Introduction

Since the introduction of “high-quality development”, The Chinese government’s focus on the economy has shifted from speed to quality. Industry plays a crucial role in enhancing efficiency and fostering sustainable development within the national economy. High-quality industrial development is an essential prerequisite for achieving high-quality overall development. It serves as a key driver for the transformation from rapid industrialization to high-quality industrialization, contributing to the construction of a prosperous modern socialist nation1.

Firstly, although China has achieved remarkable progress in industrialization, full industrialization has not been reached yet, and the problem of unbalanced and inadequate industrialization persists. Consequently, there is an urgent necessity to prioritize high—quality industrialization within the context of the “four modernizations” (new industrialization, informatization, urbanization, and agricultural modernization)2.

Secondly, Industry holds a position of paramount importance in the economic structure, and this significance has remained unchanged despite various economic developments. Industry is the cornerstone of economic activities, integrating production capabilities and knowledge accumulation. It acts as a powerful catalyst for economic growth. Its capacity to produce goods and services, underpinned by continuous knowledge enrichment, not only sustains current production but also propels the economy forward. Manufacturing, an integral part of industry, is a wellspring of technological innovation. It is where novel ideas are translated into practical products and processes. As the main originator of technological progress, manufacturing ceaselessly pursues improvements in efficiency and quality. Even in developed economies, where manufacturing’s share may seem to be shrinking, it continues to be the epicenter of most technological advancements. This reaffirms the crucial role manufacturing plays within the industrial framework and further emphasizes the importance of industry as a whole. The service industry, although distinct, has a relationship with industry that reinforces rather than diminishes industry’s significance. As industrialization advances, the service sector grows rapidly. This growth, however, is closely tied to industrial development. Producer services that are essential for production, like logistics, financial services related to manufacturing, and new digital—era services built around industrial supply chains, all rely on the foundation of industrial activities. In the digital age, the integration of manufacturing and services, such as the servitization of manufacturing and the outsourcing of manufacturing—related services, might alter the appearance of industrial sectors. But the core function of industry as an innovation engine remains unshaken. This innovation—driven nature within industry continues to fuel economic expansion, highlighting the unwavering importance of industry in the overall economic narrative.

Finally, China’s industry has achieved world-renowned developmental milestones by utilizing advantages such as institutional dividends, demographic dividends, and resource dividends. At the same time, due to the historical focus on quantity rather than quality, the industrial sector has accumulated structural risks with high input, high energy consumption, high emissions, and low efficiency. As a result, there are challenges including domestic overcapacity, resource shortages, insufficient technological innovation capacity in key sectors, environmental pollution, and ecological damage. In the current turbulent international context of economic and political fluctuations and the prevalence of anti-globalization, China’s industry is faced with significant challenges from the dual pressures: the “reshoring” of manufacturing industries in developed countries and the “diversion” of manufacturing industries because of low-cost competition in developing countries1. During this period, the significance of revitalizing the industrial economy and promoting HQID has become increasingly prominent. Meanwhile, the Party and the government have recognized the significance and urgency of HQID, and thus have proposed and enacted several policy documents aiming to promote the HQID of China3,4,5.

Extensive research on the topic of high-quality industrial development (HQID) has been conducted by scholars. The varying definitions of HQID lead to differences in evaluation index systems, evaluation objects, and evaluation methods. Existing studies that measure the level of HQID have mainly been carried out across different countries6,7, provinces3,4,5, and cities. However, most of these studies have focused on the provincial level. As a result, they fail to reveal the detailed and insufficient aspects of China’s industrial development within a more refined geographical scope. Cities, which integrate urban and rural areas, industrialization and the service industry, are crucial for driving HQID. Nevertheless, existing research has mainly concentrated on the HQID of certain urban agglomerations8 or specific cities9 in China. There is limited systematic coverage of a wide range of cities. Consequently, it is difficult to conduct a comparative analysis of HQID levels among different cities, and there is a lack of a comprehensive overall perspective. This limitation restricts the ability to identify the imbalances and inadequacies in China’s city-level industrial development. While the existing measurement index system recognizes the significance of green total factor productivity (TFP), it has overlooked the influence of crucial variables, such as the industrialization process, manufacturing servitization, and outsourcing, on the measurement of the HQID index. This has led to the omission of key variables and measurement errors.

Hence, it is imperative to define the concept of HQID, establish a rigorous measurement system for assessing its level, and specifically evaluate the HQID of cities. This will enable a precise understanding of the status quo and evolutionary trends of urban HQID in China and offer robust data support and valuable insights for further progress. Therefore, this study aims to integrate the industrialization process index of cities into the existing framework of the industrial green TFP index. It intends to build a measurement system for the HQID level, allocate weights to each index using a vertical and horizontal hierarchy method, and evaluate the HQID level of 286 cities in China. Subsequently, spatial correlation analysis and other methodologies will be used to reveal the spatial–temporal evolution patterns of HQID in China. This investigation aims to provide a theoretical foundation and policy guidance for reducing regional differences in the HQID level in China and facilitating its overall development.

This paper’s contribution is distinct from existing research in three main aspects.

Firstly, it adopts a comprehensive approach. By integrating the green TFP index and the industrialization process index, it builds a measurement system for assessing HQID level. This system highlights the driving force of innovation, the spillover of technology, and the export orientation of openness. It incorporates factors such as the number of granted patents, local governments’ science and technology expenditure, actual utilization of foreign capital, and import and export of goods, traditional factor productivity, resource utilization intensity, pollution emission intensity, industrialization level, and industrialization quality. This approach enables a more detailed examination of the imbalances and deficiencies in China’s urban industrial development. While previous studies have focused on specific urban agglomerations or cities, this study broadens its scope to include prefecture—level and higher cities, thus addressing some of the limitations in the existing research.

Secondly, existing research mainly focused on measuring and analyzing the level of urban HQID. In contrast, this paper expands its analysis by further considering the level of urban HQID. Specifically, it examines regional differences, causes, and future convergence trends in urban HQID level.

Lastly, this study uses the Dagum Gini coefficient decomposition method to examine the relative disparities in the level of urban HQID between the entire country and the eight comprehensive economic zones. This method not only reveals the magnitude, origin, and trend of regional differences but also resolves sample overlap issues and uncovers the underlying causes of regional differences in urban HQID level within the eight comprehensive economic zones. These findings provide a new perspective for balanced development in HQID.

Index system construction

Based on the essence of HQID and previous research findings, it is crucial to prioritize the emphasis on the environmental sustainability of industries and the industrialization process within cities. Additionally, careful consideration should be given to the authority, systematization, and balance of index selection. The analytic hierarchy process is used to establish an index system for assessing urban HQID level, as shown in Table 1. The target layer represents the level of urban HQID, while the first-level indicator layer includes two aspects: factor output efficiency and industrialization process. Each first-level indicator consists of 2–5 s-level indicators, resulting in a total of 7 s-level indicators. Specifically, factor output efficiency includes technological innovation intensity, factor productivity, pollution emission intensity, resource utilization intensity, and open competition intensity, whereas industrialization process comprises industrialization level and industrialization quality. Furthermore, each second-level indicator comprises 2–4 third-level indicators, resulting in a total of 18 third-level indicators. The logical correlation between these layers is explained as follows.

Table 1 Evaluation index system of level of urban HQID.
Full size table

Factor output efficiency

Innovation is the key driving force for industrial development. Capital, labor force, land, and energy constitute the fundamental input factors for economic activities, playing a crucial role in the urban industrial economy development10,11. Opening up is the only way for industry to achieve high-quality development, and represents technical cooperation and product export. As China’s future economic transformation and development focus, urban green innovation has emerged as a pivotal technological solution12,13. And green Total Factor Productivity (TFP) represents the efficiency improvements brought about by innovation, serving as a core indicator for achieving HQID in the present era14. It provides a clear and quantifiable measure of HQID. Consequently, it is both reasonable and necessary to establish an evaluation and analysis framework for assessing urban HQID based on the calculation of technical efficiency and TFP15. However, it is important to note that the green TFP index reflects the rate of change in green development efficiency rather than the efficiency itself4. It is characterized by high output and low factor input, as well as low pollution emissions and resource consumption. Therefore, to measure the industrial green TFP effectively, we establish the following second-level index based on the input–output approach:

  1. (1)

    Technological innovation intensity, in the short term, is represented by the number of granted patents by the enterprise and the science and technology expenditure of government support for innovation, and in the long term, it is reflected by human capital;

  2. (2)

    Factor productivity, which considers industrial added value (IAV) per unit of labor and IAV per unit of capital;

  3. (3)

    Pollution emission intensity, encompassing wastewater emissions, sulfur dioxide emissions, and soot emissions per unit of IAV;

  4. (4)

    Resource utilization intensity, examining IAV per unit of industrial electricity consumption and IAV per unit of industrial land use;

  5. (5)

    Open competition intensity, which is measured by the actual utilization of foreign capital and the proportion of total import and export in GDP. Among these indices, factor productivity reflects the output benefits derived from labor and capital, measured by IAV per unit of labor and IAV per unit of capital. Pollution emission intensity is indicated by the emissions of “three wastes” per unit of IAV. Higher levels of wastewater emissions, sulfur dioxide emissions, and soot emissions per unit of IAV signify a lower level of HQID. Resource utilization intensity, measured by industrial land and industrial electricity consumption, is primarily informed by current research on energy efficiency10,11,16 and industrial land use efficiency17,18,19,20. Numerous studies emphasize the importance of considering energy consumption as an input indicator when evaluating green efficiency16,21. The disorderly utilization of industrial land poses a significant concern in China22. Industrial land expansion often involves the conversion of agricultural land, resulting in reduced agricultural areas, regional vegetation coverage, and the capacity to mitigate dust and fine particles23.

Industrialization process

Industrialization process can be categorized into two aspects: industrialization index and industrialization quality.

  1. (1)

    The industrialization index measures the progress of Chinese cities in achieving industrialization based on an index system developed by Chen et al.24. This index system includes indicators such as the industrialization rate, and manufacturing industry’s location entropy. The process of economic modernization involves an interactive development between industrialization and urbanization. Industrialization provides the economic foundation and growth impetus for urbanization, while urbanization facilitates factor agglomeration and creates a broad demand market for industrialization25. The industrialization rate is computed according to established standards and formulas of the secondary and tertiary industries, as defined by Yin26. This calculation aligns with the evolutionary law of industrialization observed in long-term economic development, which involves the successive emergence of the "primary, secondary, tertiary," "secondary, tertiary, tertiary," and "secondary, tertiary, tertiary" industrial structures. This approach aims to avoid misleadingly high proportions of non-agricultural industry in cases where secondary industry is better than tertiary industry, or misleadingly high proportions of industrial structure in case where the tertiary industry is well-developed but the secondary industry is weak. These considerations are particularly relevant in the context of digitalization, which is driving the convergence of manufacturing and services. Manufacturing servitization and service outsourcing of manufacturing enterprises can lead to a statistically observable decline in manufacturing proportion and an increase in service proportion. The manufacturing location entropy is measured by the quotient of the ratio of manufacturing employees’ number to all employees of whole city and the ratio of manufacturing employees’ number to total employees of the whole country.

  2. (2)

    According to the connotation of high-quality industrialization proposed by Huang27, an index system for measuring industrialization quality has been constructed. Industrialization quality is assessed through several indicators, including real urban GDP per capita, residents’ disposable income, city labor productivity, and industrial comparative advantage. City labor productivity is measured by dividing the city’s real GDP by the total employment. Industrial comparative advantage is measured by the quotient of the IAV per employee in the city’s secondary industry and the IAV per employee in the national secondary industry.

Among the indicators in the aforementioned index system, the level of HQID exhibits a negative relationship with wastewater discharge, sulfur dioxide discharge, and soot discharge per unit of IAV. Conversely, the remaining indicators demonstrate a positive relationship with the level of HQID.

Research methods and data

Measurement method of level of urban HQID

To ensure objectivity and address the common issue of minimax weight in the entropy method, this study utilizes the vertical and horizontal grade evaluation method28 to assess and analyze the level of urban HQID. The specific steps are as follows:

  1. (1)

    The data were standardized using the range standardization method.

  2. (2)

    Following the principle of the ranking method, the n evaluated objects are treated as n points in an m-dimensional evaluation space and projected onto a specific space to ensure maximum dispersion of each projected point. The comprehensive evaluation function is formulated based on the panel data, as presented in Formula (1).

    $$y_{i} (t_{k} ) = \sum\limits_{j = 1}^{m} {{\varvec{\omega}}_{j} x_{ij} (t_{k} )}$$
    (1)

    Here, yi represents the evaluation value of evaluated unit i, ωj is weight of the evaluation index j. If \(y_{i} = (y_{1} ,y_{2} ,\ldots,y_{n} )^{T}\), \(A_{k} = \left[ \begin{array}{ccc} x_{11} (t_{k} ) & x_{12} (t_{k} )& \cdots x_{1m} (t_{k} ) \hfill \\ x_{21} (t_{k} )& x_{22} (t_{k} )& \cdots x_{2m} (t_{k} ) \hfill \\ \vdots & \vdots & \vdots \hfill \\ x_{n1} (t_{k} )& x_{n2} (t_{k} )& \cdots x_{nm} (t_{k} ) \hfill \\ \end{array} \right]\), \(\omega = (\omega_{1} ,\omega_{2} ,\cdots ,\omega_{m} )^{T}\), Formula (1) can be expressed as Formula (2).

    $$y = Aw$$
    (2)
  3. (3)

    Weight determination (ω). To maximize the distinction among evaluated objects, the linear function (y = ωTx) of index x should maximize the variance of n evaluated objects, thus determining the weight (ω). The formula for calculating the difference is presented as Formula (3).

    $$\sigma^{2} = \sum\limits_{k = 1}^{N} {\sum\limits_{i = 1}^{n} {\left[ {y_{i} (t_{k} ) - \bar{y}} \right]} }^{2}$$
    (3)

    By substituting y = Aω into Formula (3) and normalizing, we can determine that \(\overline{y} = 0\) and derive Formula (4).

    $$\sigma^{2} = \sum\limits_{k = 1}^{N} {\sum\limits_{i = 1}^{n} {\left[ {y_{i} (t_{k} )} \right]} }^{2} = \sum\limits_{k = 1}^{N} {[\omega^{T} {\varvec{H}}\omega ]} = \omega^{T} \sum\limits_{k = 1}^{N} {{\varvec{H}}_{{\varvec{k}}} \omega }$$
    (4)

    Here, ω = (ω1, ω2, … , ωm)T; \({\varvec{H}} = \sum\nolimits_{k = 1}^{N} {H_{k} }\) represents an m × m order symmetric matrix; \({\varvec{H}}_{{\varvec{k}}} = A_{k}^{T} A_{k} (k = 1,{\kern 1pt} {\kern 1pt} 2,{\kern 1pt} {\kern 1pt} \cdots ,{\kern 1pt} {\kern 1pt} N)\) is a real symmetric matrix. It is evident that when there are no constraints on d, Formula (4) can assume arbitrarily large values. However, in this case, ωTω = 1 is limited, and the maximum value of Formula (4) is determined by selecting an appropriate value for ω, resulting in Formula (5).

    $$\left\{ \begin{gathered} \max \omega^{T} {\varvec{H}}\omega \hfill \\ st.{\kern 1pt} {\kern 1pt} \omega^{T} \omega = 1 \hfill \\ \omega > 0 \hfill \\ \end{gathered} \right.$$
    (5)

    When ω represents the weight corresponding to the maximum eigenvalue of H, Formula (5) achieves its maximum value. The weight coefficient ω can be obtained by normalizing it, resulting in \(\sum\nolimits_{j = 1}^{m} {\omega_{j} = 1}\).

  4. (4)

    Compute the urban HQID index using Formula (1), which quantifies the level of urban HQID.

Spatial correlation analysis method

The level of urban HQID varies due to differences in material production, lifestyle, ecological resources, environmental quality, and geographical location among different cities. The spatial distribution among cities may exhibit spatial autocorrelation, meaning that the geographical location of a city not only impacts its own level of HQID. Furthermore, it also influences the level of HQID in its neighboring areas. Thus, it is essential to assess the spatial autocorrelation of the level of urban HQID in this scenario.

Spatial autocorrelation analysis is a widely used method for studying the spatio-temporal pattern evolution within a specific region. Moran’s index is commonly employed for measuring.

① Global spatial autocorrelation: This analysis is used to determine whether an attribute exhibits clustering within the overall space. It is represented by the global Moran index I, and the calculation formula is shown as Eq. (6).

$$I = \frac{{n\sum\nolimits_{i = 1}^{n} {\sum\nolimits_{j = 1}^{n} {W_{ij} \left| {y_{i} - \bar{y}} \right|\left| {y_{j} - \bar{y}} \right|} } }}{{\sum\nolimits_{i = 1}^{n} {\sum\nolimits_{j = 1}^{n} {W_{ij} \sum\nolimits_{i = 1}^{n} {(y_{i} - \bar{y})^{2} } } } }}$$
(6)

Here, \(n\) denotes the number of regions, \(y_{i}\) and \(y_{j}\) is the levels of regions i and j, respectively, and \(\mathop y\limits^{{{\kern 1pt} {\kern 1pt} \_}}\) represents the average level of each region. The spatial relationship is captured by the spatial weight matrix \(W_{ij}\), where the reciprocal of geographical distance is used as the spatial weight matrix.

The Moran index yields results ranging from − 1 to 1. It suggests a positive correlation among attribute values in each region, indicating spatial clustering when Moran index is greater than 0 and less than or equal to 1, When Moran’s index is close to or equal to 0, it indicates no spatial autocorrelation in attribute values. Conversely, it suggests negative correlation among attribute values, with a smaller value indicating stronger spatial differentiation when Moran index is less than 0 and greater than or equal to − 1. The Z statistic is utilized to assess the significance level of the Moran index, as follows:

$$z = \frac{I - E(I)}{{\sqrt {VAR(I)} }}$$
(7)

In Eq. (7), \(E(I)\) represents the expected value, and \(VAR(I)\) represents the variance of I. If Z is greater than 0 and passes the statistical test for z-value significance, it indicates a significant positive correlation in the spatial distribution of the attribute. Conversely, if Z is less than 0 and passes the statistical test for z-value significance, it indicates a significant negative correlation in the spatial distribution. Otherwise, it is considered not relevant.

② Local spatial autocorrelation. It decomposes the global spatial autocorrelation measured by Moran’s index into individual component units, enabling the testing of spatial agglomeration areas at the local level. The calculation formula is as follows:

$$I_{i} = \frac{{{\text{y}}_{i} - \bar{y}}}{S}\sum\limits_{j = 1}^{{\text{n}}} {W_{ij} (y_{j} - \bar{y})}$$
(8)

Here, \(I_{i}\) represents the local Moran index, when \(I_{i}\) is positive, it indicates the presence of spatial clusters with similar values, characterized by either high-high or low-low combinations, around the regional unit. Conversely, when \(I_{i}\) is negative, it indicates the existence of spatial clusters with similar values, characterized by either high-low or low–high combinations, around the regional unit.

Regional difference analysis method

The Gini coefficient decomposition method29 is employed to analyze the regional disparities in the levels of HQID in China’s eight comprehensive economic zones15. The overall Gini coefficient is calculated using the following formula:

$$G = \sum\limits_{j = 1}^{k} {\sum\limits_{h = 1}^{k} {\sum\limits_{i = 1}^{{n_{j} }} {\sum\limits_{r = 1}^{{n_{h} }} {{{\left| {y_{ji} - y_{hr} } \right|} \mathord{\left/ {\vphantom {{\left| {y_{ji} - y_{hr} } \right|} {2n^{2} \bar{y}}}} \right. \kern-0pt} {2n^{2} \overline{y}}}} } } }$$
(9)

where: \(y_{ji} (y_{hr} )\) represents the comprehensive index of HQID level; \(\overline{y}\) denotes its average HQID level; \(n\) represents cities number; \(k\) signifies the number of regions; \(n_{j} (n_{h} )\) indicates the number of cities in the region; \(G\) represents the overall Gini coefficient; \(j\) and \(h\) are different regions, \(i\) and \(r\) are different cities in region \(j(h)\).

Dagum29 decomposed the overall Gini coefficient, denoted as \(G\), into three components: the contribution of regional difference (\(G_{w}\)), the contribution of interregional net difference (\(G_{nb}\)), and the contribution of hypervariable density (\(G_{t}\)). The relationship between these components is defined as \(G = G_{w} + G_{nb} + G_{t}\). Equation (10) represents the Gini coefficient (\(G_{jj}\)) for a region (e.g. East coast), while Eq. (11) represents the Gini coefficient (\(G_{jh}\)) between regions (\(j\) and \(h\) region). Equations (12) to (14) describe the contributions of intra-regional differentials (\(G_{w}\)), inter-regional differentials (\(G_{nb}\)), and hypervariable density (\(G_{t}\)), respectively.

$$G_{jj} = \frac{1}{{2\mathop {y_{j} }\limits^{{{\kern 1pt} \_}} }}\sum\limits_{i = 1}^{{n_{j} }} {\sum\limits_{r = 1}^{{n_{j} }} {{{\left| {y_{ji} - y_{jr} } \right|} \mathord{\left/ {\vphantom {{\left| {y_{ji} - y_{jr} } \right|} {n_{j}^{2} }}} \right. \kern-0pt} {n_{j}^{2} }}} }$$
(10)
$$G_{jh} = \sum\limits_{i = 1}^{{n_{j} }} {\sum\limits_{r = 1}^{{n_{h} }} {{{\left| {y_{ji} - y_{hr} } \right|} \mathord{\left/ {\vphantom {{\left| {y_{ji} - y_{hr} } \right|} {[n_{j} n_{h} (\overline{{y_{j} }} + \overline{{y_{h} }} )}}} \right. \kern-0pt} {[n_{j} n_{h} (\overline{{y_{j} }} + \overline{{y_{h} }} )}}]} }$$
(11)
$$G_{w} = \sum\limits_{j = 1}^{k} {G_{jj} p_{j} s_{j} }$$
(12)
$$G_{nb} = \sum\limits_{j = 2}^{k} {\sum\limits_{h = 1}^{j = 1} {[G_{jh} (p_{j} s_{h} + p_{h} s_{j} )D_{jh} ]} }$$
(13)
$$G_{t} = \sum\limits_{j = 2}^{k} {\sum\limits_{h = 1}^{j - 1} {[G_{jh} (p_{j} s_{h} + p_{h} s_{j} )(1 - D_{jh} )]} }$$
(14)

Equations (12) to (14) define the variables as follows: \(p_{j} = {{n_{j} } \mathord{\left/ {\vphantom {{n_{j} } n}} \right. \kern-0pt} n}\), represents the ratio of cities number of region \(j\) to the total number of Chinese cities, \(s_{j} = {{n_{j} \overline{{y_{j} }} } \mathord{\left/ {\vphantom {{n_{j} \overline{{y_{j} }} } {n\overline{y}}}} \right. \kern-0pt} {n\overline{y}}}\), while \(D_{jh}\) denotes the relative impact of HQID level between regions. The calculation formula for \(D_{jh}\) is given by:

$$D_{jh} = {{(d_{jh} - p_{jh} )} \mathord{\left/ {\vphantom {{(d_{jh} - p_{jh} )} {(d_{jh} + p_{jh} )}}} \right. \kern-0pt} {(d_{jh} + p_{jh} )}}$$
(15)

Here, Eqs. (16)–(17) present the calculation formulas for \(d_{jh}\) and \(p_{jh}\), respectively. Specifically, \(d_{jh}\) is defined as the difference in the HQID level between region, \(p_{jh}\) is defined as the first moment of hypervariation. The variable \(F_{j} (F_{h} )\) represents the cumulative density distribution function.

$$d_{jh} = \int\limits_{0}^{\infty } {dF_{j} (y)\int\limits_{o}^{y} {(y - x)} } dF_{h} (x)$$
(16)
$$p_{jh} = \int\limits_{0}^{\infty } {dF_{h} (y)\int\limits_{o}^{y} {(y - x)} } dF_{j} (x)$$
(17)

Source of data

The macro data of 286 cities sampled in this study primarily originate from the China City Statistical Yearbook, and statistical yearbooks of provinces and cities. Missing data were addressed through linear interpolation. To ensure data completeness and availability, the sample of 286 cities excludes those in Hong Kong, Macao, and Taiwan, as well as cities with significant missing data, namely Bijie City, Chaohu City, Haidong City, Sansha City, Tongren City, Danzhou City, and various autonomous prefectures and leagues. The geographical distance between cities is determined by calculating the spherical distance using MATLAB software based on their respective longitude and latitude coordinates.

Analysis of results

Overall analysis of the level of urban HQID in China

Utilizing the aforementioned calculation formula for HQID index, we obtained the average value of the national urban HQID index, along with the corresponding scores for the factor output efficiency of the primary index and the industrialization process. These results are presented in Table 2 and Fig. 1.

Table 2 The mean value of urban HQID index and each sub-index.
Full size table
Fig. 1
figure 1

Change trend of the first-level index of HQID.

Full size image

Analysis of Table 2 and Fig. 1 reveals a gradual improvement in China’s overall HQID level, increasing from 0.1873 in 2003 to 0.3413 in 2022. Over the past 20 years, the level of development has increased by 0.8222 times, with an annual growth rate of 3.82%. But there is significant room for further growth. The level of China’s urban HQID experienced a slight decline in 2007 and 2022, along with a decrease in the industrialization process index. In 2007, China was negatively affected by the US subprime mortgage crisis, and COVID-19 occurred in China in 2020.The underlying causes of this decline warrant further in-depth research. In the remaining years, the national urban industrial high-quality development level exhibits positive growth. Considering a perfect score of 1 (100 points), the level of China’s urban HQID was 0.1873 in 2003 and gradually improved annually. However, by 2022, the average level of China’s urban HQID reached 0.3413, which remains below the threshold of 0.6 (60 points) denoting a failing level. These findings highlight the need for further improvement in the level of China’s urban HQID.

The first-level indicators, namely the factor output efficiency index and the industrialization process index, are presented in Table 2 and Fig. 1. Figure 1 reveals the following observations: (1) The annual trend of the factor output efficiency index from 2003 to 2022 is not significant, whereas the industrialization process index displays a consistent upward trend over the same period, exhibiting an average growth rate of 4.42%. In comparison to the average growth rate of the national urban HQID level (3.82%), the industrialization process index demonstrates a faster rate of growth, significantly contributing to the enhancement of the urban HQID level. (2) The industrialization process index has made a greater contribution to high-quality industrial development from 2003 to 2022 compared to the factor output efficiency index. (3) In terms of the national average, both the industrialization process index and the factor output efficiency index are at low levels, indicating significant room for development. (4) The industrialization process index has consistently surpassed the average of the factor output efficiency index, with its growth rate slightly exceeding that of the factor output efficiency index.

Spatial correlation analysis

To elucidate the disparities and spatial–temporal evolution trends in the HQID levels of individual cities, this study employed ArcGIS software to analyze the spatial–temporal evolution characteristics in the HQID level. Considering the overall low level of national HQID at present, the urban HQID level is categorized into five levels based on the maximum value (0.6184) and minimum value (0.0720): High [0.5093, 0.6184], medium high [0.4000, 0.5093), medium [0.2907, 0.4000), medium low [0.1814, 0.2907), and low [0.0720, 0.1814). A spatial distribution map is generated based on these five levels to comprehensively and visually illustrate the disparities in the HQID levels among China’s prefecture-level and higher cities and the dynamic characteristics of spatial patterns. This study examines four-time intervals (2003, 2010, 2016, and 2022) to describe and analyze the variations in HQID levels during different periods, as illustrated in Fig. 2. The white portion in the figure represents areas outside the scope of this research. Table 3 presents the distribution of cities across various levels of urban industrial high-quality development during the four-time intervals.

Fig. 2
figure 2

Spatial difference and evolution of urban HQID level in China.

Full size image
Table 3 Distribution of the number of cities at different levels of HQID.
Full size table

Figure 2 and Table 3 reveal the following characteristics regarding the changes in China’s urban HQID level: (1) During the period from 2003 to 2022, the color gradient shifted from lighter to darker, indicating a transition in urban HQID level from low and medium–low to medium. Specifically, in 2003, the level of HQID in Chinese cities was generally low, with the majority of cities categorized as medium–low or below. Only 6.3% of the regions were classified as middle and high level, with none falling into the high level. By 2010, the HQID level of China’s cities has achieved significant improvement. Notably, the number of regions in the medium–high level increased from 1 in 2003 to 6 in 2010, while the number of regions in the low level decreased from 151 in 2003 to 56 in 2010. Additionally, the number of medium–high level regions rose from 117 in 2003 to 180 in 2010. By 2016, except for nine cities that remained at a low level, 50.70% of the cities had achieved a medium level of industrial high-quality development or higher. Among them, 36 cities were classified as medium–high level. By 2022, 67.48% of the cities had attained a medium level or higher in terms of HQID. The number of cities in high-level areas and medium–high level areas reached 20 and 47, respectively, reflecting a substantial increase compared to 2016. (2) Between 2003 and 2022, there was a significant decrease in the number of areas categorized as middle-low level and below, accompanied by an increase in number of middle-high level and high-level areas. Moreover, the spatial distribution predominantly concentrated in eastern coastal areas. By 2022, low-level areas were primarily concentrated along the border between Heilongjiang and Russia.

To examine the impact of space on urban HQID level, this study analyzes the global spatial correlation using Moran index from period 2003 to 2022. Local spatial correlation was examined through the creation of thematic maps depicting the distribution of 286 cities in China for the years 2003, 2010, 2016, and 2022. Table 4 presents Moran index for the urban HQID level. From 2003 to 2022, Moran index for China’s prefecture-level and above cities is consistently positive and statistically significant, which suggests a positive correlation in the spatial distribution of China’s urban HQID level. This indicates China’s urban HQID level exhibits typical spatial agglomeration characteristics, that is the H–H spatial agglomeration pattern and the L–L spatial agglomeration pattern. This suggests that regions characterized by high levels of HQID are often adjacent to other regions exhibiting high levels. Similarly, regions with low levels of HQID are often surrounded by other cities with comparable low levels.

Table 4 Global Moran’s I Index of Urban HQID Level Based on Spatial Matrix.
Full size table

Secondly, Fig. 3 presents the results of the hot and cold spot analysis for China’s urban HQID levels in 2003, 2010, 2016, and 2022. Hot spot analysis is a commonly employed method in socioeconomic and ecological environmental analysis, wherein hot and cold spots indicate statistically significant spatial clusters of high and low values, respectively. Specifically, the hot spot analysis of ArcGIS 10.2 can be utilized to calculate the probability (P value) and the z-score, which denotes the number of standard deviations. Probabilities of 0.01, 0.05, and 0.10 are commonly used, with corresponding confidence intervals for z-scores of 90%, 95%, and 99% falling below − 1.65 or above 1.65, below − 1.96 or above 1.96, and below − 2.58 or above 2.58, respectively. A positive and significant z-score indicates a higher-than-average added value for the HQID level of unit i and its neighboring units, representing a cluster area (hot spot area) of improved industrial high-quality development. Conversely, a negative and significant z-score indicates a lower-than-average industrial high-quality development level for unit i and its neighboring units, signifying a cluster area (cold spot area) of declining industrial high-quality development. Hot and cold spots indicate statistically significant spatial clusters of high and low values, respectively.

Fig. 3
figure 3

Hotspot Analysis based on spatial distance matrix: 2003, 2010, 2016 and 2022.

Full size image

Figure 3 illustrates that, between 2003 and 2022, the prominent hot spots of Chinese cities primarily concentrated in the eastern coastal cities spanning from Shandong to Fujian, along with their adjacent central cities. Notably, the regional extent of these hot spots has been expanding. The cold spots are situated in Yunnan, Sichuan, Chongqing, Guizhou, and the northeastern province of Heilongjiang. The cold spot area resulting from the agglomeration in Yunnan, Sichuan, Chongqing, and Guizhou is diminishing, whereas the cold spot area in Northeast China is expanding.

Analysis of results by region

In June 2005, the State Council published a report on Strategy of Coordinated Regional Development. The report highlighted that China’s regional division method, which categorized regions as eastern, central, and western, not aligned with current situation. As an alternative, the report proposed eight comprehensive economic zones would be more appropriate. To depict variations in the levels of urban HQID across different regions, the entire sample was divided according to the classification standards of China’s eight comprehensive economic zones15.

Firstly, by calculating the mean value of the HQID level across the eight comprehensive economic zones, as presented in Table 5, and illustrating the evolution trend in Fig. 4, several observations can be made. First, the overall trend of HQID in these economic zones aligns closely with the national average. Secondly, regional disparities in urban HQID exist due to the influence of economic development, city size, and the industrialization process. Third, the eastern coastal areas exhibit the highest level of HQID and display the most notable growth trend. The second-highest level is observed in the northern coast. However, since 2018, the trend of HQID in the northeast and northwest areas has shown a decline. In general, the findings are consistent with the conclusions reached by Song and Zhang30 and Wang et al.3, indicating a fluctuating upward trend in the level of HQID over time.

Table 5 The average annual trend of the level of urban HQID in China’s eight comprehensive economic zones.
Full size table
Fig. 4
figure 4

Chart of Chinese Cities’ HQID level in the eight urban regions in 2003–2022.

Full size image

The Dagum Gini coefficient decomposition method is employed to assess the overall disparity in the level of HQID at the national level between 2003 and 2022. Additionally, it examines the intra-regional and inter-regional differences and identifies the sources of variation within the eight comprehensive economic zones. The results are shown in Table 6 and Fig. 5.

Table 6 Source decomposition of regional difference of HQID level.
Full size table
Fig. 5
figure 5

Trend of Gini coefficient and contribution rate.

Full size image

(1) The analysis reveals variations in the level of HQID both nationwide and within regions. Figure 5 illustrates a slight downward trend in the Gini coefficient (G) for the entire country between 2003 and 2022, suggesting that the Gini coefficient (G) has a downward trend in 2003–2018, and an upward trend in 2018–2022, and the difference of HQID in Chinese cities is increasing currently. (2) Among the eight comprehensive economic zones, Inter-regional difference undergoes the process of falling first and then rising, hypervariable density and Intra-regional differences undergo a substantial decline. (3) Based on the findings presented in Table 6 and Fig. 5, the investigation period reveals annual average contribution rates of 11.11%, 47.91%, and 40.98% for the intra and inter regional difference, and hypervariable density, respectively. These results indicate that rising inter-regional difference serves as the primary source of regional disparity, highlighting significant disparities in the contribution rates among these three factors. However, Qu et al.31 reached a contrasting conclusion through Theil index analysis, suggesting that intra-regional differences are primary drivers of regional disparities, with the largest disparities observed in the eastern region.

Discussion

Discussion on the evaluation index system of HQID

The creation of an indicator system is crucial for researching high-quality industrial development. The current literature on measuring industrial quality primarily employs either a single index method or a comprehensive index system approach. Key individual indices include value-added rate32, resource allocation efficiency33, labor productivity34, total factor productivity14, and green total factor productivity35. Recently, green total factor productivity (TFP) has been advanced by integrating land resources, environmental pollution, and energy input into the non-parametric model. This approach considers the limitations of land, energy, and the environment, aligning more closely with the principles of high-quality development. Green total factor productivity5 serves as a proxy variable to gauge regional high-quality development. Wang et al.4 and Song and Zhang30 have developed a comprehensive index system for high-quality industrial development based on five new development concepts: innovation, coordination, sustainability, openness, and sharing. Similarly, Li and Sun36 have established an evaluation index system for high-quality industrial development that focuses on the fundamental characteristics of high-quality factor input, green production process efficiency, effective output supply, and a rational internal structure.

This study is based on prior research, taking into account data availability and local conditions. After careful review and comparison, a comprehensive evaluation system was designed, adhering to principles of comprehensiveness, scientific rigor, and representativeness. This system aims to broaden and deepen the understanding of high-quality industrial development.

Discussion on the correct understanding of China’s urban HQID level and its spatial evolution dynamics

The results in this study are consistent with previous studies. The measurement results show that China’s level of HQID is comparatively low. by 2022, the average level of China’s urban HQID reached 0.3413, which remains below the threshold of 0.6 (60 points) denoting a failing level. Based on the 2022 data, China’s per capita GDP stands at 12,700 US dollars, whereas the United States’ per capita GDP is 76,400 US dollars. China’s per capita GDP amounts to approximately one-sixth of that of the United States. We believe that the measure results are reasonable and credible. These findings highlight the need for further improvement in the level of China’s urban HQID.

The results of Factor output efficiency measurement in this study show that China’s Factor output efficiency has steadily increased from 2003 to 2020, which is in line with China’s practice of green development, environmental governance, two-carbon goals, and building a resource-conserving and environment-friendly society. However, Tan and Shu’s37 findings suggest that the mean values of the industrial traditional factor productivity index, pollution emission index, and resource utilization intensity index are relatively high. Regarding the spatial distribution characteristics of high-quality industrial development, our findings concur with Jiang et al.38, indicating an overall upward trend and a spatial disparity following the pattern "East > Central and Western > Northeast". In terms of the spatial evolution of the high-quality industrial development index, our conclusions are consistent with those of Song and Zhang30 and Liu et al.14. The level of high-quality industrial development in China has steadily increased throughout the study period, yet there exists a regional discrepancy characterized by “high in the east and low in the west”. There is a discernible trend where regions with a lower quality development level are converging towards those with a higher quality development level.

Limitations and future research directions

Based on the essence of HQID and previous research findings, the study prioritizes the emphasis on the environmental sustainability of industries and the industrialization process within cities. The first-level indicator layer includes two aspects: factor output efficiency and industrialization process. Each first-level indicator consists of 2–5 s-level indicators, with a total of 7 s-level indicators. Moreover, each second-level indicator comprises 2–4 third-level indicators, leading to a total of 18 third-level indicators. Specifically, factor output efficiency involves technological innovation intensity, factor productivity, pollution emission intensity, resource utilization intensity, and open competition intensity. The industrialization process includes industrialization level and industrialization quality.

However, the index system of industrial high-quality development constructed in the study is not exhaustive, and there is still room for further improvement. Moreover, the study lacks an in-depth empirical analysis of factors influencing high-quality industrial development, an area that merits future research.

Conclusions and implications

Conclusion

The study constructs an evaluation index system to estimate urban HQID level of Chinese 286 cities from factor output efficiency and the industrialization process. Subsequently, the study calculates the level of urban HQID in China. Furthermore, it employs spatial correlation analysis, hot spot analysis, and other techniques to investigate spatial–temporal dynamic evolution characteristics of urban HQID level. The study reveals that:

  1. (1)

    The overall level of urban HQID in China has exhibited gradual improvement, rising from 0.1873 in 2003 to 0.3413 in 2022. Over the past 20 years, the level of development has increased by 0.8222 times, with an annual growth rate of 3.82%. However, there is still ample room for further improvement. From 2003 to 2022, the annual trend of the factor output efficiency index is not evident, whereas the industrialization process index demonstrates a consistent upward trend. In comparison to growth rate of the national urban HQID level (3.82%), the industrialization process index exhibits faster growth, making a substantial contribution to the enhancement of the urban HQID level. The contribution of the industrialization process index to HQID consistently surpasses that of the factor output efficiency index. However, both the industrialization process index and the factor output efficiency index remain low, leaving significant room for improvement. In contrast to the factor output efficiency index, the industrialization process index exhibits larger values and a higher growth rate.

  2. (2)

    There has been a transition from a predominantly "low and medium–low level" in 2003 to a predominantly “medium level” in 2022. Specifically, in 2003, the level of HQID in Chinese cities was generally low, with the majority of cities categorized as medium–low or below. Only 6.3% of the regions were classified as middle and high level, with none falling into the high level. By 2010, urban HQID level has been greatly improved. Notably, the number of regions in the medium–high level increased from 1 in 2003 to 6 in 2010, while the number of regions in the low level decreased from 151 in 2003 to 56 in 2010. Additionally, the number of medium–high level regions rose from 117 in 2003 to 180 in 2010. By 2016, except for nine cities that remained at a low level, 50.70% of the cities had achieved a medium level of industrial high-quality development or higher. Among them, 36 cities were classified as medium–high level. By 2022, 67.48% of the cities had attained a medium level or higher in terms of HQID. The number of cities in high-level areas and medium–high level areas reached 20 and 47, respectively, reflecting a substantial increase compared to 2016. (2) Between 2003 and 2022, there was a significant decrease in the number of areas categorized as middle-low level and below, accompanied by an increase in number of middle-high level and high-level areas. Moreover, the spatial distribution predominantly concentrated in eastern coastal areas. By 2022, low-level areas were primarily concentrated along the border between Heilongjiang and Russia.

  3. (3)

    The level of urban HQID in China exhibits typical spatial agglomeration characteristics, that is the H–H spatial agglomeration pattern and the L–L spatial agglomeration pattern. This suggests that regions characterized by high levels of HQID are often adjacent to other regions exhibiting similarly high levels. Similarly, regions with low levels of HQID are often surrounded by other cities with comparable low levels. Between 2003 and 2022, the prominent hot spots of Chinese cities primarily concentrated in the eastern coastal cities spanning from Shandong to Fujian, along with their adjacent central cities. Notably, the regional extent of these hot spots has been expanding. The cold spots are situated in Yunnan, Sichuan, Chongqing, Guizhou, and the northeastern province of Heilongjiang. The cold spot area resulting from the agglomeration in Yunnan, Sichuan, Chongqing, and Guizhou is diminishing, whereas the cold spot area in Northeast China is expanding.

  4. (4)

    Regional disparities in urban HQID exist due to the influence of economic, city size, and the industrialization process. The eastern coastal areas exhibit the highest level of HQID and display the most notable growth trend. The second-highest level is observed in the southern coast. However, since 2018, the trend of HQID in the northeast and northwest areas has shown a decline. A slight downward trend in the Gini coefficient (G) for the entire country between 2003 and 2022, suggesting that the Gini coefficient (G) has a downward trend in 2003–2018, and an upward trend in 2019–2022, and the difference of HQID in Chinese cities is increasing currently. Among the eight comprehensive economic zones, Inter-regional difference undergoes the process of falling first and then rising, hypervariable density and Intra-regional differences undergo a substantial decline. Annual average contribution rates of 11.11%, 47.91%, and 40.98% for the intra and inter regional difference, and hypervariable density, respectively. These results indicate that rising inter-regional difference serves as the primary source of regional disparity, highlighting significant disparities in the contribution rates among these three factors.

Implication

Based on the aforementioned conclusions, three implications can be derived:

Firstly, it is crucial to fully recognize the imbalanced characteristics and spatial distribution of HQID. In most cities, the level of HQID is relatively low. Considerable disparities exist among the eight comprehensive economic zones and cities, showing a Matthew effect in the trend. This implies that the approach of attracting investment to pursue industrialization and urbanization in all regions does not necessarily result in the convergence of HQID. Thus, to address the disparity in the level of urban HQID, governments at all levels should actively explore practical and effective measures to narrow the gaps in industrial structure among regions.

Secondly, considering the unbalanced characteristics of China’s urban HQID level, the state should give priority to providing policy support for urban industrial development in underdeveloped areas. Although the directions of regional support policies such as "the rise of central China," "the development of western China," and "the revitalization of Northeast China" are correct, their effectiveness in promoting high-quality industrial development in these regions is not satisfactory. This shows that it is necessary to further adjust and optimize the specific details of these relevant regional development support policies.

Thirdly, from a national perspective, it is essential to address the regional disparities in the level of HQID and promote regional coordination. Therefore, taking into account the unique circumstances of each region, cooperation and assistance should be established between cities with high-quality industrial development and those with lower levels, according to local conditions. This approach will help to implement regional policies that enable complementary advantages and coordinated development.