short-term and long-term relationships among the three variables can be explored.

3 Empirical analysis

3.1 Data sources and research methods

Because the concept of industrialization can be presented from multiple perspectives, foreign scholars have proposed a series of indexes and criteria for measuring the industrial level from different angles. The traditional industrial level measurement methods include the Hoffmann ratio method, the five-stage division of Kuznets and the six-stage division of Chenery. In these methods, the internal structure of industry, per capita GDP, the industrial distribution of the labor force and other related indexes are used for analysis. During estimation of the regional industrialization level, Chinese scholars combine the traditional theory with other factors, such as the demand structure and technological innovation. Chen et al. (2006) employed a hierarchical analysis method to determine the weights of the following five indexes: the per capita GDP, the production value structure of tertiary industry, the proportion of manufacturing value added in total added value of goods, the urbanization rate and the employment ratio of primary industry. They used the weighted synthesis approach to establish the comprehensive indexes of the industrialization level in order to conduct empirical analysis of the industrialization level of China, and they also adopted the PCA (principal component analysis) method to test the comprehensive evaluation results. Using the four main indexes of per capita GDP, the production value structure of tertiary industry, employment structure and urbanization level, Wang et al. (2012) conducted an empirical analysis of the industrialization development level of Quanzhou City according to the industrialization level measurement and classification standard. In this paper, the real development rules of industrialization are considered to reflect the actual state of the industrialization level. From the aspects of economic development level, industrial scale, employment scale, modernization degree of agriculture and government-level promotion, the indexes of the per capita gross regional product, the gross product of tertiary industry in total GDP, the proportion of labor force population from secondary and tertiary industries in the total labor force population, the total power of agricultural machinery per hectare, the proportion of R&D expenditure in GDP, and the proportions of fiscal revenue and fiscal expenditure are used to quantitatively measure the industrialization levels of ecologically vulnerable areas.
The concept and idea of industrial ecology are based on related theories of industrial economics and ecological economics, and industrial ecology integrates industrial operation and waste utilization and emission into the natural system (Cole and Neumaryer, 2004; Hauff and Wilderer, 2008). It has transitioned from an open system to a closed and cyclic industrial ecological system, the recycling of natural resources is combined with the protection of the ecological environment during the industrial development process, which has realized ecological regression of the industrial system, and comprehensive benefits are utilized to promote the coordinated development of industry and the environment (Hrenfeld, 2000; Despeisse et al., 2012). In this paper, based on its value orientation and scientific connotation, industrial ecology is considered according to the three main factors of industrial development level, energy and resource utilization and environmental pollution. The criteria are determined based on the representativeness and availability of data: the industrial added value is used to represent the development level of industrial scale, the energy consumption is adopted to reflect the energy utilization and efficiency, the resource dependence is chosen to represent resource development and utilization, the comprehensive utilization rate of general industrial solid waste is used to indicate the response to pollution regulation of the ecological environment, the industrial sulfur dioxide emission is used to represent the air pollution pressure in the ecological environment, and the industrial waste water emission is used to reflect the water pollution pressure in the ecological environment and measure the industrial ecology level of ecologically vulnerable areas (Liquete et al., 2013; Simboli et al., 2017). These indices can represent the regional pollution, resource and energy utilization and industrial development in a concentrated way. In order to prevent information overlap between indexes and the impact of subjectively set weights, the entropy weighting method is employed in this paper to determine the weights of various indexes according to the size of index variability. First of all, uniform standard processing of data is conducted to eliminate the interference of index dimensions in the processing result. The positive index is:
$ {{y}_{ij}}=({{x}_{ij}}-\min {{x}_{ij}})/(\max {{x}_{ij}}-\min {{x}_{ij}}) $
while the negative index is:
$ {{y}_{ij}}=(\max {{x}_{ij}}-{{x}_{ij}})/(\max {{x}_{ij}}-\min {{x}_{ij}}) $
in which, xij, max xij and min xij are the actual, maximum and minimum values of index j, respectively. Then, the information entropies of the six indexes of industrialization and industrial ecology are estimated, and the weights wj of the various indexes are calculated.
The urbanization index is generally measured according to the proportion of non-agricultural population or urban population in the total population. In this paper, by considering the division of administrative regions and the rules governing the spatial distribution of the population among cities and rural areas, in order to more accurately reflect the reality of economic development, the proportion of urban population in the total population is used as a measure of the urbanization degree of ecologically vulnerable areas. The panel data for the 12 western provinces during 2003-2018 used in this paper (Table 1) are from the statistical yearbooks of various provinces and the national economy and social development statistical bulletin.
Table 1 Selection of related industrialization, urbanization and industrial ecology indexes and their weights
Destination layer Criterion layer Index layer
Economic development level Per capita GDP (22.9)
Industrialization Economic scale Proportion of the gross product of tertiary industry in total GDP (18.1)
Proportion of labor force population from secondary and tertiary industries in the total labor force population (16.5)
Government promotion
Modernization level of agriculture
Proportion of R&D expenditure in GDP (14.9)
Proportions of fiscal revenue and fiscal expenditure (15.4)
Total power of agricultural machinery per hectare (12.2)
Urbanization Spatial distribution of population among cities and rural areas Proportion of urban population in total population (100)
Industrial development level Growth speed of industrial added value (21.6)
Industrial ecology Energy and resource utilization Total energy consumption (17.4)
Employment of extractive industry/ Total employment (18.2)
Environmental pollution degree Industrial waste water discharge (14.1)
Industrial sulfur dioxide emission (14.5)
Industrial solid waste utilization rate (14.2)

Note: For the weights of various indexes within the brackets, the unit is %.

After determining the index types in the comprehensive evaluation system of industrialization development level and the criteria and weights of the various indexes, the weighted synthesis method is employed to conduct normalized processing of the indexes and obtain the comprehensive indexes for industrialization level measurement. The specific formula is as follows:
$ IN={\sum\limits_{k=1}^{n}{{{\lambda }_{k}}{{W}_{k}}}}/{\sum\limits_{k=1}^{n}{{{W}_{k}}}}\; $
where, k represents the kth index; n refers to the number of evaluation indexes (n=6); λk is the evaluation value of the kth index obtained using the entropy method; Wk is the weight of the kth index obtained with the hierarchical analysis method; IN is the comprehensive evaluation value of industrialization development level obtained after weighted synthesis, and the score is within the range of 0-100. Similarly, the linear weighted method is used to calculate the industrial ecology level, and the formula is:
$ E{{C}_{ij}}=\sum\limits_{i=1}^{n}{{{\lambda }_{i}}{{W}_{ij}}} $
where, ECij represents the ecologicalization coefficient of an area in the jth year, λi is the weight of the ith index; and Wij is the ith non-dimension parameter of the jth year.
As shown in Fig. 1, the first, second and third rows, show the spatial distributions of industrialization, urbanization and industrial ecology in western China in years 2003, 2008, 2013 and 2018, respectively. According to the regional spatial distribution, the western regions are still at the middle stage of industrialization. At present, the eastern developed areas are at the later period of industrialization or the post-industrialization period, and the regional difference is significant. Especially in recent years, the excess production capacity, energy conservation and emission reduction problems are prominent, the economic growth has gradually slowed down, and as a result, the improvement of the industrialization level is not remarkable. Among the western regions, Tibet has the lowest urbanization degree. From 2003 to 2018, its overall urbanization level increased by nearly 10%, and the progress was significant. This indicates that temporally, the economic and urbanization development speed is fast in the western regions. Under related policies for the development of the western regions of China, the industrial structure of ecologically vulnerable areas has been continuously upgraded, and the labor force has been freed from traditional agricultural production in the rural areas and integrated into the non-agricultural sectors in cities. Although the urbanization level is continuously improving, there is still a distinct gap from the developed area in eastern China, and the difference is nearly 8%. The western regions generally have low industrial ecology levels, and Gansu had the lowest industrial ecology level among these areas in 2018. This indicates that the current industrial development mode mainly depends on resources with high input, high consumption and high pollution, and the development of heavy industry, nonferrous metals, oil, energy, chemical and other related industries has caused interference with the ecological resources and the environment, which has damaged the conservation and recovery ability of the ecological environment. In addition, human resources, scientific and technological innovation ability, production technology, market development degree and other factors have all restricted the development of the industrial system, the ecological system has a low resource utilization rate, the waste cannot be recycled, and the industrial development has not achieved effective symbiosis.
Fig. 1 Spatial distribution of industrialization (a), urbanization (b) and industrial ecology (c) in the western regions

3.2 PVAR model construction and parameters estimation

Smith (1993) built a vector auto-regression model to analyze the dynamic relationships of data based on a nonlinear time series. This model requires a long time series to treat variables as endogenous variables in order to prevent interference from the exogenous problem of the data, and it allows the investigation of the dynamic logic relation between a lagged-term variable and other variables. Holtz-Eakin et al. (1998) initially proposed using the vector auto-regression method to explain panel data, and they built the PVAR model to estimate the lagged-term parameters under steady panel data, so as to capture the impact on individual variables in a cross section. The model is as follows:
$ {{Y}_{it}}=\sum\limits_{j=1}^{n}{{{\beta }_{j}}{{X}_{i,t-1}}}+{{\varphi }_{i}}+{{\rho }_{t}}+{{\varepsilon }_{it}} $
where, Yit is the dependent variable; βj is the lagged-term regression coefficient of endogenous variables, Xi,t-1 is the lagged-term independent variable; i refers to the number of cross sections; t is the period number; φi represents the fixed effect of individual variables; ρt is the time effect;${{\varepsilon }_{it}}$ represents the disturbance term. This paper uses the panel data of 12 provinces in ecologically vulnerable areas during 2003-2018. The PVAR model is adopted for estimation, and the statistics software Stata15 is used to run all tests. Variables IN, UR and EC represent industrialization, urbanization and industrial ecology, respectively. In order to prevent “spurious regression”, a stationary test of data is conducted first. The Levin-Lin-Chu (LLC) and Augmented Dickey- Fuller (ADF) methods are employed for the unit root test. If the data are not stationary, a 1-order difference transformation is required until a stationary sequence is achieved, and then, the lagged order can be further determined. The test results with lagged 1-3 orders showed that when the lagged orders are 3, the amount of information of BIC, AIC and HQIC is the smallest, so the lagged orders were set at 3 (Tables 2 and 3).
Table 2 Stationary test of variables
Variable Levin-Lin-Chu (LLC)
test statistic value
Augmented Dicey-Fuller (ADF) test statistic value Result
lnIN 3.4592 8.0304 Non-stationary
lnUR 2.4235 10.0871 Non-stationary
lnEC 8.7021 11.7962 Non-stationary
d_lnIN -2.9323** -4.9669** Stationary
d_lnUR -7.4089** -10.1276* Stationary
d_lnEC 0.4426** -1.2723*** Stationary

Note: All the variables in the first three rows of the table are non-stationary. “d_” represents the transformation of variables through first-order difference. *, ** and *** represent the significance levels of statistics at 10%, 5% and 1%, respectively. The LLC test and ADF test of the first-order difference variables were significant, and the non-stationary variables became stationary sequences.

Table 3 Optimal lagged order test
Lag Bayesian Information Criterion
Akaike Information Criterion
Hannan Quinn Information Criterion (HQIC)
PVAR(1) -6.433 -5.831 -6.279
PVAR(2) -7.542 -3.955 -4.171
PVAR(3) -8.496 -6.548 -7.098
PVAR(4) -6.028 -4.103 -5.864

Note: PVAR(1), PVAR(2), PVAR(3) and PVAR(4) represent the lag orders 1-4 of the PVAR model, respectively. When the lag order is 3, the result shows BIC, AIC and HQIC with the minimum information content, so the lag order should be set as 3.

According to Table 4, when industrialization is the explained variable, the influence of industrialization on itself for a one-period lag is 0.7803, and the influence of industrial ecology on industrialization is 1.1943, which indicates that the development of industrialization can be promoted by improving the industrial ecology in the western regions. For a one-period lag, the influence of urbanization on industrialization is 0.6757. Through urbanization, the resources, human capital and industrial advantages are concentrated, which can promote the development of industrialization to a certain degree, but its influence on industrialization is more shallow than the influence of industrial ecology on industrialization. When industrial ecology is the explained variable, for a two-period lag, urbanization has an influence on the industrial ecology, and the influence coefficient is 0.1575; the influence of industrialization on industrial ecology is -0.2850, which means that a higher industrialization degree leads to a lower industrial ecology degree. For a three-period lag, urbanization has a significant influence on industrial ecology, so the higher the industrialization degree, the poorer the industrial ecology effects. When urbanization is the explained variable, for either one-period or two-period lags, urbanization is unrelated to any other variable; and for a three-period lag, industrialization and industrial ecology can better promote the development of urbanization.
Table 4 GMM estimation results of the dynamic panel
Variable h_lnIN h_lnUR h_lnEC
Coefficient P value Coefficient P value Coefficient P value
L1. h_lnIN 0.7803** 0.013 -0.2648 0.358 -0.1818 0.582
L1. h_lnUR 0.6757*** 0.001 -0.2436 0.226 0.3276 0.119
L1. h_lnEC 1.1943*** 0.000 -0.2929 0.284 0.1615 0.512
L2. h_lnIN -0.1011 0.541 -0.0655 0.715 0.0706 0.776
L2. h_lnUR 0.0173 0.917 0.1836 0.429 0.1575* 0.061
L2. h_lnEC -0.0161 0.920 -0.0430 0.825 0.1132 0.400
L3. h_lnIN -0.0916 0.346 0.0484** 0.048 -0.2850* 0.083
L3. h_lnUR 0.3333 0.114 0.1659 0.585 -0.2261*** 0.004
L3. h_lnEC -0.0203 0.831 0.4612*** 0.009 -0.0565 0.511

Note: *, ** and *** represent the significance levels of coefficients at 10%, 5% and 1%, respectively. h_lnIN, h_UR and h_EC are the sequences of lnIN, UR and EC after Helmert conversion to eliminate the individual effects, and L1, L2 and L3 respectively represent the variables of one-period, two-period and three-period lags. Coefficient and P value represent GMM estimated coefficients and estimates, respectively.

In order to ensure the stability of model, further verification is needed. As shown in Fig. 2, the absolute values of the characteristic roots of the model are all smaller than 1, which are all within the unit circle. This indicates that the PVAR model constructed above is implemented based on stable conditions.

3.3 Panel Granger Causality Test

On theoretical level, industrialization, urbanization and industrial ecology present a relationship of mutual influence, but testing with empirical data will prove this more convincingly. In this paper, the Granger causality test is adopted to further analyze the causal relationships among industrialization, urbanization and industrial ecology. Similarly, according to the influence of the lag phase, 2-order and 3- order lag phases are chosen during this test. The results show that when the lagged orders are 2, at the significance level of 5%, industrial ecology is the Granger cause of industrialization in the western regions, and an improvement in the industrial ecology level will promote the development of industrialization. At the significance level of 10%, urbanization and industrialization are the Granger causes of industrial ecology, and the interactions between and development of urbanization and industrialization have provided fundamental conditions for industrial ecology, which requires coordination and balance between people and nature. When the lagged order is 3, industrialization, urbanization and industrial ecology are the Granger causes of each other. During the urbanization process, the agricultural population is transforming into nonagricultural population, the released employment pressure is concentrated in cities, which forces related industrial sectors to transform the industrial structure in cities, the labor force requirement of tertiary industry is quickly increased, and the industrialization level is improved. During the urbanization and industrialization process, the ecologicalization problems of industries with high power consumption and high pollution are becoming increasingly prominent, and the ecological efficiency of industrial operations and the industrial ecology level can be improved through practical implementation of environmental protection policy and the effective execution of clean production processes. A good ecological basis and the overflow of various elements of industrialization (e.g., labor, technology, information, knowledge, etc.) can provide impetus to the development of urbanization to a certain extent.

3.4 Impulse response and variance decomposition

The impulse response function uses the standard deviation impact of the random disturbance term to investigate the degree of influence of the current value and future values of the three variables of industrialization, urbanization and industrial ecology in the western regions. The Monte Carlo simulation experiment was run 500 times, the impact time includes 12 periods, and the impulse response function of industrialization and other variables is obtained (Li and Mu, 2011). As shown in the first line in Fig. 3, industrialization does not have a significant impact on itself, and with the continuation of the time periods, the impact becomes milder and converges to 0. The impact of industrial ecology on industrialization is not obvious either, a slow growth trend continues until the 8th period, it then presents a gradual decline with a slow decline speed, which gradually drops to near 0.008. This indicates that the impact of industrial ecology on industrialization is positive at first, and then it becomes negative. In the period with a low degree of industrialization, the effects of industrial ecology are significant; however, with the deepening of industrialization, the requirement for industrial ecology becomes higher, and industrial ecology may even be able to catch up with the speed of industrialization evolution. As for the impact of urbanization on industrialization, during the 0-2 periods, the response between urbanization and industrialization is one of reciprocal transformation; during the 2-4 periods, it is a synthetic transformation; however, after the 4th period, it is all negative response, which indicates that the urbanization development has a significantly negative effect on the improvement of industrialization. After the 4th period, urbanization promotes the development of industrialization, which reflects its positive effect. For the variable of industrial ecology, the impact of industrialization is positive before the 5th period and negative after the 5th period. This means that the industrial ecology shows high ecological efficiency during the initial period of industrialization development, but after it has developed to a certain stage, industrialization has a negative impact on industrial ecology. The impact of urbanization on industrial ecology is similar to the impact of urbanization on industrialization, and the response also changes from negative to positive. The third line of Fig. 3 shows the impacts of other variables on the variable of urbanization. The impact of industrialization on urbanization is similar to the impact of industrialization on industrial ecology, the difference lies in the periods, and the impact of industrialization on urbanization shows a positive response before the 3rd period and a negative response after the 3rd period. This indicates that during the early stage of urbanization, industrialization can effectively promote the development of urbanization, but the progress of industrialization will affect and restrict the development of urbanization to a certain extent. The impact of industrial ecology on urbanization is always one of positive feedback on a long-term basis, and with advanced development of industrial ecology, the improvement of urbanization will also be fast, but it will also experience the dynamic evolution process with both fast and slow growth.
Fig. 3 Impulse response of the impacts of industrialization on urbanization and industrial ecology
Table 5 Panel Granger Causality Test results
Variable Causality 2-order lag 3-order lag
Estimated value Coefficient Result Estimated value Coefficient Result
lnIN lnEC non-Granger cause 2.698** 0.041 Reject 7.452* 0.084 Reject
lnUR non-Granger cause 1.819 0.611 Not reject 6.308* 0.092 Reject
ALL non-Granger cause 3.952 0.683 Not reject 9.632* 0.067 Reject
lnUR lnIN non-Granger cause 2.227 0.527 Not reject 5.034* 0.058 Reject
lnEC non-Granger cause 10.643** 0.014 Reject 3.011*** 0.000 Reject
ALL non-Granger cause 11.723* 0.068 Reject 2.654*** 0.008 Reject
lnEC lnIN non-Granger cause 5.755 0.324 Not reject 1.457*** 0.003 Reject
lnUR non-Granger cause 2.085 0.555 Not reject 1.019** 0.021 Reject
ALL non-Granger cause 7.389* 0.086 Reject 5.117** 0.029 Reject

Note: *, ** and *** represent the significance levels of 10%, 5% and 1%, respectively.

Table 6 shows the variance decomposition results of the error term for the three variables in periods 7-10. In the 7th period, during the decomposition of the error term of industrialization, its own interpretability accounts for 62.31%, and the interpretabilities of urbanization and industrial ecology are 4.68% and 33.01%, respectively; while in the 8th and 10th periods, the interpretabilities of industrial ecology increase to 40.03% and 50.04%, respectively, which means that the impact of industrial ecology on industrialization has experienced a significant change. The urbanization’s interpretability of industrialization does not have a significant change, and remains around 4%. The error term decomposition results of industrial ecology show that its own interpretability reaches 84.86% in the 7th period, while the interpretabilities of other variables are relatively small. In the 10th period, its own interpretability is enhanced, while the other variables’ interpretabilities of industrial ecology have decreased, but such a decrease is not significant. In the error term decomposition of urbanization, the industrial ecology’s interpretability of urbanization is significantly higher than the urbanization’s own interpretability and the interpretability of industrialization. In the 7th and 10th periods, the interpretabilities of industrial ecology and industrialization to urbanization have slightly increased, while urbanization’s own interpretability has declined. In summary, from the long-term perspective, industrialization and industrial ecology have strong self-accumulation development mechanisms, while the self-accumulation mechanism of urbanization is not significant, and it may even present degradation. The possible reason is that the marketization progress of the western regions is not remarkable, and the marketization degree depends on the efficiency of the market economy and the free flow between elements. The western regions have always been the national hub of heavy industry, nonferrous metals, energy and chemical bases, the management system is outdated, the state-owned enterprises have low efficiency, it is difficult for the market to carry out its decisive role, and the interaction between industrialization and urbanization is weakened. Industrial ecology and industrialization are “completely incompatible” contradictions, the development of industrialization will definitely pollute the environment and damage the ecological resources, while the recovery qualities of the ecological system indirectly affect the urbanization construction.
Table 6 Variance decomposition of the panel error term
Variable Period lnIN lnUR lnEC
lnIN 7 0.6231 0.0468 0.3301
lnUR 7 0.0993 0.2554 0.6453
lnEC 7 0.1011 0.0503 0.8486
lnIN 8 0.5549 0.0448 0.4003
lnUR 8 0.1027 0.2431 0.6542
lnEC 8 0.0942 0.0450 0.8608
lnIN 9 0.4987 0.0442 0.4571
lnUR 9 0.1084 0.2335 0.6581
lnEC 9 0.0914 0.0412 0.8673
lnIN 10 0.4544 0.0452 0.5004
lnUR 10 0.1146 0.2270 0.6583
lnEC 10 0.0911 0.0391 0.8698

4 Discussion

At present, the western regions have low industrialization levels, their industry and economy are both less than developed, and the industrial structure requires improvement. Efforts should be made to strengthen the industrial development, and a series of leading companies should be developed based on the rich coal resources, characteristic energy, tourism resources and characteristic agriculture. These leading companies should mobilize various local companies in related fields, and scale the industrial cluster and eco-friendly industrial parks should be established to introduce advanced management experiences, techniques and methods from the eastern area and abroad. The industrial chain should be improved, and competitive large companies should be established (Desrochers, 2002). These companies should actively participate in economic and trade cooperation with countries joining the Belt and Road Initiative, and carry out radiation and polarization effects throughout the area. They should carry out the role of leading industry in promoting the industrialization in the western regions, and also the progress of the entire region.
During the development of urbanization, the unbalanced internal development of the regions should be avoided, and mutual promotion between industrialization and urbanization should be realized. Against the background of rural revitalization, the leading role of medium and small towns in the west should be carried out to realize population agglomeration, share resources and public service facilities, make effective utilization of land and effectively handle the relationship between industrial development and ecological environment. The overall management level of cities should be improved through reasonable layout of infrastructure construction, covering the fields of transportation, water conservancy, energy, communication and environmental protection. In areas with a high urbanization level, efforts should be made to actively introduce new industry, transform the local employment structure, promote the development of tertiary industry, and improve the quality and level of industrialization. In areas with a low urbanization level, local advantages should be fully carried out, actively undertaking industrial transfer and accelerating the flow of rural population to cities, so as to promote the development of industrialization and industrial transfer and establish positive coupling development among these three aspects.
Facing the situation of a low level of industrial ecologicalization in west, the local government should take steps to improve the utilization rate of resources and energy, implement “green production”, vigorously promote industrial ecology in west regions, encourage recycling production within the industry, provide guidance with related theories of industrial ecology, and optimize the product and industrial structure (Lu and Sheng, 2012). Furthermore, they should also promote the transformation of industrial production model toward the direction with low pollution and low energy consumption, improve the overall quality of employees of the industry, manage companies with heavy pollution and low added value in traditional industry, increase the number of technology-intensive and capital-intensive companies with high added value, accelerate the transformation and upgrading of traditional manufacturing companies, and introduce creative and cultural industry with less environmental pollution and great development prospects. In major energy, chemical, resource and construction material industries, efforts should be made to build demonstration projects and industry, industrial parks and agricultural parks, adopt the modes of industrial symbiosis, clean production and circular economy, reduce pollution and energy consumption, and fully carry out the overflow effect of industrial agglomeration and industrial ecology. In the industrial park, they should adopt different construction methods by adjusting measures to local conditions, and promote the construction of industrial ecology (Denhodnd, 2000). Furthermore, the government should actively cooperate and coordinate with various departments to establish incentive policies and constraint mechanisms for industrial ecology. From investment direction to preferential tax policy, they should encourage the company to cooperate with other symbiotic enterprises by using the market. The government should establish a symbiosis network, and utilize the preferential policies for energy conservation, emission reduction and fiscal levy, and improve various companies’ ecological awareness. In addition, the local government should alsoformulate compulsory conservation policy, provide funding to support the company to improve outdated technology, and develop environment-friendly and efficient production technology with energy conservation, encourage local universities to strengthen collaboration between enterprises, universities and research institutes, and carry out the advantages of universities in combining scientific and technological innovation with practical applications.
Industrialization, urbanization and industrial ecology are all important components of the large and complex system of society and economy. Just like the famous “EKC” curve, in the period with low economic development (i.e., the early stage in the development of industrialization and urbanization), the ecological environment suffers from severe damage, and the industrialization, urbanization and industrial ecology present conflicting relationships. However, with the improvement of the industrialization and urbanization level, the transformation of industrial structure and production methods and the environment protection by the government, industrialization, urbanization and industrial ecology will present harmonious development, which will reflect the effective integration between industrial civilization and ecological civilization. By gradually overcoming the defects of the industrial system during the development process of industrialization and urbanization, the western regions of China will break the bottleneck constraint of the “resource curse”, reduce damage to the ecological resources and environment, improve the industrial ecology level, and strengthen ecological compensation and construction of the ecological civilization. On the other hand, industrial ecology is a development mode which can further promote the high-quality growth of the local economy, which is also an effective approach to realize sustainable development of the economy and society. It can coordinate the interactive relationship among industrialization, urbanization and industrial ecology, and achieve high-quality economic and social development of the western regions.

5 Conclusions

Our analysis of western China’s inter-provincial panel data of during 2003-2018, using the PVAR model to investigate the long-term interactions among industrialization, urbanization and industrial ecologicalization, leads to the following conclusions: 1) The western region is at the mid-stage of industrialization development, its urbanization level is lower than that of the developed area, its industrial ecology is low, and from the perspective of spatial distribution, the three variables of industrialization, urbanization and industrial ecology present certain spatial differences. 2) In the western region, it is difficult to demonstrate the short-term causal relationship among industrialization, urbanization and industrial ecology. In a 2-period lag, industrial ecology is the Granger cause of industrialization, and urbanization and industrialization are the Granger causes of industrial ecology. In a 3-period lag, the industrialization, urbanization and industrial ecology present mutual Granger causality. This indicates that during the industrialization development process, the improvement of the industrialization level can be better promoted by strengthening the spread in the fields of urbanization and industrial ecology. 3) Similarly, in a 3-period lag, industrial ecology and industrialization have promoted urbanization. The coefficient of influence of industrial ecology on urbanization is 0.4612, but industrialization and urbanization have negative impacts on industrial ecology. For each 1% increase in urbanization and industrialization, the industrial ecology declines by 0.2261% and 0.2850%, respectively. From the long-term perspective, the interpretability of industrial ecology is superior to that of industrialization and urbanization in the western regions. Industrialization and industrial ecology have strong self-accumulation development mechanisms, while the self-accumulation mechanism of urbanization is not significant, and it may even present degradation. Strong efforts should be made to promote urbanization construction, which can also help improve the level of industrialization and industrial ecology. The improvement of industrial ecology effects can also promote the development of industrialization and urbanization.
Balasubramanyam V N, Sanjaya L. 2000. Current issues in development economics (Translated by Liang X M). Beijing: China Tax Publishing House Press, 79-84. (in Chinese)

Chen J G, Huang Q H. 2006. Comprehensive evaluation and feature analysis of regional industrialization progress in China. Economic Research, 52(6):4-15. (in Chinese)

Cole M A, Neumayer E. 2004. Examining the impact of demographic factors on air pollution. Population and Environment, 26(1):17-21.

Denhodnd F. 2000. Industrial ecology: A review. Regional Environmental Change, 2(2):63-83.

Despeisse M, Ball P D, Evans S, et al. 2012. Industrial ecology at factory level: A conceptual model. Journal of Cleaner Production, 31(10):30-39.


Desrochers P. 2002. Cities and industrial symbiosis: Some historical perspectives and policy implications. Journal of Industrial Ecology, 5(4):29-44.


Fu Q, Zhuo C F. 2018. Urbanization, industrialization and regional innovation ability—Empirical analysis based on spatial econometrics model. Science and Technology Management Research, 38(1):30-34. (in Chinese)

Guo F Y, Tong L J, Liu Z G, et al. 2019. Spatial-temporal variation characteristics and influencing factors of industrial ecology in Shandong Province—Based on the spatial-temporal panel data of prefecture-level cities. Geographical Research, 38(9):2226-2238. (in Chinese)

Guo J H, Xu J Y. 2017. Measurement and countermeasure research of the synchronous development of industrialization, informatization, urbanization and agricultural modernization—With Shaanxi as example. Journal of Northwestern University, 105(7):32-39. (in Chinese)

Hauff M V, Wilderer P A. 2008. Industrial ecology: Engineered representation of sustainability. Sustainability Science, 3(1):103-115.


Hololtz-Eakin D, Newey W, Rosen H. 1998. Estimating vector auto-regressions with panel data. Econmetrica, 56(6):1371-1395.

Hrenfeld J R. 2000. Industrial ecology: Paradigm shift or normal science? The American Behavioral Scientist, 44(2):229-244.

Korhonen J. 2002. Two paths to industrial ecology: Applying the product-based and geographical approaches. Journal of Environmental Planning and Management, 45(1):39-47.


Kuznets S S. 1964. Economic growth and the contribution of agriculture: Notes on measurement. Agriculture in Economic Development, 3(1):102-119.

Li H N, Mu H L. 2011. Analysis on influence factors of China’s CO2 emissions based on path-stirpat model. Energy Policy, 39(1):6906-6911.


Li T Z, Wang W W. 2018. Analysis of urbanization, industrialization and population agglomeration in Northeast area based on the PVAR model. Population Journal, 38(6):75-85. (in Chinese)

Liquete C, Zulian G, Delgado I, et al. 2013. Assessment of coastal protection as an ecosystem service in Europe. Ecological Indicators, 30(2):205-217.


Lu G Y, Sheng L. 2012. Static and dynamic analysis of the industrial ecology level of China—Empirical research based on the inter-provincial data. China Industrial Economics, 30(3):147-159. (in Chinese)

Marx. 1995. Marx and engles anthology (Translated by Compilation and Translation Bureau of the CPC Central Committee). Beijing: People’s Publishing House Press, 217-223. (in Chinese)

Mi G. 2014. Research on the urbanization system for rural collective land—From the perspective of integration of urban and rural land use systems. Diss., East China University of Political Science and Law. (in Chinese)

Odum H T, Elisabeth C. 2000. Modeling for all scales: An introduction to system simulation. San Diego, USA: Academeic Press, 116-125.

Simboli A, Taddeo R, Raggi A. 2017. The multiple dimensions of urban contexts in an industrial ecology perspective: An integrative framework. International Journal of Life Cycle Assessment, 42(1):1-12.

Smith A. 1993. Estimating nonlinear time-series models using simulated vector auto-regressions. Journal of Applied Econometrics, 8(1):63-84.

Smith R L, Sengupta D, Takkellapati S, et al. 2015. An industrial ecology approach to municipal solid waste management: Methodology. Resources Conservation and Recycling, 104(5):311-316.


Thaker S Y. 1986. Industriallization and economic development. Delhi, India: South Asia Press, 17-21.

Wang H B, Li C P. 2012. Measurement and empirical analysis of industrialization level—With Quanzhou City as example. Academic Communication, 28(1):71-75. (in Chinese)

Xiong X, Yu X H. 2017. Development of urbanization and industrialization under government leadership and their interactive relationship. Journal of Guangdong University of Finance and Economics, 32(2):100-112. (in Chinese)

Ye A Z, Chen T. 2017. Empirical research on the relationship among urbanization, industrialization and agricultural modernization from the spatial perspective—Based on the semi-parameter spatial panel VAR model. Journal of Soft Science, 31(7):54-60. (in Chinese)

Zhang P G. 1984. Agriculture and industrialization. Wuhan: Huazhong University of Sceince & Technology Press, 23-28. (in Chinese)

Zhang P G. 1991. General theory of development economics—Industrialization problem of agricultural country. Changsha, China: Hunan Press, 56-60. (in Chinese)