How Can New Quality Productive Forces Reshape the Industrial Landscape?—The Dual Enabling Effects of Factor Endowments and Synergies

Abstract

China’s economy has shifted from high-speed development to high-quality development as the marginal benefits of the traditional growth model decline. This study explores how new quality productive forces (NQPFs) drive industrial structure upgrading via factor endowment upgrading and factor synergy, with the goal of addressing development bottlenecks. Using 2013–2024 provincial panel data, NQPFs are measured through the entropy weight-TOPSIS method (factor endowment upgrading) and a coupled coordination model (factor synergy), integrated into a comprehensive index, and analyzed via two-stage least squares (2SLS) regression. The empirical results show that China’s NQPFs are in their initial stage, with factors still in the process of accumulation and adaptation. NQPFs significantly promote industrial structure advancement and rationalization, with factor synergy outperforming factor endowment upgrading. Its promotion of advancement is universal, but its impact on rationalization is regionally heterogeneous, with inadequate transmission in underdeveloped and high-environmental-regulation provinces and ineffective dual mechanisms in manufacturing-led provinces. Notably, the driving effect of NQPFs on manufacturing upgrading is stronger than that on services. This study provides guidance for NQPF cultivation and contributes to the theory and Chinese practice of industrial structure upgrading from a dual-factor perspective.

1. Introduction

At present, China’s economy is undergoing a critical transition from high-speed growth to high-quality development, thus making industrial restructuring imperative. Domestically, China’s industrial structure has long been characterized by structural imbalances. According to data from China’s National Bureau of Statistics, 2024 Chinese manufacturing value added accounted for 24.9% of GDP—its share has ranked globally for 15 straight years. However, high-technology manufacturing value added makes up just 18.5% of total manufacturing, far below the levels of Germany and the U.S., indicating insufficient competitiveness in core technologies and high-end segments. Separately, 2024 productive services accounted for 52.9% of total service output, notably lagging behind the average of developed countries (J. Zhao & Tang, 2018; T. Zheng et al., 2019). Internationally, the reconstruction of global value chains is accelerating. Developed countries erect technical barriers through strategies such as technological blockades and a “small yard, high fence” approach; in turn, emerging economies leverage their advantages of lower factor costs to accelerate the acceptance of low- and medium-end industrial transfers. This dynamic has trapped China’s industrial sector in a competitive pattern of “being squeezed from both above and below” (Barro, 2016).

New Quality Productive Forces (NQPFs) are advanced productive forces driven by revolutionary technological breakthroughs and the rise in emerging industries, which transcend traditional growth boundaries and innovate value creation modes via the reconfiguration of production factor combinations (F. Gao, 2023; Huang et al., 2024). Distinct from old productive forces reliant on the extensive accumulation of traditional tangible factors, NQPFs are characterized by factor endowment upgrading and factor synergy enhancement: it takes scientific and technological innovation as the core driver, highlights data’s value as a new production factor and marks a fundamental shift in core production input composition (Oztemel & Gursev, 2020; X. Zhao & Wang, 2020), and leverages digital and intelligent technologies to build an inter-factor “strong linkage” network, thus unleashing multiplicative synergistic effects (Heo & Lee, 2019).

New Structural Economics serves as the core theoretical foundation for understanding NQPF’s generation and evolution. A central proposition of NSE is that an economy’s industrial structure and technological level are endogenously determined by its factor endowment structure, and sustainable growth is achieved only when industrial and technological choices align with the comparative advantages embedded in this structure (J. Y. Lin, 2009, 2011). A key insight is that productivity upgrading hinges on factor endowment upgrading. NSE further emphasizes that factor endowments cannot be directly converted into productive capacity, with the efficient linkage and allocation of factors acting as the critical intermediary (J. Lin & Monga, 2011)—a principle that underpins the efficient linkage of new factors in NQPF cultivation.

While partial consensus has been reached on NQPFs and industrial structure upgrading, D. Zhang and Li (2025) have demonstrated via threshold regression a significant nonlinear relationship between NQPFs and export structure upgrading, underscored by the lag and asynchrony inherent in industrial structure upgrading. In practice, talent mismatches and inefficient technology transfer constrain the driving efficiency of NQPFs (Bai et al., 2020; Jiang & Guo, 2022), and China’s regional imbalance presents both nationwide empowerment challenges and differentiated upgrading opportunities. Accordingly, this study addresses three core questions. What is the actual level of China’s NQPFs? Can the dual mechanism of factor endowment upgrading and synergy effectively drive industrial structure upgrading? What heterogeneities exist in the effects of this dual mechanism across regional contexts?

The literature falls primarily into two categories. The first category examines the drivers of industrial structure upgrading across multiple dimensions. With respect to capital factors, the accumulation scale and allocation efficiency are key enablers—facilitating industrial shifts toward capital- and technology-intensive sectors via equipment renewal, scale expansion, and high-value-added area layout. Policy factors operate via environmental regulation, industrial support, and factor market reform. Technological upgrading is the core driver—disruptive breakthroughs catalyze emerging industries, with technological transformation and spillovers reshaping traditional sectors and driving full industrial chain upgrading and intelligence. External factors focus on global value chain restructuring and international demand shifts—global division of labor adjustments drive local industries toward independent, controllable, high-value-added transformation, while international demand and regional industrial transfer guide industrial structures to match the external environment and comparative advantages.

The second category examines the mechanisms by which NQPFs influence industrial structure upgrading. Scholars argue that NQPF-driven industrial structure upgrading has the following three key characteristics: first, a shift in the interindustry structure from being dominated by traditional manufacturing to coordinated advanced manufacturing–modern service development, with strategic emerging and future industries expanding quickly (F. Gao, 2023); second, the formation of a high-efficiency, low-consumption intensive model by industrial production systems through deep integration of intelligent equipment and clean production tools (B. Zhu et al., 2019); and third, the digital-intelligent transformation of traditional industries and backward capacity elimination drive structural rationalization and upgrading (Y. Xu et al., 2024). From a research perspective, most existing theories focus on factor endowment upgrading (Shi et al., 2025) and local government attention allocation (Y. Xu et al., 2024). The importance of factor synergy for industrial upgrading is underscored by the reshaping of factor links by digital intelligence technologies (Ghasemaghaei & Calic, 2020). However, existing studies overlook inter-factor coupled relationships, rendering them ill-suited for complex system analysis.

To summarize, the literature has three key limitations. First, most studies focus only on factor endowment upgrading and lack a two-dimensional integrated “factor endowment–factor synergy” framework, making it difficult to explain the complex systematic logic of NQPF-driven industrial structure upgrading; second, methodologically, when measuring NQPFs, the effects of factor synergy are excluded, causing measurement bias and undermining mechanism testing accuracy; and third, contextually, inattention to regional development imbalance results in failure to identify the dual mechanism’s heterogeneous characteristics, causing misalignment with China’s regional development realities.

This study differs from and extends prior research in the following three key ways: (1) theoretically, factor endowment upgrading and synergy are first integrated into a unified framework, revealing the dual-mechanism logic of NQPFs for industrial structure upgrading and addressing single-mechanism limitations; (2) entropy weight-TOPSIS and coupled coordination models are innovatively combined, overcoming prior neglect of factor synergy during measurement to enable precise quantification of NQPFs; and (3) empirically, the elasticity coefficients of the two factors on industrial structure upgrading are estimated to clarify effect magnitudes aiding countries in overcoming traditional factor-driven bottlenecks, and the heterogeneous regional effects of the dual mechanism are examined, enhancing empirical generalizability.

2. Theoretical Analysis and Research Hypotheses

The essence of NQPFs lies in the two-dimensional superimposed effect of factor endowment upgrading and factor synergy enhancement, and its mechanism of action in driving industrial structure upgrading depends on the coordinated development of these two dimensions (Figure 1).

2.1. Mechanism of Upgrading NQPF Factor Endowments for Industrial Structure Upgrading

According to factor endowment theory and new structural economics, industrial structure upgrading evolves in tandem with changes in factor endowments. The scope of factor endowments has expanded to include profound shifts in their internal advantage structures, with the driving logic transitioning from factor quantity accumulation to the leapfrog improvement of factor quality and combination efficiency.

Upgrading the labor quality structure is pivotal to industrial restructuring. Historically, a large low-quality labor force has fueled China’s “demographic dividend,” enabling labor-intensive and rough-processing industries (Fu et al., 2020). Today, digitally skilled workers drive the quality and efficiency of traditional industries (Acemoglu & Restrepo, 2020; Peng & Feng, 2025); creative-thinking workers foster strategic emerging and future industries (Georgieff & Hyee, 2022); and highly knowledgeable, skilled workers handle complex professional tasks, propelling the rapid development of technology- and knowledge-intensive sectors (Acemoglu & Restrepo, 2018; Sheng & Montgomery, 2025).

New types of labor objects advance industrial structure upgrading through the following three key pathways: first, enterprises use in-depth data analysis to shift production from standardized to flexible, customized models—increasing resource utilization (Ghasemaghaei & Calic, 2020) and extending value chains toward “smile curve” ends to facilitate servitization (Chen & Srinivasan, 2024); second, digital products and services, with near-zero marginal costs and strong radiation capabilities, overcome traditional business logic and address “Baumol’s cost disease” in services (Paiola & Gebauer, 2020; Rong & Luo, 2023); and third, clean, renewable new factors (e.g., new energy and new materials) drive industries toward intensification, energy efficiency, and green development (Fan et al., 2019).

Intelligent production materials drive industrial upgrading multidimensionally—intelligent equipment optimizes processes and enhances efficiency/quality via autonomous regulation (Raisch & Krakowski, 2021) while helping enterprises cope with changes in the external environment, anticipate the direction of innovation, and cultivate new industries (Duan et al., 2019); high-precision equipment (e.g., integrated circuits) enables the deep processing of labor objects and the upgrading of value chains (Yuan et al., 2025); and digital platforms (e.g., the industrial internet) allow for building efficient information networks, fostering of virtual industrial clusters (L. Zhang et al., 2022), and strengthening of service penetration for productive services (Coreynen et al., 2017).

H1. 

Factor endowment upgrading by new quality productive forces (NQPFs) significantly promotes industrial structure upgrading.

2.2. Mechanism Through Which New Quality Productivity Factors Create Synergy to Promote Industrial Structure Upgrading

Productivity level and growth rate are not simple sums of factor endowments’ production capacity but are the result of factor aggregation and adaptation. Factor synergy is essentially a modern manifestation of Marx’s “wholeness of the labor process.” Unlike traditional productivity’s extensive growth via scale expansion, NQPFs rely on cutting-edge technological integration, which reshapes factor connections and allocation through new technologies, imposing greater demands on factor synergy.

First, technology acts as a key hub linking factors, building a multifactor collaborative industrial network for dynamic adaptation and precision production. Unlike traditional loose, linear factor connections, where technology merely serves as a production link (Acemoglu & Zilibotti, 2001), enterprises use digital–intelligent tools to digitize business information; laborers analyze these data to adaptively regulate production factors, uniting dispersed new factors into synergy (Ghasemaghaei & Calic, 2020). This situation shifts the factors from passive use to active adaptation, driving high-quality industrial upgrading.

Second, platform openness creates new factor synergy scenarios. Digital platforms (e.g., the industrial internet) break enterprise boundaries (Oztemel & Gursev, 2020); shared manufacturing allows idle factors to be leased on demand, transcending geographical/organizational limits, accelerating matching, reducing mismatches, and promoting industrial structure rationalization.

Finally, cross-domain convergence deepens factor synergy and reconfiguration. Single-domain factor combinations are replaced by interdisciplinary architectures, with cross-domain integration enabling functional complementarity (Han & Sohn, 2016)—e.g., manufacturing–information technology integration, spawning intelligent manufacturing. This approach overcomes disciplinary/industrial boundaries, shifting factors from single-functional application to multi-value creation and pushing industries up the value chain, facilitating industrial structure advancement (Cheng et al., 2025).

H2. 

The factor synergy of new quality productive forces (NQPFs) promotes industrial structure upgrading.

Figure 1. Mechanism of the effect of new quality productive forces on industrial structure upgrading.

Figure 1. Mechanism of the effect of new quality productive forces on industrial structure upgrading.

3. Methodology

3.1. Econometric Model Setting

To address the endogeneity between NQPFs and industrial structure upgrading, research by Qiao et al. (2024) is drawn on, and an instrumental variable (IV) is used as the interaction of a province’s lagged internet users and 1984 telephone sets per 10,000 people. For relevance, this IV integrates historical communication endowments with current digital penetration, closely aligning with NQPF’s connotation. For exogeneity, the 1984 telephone data are considered a historical exogenous variable, satisfying the exclusion restriction. Two-stage least squares (2SLS) regression reveals the causal effect, with the model specifications detailed below:

In the first stage, ordinary least squares (OLS) regression is employed to isolate the exogenous component of the core explanatory variables. The regression equation is specified as follows:

$${NQP}_{it}^{\prime }={\alpha }_0+{\gamma }_1{IV}_{it}+{\gamma }_c{X}_{it}+u_i+{\delta }_t+{\epsilon }_{it}$$

(1)

where i denotes the region, t denotes time, and ${IV}_{it}$ is the instrumental variable. ${NQP}_{it}^{\prime }$ represents the original values of the core explanatory variables, including the comprehensive NQPF level ${nqp}_{it}$, factor endowment upgrading ${fe}_{it}$, and factor synergy ${fs}_{it}$. $X_{it}$ denotes other control variables; $u_i$ and ${\delta }_t$ denote province and year fixed effects, respectively; and ${\epsilon }_{it}$ is the random error term.

In the second stage, an equation is subsequently constructed based on the fitted value $\widehat{{NQP}_{it}}$ to estimate the impact on industrial structure upgrading as follows:

$${ind}_{1,it}={\alpha }_0+{\alpha }_1\widehat{{NQP}_{it}}+{\alpha }_c{X}_{it}+u_i+{\delta }_t+{\epsilon }_{it}$$

(2)

$${ind}_{2,it}={\alpha }_0+{\beta }_1\widehat{{NQP}_{it}}+{\beta }_c{X}_{it}+u_i+{\delta }_t+{\epsilon }_{it}$$

(3)

where ${ind}_1$ denotes industrial structure advancement and ${ind}_2$ denotes industrial structure rationalization. Coefficients ${\alpha }_1$ and ${\beta }_1$ are used to measure the effects of the core explanatory variables on industrial structure upgrading.

3.2. Variable Selection

3.2.1. Core Explanatory Variables

In most studies, NQPFs are measured solely as factor endowment upgrading, failing to fully capture the NQPFs’ intrinsic characteristics and leading to potential measurement bias. To mitigate this bias, research by Y. Gao et al. (2025) and Liu and He (2024) is drawn on in this study to construct an NQPF indicator system (encompassing labor, means of production, and objects of production) and comprehensively measure NQPFs from the dual dimensions of factor endowment upgrading and synergy. Following Shao et al. (2024), the entropy weight-TOPSIS method is employed to measure factor endowment upgrading (fe), in which indicators are objectively weighted by variability and their enabling effect on productivity is reflected. Drawing on the methods of L. Guo and Liu (2025), the three factors are treated as subsystems; on the basis of entropy-derived subsystem scores, factor synergy (fs) is measured via the coupled coordination model, and the impact of inter-subsystem interactions and synergy on productivity system performance is quantified. Finally, the two indicators are multiplied (nqp = fe × fs) to obtain the comprehensive NQPF level (nqp), with the economic connotation of synergizing factor quality improvement and allocation efficiency optimization. This value reflects the regional NQPF’s actual capacity to drive industrial upgrading—a high fe with a low fs limits the effectiveness of new quality factors (low nqp), and a simultaneous improvement in fe-fs results in a significant increase in nqp, corresponding to the NQPF that truly drives industrial upgrading (

Table 1. Indicator system used to measure the level of new quality productive forces.

Factor	Subfactor	Indicator	Description	Features
Laborer	Worker skills	Educational attainment	Average number of years of education per capita	+
		Human capital structure	Average number of students enrolled in higher education per 100,000 people	+
			Number of R&D personnel in industrial enterprises above scale	+
			Employed persons in urban units in information transmission, software and information technology services/employed persons in urban units	+
	Labor efficiency	Per capita income	Average wage of private sector workers	+
	Worker awareness	Employment structure	Tertiary employment/total employment	+
		Enterprising spirit	Innovation and Entrepreneurship Index	+
Labor resources	Digital infrastructure	Digital infrastructure	Length of fiber optic cable lines/area	+
			Number of cell phone base stations/area	+
	Digital penetration	Digital user	Number of mobile internet users per 10,000 people	+
		Digital platform	Number of domains per 10,000 people	+
		Degree of enterprise digitization	Number of enterprises with e-commerce trading activities	+
	New production tools	Industrial robot penetration	Number of robots installed in each industry nationwide $\times$ percent of employment in each province	+
		Electronic information manufacturing	Integrated circuit (IC) production	+
	Energy consumption	Energy intensity	Energy consumption/GDP	−
		Energy consumption structure	New energy generation/total generation, where new energy includes nuclear, wind, and solar	+
Labor object	Data elements	Big data generation	Mobile internet access data traffic	+
			Number of internet broadband access ports per 10,000 people	+
	Ecological environment	Green coverage	Percentage of forest cover	+
		Air quality	Air quality index (AQI)	+
		Industrial waste management	Comprehensive utilization rate of general industrial solid waste	+

Note: The “+” sign indicates that the indicator is positively correlated with the level of new quality productive forces, while the “−” sign indicates a negative correlation.

).

3.2.2. Dependent Variable: Industrial Structure Upgrading

This study examines industrial structure upgrading from two dimensions: industrial structure advancement and industrial structure rationalization. With respect to industrial structure advancement, as factor quality improves, the industrial structure not only undergoes overall upgrading but also experiences enhanced internal efficiency across individual industries. Traditional indicators, such as changes in the proportion of the three major industries, fail to accurately capture this dynamic process. Therefore, in accordance with the work of Xia et al. (2024), the structural efficiency index is used as the indicator for measuring industrial structure advancement and is represented by Equation (4). For industrial structure rationalization, the methodology of J. Zheng et al. (2021) is followed, and the Theil index is used for measurement. Since the Theil index is a negative indicator, its inverse is used in subsequent analyses to facilitate interpretation and is presented in Equation (5).

$${ind}_{1i,t}={\sum }_{m=1}^3{\displaystyle \frac{Y_{i,m,t}}{Y}}\times {\displaystyle \frac{Y_{i,m,t}}{L_{i,m,t}}},m=\mathrm{1,2},3$$

(4)

$${ind}_{2i,t}=1-1/3\left({\sum }_{m=1}^3\left|y_{imt}-l_{imt}\right|\right)$$

(5)

where $y_{imt}$ denotes the share of value added of industry m in province i in period t and $l_{imt}$ denotes the share of employment in industry m in province i in period t.

3.2.3. Control Variables

Drawing on the prior literature, the following control variables are included to mitigate confounding effects: foreign direct investment (FDI) as a share of GDP, which measures openness (fdi); the urban population as a share of the total population, which proxies the urbanization level (urb); the logarithm of per capita GDP, which captures the economic development level (lngdp); government public expenditure as a share of GDP, which reflects the government intervention level (gov); financial sector GDP as a share of regional GDP, which gauges the financial development level (fin); and nonstate economy fixed asset investment as a share of total fixed asset investment, which reflects the marketization level (mar).

3.3. Data Sources and Descriptive Analysis

Thirty Chinese provinces (excluding Tibet, Hong Kong, Macao, and Taiwan) are selected for this study, with a sample period of 2013–2024. The data are sourced from the CSMAR Database, China Statistical Yearbook, China Environmental Statistical Yearbook, China Industrial Statistical Yearbook, China Tertiary Industry Statistical Yearbook, High-Technology Industry Statistical Yearbook, China Regional Innovation Capability Evaluation Report, International Federation of Robotics (IFR), and Air Quality Online Monitoring and Analysis Platform.

For the 2024 provincial panel data, linear interpolation is used to estimate select variables for which official figures have not yet been released. Interpolated observations account for 0.49% of the full sample. Descriptive statistics are presented in

Table 2. Descriptive statistics.

Variable	Observations	Mean	Std. Dev.	Min	Max
nqp	360	0.023	0.032	0.002	0.223
fe	360	0.117	0.099	0.029	0.610
fs	360	0.151	0.057	0.057	0.365
ind1	360	1.202	0.499	0.382	3.348
ind2	360	14.691	21.032	2.075	163.898
manu	360	0.212	0.109	0.041	0.581
serv	360	0.094	0.061	0.032	0.377
lngdp	360	10.942	0.453	9.849	12.207
gov	360	0.258	0.110	0.105	0.758
fin	360	0.073	0.032	0.020	0.201
fdi	360	0.016	0.016	0.000	0.121
mar	360	0.597	0.095	0.347	0.860
urb	360	0.612	0.116	0.364	0.896

, with variables showing a reasonable distribution and no significant outlier issues.

4. Results

4.1. Measurement of the Level of NQPFs

Table 3. Levels of new quality productive forces in China.

Province	fe	fs	nqp	Sorting Changes
Shanxi	0.057	0.118	0.007	4
Hebei	0.077	0.139	0.011	2
Anhui	0.094	0.200	0.015	2
Inner Mongolia	0.053	0.104	0.006	1
Liaoning	0.088	0.150	0.014	1
Hubei	0.095	0.156	0.016	1
Hunan	0.087	0.145	0.014	1
Guizhou	0.057	0.111	0.007	1
Shaanxi	0.085	0.145	0.013	1
Beijing	0.352	0.263	0.095	0
Tianjin	0.112	0.164	0.019	0
Jilin	0.064	0.122	0.008	0
Heilongjiang	0.059	0.115	0.007	0
Shanghai Municipality	0.239	0.235	0.058	0
Jiangsu	0.286	0.252	0.077	0
Zhejiang	0.203	0.222	0.048	0
Fujian	0.136	0.173	0.025	0
Jiangxi	0.074	0.131	0.010	0
Shandong	0.157	0.189	0.032	0
Henan	0.102	0.154	0.017	0
Guangdong	0.381	0.276	0.114	0
Guangxi Zhuang Autonomous Region	0.067	0.117	0.008	0
Chongqing Municipality	0.078	0.140	0.011	0
Sichuan	0.095	0.150	0.015	0
Xinjiang Uygur Autonomous Region	0.049	0.094	0.005	0
Yunnan	0.058	0.108	0.007	−1
Ningxia Hui Autonomous Region	0.053	0.101	0.006	−1
Hainan	0.082	0.122	0.011	−2
Qinghai	0.060	0.090	0.006	−4
Gansu	0.098	0.109	0.011	−6
Average	0.117	0.153	0.023

shows NQPFs, factor endowment upgrading (fe), and factor synergy (fs), which are all less than 0.5. In accordance with L. Guo and Liu’s (2025) factor synergy classification, China’s NQPFs remain in early development, with low factor synergy and factor endowment in an adjustment phase.

NQPF ranking shifts are concentrated in central–western provinces—those with improved rankings (e.g., Shanxi and Hebei) have strong fs despite moderate fe. These shifts are supported by central fiscal transfers for digital infrastructure (e.g., Shanxi’s 5G/base stations for coal mine digitalization) and regional policies enabling precise factor synergy (e.g., Hebei’s talent/equipment sharing via the Beijing–Tianjin–Hebei Strategy) that offset endowment shortcomings.

In contrast, central–western provinces with declining rankings (e.g., Gansu and Qinghai) face dual-fs constraints: (1) inadequate digital infrastructure with supply-demand mismatches (e.g., Qinghai’s capital-concentrated infrastructure misaligning with industrial needs); (2) weak industrial foundations (e.g., Gansu’s underdeveloped integrated circuit industry lacking core clusters), hindering synergy policies and limiting fe’s amplification effect.

Eastern coastal provinces exhibit balanced development, strong new quality factor absorption, sound institutions, and targeted policies. Fe and fs advance in tandem, ensuring stable rankings (Liu & He, 2024).

In conclusion, the asymmetry between fe and fs constitutes the key constraint on NQPF-driven industrial structure upgrading effectiveness.

NQPF factor development in 30 Chinese provinces is shown in Figure 2. In-depth analysis reveals significant disparities in factor levels, again indicating that factor synergy remains in a transitional phase.

A further analysis of the development rates of each factor reveals distinct disparities: the absolute levels of all three factors exhibit a sustained growth trend. Among these, the means of labor register the most prominent growth, rising steadily from approximately 0.008 in 2013 to 0.047 in 2024, a nearly 4.9-fold increase, and thus are identified as the most dynamic factor in the development of New Quality Productive Forces (NQPFs). The subject of labor shows the second-highest growth margin, surging from around 0.019 in 2013 to 0.058 in 2024, an increase of about 2.1 times. In contrast, the labor force records a relatively moderate growth, climbing from roughly 0.012 in 2013 to 0.028 in 2024, a mere 1.3-fold rise, which indicates a significant growth lag in the labor force.

This phenomenon stems from the following three key factors: (1) skilled labor cultivation’s long “education–practical training–on-the-job adaptation” cycle lags digital technology/smart manufacturing iteration (Graetz & Michaels, 2018); (2) insufficient vocational education investment and low social recognition limit skilled personnel scale/quality (Zeng, 2021); and (3) regional household registration/social security divisions worsen labor mobility imbalances and skilled talent outflows from central/western regions, trapping local labor skill upgrading in “unable to retain or attract talent (Stainback & Tang, 2019).

4.2. Benchmark Regression

Table 4. Impact of new quality productive forces on advanced industrial structure.

Variable	nqp	fe	fs	ind1	ind1	ind1	ind2	ind2	ind2
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
IV	1.089 ***
	(0.035)
IV		1.017 ***
		(0.032)
IV			1.015 ***
			(0.021)
nqp				3.158 ***			195.551 ***
				(0.720)			(42.456)
fe					1.261 ***			80.376 ***
					(0.265)			(17.644)
fs						5.026 ***			274.640 ***
						(0.784)			(58.307)
Province Control	YES	YES	YES	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES	YES	YES	YES
KP-LM	20.676	27.190	49.473	-	-	-	-	-	-
KP-F	969.887	999.096	2377.757	-	-	-	-	-	-
N	360	360	360	360	360	360	360	360	360
R2	-	-	-	0.953	0.952	0.955	0.762	0.760	0.764

Note: *** indicate significance at 1% levels, respectively.

presents the results for empirical models (1)–(3). The Kleibergen–Paap rk Wald F statistic (16.38) from the first-stage instrumental variable (IV) regression confirms the absence of weak IV issues.

Columns (4)–(6) show significantly positive coefficients for industrial structure upgrading at the 1% level—comprehensive NQPF (nqp = 3.158), factor endowment upgrading (fe = 1.261), and factor synergy (fs = 5.026). A one-standard-deviation increase in each drive increases by 0.20, 0.25, and 0.58 standard deviations, respectively. These results support Hypothesis 1, with the coefficient of fs being significantly greater than that of fe, indicating that factor synergy has a more prominent marginal contribution.

For industrial structure rationalization, Columns (7)–(9) report 1% significantly positive coefficients as follows: nqp = 195.551, fe = 80.376, and fs = 274.640. The corresponding one-standard-deviation increases drive rationalization by 0.30, 0.38, and 0.75 standard deviations. These findings confirm Hypothesis 2 and reinforce factor synergy as the key driver of industrial structure optimization, given its largest coefficient and standard deviation effect.

A comparison of the core variables reveals that the marginal contribution of factor synergy outperforms that of fe in terms of both upgrading and rationalization. Comprehensive NQPF is constrained by asynchrony between fe and fs development, which highlights that addressing this asymmetry is critical to unlocking the effectiveness of NQPFs.

4.3. Robustness Tests

(1) The core explanatory variables are replaced with their first-order lags, and OLS regression is performed. The results are not significantly different from those of the benchmark, confirming the robustness of the main conclusions (

Table 5. Regressions using lagged terms of new quality productive forces as explanatory variables.

Variable	ind1	ind1	ind1	ind2	ind2	ind2
	(1)	(2)	(3)	(4)	(5)	(6)
nqp	3.518 ***			224.714 ***
	(0.889)			(57.748)
fe		1.336 ***			88.951 ***
		(0.308)			(22.743)
fs			5.363 ***			290.423 ***
			(0.896)			(71.643)
Controls	YES	YES	YES	YES	YES	YES
Province Control	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES
N	330	330	330	330	330	330
R2	0.954	0.953	0.958	0.781	0.779	0.782

Note: *** indicate significance at 1% levels, respectively.

).

(2) OLS regression with all explanatory variables lagged one period yields results consistent with the benchmark (

Table 6. All explanatory variables lagged one period.

Variable	ind1	ind1	ind1	ind2	ind2	ind2
	(1)	(2)	(3)	(4)	(5)	(6)
nqp	2.892 ***			226.801 ***
	(0.776)			(54.368)
fe		1.091 ***			92.028 ***
		(0.279)			(22.077)
fs			4.042 ***			320.776 ***
			(0.800)			(72.852)
Controls	YES	YES	YES	YES	YES	YES
Province Control	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES
N	330	330	330	330	330	330
R2	0.956	0.955	0.957	0.772	0.772	0.776

Note: *** indicate significance at 1% levels, respectively.

).

(3) Instrumental variable placebo test: Historical communication infrastructure may directly affect industrial upgrading through unobserved regional advantages. Coastal provincial and municipal samples are thus excluded, and IV regression is re-estimated for inland provinces. The results are highly consistent with the baseline regression, effectively invalidating the alternative explanation that instrumental variables impact industrial upgrading via omitted factors such as coastal advantages. The exogeneity of the 1984 telephone data is further verified: the National Bureau of Statistics of China reports a national average of only 0.53 telephones per 100 people in 1984. The highly administrative allocation of fixed-line telephones and their layout reflecting planned-economy resource priorities render this historical data largely uncorrelated with contemporary unobserved drivers of industrial structure upgrading, thereby satisfying the exclusion restriction (

Table 7. Instrumental variables falsification test.

Variable	ind1	ind2	ind1	ind2	ind1	ind2
	(1)	(2)	(3)	(4)	(5)	(6)
nqp	4.654 ***	72.720 *
	(0.890)	(42.539)
fe			22.509 **	44.647 ***
			(20.734)	(12.367)
fs					8.337 ***	1.528 **
					(1.956)	(0.604)
Controls	YES	YES	YES	YES	YES	YES
Province Control	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES
KP-LM	47.763		56.948		66.312
KP-F	507.671		140.057		1008.832
N	204	204	204	204	204	204
R2	0.958	0.817	0.957	0.826	0.960	0.818

Note: *, **, *** indicate significance at 10%, 5%, and 1% levels, respectively.

).

(4) Variable replacement: Core explanatory variables are remeasured via criterion importance through the intercriteria correlation (CRITIC) weighting method, with 2SLS re-regression. The core coefficients are positive and 1% significant, confirming robustness (

Table 8. Regressions after replacing the new quality productive forces measures.

Variable	ind1	ind1	ind1	ind2	ind2	ind2
	(1)	(2)	(3)	(4)	(5)	(6)
nqp	8.439 ***			522.571 ***
	(1.837)			(110.336)
fe		6.448 ***			411.128 ***
		(1.466)			(92.824)
fs			10.572 ***			577.622 ***
			(1.641)			(125.288)
Province Control	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES
KP-LM	24.020	22.885	51.415	24.020	22.885	51.415
KP-F	300.446	203.826	669.859	300.446	203.826	669.859
N	360	360	360	360	360	360
R2	0.955	0.945	0.952	0.762	0.732	0.748

Note: *** indicate significance at 1% levels, respectively.

).

(5) Estimation method replacement: The generalized method of moments (GMM) addresses potential 2SLS heteroskedasticity. The results align with the benchmark, passing robustness tests (

Table 9. GMM regression with lagged terms of new quality productive forces as instrumental variables.

Variable	ind1	ind1	ind1	ind2	ind2	ind2
	(1)	(2)	(3)	(4)	(5)	(6)
nqp	3.158 ***			195.551 ***
	(0.720)			(42.456)
fe		1.261 ***			80.376 ***
		(0.265)			(17.644)
fs			5.026 ***			274.642 ***
			(0.784)			(58.307)
Province Control	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES
N	360	360	360	360	360	360
R2	0.953	0.952	0.955	0.762	0.760	0.764

Note: *** indicate significance at 1% levels, respectively.

).

(6) Descriptive statistics show generally reasonable variable distributions, but ind₂’s maximum (163.898) may affect regression robustness. To address this concern, 1% and 2% winsorization are performed, as presented in

Table 10. Regression with sample outliers removed.

Variable	1% Before and After Indent			2% Before and After Indent
	ind2	ind2	ind2	ind2	ind2	ind2
	(1)	(2)	(3)	(4)	(5)	(6)
nqp	184.169 ***			133.180 ***
	(35.796)			(21.077)
fe		75.638 ***			52.755 ***
		(15.118)			(9.416)
fs			264.217 ***			174.765 ***
			(53.258)			(28.949)
Province Control	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES
KP-LM	20.676	27.190	49.473	20.676	27.190	49.473
KP-F	969.887	999.096	2377.757	969.887	999.096	2377.757
N	360	360	360	360	360	360
R2	0.797	0.795	0.800	0.911	0.909	0.913

Note: *** indicate significance at 10%, 5%, and 1% levels, respectively.

, and reverse winsorization is also conducted, as shown in

Table 11. Logarithmic regression on ind₂.

Variable	(1)	(2)	(3)
	ln(ind2)	ln(ind2)	ln(ind2)
Nqp	3.488 ***
	(1.036)
Fe		1.294 ***
		(0.480)
Fs			4.861 ***
			(1.303)
Province Control	YES	YES	YES
Year Control	YES	YES	YES
KP-LM	20.676	27.190	49.473
KP-F	969.887	999.096	2377.757
N	360	360	360
R2	0.952	0.951	0.952

Note: *** indicate significance at 1% levels, respectively.

. The core coefficients align with the baseline in sign and significance, with no substantial outlier distortion, further confirming NQPFs’ robustness in promoting industrial structure upgrading.

(7) Robustness to interpolated data: To gauge the potential influence of interpolated values on the core results, the 2024 observations are excluded from the analysis, and the model is re-estimated. The coefficients for the core explanatory variables remain consistent with the baseline regression in terms of sign, statistical significance, and economic magnitude. These results confirm that the smoothing bias from linear interpolation does not materially alter the conclusions (

Table 12. Smoothing deviation test for interpolation.

Variable	ind1	ind1	ind1	ind2	ind2	ind2
	(1)	(2)	(3)	(4)	(5)	(6)
nqp	3.652 ***			231.419 ***
	(0.854)			(62.789)
fe		1.487 ***			97.020 ***
		(0.317)			(25.773)
fs			5.943 ***			334.244 ***
			(0.988)			(87.162)
Controls	YES	YES	YES	YES	YES	YES
Province Control	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES
KP-LM	15.429	19.971	37.647	15.429	19.971	37.647
KP-F	457.485	520.140	1269.995	457.485	520.140	1269.995
N	330	330	330	330	330	330
R2	0.954	0.953	0.957	0.772	0.770	0.774

Note: *** indicate significance at 1% levels, respectively.

).

4.4. Heterogeneity Test

4.4.1. Heterogeneity Analysis at the Level of Economic Development

The level of economic development influences NQPFs, which in turn impacts industrial structure upgrading. Using the provincial average per capita GDP median and average (sample period) as a cutoff, the sample is split into developed and less developed sub-samples (

Table 13. Heterogeneity analysis of regional economic development.

Variable	Median Grouping						Average Value Grouping
	Economically Developed Provinces		Less Economically Developed Provinces				Economically Developed Provinces		Less Economically Developed Provinces
	ind1	ind2	ind1	ind2	ind2	ind2	ind1	ind2	ind1	ind2	ind2	ind2
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
nqp	2.669 ***	182.330 ***	5.847 **	61.975			2.534 ***	206.632 ***	7.581 ***	−59.316
	(0.774)	(58.276)	(2.553)	(78.712)			(0.722)	(56.795)	(2.290)	(85.466)
fe					−20.890						−44.177
					(19.028)						(20.164)
fs						133.696 ***						104.614 ***
						(33.403)						(36.726)
Controls	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Province Control	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
KP-LM	23.190	23.190	35.611	35.611	41.846	58.965	22.845	22.845	34.676	34.676	41.336	56.174
KP-F	824.406	824.406	211.700	211.700	106.916	604.329	844.411	844.411	173.596	173.596	98.271	490.323
N	180	180	180	180	180	180	192	192	168	168	168	168
R2	0.943	0.747	0.962	0.839	0.839	0.858	0.944	0.750	0.967	0.847	0.852	0.858

Note: **, *** indicate significance at 5% and 1% levels, respectively.

).

For industrial structure advancement, the coefficient of nqp on ind₁ in less developed provinces (5.847, p < 0.05) is 2.2 times greater than that in developed provinces (2.669, p < 0.01), indicating a stronger marginal effect of NQPFs. This finding may stem from the following two factors: first, less developed provinces have a low industrial upgrading base, and thus, new industries driven by NQPFs can more rapidly alter the composition of industrial structure, generating a more pronounced marginal improvement, and second, when undertaking industrial transfers from eastern regions, these provinces can leverage the NQPFs to skip certain traditional upgrading stages, accelerating the shift to high-value-added industries and fostering a “latecomer advantage” (Z. Gao et al., 2024).

With respect to rationalization, nqp has a significant positive effect in developed provinces but a nonsignificant effect in less developed provinces. Mechanistically, the coefficient of fe (−20.890) suggests that factor endowment upgrading hinders rationalization in less developed provinces, reflecting a weak foundation for NQPF cultivation that limits the realization of its benefits (Z. Guo et al., 2026).

4.4.2. Heterogeneity Analysis at the Level of Dominant Types of Regional Industries

The existing foundation of regional industrial development serves as the fundamental driver of industrial structure upgrading. The GDP ratio of the secondary and tertiary industries in each province is used as the sample classification criterion (Bi et al., 2025)—provinces with the highest share of secondary industry are classified as manufacturing-led, while those with the highest share of tertiary industry are considered service-led (

Table 14. Heterogeneity analysis of the dominant types of regional industries.

Variable	Median Grouping						Average Value Grouping
	Manufacturing-Led Provinces				Service-Led Provinces		Manufacturing-Led Provinces				Service-Led Provinces
	ind1	ind2	ind2	ind2	ind1	ind2	ind1	ind2	ind2	ind2	ind1	ind2
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
nqp	13.693 ***	−22.821			2.699 ***	191.988 ***	12.162 ***	−42.953			2.479 ***	180.546 ***
	(2.237)	(50.501)			(0.676)	(50.476)	(2.090)	(47.850)			(0.666)	(51.560)
fe			−13.455						−18.002
			(12.784)						(12.333)
fs				11.861						−2.072
				(25.845)						(24.212)
Controls	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Province Control	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
KP-LM	35.839	35.839	29.516	50.031	20.146	20.146	47.455	47.455	39.341	63.981	19.819	19.819
KP-F	170.715	170.715	98.672	576.818	992.009	992.009	218.296	218.296	124.802	762.622	1016.871	1016.871
N	180	180	180	180	180	180	204	204	204	204	156	156
R2	0.962	0.869	0.869	0.869	0.956	0.748	0.961	0.869	0.869	0.868	0.954	0.745

Note: *** indicate significance at 1% levels, respectively.

).

For industrial structure advancement, nqp’s coefficients on ind₁ are positive and significant for both subsamples, with a stronger marginal effect observed in manufacturing-led provinces. In contrast, service-led provinces exhibit a smaller marginal effect because of their already higher overall industrial level, which limits the space for NQPF-driven improvement.

For industrial structure rationalization, nqp’s coefficient on ind₂ in manufacturing-led provinces is −22.821 (p > 0.1), with similarly insignificant coefficients for fe (−13.455) and fs (11.861). This finding contrasts with service-led provinces, where the nqp coefficient is positive and significant (191.988, p < 0.01), indicating that factor endowment upgrades and factor synergy mechanisms do not operate effectively in manufacturing-led provinces. The reason for this is likely that long-term manufacturing dominance has rendered their industrial structure and production organization mature and rigid, with high adjustment costs that hinder the rapid integration of new technologies and production tools, limiting the marginal effect of the release of NQPFs (Jia et al., 2025).

4.4.3. Heterogeneity Analysis at the Level of Regional Environmental Regulatory Intensity

Environmental regulation, a key government industrial intervention tool, varies regionally in terms of intensity, influencing industrial survival thresholds and new quality factor input directions, and serves as a critical contextual variable for industrial upgrading heterogeneity. Drawing on Pan et al. (2019), environmental regulation intensity is measured as the ratio of regional industrial pollution control investment to industrial value added. The sample is split into provinces with high and low levels of environmental regulation on the basis of the indicator’s median and average accordingly (

Table 15. Heterogeneity analysis of regional environmental regulatory intensity.

Variable	Median Grouping						Average Value Grouping
	Provinces with High Levels of Environmental Regulation				Provinces With Low Levels of Environmental Regulation		Provinces with High Levels of Environmental Regulation				Provinces with Low Levels of Environmental Regulation
	ind1	ind2	ind2	ind2	ind1	ind2	ind1	ind2	ind2	ind2	ind1	ind2
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
nqp	2.779 **	23.612			1.876 ***	253.714 ***	4.639 ***	8.989			2.242 ***	214.349 ***
	(1.233)	(49.964)			(0.586)	(55.796)	(1.498)	(48.408)			(0.602)	(51.732)
fe			−14.417						−3.996
			(14.456)						(14.377)
fs				110.531 **						49.817 *
				(45.213)						(30.088)
Controls	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Province Control	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
KP-LM	15.578	15.578	19.945	17.208	20.097	20.097	19.176	19.176	23.524	33.210	19.628	19.628
KP-F	402.156	402.156	468.795	430.741	882.264	882.264	390.387	390.387	476.800	845.217	975.196	975.196
N	180	180	180	180	180	180	168	168	168	168	192	192
R2	0.964	0.890	0.891	0.893	0.971	0.763	0.960	0.897	0.898	0.898	0.970	0.773

Note: *, **, *** indicate significance at 10%, 5%, and 1% levels, respectively.

).

With respect to industrial structure advancement, the nqp coefficients are significantly positive across both subsamples, with a stronger marginal effect in provinces with high levels of environmental regulation. The reason for this is that strict environmental standards in high-regulation provinces impose pressure on the survival of high-energy-consumption, high-pollution, and low-end industries, accelerating the phase-out of low-end capacity and freeing up resource space for high-value industries (Dagestani et al., 2023).

With respect to industrial structure rationalization, nqp has a positive effect only in provinces with low levels of environmental regulation and has no significant effect in provinces with high levels of regulation. Specifically, the fe coefficient on ind₂ in high-regulation provinces is −14.417 (p > 0.1). The reason for this is likely that low-regulation provinces face fewer constraints, with active innovation and flexible industrial development paths (Yu et al., 2017), which provide space for the convergence and synergy of new quality factors, effectively promoting structural rationalization. In contrast, enterprises in high-regulation provinces must invest substantial resources to meet environmental requirements; rising compliance costs crowd out new factor input allocation, weakening the effectiveness of NQPFs in driving structural rationalization (You et al., 2019).

4.5. Further Analysis: Industry Heterogeneity

Having confirmed the ability of NPQFs to promote industrial structure upgrading, it is noted that the role of NPQFs extends beyond interindustry transitional leaps to intra-industry structural dynamic evolution. To fully reveal the effectiveness of industrial structure upgrading, the heterogeneous impact of the NQPFs on intra-industry upgrading is examined. Notably, the primary industry (agriculture, forestry, animal husbandry, and fisheries) lacks a clear structural hierarchy, with NQPFs manifesting more production efficiency improvement than does intra-industry upgrading. Thus, this study focuses on manufacturing and services, builds empirical models (6) and (7), and employs 2SLS regression.

$${manu}_{it}={\alpha }_0+{\phi }_1\widehat{{NQP}_{it}}+{\phi }_c{X}_{it}+u_i+{\delta }_t+{\epsilon }_{it}$$

(6)

$${serv}_{it}={\alpha }_0+{\tau }_1\widehat{{NQP}_{it}}+{\tau }_c{X}_{it}+u_i+{\delta }_t+{\epsilon }_{it}$$

(7)

Here, ${manu}_{it}$ denotes manufacturing sector structural upgrading, and ${serv}_{it}$ denotes service sector structural upgrading. Coefficients ${\phi }_1$ and ${\tau }_1$ are the core estimators of interest and reflect the impact of NQPFs on structural upgrading in the manufacturing and service sectors, respectively.

With respect to the selection of dependent variables, the work of F. Xu and Hu (2025) is drawn on in this study, and the share of operating income of technology-intensive manufacturing industries is used to measure manufacturing structural upgrading. Following the suggestions of C. Guan et al. (2023), the share of employment in producer services relative to the tertiary industry is adopted to measure service sector structural upgrading (

Table 16. Role of new quality productivity in structural upgrading within industries.

Variable	manu	Manu	manu	serv	serv	serv
	(1)	(2)	(3)	(4)	(5)	(6)
nqp	0.472 ***			0.141 ***
	(0.145)			(0.040)
fe		0.194 ***			0.042 **
		(0.052)			(0.017)
fs			0.761 ***			0.305 ***
			(0.164)			(0.057)
Province Control	YES	YES	YES	YES	YES	YES
Year Control	YES	YES	YES	YES	YES	YES
KP-LM	20.676	27.190	49.473	20.676	27.190	49.473
KP-F	969.887	999.096	2377.757	969.887	999.096	2377.757
N	360	360	360	360	360	360
R2	0.945	0.945	0.946	0.976	0.975	0.977

Note: **, *** indicate significance at 5% and 1% levels, respectively.

).

The regression results indicate that the coefficients of the NQPF composite level (nqp) on manufacturing industry structural upgrading (manu) and service industry structural upgrading (serv) are 0.472 and 0.141, respectively (both p < 0.01). This finding confirms that NQPFs promote the internal structure upgrading of both industries, with a more pronounced effect observed in manufacturing.

This disparity stems primarily from two factors. First, manufacturing features clear production processes and technologies, enabling more readily achievable automated and intelligent transformation. Its strong upstream–downstream industrial chain linkages also facilitate the diffusion of upgrading effects (T. Zhu et al., 2025). Second, the service industry adopts more flexible service models, as NQPFs tend to enhance service efficiency rather than alter internal structural proportions. Additionally, producer services rely on manufacturing demand, with their upgrading pace constrained by manufacturing development, further limiting the magnitude of NQPFs’ impact on the internal structural upgrading of the service industry (Schiavone et al., 2022).

5. Discussion

5.1. Theoretical Framework Breakthroughs and Paradigm Contributions

Most of the literature focuses on a single dimension of factor endowment upgrading, reducing the NQPF mechanism to a linear logic of “factor quality upgrading drives industrial upgrading” (Shao et al., 2024). This finding fails to explain why industrial upgrading lags despite significant progress in factor endowment upgrading. Drawing on Marx’s holistic labor process theory and the factor allocation logic of new structural economics, this study incorporates factor synergy into the analytical framework for the first time, proposing a two-dimensional mechanism of “factor quality enhancement and factor relationship optimization.” The former serves as the foundational support for NQPFs, while the latter acts as its amplification mechanism—its degree of synergy determines the level of industrial upgrading momentum released. This framework aligns with the inherent nature of NQPFs and overcomes the limitations of existing studies that focus solely on factor quantity and quality, providing a systematic paradigm for understanding the complex relationship between NQPFs and industrial structure upgrading.

5.2. Commonalities and Differences in Empirical Findings

The core conclusion—that NQPFs significantly promote industrial structure upgrading and rationalization—aligns with the literature, confirming the universal value of NQPF.

Nevertheless, the following three core differences emerge in the empirical findings. (1) The literature simplifies the mechanism to “factor endowment upgrading drives industrial upgrading” (Shao et al., 2024), while this study confirms that the marginal contribution of factor synergy significantly outperforms that of factor endowment upgrading. These findings reveal that insufficient factor synergy is the key reason industrial upgrading lags despite advances in factor endowment upgrading. (2) Y. Zhu et al. (2025) only report that NQPF levels are higher in eastern China, but this study further reveals that the core distinction across the eastern, central, and western regions lies in the alignment between factor endowment upgrading and factor synergy. The two factors advance in tandem in the eastern region, whereas the central and western regions face greater asymmetric development rather than merely an insufficient single-factor scale. (3) Most prior studies argue that environmental regulation can drive industrial structure upgrading (S. Guan et al., 2022), but this study reveals that the impact of NQPFs is nonsignificant in provinces with high environmental regulation intensity. This finding is rooted in the lack of alignment between environmental and industrial policies—high compliance costs crowd out enterprise investments in factor cultivation, thereby weakening the effectiveness of institutional support.

5.3. Implications for Non-Chinese Emerging Economies

This study’s framework and conclusions provide the following key path guidance for emerging economies grappling with fading factor dividends and industrial low-end lock-in: these economies should abandon the traditional single-factor catch-up path and adopt a two-pronged “factor endowment upgrading and factor synergy” driving model. While new factors accumulate, they should concurrently establish factor synergy platforms to avoid the predicament of “possessing factors but failing to unlock their effective potential.”

This path is particularly critical for low-cost factor-dependent economies such as those in Southeast Asia and Latin America. These economies have long been trapped in low-end manufacturing (Kikuchi, 2018), with the root cause likely lying in their focus solely on the quantitative accumulation of factors (e.g., labor and capital) while neglecting the need to adapt and link these factors. The dual-mechanism framework demonstrates that only when new factor endowment upgrading and synergy mechanisms are implemented in tandem can factor advantages be converted into momentum for industrial upgrading—providing both theoretical and practical guidance for breaking free from “low-end lock-in”.

6. Conclusions

6.1. Conclusions and Recommendations

This study presents a two-dimensional integrated framework of factor endowment upgrading and factor synergy, filling the gap in systematic theories of NQPF-driven industrial structure upgrading and offering insights for emerging economies facing fading factor dividends and low-end lock-in. The key findings are as follows. (1) China’s NQPF, factor endowment upgrading, and factor synergy scores are all below 0.5, remaining in the initial development stage; thus, the development of labor lags behind that of objects of labor and means of production. (2) NQPFs significantly promote industrial structure upgrading and rationalization, with factor synergy outperforming factor endowment upgrading in terms of effectiveness. (3) There is regional heterogeneity in the effect of industrial structure rationalization—the factor endowment upgrading path is uneven in economically underdeveloped and highly environmentally regulated provinces, while the dual mechanism fails to operate effectively in manufacturing-led provinces. (4) NQPFs have a relatively stronger marginal effect on manufacturing structure upgrading, but the long-term potential of service industry upgrading remains to be further explored.

Relevant policy implications are proposed accordingly.

First, governments should promote supply-side reforms in education and skills training; establish a tripartite collaborative mechanism involving governments, enterprises, and educational institutions; develop talent catalogs aligned with industrial planning; design modular, short-cycle training programs for key industries; and introduce incentives and support policies, such as by providing tuition subsidies for individuals who participate in skills training for the labor force in short supply.

Second, cross-industry and cross-regional platforms for the dynamic matching of factor supply and demand, alongside dynamic evaluation systems for factor synergy indices, should be established to form a closed loop of early warning, deployment, and optimization. Early warnings are triggered when an index falls below a specified threshold, government-led joint platforms initiate cross-regional factor deployment, and evaluation outcomes are incorporated into the assessment of local industrial policies.

Third, region-specific measures should be implemented to address transmission barriers. For economically less developed provinces, financial transfer payments should increase, infrastructure development in the transportation and digital sectors should be enhanced, and factor collaboration channels should be established to introduce high-quality factor resources and technical expertise from eastern regions. For provinces with strict environmental regulations, the design of environmental regulation policies should be optimized—industrial rationalization should be incorporated into policy assessment frameworks, differentiated regulatory measures should be implemented, and subsidy schemes should be established to support factor upgrading. For manufacturing-dominated provinces, the expansion of new infrastructure, such as industrial internet and intelligent computing centers, to key industrial clusters should be accelerated; factors should be guided to concentrate in high-end manufacturing; and gradual transformation incentive policies should be implemented to encourage the shortening of industrial adjustment cycles.

6.2. Research Limitations and Future Perspectives

China’s industrial structure upgrading is deeply embedded in the global production network; however, this study has four key limitations. First, in this work, a domestic-only perspective is adopted, and international factors (e.g., GVC position and technological dependence) are excluded, leading to an incomplete portrayal of NQPF’s external constraints. Second, endogeneity handling needs refinement—the interaction between digital economy support policies and NQPFs is unaddressed, potentially biasing core coefficients and hindering accurate identification of the net effect. Third, this study is limited to a single country without cross-country comparisons and fails to reveal international commonalities/heterogeneities, thus restricting its reference value internationally. Finally, in this study, between-group heterogeneity is identified via median split, but its single-threshold reliance must be noted, as it fails to fully capture intensity’s continuous properties.

Future research could be extended in three directions. First, international factors could be incorporated to analyze the moderating effect of global production network changes on the NQPF–industrial structure upgrading nexus, and the effects of GVC position and technological dependence could be explored. Second, the cross-country comparative perspective could be expanded; countries with diverse income levels and industrial bases could be selected as samples, the institutional environment’s moderating role in the dual mechanism could be tested, typical paths could be refined, and targeted insights could be offered for emerging economies trapped in “low-end lock-in.” Third, continuous interaction terms—e.g., interaction regressions between core explanatory and moderating variables—could be applied to precisely capture heterogeneity nuances and enhance the robustness of the results.