Research on Ecological Niche Characteristics and Associated Pathways of Urban New Quality Productive Forces

Abstract

Developing new quality productive forces (NQPF) is an important direction for accelerating China’s high-quality development and advancing Chinese-style modernization. Based on the connotation and characteristics of NQPF, this study constructs an urban evaluation index system from three aspects of niche width, height, and overlap, and uses panel data of 283 Chinese prefecture-level cities over 2010–2023 to examine the spatiotemporal evolution patterns and key influencing factors of urban NQPF. Results are as follows. Firstly, urban NQPF exhibits an overall upward trajectory with an average annual growth rate of about 4.89%, showing phased features of rapid growth, moderated growth, and slight decline, and a spatial gradient of Eastern leadership, Central catch-up, and relatively low levels in the Western and Northeastern zones, with high-level cities growing into the largest group. Secondly, niche width, height, and overlap all show long-term upward trends. The energy level of factor resources and external resource acquisition capacity grow persistently. Inter-city competition evolves from moderate to high intensity, and regional disparities gradually converge. Thirdly, niche width and height are significantly and positively associated with NQPF development, whereas overlap is significantly and negatively associated, with marked regional heterogeneity. Combinatorial pathway analysis reveals that dominant pathways shift from “LLL + MLM” to “HHH + MMH” types, with HHH dominating high-level cities, MMH prevailing among medium-level ones, and low-level cities shifting from LLL to LLH. Improving energy level of urban new qualitative factor resources, strengthening spatial interaction level of new qualitative factors among cities, and avoiding adverse effects caused by excessive competition among cities should become an important content of optimizing the spatial layout of NQPF factor resources in China.

1. Introduction

New quality productive forces (NQPF) represent an advanced qualitative state of productivity characterized by high technology, high efficiency, and high quality [1], serving as an inherent requirement for driving high-quality economic development in the context of the new era [2]. At the Third Plenary Session of the 20th Central Committee of the Communist Party of China in 2024, the strategic directive to “develop new quality productive forces according to local conditions” was formally proposed. In March 2025, the Government Work Report of the State Council explicitly listed this objective as the second of ten annual priority tasks, underscoring the critical position and significant importance of vigorously developing new quality productive forces within China’s high-quality economic development framework. Since the concept of new quality productive forces was first introduced in September 2023, theoretical research has progressively deepened, encompassing connotation and characteristics [3,4], cultivation pathways [5,6], and empowerment mechanisms [7,8]. Quantitative research remains in its nascent stage, yet scholarly output concerning regional and industrial new quality productive forces at the meso-level has grown considerably. Regional-dimension studies have predominantly focused on provincial-level samples [9,10,11,12,13], while city-level research has only recently emerged [14,15,16,17,18]. Among these, Shi et al. (2025), employing the entropy weight method to measure NQPF across China’s five major urban agglomerations, utilized kernel density estimation, the Gini coefficient, and Markov chain analysis to explore spatial and temporal dynamics and evolutionary trends [14]. Mi et al. (2025) examined the Yangtze River Delta urban agglomeration, measuring and analyzing the urban association network characteristics of NQPF development and the effects of network structure on environmental improvement [15]. Gu and Liu (2025) measured the impact of innovation pilot policy implementation on urban NQPF [17]. Zhang et al. (2026) employed the Gini coefficient, kernel density distribution, and variance decomposition to measure regional disparities and dynamic evolutionary patterns in NQPF development among Chinese cities, subsequently identifying driving factors using the obstacle-factor method [18]. Cities play an indispensable role in China’s participation in international competition and cooperation, in sustaining national economic growth, and in promoting coordinated economic development. As China enters an accelerated phase of implementing the new-type urbanization strategy, urban development confronts a series of emerging opportunities and challenges, making inter-city disequilibrium and coordinated development equally crucial [19,20]. Consequently, vigorously cultivating urban new quality productive forces carries substantial practical value. Conducting an in-depth and detailed analysis of the dynamic evolution, spatial layout, and associated-pathway characteristics of NQPF development in Chinese cities constitutes a pivotal research agenda for thoroughly implementing the strategy of “developing new quality productive forces according to local conditions” and optimizing the spatial allocation of new-quality factor resources.

As China’s primary social contradiction has evolved, the issues of unbalanced and inadequate development in economic growth have drawn increasing attention. Urban development models have also exhibited a more diversified trajectory, transitioning from traditional monocentric patterns toward polycentric configurations, with spatial structures currently undergoing accelerated restructuring [21]. The intensification of urban economic activities serves as the fundamental driver of evolutionary change in urban agglomeration network structures [22], while urban innovation linkages formed through knowledge flows, scientific and technological collaboration, and other channels constitute a critical momentum for reshaping spatial relationships. Consequently, optimizing the spatial allocation of factor resources has become an essential component of enhancing regional coordinated development. Existing research in this domain has primarily focused on spatial disparities, spatial spillovers, and social network perspectives [23,24,25], deriving findings from inter-city differences, association intensity, and network centrality. Concurrently, inter-city competition and cooperation arising from endowment disparities, geographic location, and development policies also represent important elements in the holistic optimization of spatial layouts. The concept of “ecological niche” originated in biological evolutionary theory and matured within ecology. Charles Robert Darwin, in On the Origin of Species, observed that natural selection operates on those variations that are “better adapted to certain positions” within the natural system of a region [26]. The “positions” referred to herein denote “ecological niches.” Niche theory is progressively emerging as a new frontier in regional economics, with related concepts such as “regional ecological niche” and “urban ecological niche” having been proposed to characterize the “positions” and “functions” of distinct spatial entities within spatial systems [27,28,29]. An ecological niche denotes an N-dimensional hypervolume system in which biological species survive and undergo self-iteration, serving as a framework for evaluating competitive, predatory, and other interactive relationships among different biological populations under consideration of physiological characteristics and habitat conditions [30,31]. Cities exhibit certain similarities to species; when a city is conceptualized as an individual population, its “position” in space—given that it occupies only a portion of the resource space—is termed its urban ecological niche. Existing measurements of urban ecological niches have primarily focused on niche width and niche overlap, with niche width used to characterize a city’s capacity to acquire external resources, and niche overlap employed to reflect the degree of resource competition among cities. However, research on niche height remains relatively scarce. Distinct from niche width and overlap, niche height is used to characterize the energy-level state of a city’s internal resource endowment [32,33]. Given the limited body of work on niche height characteristics and the absence to date of urban niche height measurements grounded in new quality productive forces, research on the competitive advantages of urban new-quality factor resources remains in its nascent stage.

Guided by this framework, this study addresses three questions: (1) how urban NQPF levels changed across 283 prefecture-level cities and China’s four official economic regions from 2010 to 2023; (2) how niche width, height, and overlap evolved over the same period; and (3) which conditional associations and dominant configurations connect these niche characteristics to NQPF levels. Behind these questions lies a practical issue: as local governments mobilize strikingly similar packages of digital investment, high-tech talent, and strategic projects, each city must strengthen external connections and internal capability without intensifying homogeneous competition—otherwise, “developing according to local conditions” risks degenerating into “developing identically.” Academically, the study asks whether ecological niche measures can render this balance transparent at the city level, which existing designs cannot achieve because they conflate external breadth, internal endowment, and competitive similarity in a single construct. Although the empirical setting is China, this tension is not country-specific; the framework therefore offers a transferable descriptive vocabulary for economies transitioning toward innovation-driven development.

This study makes three contributions relative to existing literature. First, whereas prior city-level studies either measure NQPF directly or examine niche width and overlap in isolation, it links the NQPF with a three-dimensional niche framework, producing a unified city-year measurement system that separately identifies resource level, resource breadth, and inter-city similarity. Second, it extends city-level evidence beyond single urban agglomerations by examining 283 cities over fourteen years under a reproducible official regional classification. Third, it employs multiple regression models and combinatorial pathway analysis to characterize how different niche characteristics are associated with urban NQPF, both individually and in combination, thereby providing structured evidence to inform regional differentiation strategies.

2. Materials and Methods

2.1. Theoretical Linkage Between Ecological Niche and Urban NQPF

New quality productive forces are treated as a city-level productive system formed by the interaction of advanced laborers, upgraded subjects of labor, and material and intangible means of labor. Regional innovation systems research indicates that productivity and innovation depend on the local stock of skills and institutions as well as the connections through which knowledge, capital, and technology circulate [34]. From this perspective, an urban ecological niche is not a direct biological equivalence. It is an analytical representation of a city’s position in a shared resource space.

Niche width and niche height describe different aspects of that position. Width summarizes the diversity and relative reach of the external resource portfolio. Access to a broader set of knowledge, capital, talent, and industrial inputs can enlarge opportunities for recombination and reduce dependence on a narrow development path. Height summarizes the internal endowment level and reflects the capacity to absorb, transform, and retain these resources. Regional innovation systems and evolutionary economic geography research therefore support a positive expected association of both width and height with NQPF, while recognizing that accumulated capabilities can produce persistent regional differences [34,35].

Niche overlap summarizes similarity between city resource portfolios. Similarity can support learning when cities have related capabilities and sufficient channels for knowledge exchange [36,37]. It can also intensify competition for the same talent, finance, projects, and policy resources. When overlap becomes high, repeated industrial positioning and duplicated investment can crowd out differentiated specialization. The empirical coefficient on overlap is consequently interpreted as the net association in this sample, not as evidence that every form of proximity or competition is harmful.

2.2. Research Methods

2.2.1. Construction of the Indicator System for Measuring Urban NQPF

The construction of the indicator system is grounded in the three-factor theory of productive forces [2,3], which decomposes new quality productive forces into three primary dimensions: new-quality laborers, new-quality subjects of labor, and new-quality means of labor. These primary dimensions are further broken down into secondary and tertiary indicators, with specific operational indicators as follows.

For new-quality laborers, the dimension reflects the human capital perspective of new quality productive forces and is decomposed into three aspects: laborer quality, labor productivity, and laborer consciousness [38]. Laborer quality focuses on continuous learning to form high-quality human capital, measured by the average educational attainment of residents and the proportion of higher education students. Laborer consciousness focuses on the innovation atmosphere, captured through employment concept and entrepreneurial spirit, using the proportion of employees in the tertiary industry and entrepreneurial activity (number of newly established enterprises per 100 people) as indicators. Labor productivity focuses on improving labor productivity with “low input, high output,” measured by per capita output value and per capita income [39], both of which reflect efficiency enhancement and value return.

For new-quality subjects of labor, the dimension focuses on the expansion of labor scope and quality upgrading. It is constructed from two dimensions: new-quality industries and ecological environmental protection [40]. In terms of new-quality industries, strategic emerging industries and future industries are the core carriers of new quality productive forces; this study selects the number of strategic emerging industry enterprises and robot installation density to measure their development level [40]. In terms of ecological environmental protection, green development and pollution abatement are important components of new quality productive forces; this study selects green coverage rate, environmental protection intensity, pollutant emission level, and pollution treatment capacity to reflect their actual effectiveness.

For new-quality means of labor, the dimension includes both material tools that laborers use to transform labor objects and intangible technological and informational resources that support creativity [41]. Thus, this study constructs indicators from two dimensions: material means of labor and intangible means of labor. For material means of labor, modern infrastructure and high-efficiency new production tools enhance resource circulation efficiency and product quality, supporting urban digital operation and green transformation; this study selects infrastructure and energy utilization levels to measure their infrastructure level. For intangible means of labor, scientific and technological innovation and digitalization are key representations; this study selects innovation output, innovation input, and the digital economy index to assess their supporting capacity.

Table 1. Indicator system for measuring urban new quality productive forces (NQPF).

Primary Indicator	Secondary Indicator	Tertiary Indicator	Specific Indicator	Measurement Formula	Polarity
New-quality laborers	Laborer quality	Human resource potential	Proportion of higher education students	Number of university students/Total population	Positive
		Higher education penetration	Average educational attainment	Average years of schooling	Positive
	Labor productivity	Economic output	Per capita output value	Per capita gross regional product	Positive
		Economic income	Per capita income	Average wage of employees	Positive
	Laborer consciousness	Employment concept	Employment industrial structure	Proportion of employees in tertiary industry	Positive
		Entrepreneurial spirit	Entrepreneurial activity	Number of newly established enterprises per 100 people	Positive
New-quality subjects of labor	New-quality industries	Strategic emerging industries	Development of strategic emerging industries	Number of strategic emerging industry enterprises	Positive
		Future industries	Robot count	Robot installation density	Positive
	Ecological environmental protection	Green development	Green coverage rate	Green coverage area/Total area	Positive
			Environmental protection intensity	Environmental regulation intensity	Positive
		Pollution abatement	Pollutant emission level	Industrial wastewater discharge	Negative
				Industrial SO2 emissions	Negative
				Industrial soot emissions	Negative
			Pollution treatment capacity	Comprehensive utilization rate of industrial solid waste	Positive
				Harmless treatment rate of household waste	Positive
New-quality means of labor	Material means of labor	Infrastructure	Traditional infrastructure	Highway mileage	Positive
			Digital infrastructure	Share of digital infrastructure-related terms in government work reports	Positive
		Energy utilization	Total energy consumption	Total energy consumption	Negative
			Clean energy consumption	Clean energy consumption	Positive
	Intangible means of labor	Scientific and technological innovation	Innovation output	Number of granted patents/Total population	Positive
			Innovation input	Science and technology expenditure/Fiscal expenditure	Positive
		Digitalization level	Digital economy development level	Digital economy index	Positive

reports the operational indicators and their polarity.

2.2.2. Measurement of Urban New Quality Productive Forces Level: Entropy Weight Method

The entropy weight method is an objective weighting approach that determines the weights of individual indicators in a multi-indicator evaluation system based on information entropy. Its core principle is to use information entropy to measure the degree of dispersion of indicators, thereby determining their weights and objectively reflecting the importance of each indicator within the overall evaluation system.

The measurement procedure is as follows.

Step 1: The original indicators are standardized to eliminate dimensional effects using the range standardization method. In Equation (1), x denotes the original indicator value, x’ denotes the standardized value, i denotes the indicator, and j and t denote city and year, respectively.

$${x^{\prime }}_{ijt}=\left\{\begin{array}{c}{\displaystyle \frac{x_{ijt}-\min \left(x_{it}\right)}{\max \left(x_{it}\right)-\min \left(x_{it}\right)}},for positive indicators\\ {\displaystyle \frac{\max \left(x_{it}\right)-x_{ijt}}{\max \left(x_{it}\right)-\min \left(x_{it}\right)}},for negative indicators\end{array}\right.$$

(1)

Step 2: The proportion $p_{ijt}$ and information entropy $e_i$ are calculated as shown in Equation (2).

$$p_{ijt}={x^{\prime }}_{ijt}/{\displaystyle \sum _j{x^{\prime }}_{ijt}}; e_i=-{\displaystyle \frac{1}{\ln n}}\times {\displaystyle \sum _j{\displaystyle \sum _t\left(p_{ijt}\times \ln p_{ijt}\right)}}$$

(2)

where n is the number of years in the study period.

Step 3: The redundancy $d_i$ and weight $w_i$ of each indicator are obtained as shown in Equation (3).

$$d_i=1-e_i; w_i=d_i/{\displaystyle \sum _i{d}_i}$$

(3)

Step 4: Based on the standardized indicator values and the calculated weights, the new quality productive forces level (NQPF) for each city over the study period is derived as shown in Equation (4).

$$NQP{F}_{jt}={\displaystyle \sum _i\left({x^{\prime }}_{ijt}\times w_i\right)}$$

(4)

2.2.3. Urban Ecological Niche Measurement Model

The concept of ecological niche, originating in ecology, describes the position and function occupied by a particular population within an ecosystem, revealing the competitive and coexistent relationships among populations through the characterization of resource utilization patterns and intensity [42,43,44]. Introducing niche theory into the study of urban layout for NQPF allows for a multidimensional deconstruction of the competitive, cooperative, and symbiotic relationships among cities in utilizing NQPF resources, thereby providing a reference for optimizing the spatial allocation of these resources and examining associated pathways. Currently, a “trinity” analytical framework encompassing niche width, niche height, and niche overlap has begun to enter the field of urban ecological niche analysis [33]. Specifically, niche width characterizes a city’s capacity to acquire external resources; niche height evaluates the energy-level state of a city’s internal resource endowment, enabling comparative advantage analysis across cities; and niche overlap measures the similarity in resource endowments between two cities, thereby reflecting the intensity of inter-city competition within or across urban agglomerations. Through this three-dimensional analytical system, this study provides a detailed assessment of inter-city competitive and cooperative patterns in the domain of NQPF resources in China.

First, niche width (width). Drawing on the Shannon–Wiener index [33], this study constructs an urban niche width indicator for NQPF, as specified in Equation (5). In Equation (5), width denotes the niche width; k is the number of resource dimensions (indicators); and p represents the proportion of the j-th NQPF resource (i.e., new-quality laborers, subjects of labor, and means of labor) utilized by city i relative to the total across all cities. A larger niche width value indicates that a city can acquire more external new-quality factor resources in developing NQPF, implying stronger external adaptability in the process [45].

$$widt{h}_{jt}=-{\displaystyle \sum _{i=1}^k{p}_{ijt}}\ln p_{ijt}$$

(5)

Second, niche height (height). This study employs the entropy weight–catastrophe progression method [33], with the measurement procedure as follows. Step 1: The relative weights of indicators are determined and ranked using the entropy weight method. Step 2: The type of catastrophe system is determined according to the number of indicators at each hierarchical level: the cusp catastrophe model is selected when there are two indicators, the swallowtail catastrophe model when there are three, and the butterfly catastrophe model when there are four. Step 3: The normalized formulas are applied to obtain the niche height value of NQPF for each city. The specific mathematical forms are presented in Equations (6)–(8). In these equations, x denotes the state variable, V(x) represents the potential function, and a, b, c, and d are control variables.

Cusp catastrophe model:

$$\begin{array}{c}V(x)=x^4+a{x}^2+bx, \\ Normalization formula is x_a=\sqrt{a},x_b=\sqrt[3]{b}\end{array}$$

(6)

Swallowtail catastrophe model:

$$\begin{array}{c}V(x)={\displaystyle \frac{1}{5}}{x}^5+{\displaystyle \frac{1}{3}}a{x}^3+{\displaystyle \frac{1}{2}}b{x}^2+cx, \\ Normalization formula is x_a=\sqrt{a},x_b=\sqrt[3]{b},x_c=\sqrt[4]{c}\end{array}$$

(7)

Butterfly catastrophe model:

$$\begin{array}{c}V(x)={\displaystyle \frac{1}{6}}{x}^6+{\displaystyle \frac{1}{4}}a{x}^4+{\displaystyle \frac{1}{3}}b{x}^3+{\displaystyle \frac{1}{2}}c{x}^2+dx, \\ Normalization formula is x_a=\sqrt{a},x_b=\sqrt[3]{b},x_c=\sqrt[4]{c},x_d=\sqrt[5]{d}\end{array}$$

(8)

The final niche height value for each city is obtained by applying the corresponding normalization formulas in a bottom-up hierarchical manner, with the state variable x representing the composite niche height index.

Third, the calculation formula for niche overlap (C) is presented in Equation (9). In Equation (9), C denotes the niche overlap between cities j and f; k is the number of resource dimensions (indicators). The indicator ranges from 0 to 1. A larger value indicates a higher degree of overlap in NQPF factor resources between two cities, implying greater resource similarity and, consequently, more intense inter-city competition. Specifically, when C = 0, the ecological niches of the two cities are completely separated; when C = 1, the new-quality factor resources of the two cities are in a state of complete overlap, representing the most intense level of competition. Subsequently, the mean niche overlap value for each city is derived.

$$C_{jf}={\displaystyle \sum _{i=1}^k\left(p_{ij}{p}_{if}\right)}/\sqrt{\left({\displaystyle \sum _{i=1}^k{p}_{ij}^2}\right)\left({\displaystyle \sum _{i=1}^k{p}_{if}^2}\right)}$$

(9)

2.3. Sample Selection, Data Collection, and Processing

This study selects prefecture-level and above cities as the research sample. The spatial scope of this study covers prefecture-level and above cities in mainland China from 2010 to 2023. According to the latest data released by the administrative division website (www.xzqh.org (accessed on 20 August 2026)), mainland China comprised 333 prefecture-level administrative units and 4 municipalities directly under the central government at the end of 2025, totaling 337 prefecture-level and above units. These 337 units constitute the initial sampling frame of this study. On this basis, three screening rules were applied sequentially: (1) special-function cities lacking a conventional statistical reporting system; (2) cities newly established during the study period or subjected to major boundary adjustments that would break panel comparability; and (3) cities with consecutive multi-year missing values on core indicators that could not be reliably imputed. Accordingly, data availability was a necessary but not sufficient condition for inclusion; the completeness, comparability, and definitional consistency of the panel over the entire study period constituted equally important screening criteria. The final balanced panel comprises 283 prefecture-level cities, accounting for over 84% of the total—a predominant share that ensures sample representativeness. Regarding regional division, the 283 cities are classified by province into four major economic regions—Eastern, Central, Western, and Northeastern—in accordance with the official statistical classification criteria issued by the National Bureau of Statistics of China. This classification scheme has been consistently adopted in national planning documents and statistical yearbooks, and is also a widely used inter-regional comparison framework in the Chinese regional economics literature; it possesses administrative authority and reproducibility, corresponds directly to China’s differentiated regional policy system, and ensures a sufficient number of observations within each region for reliable estimation. The specific sample sizes are 86 cities in the Eastern region, 80 in the Central region, 83 in the Western region, and 34 in the Northeastern region.

Regarding data sources and processing, the primary socioeconomic data for prefecture-level and above cities are obtained from the China City Statistical Yearbook, provincial statistical yearbooks of all provinces, municipalities, and autonomous regions, and the EPS database. In addition, the average years of schooling indicator is compiled from the results of the Fifth, Sixth, and Seventh National Population Censuses, with linear interpolation applied to derive data for intermediate years. The entrepreneurial activity indicator is sourced from the Regional Innovation and Entrepreneurship Index of China (IRIEC), compiled by Peking University (https://opendata.pku.edu.cn/dataset.xhtml?persistentId=doi:10.18170/DVN/NJIVQB (accessed on 1 November 2025)). The classification of strategic emerging industries follows the Catalog for Classification of Strategic Emerging Industries (2023) and the processing approach of Li et al. (2022) [46], with the number of enterprises retrieved from Tianyancha (https://www.tianyancha.com (accessed on 20 January 2025)), a Chinese commercial enterprise information platform. Environmental protection intensity is measured by environmental regulation intensity, following the approach of Shao et al. (2024) [47]. Pollutant emission level indicators include three emission sources: industrial wastewater, industrial SO₂, and industrial soot. Pollution treatment capacity is composed of the comprehensive utilization rate of industrial solid waste and the harmless treatment rate of household waste. Digital infrastructure level is constructed following the method of Chao et al. [48], using the frequency share of digital-facility terms in government work reports. The digital economy development level is measured from two aspects—internet development and digital financial inclusion—based on city-level data availability, following established practices [49]. Urban internet development level is captured by three indicators: internet broadband access users per 100 people, per capita total telecommunication services, and mobile phone subscribers per 100 people [50]. Digital financial development is measured using the China Digital Financial Inclusion Index, jointly compiled by the Digital Finance Research Center of Peking University and Ant Group [51]. Ultimately, the comprehensive digital economy development index is derived through principal component analysis, in which the above four indicators are standardized and then dimensionally reduced [49].

3. Analysis of Urban Ecological Niche Characteristics of NQPF

3.1. Evolutionary Characteristics of Urban NQPF

First, based on Equations (1)–(4), the NQPF level values for all cities are obtained. Figure 1 presents the mean value curves for the national sample and the four economic zones, and

Table 2. Distribution statistics of cities by NQPF type.

Year	Number of Cities			Proportion of Cities (%)
	High-Level (H)	Medium-Level (M)	Low-Level (L)	High-Level (H)	Medium-Level (M)	Low-Level (L)
2010	34	54	195	12.01	19.08	68.90
2011	38	62	183	13.43	21.91	64.66
2012	45	62	176	15.90	21.91	62.19
2013	57	68	158	20.14	24.03	55.83
2014	72	90	121	25.44	31.80	42.76
2015	74	101	108	26.15	35.69	38.16
2016	88	104	91	31.10	36.75	32.16
2017	127	102	54	44.88	36.04	19.08
2018	131	108	44	46.29	38.16	15.55
2019	132	106	45	46.64	37.46	15.90
2020	134	112	37	47.35	39.58	13.07
2021	135	118	30	47.70	41.70	10.60
2022	131	116	36	46.29	40.99	12.72
2023	123	117	43	43.46	41.34	15.19

reports the statistical results of the typological distribution of NQPF. According to the evolution curve of the mean NQPF across all cities in Figure 1, the NQPF exhibits a sustained upward trend over the period 2010–2023. In 2010, the mean value was 0.0511, rising to 0.0950 by 2023, representing a cumulative increase of 85.91% and an average annual geometric growth rate of 4.89%. This indicates that, during the study period, China’s urban NQPF maintained a continuous growth trajectory at an annual rate of 4.89%. A phased breakdown reveals that during the 12th Five-Year Plan period (2011–2015), the average annual geometric growth rate was approximately 8.09%, representing a relatively high growth phase, which is defined as the “rapid growth period.” During the 13th Five-Year Plan period (2016–2020), the average annual geometric growth rate moderated to approximately 5.19%. Although still in a growth phase, the pace slowed notably, termed the “moderated growth period.” Entering the 14th Five-Year Plan period (2021–2025), by 2023, the average annual growth rate adjusted to approximately −0.0073%, hovering near zero. Specifically, a slight increase occurred from 2020 to 2021, followed by negative growth in 2021–2022 and 2022–2023. This indicates that, during the 14th Five-Year Plan period, urban NQPF development has essentially entered a “moderate decline period.” In summary, China’s urban NQPF has exhibited a sustained upward trajectory, with an average annual growth rate of approximately 4.89% over the study period. Concurrently, distinct phased characteristics are evident, transitioning from “rapid growth” to “moderated growth” and then to “moderate decline” across the 12th, 13th, and 14th Five-Year Plan periods. Since the onset of the 14th Five-Year Plan period, NQPF has entered a phase of moderate decline, making urban NQPF development a critical challenge for high-quality development.

Second, as shown in Figure 1, the mean value of NQPF in the Eastern economic zone remains at the highest level. It is significantly higher than the mean values of the Central, Western, and Northeastern regions, and also markedly exceeds the national average across all cities. Moreover, comparing the NQPF curves of the Eastern region with the national sample reveals a widening gap over time: in 2010, the Eastern mean exceeded the national average by approximately 0.0211, and by 2023, this gap had increased to 0.0455. Thus, the Eastern economic zone occupies a “leading” position in urban NQPF development. Relative to the national city sample, the Central economic zone’s NQPF development was generally slightly below the national average during the study period. However, since 2020, the gap has narrowed significantly, and by 2023, the two values were essentially on par. This suggests that the Central economic zone represents an “actively catching up” type. Meanwhile, the mean NQPF levels of the Western and Northeastern regions consistently remained below the national average, indicating that both regions have maintained a “relatively low level” of NQPF development. Taking 2017 as a turning point, the Northeastern region shifted from being slightly above the Western region to slightly below it, with the Western region’s growth rate slightly outpacing that of the Northeast. Consequently, the four major economic zones exhibit a spatial gradient distribution pattern characterized by “Eastern leadership, active Central catch-up, and relatively low levels in the West and Northeast.”

Third, employing the tercile method, urban NQPF development is classified into three types: high-level, medium-level, and low-level, with Table 2 presenting the distribution statistics of cities across these types for the corresponding years. Historically, the number of low-level cities has declined rapidly and significantly. In 2010, there were 195 low-level cities, accounting for 68.90% of the total and occupying an absolutely dominant position among the three types, but this number decreased to 108 by 2015, 37 by 2020, and 43 by 2023. This indicates that low-level cities have withdrawn from their dominant position, now covering only approximately 15.19% of prefecture-level cities in recent years. Conversely, high-level cities have experienced a dramatic and sustained expansion. Starting with only 34 prefecture-level and above cities in 2010, this number rose to 127 by 2017, surpassing low-level cities in absolute quantity. By 2023, the number of high-level cities reached 123, accounting for 43.46% of the total, making them the largest group and the core force supporting urban NQPF development. Meanwhile, medium-level cities have shown transitional growth, with their number increasing from 54 (19.08%) in 2010 to 117 (41.34%) in 2023. Although they no longer constitute the dominant group, they play a crucial transitional role between the high and low levels. Ultimately, the pattern of “high-level expansion, medium-level transition, and low-level shrinkage” has now taken shape.

In summary, during the period 2010–2023, urban NQPF has generally maintained an upward trajectory. However, the phased, regional, and typological evolution characteristics should not be overlooked. Specifically, the evolution exhibits a phased pattern of “rapid growth → moderated growth → slight decline,” with a modest decline observed since the onset of the 14th Five-Year Plan period. Across the four major economic zones, a gradient distribution pattern of “Eastern leadership, Central catch-up, and relatively low levels in the West and Northeast” has emerged. The number of low-level cities has declined sharply, while high-level cities have grown rapidly to become the largest group, establishing the current typological pattern of “high-level dominance.”

3.2. Measurement and Evolutionary Patterns of Niche Width Characteristics

Based on Equation (5), the urban niche width measurement results for 2010–2023 are obtained. Figure 2 presents the mean value curves for the national sample and the four major economic zones.

First, as shown in Figure 2, the national mean width exhibits an overall upward trend over the observation period. In 2010, the mean niche width was 2.5472, rising to 2.6019 by 2023, representing a cumulative increase of 2.15%. Specifically, the period 2010–2011 experienced a notable short-term decline of 2.07%, defined as the “short-term adjustment period.” This dip reflects the structural growing pains and resource reallocation frictions as China’s urban economy transitioned from traditional extensive growth to a new innovation-driven development model. From 2011 onward, during the 12th and 13th Five-Year Plan periods, the national mean width grew steadily, reaching a cumulative increase of 4.25% by 2021, with an average annual compound growth rate of 0.42%. This phase is defined as the “long-term growth period,” indicating that cities gradually adapted to the new development paradigm and expanded their capacity to acquire external NQPF factors. Entering the 14th Five-Year Plan period, growth momentum began to decelerate from 2021 to 2022, and in 2023, negative growth (−0.10%) emerged, termed the “momentum attenuation period,” reflecting the mounting pressure of resource bottlenecks and diminishing marginal returns in external factor acquisition. In summary, from 2010 to 2023, the national mean niche width of urban NQPF in China increased modestly, while exhibiting a three-phase pattern of “short-term adjustment → long-term growth → momentum attenuation.”

Second, as shown in Figure 2, the Eastern economic zone has maintained a sustained lead in niche width, i.e., the external acquisition capacity for NQPF factor resources. This persistent lead is deeply rooted in its first-mover advantage in institutional innovation and its concentration of high-end innovation platforms, which generate a strong “siphon effect” for external resources. However, its growth momentum has gradually weakened, and its first-mover advantage is being progressively narrowed by the Central economic zone, with the gap shrinking by 28.22%. In 2022, the Eastern region became the first to experience negative growth, indicating that its resource acquisition capacity is approaching a saturation point under existing structural constraints. The Central economic zone has demonstrated a robust catching-up trajectory, with a cumulative width increase of 2.56%. This “counter-trend overtaking” aligns with the theoretical framework of regional gradient transfer, where the Central region actively undertakes industrial and factor spillovers from the East, leveraging its geographic hub status. Since 2018, the Western economic zone has surpassed the Northeast, ending the former pattern of “Northeast > West.” This shift is largely attributable to national strategic support, such as the Western Development Strategy, which has effectively enhanced the West’s institutional capacity to attract emerging productive forces. Conversely, the Northeastern economic zone has experienced a cumulative width increase significantly lower than other regions, with its regional standing in continuous decline. This reflects a severe “path dependence” on traditional heavy industries and institutional rigidities, which hinder the inflow of new quality factors. Taking 2018 as the dividing point, the regional ordering shifted from “East > Central > Northeast > West” to “East > Central > West > Northeast.”

Third, employing the tercile method, all cities are classified into three types of niche width: high-level, medium-level, and low-level, with

Table 3. Distribution statistics of cities by niche width type.

Year	Number of Cities			Proportion of Cities (%)
	High-Level (H)	Medium-Level (M)	Low-Level (L)	High-Level (H)	Medium-Level (M)	Low-Level (L)
2010	83	105	95	29.33	37.10	33.57
2011	41	52	190	14.49	18.37	67.14
2012	50	61	172	17.67	21.55	60.78
2013	56	67	160	19.79	23.67	56.54
2014	58	84	141	20.49	29.68	49.82
2015	69	81	133	24.38	28.62	47.00
2016	75	90	118	26.50	31.80	41.70
2017	85	106	92	30.04	37.46	32.51
2018	92	121	70	32.51	42.76	24.73
2019	105	125	53	37.10	44.17	18.73
2020	129	121	33	45.58	42.76	11.66
2021	156	103	24	55.12	36.40	8.48
2022	161	104	18	56.89	36.75	6.36
2023	161	100	22	56.89	35.34	7.77

presenting the distribution statistics of cities across these types. From a dynamic evolution perspective, the number of low-level cities first increased and then decreased, reflecting the cyclical shocks of macroeconomic restructuring on resource allocation. The proportion of low-level cities surged to 67.14% in 2011 during the initial structural adjustment phase, before declining rapidly to 6.36% by 2022. However, in 2023, the proportion rebounded slightly to 7.77%, suggesting mounting upward mobility barriers for tail-end cities that suffer from geographic and institutional marginalization. Meanwhile, medium-level cities have played a crucial buffering role in the transition. Their proportion peaked at 44.17% in 2019 before gradually declining to 35.34% in 2023, serving as a transitional reservoir for cities upgrading to higher levels. Most notably, high-level cities have experienced dramatic quantitative expansion, becoming the absolute dominant group. The number of high-level cities grew continuously from 83 (29.33%) in 2010 to 161 (56.89%) in 2023. This structural shift is theoretically consistent with the “agglomeration economy” effect: as NQPF heavily relies on knowledge spillovers and digital networks, factor resources increasingly concentrate in cities with high niche width, creating a “rich-get-richer” cumulative causation process. Overall, the typological distribution of urban niche width has transitioned from a decentralized “pyramid-shaped” structure to a “high-level dominant” inverted-pyramid structure, though the marginal rebound of low-level cities warrants policy attention regarding inclusive development.

In summary, the niche width of NQPF factor resources in Chinese cities has exhibited a significant overall upward trend, with phased changes characterized by “short-term fluctuation → stable growth → recent attenuation.” Across the four major urban agglomerations, the regional pattern has shifted from “East > Central > Northeast > West” to “East > Central > West > Northeast,” with the Western region’s catch-up momentum driven by strategic support contrasting sharply with the Northeast’s institutional lock-in. Meanwhile, the dominant position of high-level cities has steadily strengthened, reflecting the inherent agglomeration logic of new quality productive forces, and the distribution has evolved into a “high-level dominant” structure.

3.3. Measurement and Evolutionary Patterns of Niche Height Characteristics

First, based on Equations (6)–(8), the urban niche height measurement results for all cities are obtained. Figure 3 presents the mean value curves for the national sample and the four major economic zones. As shown in Figure 3, during the period 2010–2023, urban niche height in China exhibited a sustained upward trend. In 2010, the mean niche height was 0.9201, reaching a peak of 0.9362 in 2021, with a cumulative increase of 1.75%. This was followed by two consecutive years of decline from 2021 to 2022 and from 2022 to 2023, with a cumulative decrease of 0.16% over the 2021–2023 period. This initial rise and subsequent decline closely mirror China’s macroeconomic transition: the early gains reflect the demographic and structural dividends of shifting from traditional extensive growth to innovation-driven development under the supply-side structural reforms, while the recent decline since 2021 points to the mounting external pressures of technological “chokepoints” and the diminishing marginal returns of domestic R&D investment in high-quality factor cultivation. A phased breakdown reveals that during the 12th Five-Year Plan period (2011–2015), the average annual growth rate was 0.23%, characterized as a “steady improvement period.” During the 13th Five-Year Plan period (2016–2020), the growth rate slowed slightly to 0.12%, entering a “moderated growth period.” In the first three years of the 14th Five-Year Plan period (2021–2023), the mean value shifted from increase to decrease, with an average annual decline of 0.04%, characterized as a “moderate decline period.” Therefore, while urban niche height in China’s NQPF development has maintained a long-term upward trajectory, it exhibits a phased pattern of “sustained rise followed by a modest decline in recent years.”

Second, a comparison of the niche height curves for the Eastern, Central, Western, and Northeastern economic zones in Figure 3 reveals that the Eastern region has long maintained an absolute leading position. Its energy-level state of NQPF factor resources is significantly higher than the national average, positioning it as the most favorable urban agglomeration for NQPF factor resources and occupying a core status. This sustained leadership is theoretically grounded in the “Matthew effect” of regional innovation systems: the Eastern region’s entrenched advantages in top-tier universities, national laboratories, and headquarters economies create a self-reinforcing cycle of high-end talent attraction and proprietary technology generation. The Central economic zone’s energy-level state of NQPF factor resources is slightly below the national average, while the Western and Northeastern economic zones occupy the third tier, significantly below the national average, representing the regions with the lowest energy levels of NQPF factor resources in China. The lag of the Northeast can be attributed to institutional lock-in and severe brain drain, whereas the West, despite strategic state support, still lacks the dense innovation networks necessary to autonomously cultivate high-energy factors. Thus, the energy-level state of NQPF factor resources across Chinese cities exhibits a gradient distribution pattern of “Eastern sustained leadership, Central secondary, Western and Northeastern lagging behind,” with the Eastern economic zone’s concentration of high-energy NQPF factor resources being particularly prominent.

Third, employing the tercile method, all cities are classified into three types of niche height: high-level, medium-level, and low-level.

Table 4. Distribution statistics of cities by niche height type.

Year	Number of Cities			Proportion of Cities (%)
	High-Level (H)	Medium-Level (M)	Low-Level (L)	High-Level (H)	Medium-Level (M)	Low-Level (L)
2010	38	25	220	13.43	8.83	77.74
2011	53	35	195	18.73	12.37	68.90
2012	58	41	184	20.49	14.49	65.02
2013	64	67	152	22.61	23.67	53.71
2014	78	83	122	27.56	29.33	43.11
2015	86	93	104	30.39	32.86	36.75
2016	90	111	82	31.80	39.22	28.98
2017	108	122	53	38.16	43.11	18.73
2018	113	131	39	39.93	46.29	13.78
2019	119	123	41	42.05	43.46	14.49
2020	132	120	31	46.64	42.40	10.95
2021	134	125	24	47.35	44.17	8.48
2022	129	123	31	45.58	43.46	10.95
2023	119	121	43	42.05	42.76	15.19

presents the distribution statistics of cities across these three types. The dynamic evolution reveals a fundamental structural reshaping of the energy-level state of NQPF factor resources. Historically, in 2010, there were 220 low-level cities, accounting for an overwhelming 77.74% of the total, representing a classic “pyramid” structure heavily skewed toward the bottom. However, this landscape underwent a rapid hollowing-out process at the base. The proportion of low-level cities plummeted to 15.19% (43 cities) by 2023. Concurrently, medium-level cities experienced substantial expansion, growing from just 25 cities in 2010 to 121 cities (42.76%) in 2023, thereby establishing themselves as the largest single group. Furthermore, high-level cities demonstrated a remarkable trajectory of persistent ascent, with their numbers increasing more than threefold from 38 in 2010 to 119 (42.05%) in 2023. By 2023, the combined proportion of medium- and high-level cities reached 84.81%, definitively ending the historical dominance of low-level cities. This dual transition is theoretically consistent with the “threshold effect” in regional economics: only after surpassing a critical mass of innovation infrastructure and human capital can cities upgrade their factor energy levels, moving from a low-quality resource pool to medium- and high-level ecosystems. This structural shift indicates that while the overall energy-level state of NQPF factor resources in Chinese cities has been steadily rising, the “medium-level to high-level” transition has become the critical descriptive and mechanistic pathway for future high-quality development, highlighting the urgency of overcoming technological bottlenecks to sustain this upward mobility.

In summary, during the study period, the urban energy-level state of NQPF factor resources in China generally maintained a stable upward trajectory, with a modest decline observed since 2021. In terms of spatial distribution, the Eastern economic zone has consistently remained the center of factor resource agglomeration, followed by the Central urban agglomeration, while the Western and Northeastern regions remain far below the national average. Typologically, the tercile method reveals a profound structural transition from a “low-level dominant” pyramid to a “medium-high level dual-peak” structure, with medium- and high-level cities accounting for 84.81% of the total by 2023, making the transition toward higher energy levels a critical evolutionary pathway.

3.4. Measurement and Evolutionary Patterns of Niche Overlap Characteristics

First, based on Equation (9), the urban niche overlap values for 2010–2023 are obtained. Figure 4 presents the mean niche overlap values for all cities and for the four major economic zones. According to the overall mean curve shown in Figure 4, the mean niche overlap increased from 0.8651 to 0.9023 over the study period, exhibiting a fluctuating upward trend with an average annual growth rate of 0.32%. A phased breakdown reveals that during the 12th Five-Year Plan period (2011–2015), the overlap increased from 0.8431 to 0.8519, with the average annual growth rate moderating to approximately 0.26%. This indicates that the intensity of inter-city competition for NQPF factor resources continued to rise, albeit at a decelerating pace. This intensifying competition in the early stages reflects the “convergent” industrial policies adopted by local governments rushing to seize emerging technology tracks, leading to homogenized resource allocation. During the 13th Five-Year Plan period (2016–2020), the average annual growth rate accelerated to 0.81%, marking a significant increase in the pace of competition. Entering the 14th Five-Year Plan period, the overlap value declined moderately to 0.9023 over 2021–2023, exhibiting a modest downward trend. The recent downturn suggests an initial optimization of resource allocation, potentially driven by the construction of a unified national market and central guidance for differentiated development to alleviate excessive homogenization. Therefore, the intensity of inter-city competition for NQPF factor resources in China exhibits an evolutionary pattern of “low-speed growth → medium- to high-speed growth → moderate decline.”

Second, as shown in the evolution curves of niche overlap for the four major economic zones in Figure 4, the overlap values for the Eastern, Central, and Northeastern regions all lie close to the national average. Prior to 2015, the curves for the three regions exhibited notable differentiation, with the Eastern and Northeastern economic zones significantly higher than the Central economic zone. From 2016 onward, the niche overlap values of the three regions essentially converged, aligning closely with the national average. Concurrently, over the study period, the Western economic zone’s niche overlap remained significantly below the national average; however, since 2015, the gap has narrowed considerably, with a pronounced catch-up effect. The persistently lower overlap in the West indicates that it has not yet been fully integrated into the core network of national factor competition, but its catch-up trajectory shows it is increasingly drawn into the competitive arena. Thus, in terms of inter-zone comparison for NQPF factor resources, all four economic zones—Eastern, Central, Northeastern, and Western—have demonstrated a steady upward trend. The spatial pattern, characterized by slightly higher overlap in the Eastern, Central, and Northeastern regions and slightly lower overlap in the Western region, remains evident, yet regional disparities have narrowed significantly.

Third, employing the tercile method, all cities are classified into three types of niche overlap: low-level, medium-level, and high-level, with

Table 5. Distribution statistics of cities by niche overlap type.

Year	Number of Cities			Proportion of Cities (%)
	High-Level (H)	Medium-Level (M)	Low-Level (L)	High-Level (H)	Medium-Level (M)	Low-Level (L)
2010	47	132	104	16.61	46.64	36.75
2011	8	119	156	2.83	42.05	55.12
2012	50	129	104	17.67	45.58	36.75
2013	22	114	147	7.77	40.28	51.94
2014	36	105	142	12.72	37.10	50.18
2015	28	109	146	9.89	38.52	51.59
2016	74	97	112	26.15	34.28	39.58
2017	110	93	80	38.87	32.86	28.27
2018	122	91	70	43.11	32.16	24.73
2019	126	88	69	44.52	31.10	24.38
2020	139	82	62	49.12	28.98	21.91
2021	162	72	49	57.24	25.44	17.31
2022	210	43	30	74.20	15.19	10.60
2023	187	46	50	66.08	16.25	17.67

presenting the distribution statistics for the three types over the study period. The dynamic evolution reveals a profound structural shift toward high-intensity competition. In 2010, the distribution was relatively dispersed, with medium-level cities dominating at 132 (46.64%) and low-level cities closely following at 104 (36.75%). However, the landscape underwent a dramatic hollowing-out of the middle and bottom tiers over the decade. By 2023, high-level cities surged to 187, accounting for an overwhelming 66.08% of the total, while medium- and low-level cities shrank to 46 (16.25%) and 50 (17.67%), respectively. This structural transition toward a “high-level dominant” overlap pattern indicates that inter-city competition for NQPF factors has entered a state of “homogenized involution.” Cities are increasingly competing for identical strategic resources—such as AI talent and new energy projects—within the same ecological niche, elevating the risk of zero-sum competition. From the perspective of niche theory, this excessive overlap necessitates urgent policy guidance for “niche differentiation” to avoid redundant investments and foster complementary, synergistic development pathways. How to avoid excessive competition among cities in this resource domain is gradually becoming an important issue in the spatial planning of China’s NQPF development.

In summary, inter-city competition for NQPF factor resources in China has maintained a steady upward trend over the study period, with regional disparities in competitive intensity continuing to narrow and converging over time. This evolution culminated in a structural transition toward “high-intensity competition dominance.” Avoiding excessive competition arising from homogenization and guiding differentiated urban development are increasingly becoming critical issues in accelerating NQPF development.

4. Niche-Related Determinants and Combinatorial Pathway Analysis of Urban NQPF

Building on the findings of Section 3, this section further employs regression analysis and combinatorial pathway approaches to identify the key influencing factors of urban NQPF development. The detailed procedures are as follows.

4.1. Empirical Analysis of the Nexus Between Niche Characteristics and Urban NQPF

4.1.1. Benchmark Regression and Robustness Checks

A multiple regression model is constructed with NQPF as the dependent variable and niche characteristics as the explanatory variables, as specified in Equation (10).

$$NQP{F}_{it}={\beta }_0+{\beta }_{1}widt{h}_{it}+{\beta }_{2}heigh{t}_{it}+{\beta }_3{C}_{it}+\gamma X_{it} +{\mu }_i +{\lambda }_t +{\epsilon }_{it}$$

(10)

where $NQP{F}_{it}$ denotes the new quality productivity level of city $i$ in year $t$; $widt{h}_{it}$, $heigh{t}_{it}$, and $C_{it}$ denote niche width, height, and overlap, respectively. $X_{it}$ is a vector of city-level control variables, including economic development level (log of per capita GDP), urbanization rate (urban population/total population), government intervention (fiscal expenditure/GDP), financial development (bank loans and deposits/GDP), population density (total population/administrative area), and openness to foreign investment (total imports and exports/GDP). ${\mu }_i$ and ${\lambda }_t$ are city fixed effects and year fixed effects, respectively. ${\epsilon }_{it}$ is the error term.

For model selection, the Hausman test (χ² = 56.78, p < 0.01) supports the fixed-effects model over the random-effects model; the F-test (Chow test) and Breusch-Pagan LM test both reject the pooled OLS model (p < 0.01). The Variance Inflation Factor (VIF) test shows that all VIF values are well below the critical threshold of 10 (mean VIF = 2.34, max VIF = 3.67), indicating that multicollinearity is not a serious concern.

For residual diagnostics, the Modified Wald test (χ² = 333,522.24, p < 0.01) indicates significant groupwise heteroskedasticity, and the Pesaran CD test (CD = 35.586, p < 0.01) indicates significant cross-sectional dependence, with an average absolute correlation coefficient of 0.506 among residuals, suggesting a relatively strong degree of cross-sectional association. To address these issues, we report cluster-robust standard errors at the city level in the main regressions and further employ Driscoll–Kraay standard errors in the robustness checks to simultaneously account for cross-sectional dependence, heteroskedasticity, and serial correlation.

Table 6. Benchmark regression and robustness check results.

Variable	(1)	(2)	(3)	(4)	(5)
	Cluster-Robust SE	Driscoll–Kraay SE	Lagged	Winsorized	Excl. Municipalities
width	0.122 ** (0.048)	0.122 *** (0.010)	0.148 *** (0.034)	0.158 *** (0.031)	0.173 *** (0.031)
height	1.385 *** (0.161)	1.385 *** (0.087)	0.971 *** (0.153)	1.362 *** (0.171)	1.320 *** (0.175)
C	−0.012 ** (0.006)	−0.012 * (0.006)	−0.011 ** (0.005)	−0.017 ** (0.008)	−0.017 ** (0.007)
pgdp	−0.014 *** (0.004)	−0.014 ** (0.005)	−0.018 *** (0.005)	−0.017 *** (0.004)	−0.017 *** (0.004)
urban	−0.045 *** (0.018)	−0.045 *** (0.004)	−0.026 ** (0.010)	−0.034 *** (0.012)	−0.025 ** (0.011)
gov	−0.047 *** (0.012)	−0.047 ** (0.019)	−0.055 *** (0.013)	−0.054 *** (0.012)	−0.051 *** (0.012)
finance	−0.001 ** (0.001)	−0.001 (0.001)	−0.002 * (0.001)	−0.002 (0.001)	−0.002 (0.001)
popden	0.000 *** (0.000)	0.000 *** (0.000)	0.000 (0.000)	0.000 *** (0.000)	0.000 *** (0.000)
open	−0.038 *** (0.014)	−0.038 *** (0.006)	−0.017 ** (0.007)	−0.029 *** (0.009)	−0.020 *** (0.006)
Constant	−1.389 *** (0.117)	−1.389 *** (0.081)	−1.005 *** (0.111)	−1.429 *** (0.127)	−1.434 *** (0.128)
Year FE	YES	YES	YES	YES	YES
City FE	YES	YES	YES	YES	YES
Observations	3962	3962	3679	3962	3906
Adjusted R2	0.734	0.734	0.692	0.749	0.759

Notes: (1) ***, **, and * denote significance at the 1%, 5%, and 10% confidence levels, respectively; (2) Driscoll–Kraay standard errors (maximum lag = 2) are reported in parentheses in Column (2); (3) Cluster-robust standard errors are in parentheses in all other columns.

reports the benchmark regression and robustness check results. Column (1) employs cluster-robust standard errors; Column (2) employs Driscoll–Kraay standard errors; Column (3) uses one-year-lagged explanatory variables to mitigate reverse-causality concerns; Column (4) applies 1% winsorization to all continuous variables to exclude the influence of extreme values; Column (5) excludes the four municipalities directly under the central government to test whether the results are driven by these special administrative cities. The results in Column (1) show that the estimated coefficients of niche width and niche height are 0.122 and 1.385, respectively, with niche width significant at the 5% level and niche height significant at the 1% level, indicating that both variables exert significantly positive effects on NQPF development. The coefficient of niche overlap is −0.012, significant at the 5% level, indicating that homogenized inter-city competition has an inhibitory effect on NQPF. Regarding control variables, population density exhibits a significant positive effect on NQPF (coefficient = 0.0001, p < 0.01), indicating that the scale effects and knowledge spillover effects arising from urban population agglomeration facilitate NQPF development. The coefficients of per capita GDP and government intervention are both significantly negative, reflecting the stage-specific characteristics of NQPF development—developed cities experience slower growth due to their higher baseline levels, while excessive government fiscal expenditure may crowd out NQPF development.

The robustness check results in Columns (2) through (5) further validate the reliability of the above conclusions. After employing Driscoll–Kraay standard errors, lagged explanatory variables, winsorization, and excluding municipalities, the coefficients of niche width and niche height remain significantly positive at the 1% level, and the coefficients of niche overlap remain negative with 5% or 10% significance levels. These results indicate that the promoting effects of niche width and height and the inhibiting effect of overlap are not driven by model specifications or particular samples, confirming the robustness of our findings.

4.1.2. Regional Heterogeneity Analysis

Given the significant disparities among China’s four major economic regions in terms of economic development levels, factor endowment structures, and policy support intensities, the impact of niche characteristics on NQPF may exhibit regional heterogeneity. To further reveal such heterogeneity, this study divides the full sample into four major economic zones—Eastern, Central, Western, and Northeastern—and conducts subsample regressions for each region.

Table 7. Regional heterogeneity regression results.

Variable	(1)	(2)	(3)	(4)
	Eastern Cities	Central Cities	Western Cities	Northeastern Cities
width	0.171 (0.146)	0.271 *** (0.048)	0.099 *** (0.024)	0.080 *** (0.057)
height	1.241 *** (0.300)	0.972 *** (0.212)	0.848 *** (0.143)	1.094 *** (0.166)
C	0.003 (0.023)	−0.000 (0.007)	−0.006 * (0.005)	−0.013 (0.013)
Constant	−1.111 *** (0.340)	−1.284 *** (0.165)	−0.807 *** (0.115)	−1.101 *** (0.169)
Controls	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
City FE	YES	YES	YES	YES
Observations	1204	1120	1162	476
R2	0.758	0.826	0.882	0.729

Notes: *** and * denote significance at the 1% and 10% confidence levels, respectively. Cluster-robust standard errors are in parentheses.

reports the regional subsample regression results.

For the Eastern region, the coefficient of width is 0.171, positive but statistically insignificant, while the coefficient of height is 1.241, significant at the 1% level. This indicates that further advances in NQPF among Eastern cities currently derive primarily from upgrading the internal energy level of factor resources rather than from expanding external acquisition channels. A plausible explanation is that, constrained by existing technological and institutional frontiers, the niches of Eastern cities are already relatively broad, so the marginal contribution of additional external resource acquisition tends to diminish, whereas the marginal return to raising factor energy levels remains substantial. This benefits from the region’s well-developed market systems, high-level openness, and dense agglomeration of innovation resources. For the Central region, the coefficient of width is 0.271, the strongest among the four regions, and the coefficient of height is 0.972, both significant at the 1% level. As a crucial hub connecting the East and the West and a major recipient of industrial transfers, the Central region exhibits the most prominent width effect, indicating that expanding external resource acquisition channels yields the greatest marginal contribution to NQPF development. The Central region is in a rapid catch-up phase, where the introduction of external resources plays a particularly critical role in driving NQPF growth. For the Western region, the coefficient of width is 0.099 and the coefficient of height is 0.848, both significant at the 1% level, and the coefficient of overlap is −0.006, weakly significant at the 10% level. The Western region as a whole is still in the early stages of catch-up development, where the marginal effects of niche dimensions have not yet been fully realized; however, preliminary signs of homogenized competition have emerged and warrant attention. For the Northeastern region, the coefficient of width is 0.080, the lowest among the four regions, but still significant at the 1% level; the coefficient of height is 1.094, significant at the 1% level, ranking second only to the Eastern region. The Northeastern region exhibits a relatively prominent height effect while the width effect is comparatively insufficient. This combination reflects the practical dilemma of the old industrial base in Northeast China—deep industrial foundations and research capabilities coexist with insufficient external resource acquisition and market vitality, constraining the rapid development of NQPF.

In summary, the regional subsample regression results exhibit significant heterogeneity. The width effect attains statistical significance only in the Central, Western, and Northeastern regions rather than in the Eastern region, among which the Central region exhibits the largest marginal effect of external resource acquisition. The ordering of height effects is “Eastern > Northeastern > Central > Western,” with internal factor energy level improvements playing a more pronounced role in promoting NQPF in the Eastern and Northeastern regions. The overlap effect is only weakly negative in the Western region and insignificant in other regions. These results indicate that the impacts of niche characteristics on NQPF differ significantly across regions, and each region should adopt differentiated development strategies based on its specific structural characteristics of niche endowments.

4.2. Dominant Combinatorial Pathways of Multi-Dimensional Niche Structures

The “Combinatorial Pathway Distribution” method is grounded in niche theory. It involves cross-classifying the discrete levels (low, medium, high) of the three niche characteristics—width, height, and overlap—to generate 3 × 3 × 3 = 27 combinatorial pathways. This method is highly applicable to this study because NQPF represents a multi-dimensional synergistic evolutionary process. Evaluating single dimensions in isolation cannot capture the complex structural configurations of urban development. By statistically examining combinatorial pathways, we can identify the dominant structural typologies associated with different NQPF gradients. This approach transforms complex multi-dimensional data into structural typologies that facilitate differentiated policy design, effectively complementing the associational findings from the panel regression by revealing how multiple variables co-occur in structural patterns.

Based on the findings from Section 3.1, Section 3.2, Section 3.3 and Section 3.4, the three niche characteristics of urban NQPF are matched to form typological combinations, yielding a total of 27 possible combinatorial pathways.

Table 8. Distribution statistics of urban NQPF niche combination pathways.

No.	Combination Type	Number of Cities in 2010			Number of Cities in 2023
		High-Level	Medium-Level	Low-Level	High-Level	Medium-Level	Low-Level
1	HHH	0	1	0	70	4	0
2	HHM	16	6	0	13	1	0
3	HHL	5	3	0	21	0	0
4	HMH	1	3	0	7	25	0
5	HMM	6	11	0	2	5	0
6	HML	1	2	0	5	5	0
7	HLH	0	5	1	0	1	0
8	HLM	3	6	6	0	1	0
9	HLL	0	5	2	0	0	0
10	MHH	0	0	1	3	4	0
11	MHM	0	1	2	1	0	0
12	MHL	0	1	2	1	1	0
13	MMH	0	0	0	0	45	2
14	MMM	0	0	0	0	12	2
15	MML	0	0	0	0	8	0
16	MLH	0	1	21	0	3	10
17	MLM	2	5	40	0	0	6
18	MLL	0	4	25	0	1	2
19	LHH	0	0	0	0	0	0
20	LHM	0	0	0	0	0	0
21	LHL	0	0	0	0	0	0
22	LMH	0	0	0	0	1	1
23	LMM	0	0	1	0	0	0
24	LML	0	0	0	0	0	1
25	LLH	0	0	13	0	0	11
26	LLM	0	0	27	0	0	3
27	LLL	0	0	54	0	0	5
Total	—	34	54	195	123	117	43

Notes: The codes H/M/L in the combination types denote high/medium/low levels of niche width, height, and overlap, respectively, in that order (e.g., HHM = high width, high height, medium overlap).

presents the distribution statistics of the pathway types observed in 2010 and 2023.

First, according to the distribution of combinatorial pathways for the 283 cities in 2010 shown in Table 8, there were 34 high-level, 54 medium-level, and 195 low-level cities. Among the 34 high-level cities, the predominant pathways were HHM (16 cities), HMM (6 cities), and HHL (5 cities). Thus, high width and height were the key pathways associated with high-level NQPF development, though overlap varied. Among the 54 medium-level cities, the most prevalent pathways were HMM (11 cities), HLM (6 cities), and HHM (6 cities), indicating a certain divergence in niche characteristics. Meanwhile, among the 195 low-level cities, the four most frequent pathway types were LLL (54 cities), MLM (40 cities), MLL (25 cities), and LLM (27 cities), together accounting for 74.87% of low-level cities. This indicates a widespread “low-level lock-in” effect—where cities find it difficult to break out of their current low-level development patterns—particularly struggling to advance their niche height and overlap. In summary, the LLL pathway contained the largest number of cities, accounting for 19.08% of the total.

Second, according to the 2023 distribution shown in Table 8, there were 123 high-level, 117 medium-level, and 43 low-level cities. Among the 123 high-level cities, the predominant pathways were HHH (70 cities), HHL (21 cities), and HHM (13 cities), together accounting for approximately 84.55% of high-level cities. This indicates that high-level cities have achieved a synergistic leap in niche width, height, and overlap. For medium-level NQPF cities, among the 117 cities, 45 belonged to the MMH pathway, ranking first, followed by HMH with 25 cities. Thus, MMH and HMH became the primary evolutionary pathways associated with medium-level NQPF development, underscoring that the elevation of niche overlap is a key associated factor. For low-level NQPF cities, the dominant pathways shifted to LLH (11 cities) and MLH (10 cities). Compared with 2010, the emergence of high overlap (H) in these low-level cities is particularly notable. Furthermore, in the 2023 distribution across all 283 cities, the HHH pathway contained the largest number of cities (74), followed by MMH (47) and HMH (32), with high-level NQPF cities dominating these groups. Compared with 2010, the transition from LLL and MLM dominance to HHH and MMH dominance represents an important characteristic associated with changes in urban NQPF pathways.

Therefore, during the study period, as China’s urban NQPF typological distribution transitioned from low level to medium-high level dominance, the main evolutionary pathway changes are as follows. First, the shift from “LLL and MLM” to “HHH and MMH” as the dominant combinatorial pathways is a critical change, with the transition of niche overlap from low/medium to high level being particularly noteworthy. Second, the key pathways associated with high-level NQPF cities have accumulated and simplified, evolving from the multi-pathway coexistence of “HHM + HMM” in 2010 to the absolute dominance of “HHH” in 2023, indicating a period of continuous agglomeration of high-level NQPF factors. Third, the rapid growth of the MMH pathway (from 0 to 45 cities) in medium-level cities reflects that the transition of niche overlap from medium to high is highly correlated with their gradient upgrading. Fourth, the dominant pathways for low-level NQPF cities shifted from LLL to LLH, confirming a process of quality improvement where overlap is continuously enhanced even at the low level.

5. Discussion

5.1. Interpretation and International Relevance

This study systematically analyzes the ecological niche characteristics and evolutionary pathways of urban NQPF in China. To clarify the novelty and positioning of our contribution, we situate these findings within the broader international scholarship on regional innovation systems, agglomeration economies, and evolutionary economic geography [34,35,36,37], to which the ecological-niche framing is closely related.

The empirical results reveal several important findings. First, the long-term upward trend in urban NQPF and the persistent Eastern lead are consistent with recent city-level and province-level measurements in China. The niche framework adds a different lens by separating external resource breadth, internal endowment, and inter-city similarity. The positive associations for niche width and niche height align with regional innovation systems research [34], which emphasizes that innovation depends on both local capability and connections to external knowledge and institutions. Niche width, measuring a city’s capacity to acquire external resources, corresponds to the concept of regional absorptive capacity and urban network externalities. Broader access to knowledge, capital, talent, and industrial inputs enlarges opportunities for recombination and reduces dependence on a narrow development path. Niche height, reflecting the internal energy-level state of factor resources, corresponds to agglomeration economies and regional path creation, where high-quality factor endowments drive cumulative causation through self-reinforcing cycles of innovation and investment. The observed spatial gradient pattern, characterized by Eastern leadership, Central catch-up, and Western and Northeastern lagging, is not merely an administrative phenomenon but a spatial manifestation of regional path dependence and core-periphery dynamics [35].

Second, regarding the negative association between niche overlap and NQPF, it is necessary to further explore its underlying mechanisms at a conceptual level. Although the regression design establishes association rather than strict causation, the consistent negative sign across multiple robustness checks invites plausible mechanism-based interpretation. The suppression of NQPF by high niche overlap can be interpreted through three distinct channels. The first is the crowding-out channel. High overlap implies intense competition among cities for identical new-quality factors, such as high-tech talent, specific digital investments, and strategic emerging industry projects, leading to diminishing marginal returns and crowding out productive investments. The second is the homogenization channel. High overlap reflects structural isomorphism. Instead of pursuing place-based differentiated pathways, cities converge on similar “hot” industries, leading to homogenized competition, price wars, and a loss of differentiation advantages. The third is the knowledge-spillover channel. While agglomeration economies typically suggest positive knowledge spillovers, when niche overlap is excessively high, local knowledge spillovers may become redundant rather than complementary. Breakthrough innovations require a degree of technological distance and related variety [36,37]. Without it, the marginal benefit of localized knowledge exchange drops, stifling true novelty. These interpretations remain mechanism-based explanations rather than tested mediation results, and future work should distinguish these channels with direct measures.

Third, the configuration analysis reveals a structural transformation in the combinatorial pathways of urban NQPF. The shift from LLL and MLM dominance in 2010 to HHH and MMH dominance in 2023 indicates that the evolution of NQPF is not a single-factor process but a multi-dimensional synergetic transition. Cities that successfully upgraded their NQPF levels achieved coordinated improvements across width, height, and overlap simultaneously. The emergence of MMH as a dominant pathway in medium-level cities further suggests that the transition of niche overlap from medium to high is a critical gateway for cities moving toward high-level NQPF. This finding aligns with the evolutionary economic geography perspective that regional development trajectories are shaped by the co-evolution of multiple dimensions rather than any single factor in isolation.

Furthermore, although the empirical context is specific to China’s transitional economy, the theoretical implications of these findings have broader relevance. The spatial evolution from low-level to medium-high-level dominance and the shift in combinatorial pathways represent a spatial manifestation of the transition from extensive growth to innovation-driven development. The challenge of optimizing resource allocation while avoiding homogenized competition is a universal dilemma faced by urban systems globally undergoing structural transformation. Regions in many countries face a related problem: they need access to mobile knowledge, capital, and talent while maintaining local capabilities and avoiding uniform industrial strategies. The width-height-overlap framework offers a transferable descriptive vocabulary for this balance. The framework proposed in this study does not require other countries to mechanically replicate China’s indicator system. Instead, researchers and policymakers in different national contexts may combine the ecological-niche framework with their own institutional arrangements, urban development stages, and data availability. In this sense, the China-based evidence presented in this study provides a reference for constructing locally adapted measures of innovation-driven urban productivity and for analyzing how such productivity transformation reshapes inter-city competitive and cooperative relationships in other developing economies.

5.2. Limitations and Future Research

While this study offers valuable insights, a candid acknowledgment of its boundaries is necessary to help readers calibrate the policy recommendations and guide future extensions.

First, regarding data aggregation, the study relies on prefecture-level aggregate data. The exclusion of county-level dynamics limits the analysis of intra-regional heterogeneity and micro-level spatial interactions. Future research could integrate county-level or firm-level data to explore the micro-driving factors of NQPF. Second, concerning the descriptive nature of the niche measures, although this study constructs a three-dimensional structural analysis framework, these measures primarily capture the macro-level states of resource acquisition, energy levels, and competitive intensity. They do not fully uncover the underlying micro-mechanisms or the dynamic interactions among specific agents within the region. Third, regarding causal identification, while the multiple regression and robustness check effectively identify the key driving effects, the study does not employ strict quasi-experimental designs (e.g., instrumental variables or difference-in-differences) to fully rule out potential endogeneity. However, given that the primary objective of this study is to characterize evolutionary patterns and identify major driving pathways rather than evaluate a specific policy’s causal impact, this limitation does not undermine the core conclusions of the study.

To address these constraints, we propose concrete extensions for future research: (1) Spatial econometric modeling: Future studies should employ Spatial Durbin Models (SDM) or Spatial Error Models (SEM) to disentangle the direct and indirect spatial spillover effects of NQPF dimensions across city networks. (2) Explicit mechanism testing: Future research could use mediation analysis to empirically test the three channels (crowding-out, homogenization, knowledge-spillover) through which niche overlap affects productivity. (3) Comparison with alternative regional schemes: Researchers could compare results using alternative regional classifications (e.g., mega-city regions like the Yangtze River Delta) to test the sensitivity of the framework.

6. Research Conclusions and Policy Implications

From an ecological niche perspective, this study evaluates city resource endowments, external access, and inter-city similarity and describes their associations and dominant configurations in relation to NQPF. The main conclusions are as follows.

6.1. Main Research Conclusions

First, the spatiotemporal evolution of urban NQPF in China exhibits multidimensional complexity. Temporally, NQPF maintained an overall upward trend from 2010 to 2023 with an average annual growth rate of 4.89%, yet the trajectory shifted from rapid growth during the 12th Five-Year Plan period to moderated growth in the 13th, and then to a slight decline in the initial years of the 14th. Spatially, a stable gradient pattern persists in which the Eastern region leads with a widening advantage, the Central region actively catches up and has essentially reached parity with the national average, while the Western and Northeastern regions remain relatively low. Typologically, low-level cities have withdrawn from their dominant position, high-level cities have become the largest group, and a pattern of high-level expansion, medium-level transition, and low-level shrinkage has taken shape.

Second, the niche characteristics of urban NQPF exhibit significant dynamic evolution and spatial differentiation. Niche width shows an overall upward trend following a three-phase trajectory of short-term adjustment, long-term growth, and recent momentum attenuation, with its typological distribution transforming from a pyramid-shaped into a high-level dominant structure. Regionally, the ordering shifted after 2018 from East above Central above Northeast above West to East above Central above West above Northeast. Niche height rose steadily before a modest recent decline, remaining highly concentrated in the Eastern region; medium- and high-level cities now jointly account for the vast majority, marking a transition toward a medium–high dual-peak configuration. Niche overlap increased in a fluctuating manner with narrowing regional disparities, and high-overlap cities have come to dominate, raising concerns about homogenized competition.

Third, urban NQPF development exhibits a dynamic pattern of niche linkage and pathway differentiation. Regression results indicate that niche width and height are significantly and positively associated with NQPF levels, whereas overlap exhibits a significant negative association. These associations display marked regional heterogeneity: the width association is strongest in the Central region, the height associations are most prominent in the Eastern and Northeastern regions, and the negative overlap association appears only weakly in the Western region. Combinatorial pathway analysis reveals a structural shift from LLL and MLM types to HHH and MMH types as the dominant configurations. For high-level cities, the HHH pathway has become absolutely dominant; for medium-level cities, MMH has grown into the primary route of gradient ascent; for low-level cities, the dominant pathway shifted from LLL to LLH. Overall, China’s urban NQPF follows a gradient upgrading trajectory characterized by high-level expansion, medium-level reinforcement, and low-level quality improvement.

6.2. Policy Implications

First, the spatial optimization of NQPF factor resource allocation should be prioritized, given that both niche width and height are positively associated with NQPF levels, yet currently exhibit a spatial hierarchy of East highest, Central second, and West and Northeast lowest. For the Eastern economic zone, which holds the most favorable position in both dimensions, policy should continue to support advanced labor force cultivation and cutting-edge technological R&D, reinforcing its leading role in high-quality factor agglomeration—while remaining alert to saturation signals reflected in its attenuating width growth. For the Central economic zone, where the width association is strongest nationally, improving the investment environment and strengthening spatial diffusion channels to attract factor spillovers from the East may yield greater marginal returns and thus merit priority in policy design. For the Western and Northeastern zones, accelerated catch-up should be pursued through factor flow attraction, preferential policy support, and talent and industrial transfer. Notably, the Northeastern region displays a comparatively strong height association alongside its weakest width association, implying that policies prioritizing external openness and market vitality, so as to convert its substantial innovation foundations into effective resources, may be especially valuable.

Second, differentiated development pathways should be encouraged to mitigate the risks arising from high niche overlap. Given the significant negative association between niche overlap and NQPF, and given that high-overlap configurations had come to dominate the national landscape by 2023, reducing inter-city resource redundancy deserves to be treated as a strategic direction. Cities at all levels should conduct comprehensive assessments of their NQPF factor resources, including systematic SWOT analyses, to formulate development plans aligned with their unique endowments. R&D-intensive cities should consolidate positions in high-end talent cultivation, commercialization of high-tech outcomes, and specialized enterprise incubation. Beyond such pathways, differentiated models such as green-factor-driven, high-tech-industry-driven, and digitalization-driven routes should be cultivated to achieve staggered, complementary development. Special caution is warranted in the Western region, where early signs of homogenized competition have begun to appear, and preemptive guidance against blindly chasing trending industries may help avoid redundant investment.

Third, inter-regional spillover channels and coordinated incentive policies should be reinforced to facilitate gradient upgrading and holistic improvement of urban NQPF. Enhanced support for the Central, Western, and Northeastern regions through fiscal transfers, tax incentives, and industrial policies should guide resources toward these areas, accelerating digital infrastructure construction and technological capacity building to narrow gaps with the East. Dedicated funds could support R&D in key core technologies, while collaboration between innovative enterprises in the East and research institutions in the Central and Western regions should be encouraged to achieve technology sharing and innovation synergy. Inter-regional education cooperation and talent exchange should be promoted through targeted assistance and joint training programs to cultivate and retain high-quality talent. Additionally, industrial transfer and absorption should be strengthened with improved supporting systems that facilitate chain extension and upgrading. Closer industrial synergy mechanisms within the Yangtze River Delta and Pearl River Delta urban agglomerations may serve as replicable reference models for raising regional NQPF levels and cultivating globally influential science and technology innovation centers.