The Impact of New Quality Productive Forces on High-Quality Development of Higher Education: Evidence from China

Abstract

Amid accelerating technological change and structural transformation, advanced productivity regimes characterized by technological innovation, digital transformation, and green upgrading have become key drivers of economic restructuring. In China, this transformation is conceptualized as new quality productive forces (NQPFs). However, their implications for higher education systems remain insufficiently explored. This study examines how NQPFs influence the high-quality development of higher education. Using panel data from 30 Chinese provinces from 2015 to 2022, composite indices constructed with the entropy method are used to measure NQPFs and the high-quality development of higher education, and mediation effect models, together with a Spatial Durbin Model, are employed to analyze the underlying mechanisms and spatial interactions. The results show that NQPFs significantly promote the high-quality development of higher education. This effect operates mainly through industrial collaborative agglomeration and digital infrastructure development and also generates positive spatial spillover effects across regions. These findings highlight the role of productivity transformation in shaping the structural foundations of higher education development in the digital era.

1. Introduction

Across the globe, higher education systems are facing growing pressure to adapt to rapid technological and economic transformation. Advances in artificial intelligence, digital technologies, and green innovation are reshaping production systems, labor markets, and regional development trajectories [1,2], fundamentally altering the skills and knowledge required in modern economies [3]. These changes place universities, students, and regional innovation systems at the center of structural adjustment, as higher education institutions must continuously update teaching, research, and knowledge transfer functions to remain aligned with evolving technological landscapes [4]. Failure to adapt may result in mismatches between talent supply and industrial demand, weakening regional innovation capacity and long-term economic competitiveness [5,6].

These pressures reflect a broader transformation in the underlying productivity regimes of modern economies. In rapidly transforming economies, this shift is characterized by the integration of technological innovation, digital transformation, industrial upgrading, and green development. In China, this paradigm is conceptualized as new quality productive forces (NQPFs), which represent a systemic leap in total factor productivity (TFP) driven by revolutionary technological breakthroughs and the sophisticated integration of production factors [7]. Unlike traditional productivity models reliant on extensive labor and capital inputs, NQPFs are an advanced form of productive forces that depart from conventional economic growth models and traditional development paths, and are distinguished by high technology, high efficiency, and high quality [8]. Although the terminology originates in China’s policy discourse, its substantive connotation closely aligns with broader theoretical strands in innovation economics, digital economy studies, and sustainability transition research [9]. Compared with narrower concepts such as innovation-driven development or digital economy transformation, NQPFs represent a more integrated productivity framework that simultaneously emphasizes technological breakthroughs, factor reallocation efficiency, industrial structural upgrading, and ecological sustainability [10].

Such productivity transformation has profound implications for higher education systems. By reshaping the structure of skill demand, advanced productivity regimes increase the need for interdisciplinary competencies, digital literacy, and innovation-oriented capabilities, prompting universities to adjust disciplinary structures, redesign curricula, and enhance research capacity to remain aligned with evolving technological and industrial landscapes [11]. At the same time, regions characterized by stronger innovation capacity and digital foundations tend to provide more favorable material and institutional conditions for higher education development through enhanced fiscal capacity, policy coordination, and digital infrastructure investment [12]. Beyond these functional and resource adjustments, productivity transformation also reconfigures the institutional positioning of universities within regional innovation ecosystems, deepening university–industry linkages and reinforcing higher education as a central actor in knowledge exchange, collaborative innovation, and sustainable regional development [13].

Despite these theoretical linkages, empirical evidence connecting productivity transformation to higher education quality remains limited. Existing studies predominantly examine the impact of NQPFs on economic growth, industrial upgrading, or environmental performance [14,15,16]. Higher education is often treated as a background condition rather than as an adaptive institutional system co-evolving with structural transformation. Moreover, the spatial dimension of this relationship remains underexplored. Productivity transformation is inherently spatial, characterized by knowledge diffusion, talent mobility, and cross-regional spillovers [17]. Whether and how such dynamics influence the high-quality development of higher education across regions is an open empirical question with significant implications for balanced and sustainable development.

This study aims to examine whether and how NQPFs influence the high-quality development of higher education across Chinese provinces, where higher education planning, policy implementation, and resource allocation are largely organized at the provincial level. Using panel data from 30 provinces in mainland China during 2015–2022, the study constructs multidimensional composite indices to measure both NQPFs and the high-quality development of higher education. Mediation effect models are employed to identify potential transmission channels, focusing on industrial collaborative agglomeration and digital infrastructure as key mechanisms. Furthermore, a Spatial Durbin Model is used to test whether productivity transformation generates cross-regional spillover effects on the high-quality development of higher education.

This study contributes to the literature in three ways. First, it conceptually bridges productivity transformation and higher education development by positioning universities as endogenous components of innovation-driven and sustainability-oriented economic systems. Second, it identifies concrete structural channels—industrial collaborative agglomeration and digital infrastructure—through which advanced productivity regimes translate into educational upgrading. Third, by incorporating spatial econometric analysis, it reveals that productivity transformation exerts not only local but also interregional spillover effects on the high-quality development of higher education, highlighting the importance of coordinated regional development in sustaining educational advancement.

By integrating innovation economics, digital transformation theory, and higher education research, this study provides a theoretically grounded and empirically robust analysis of how advanced productivity regimes reshape higher education systems in a context of rapid structural change. Although based on Chinese provincial data, the findings offer broader insights for economies seeking to align higher education systems with innovation-driven and sustainability-oriented development trajectories.

2. Theoretical Analysis and Research Hypotheses

2.1. NQPFs and High-Quality Development of Higher Education

The high-quality development of higher education is closely conditioned by the broader productivity regime within which universities operate. From the perspectives of regional innovation systems and endogenous growth theory, universities are not merely educational organizations, but also important institutional actors in knowledge creation, human capital accumulation, and technological adaptation [18]. Their developmental trajectory therefore depends not only on internal educational reforms but also on the wider transformation of regional economic structures, innovation environments, and factor allocation patterns.

In this context, NQPFs represent a new stage of productivity development characterized by technological innovation, digital transformation, industrial upgrading, and ecological sustainability [7,15]. Unlike traditional growth models centered on factor accumulation and scale expansion, NQPFs emphasize qualitative improvements in innovation capacity, factor allocation efficiency, structural optimization, and long-term sustainable development. Although NQPFs share certain features with concepts such as innovation-driven development, digital economy transformation, and green productivity, they are broader in scope and more integrative in orientation. Innovation-driven development mainly highlights technological progress as the engine of economic growth [19]; digital transformation emphasizes the restructuring of production, exchange, and organizational processes through digital technologies [20]; and green productivity focuses on improving environmental efficiency and reducing ecological costs [21]. By contrast, NQPFs incorporate these dimensions into a unified productivity framework that reshapes technological capability, factor allocation, industrial organization, and sustainability pathways in a coordinated manner [22].

This systemic transformation has important implications for higher education. First, NQPFs reshape the demand structure for knowledge and talent. As regional economies become more innovation-intensive, digitalized, and sustainability-oriented, the demand for interdisciplinary capabilities, digital literacy, and research-based skills correspondingly increases. Universities, as major suppliers of advanced human capital, are therefore under endogenous pressure to adjust disciplinary structures, redesign curricula, and strengthen research capacity in response to evolving technological and industrial conditions [23,24].

Second, NQPFs alter the external environment in which higher education institutions operate. Regions with stronger NQPFs generally exhibit higher innovation capacity, more efficient factor allocation, and more advanced digital foundations, all of which provide more favorable conditions for higher education development. Such environments help expand fiscal and policy support, strengthen collaborative networks, and improve the availability of educational and research resources, thereby enhancing the institutional foundations of higher education quality [12].

Third, NQPFs deepen the embeddedness of universities within regional innovation systems. As industrial upgrading advances and digital connectivity intensifies, universities become more closely integrated into collaborative networks involving firms, governments, and research institutions [5]. This institutional embeddedness strengthens knowledge exchange, facilitates joint innovation, and improves the capacity of higher education systems to generate high-quality research and cultivate socially relevant talent. From a sustainability perspective, such co-evolution between advanced productive forces and higher education supports resilient development and long-term human capital accumulation.

Taken together, NQPFs should be understood not only as a driver of economic growth but also as a structural force that reshapes the functional orientation, resource conditions, and institutional environment of higher education systems. Regions with higher levels of NQPFs are therefore more likely to achieve stronger performance in the high-quality development of higher education. Accordingly, this study proposes the following hypothesis:

H1. 

NQPFs positively promote the high-quality development of higher education.

2.2. The Mediating Effect of Industrial Collaborative Agglomeration

As NQPFs advance, regional industrial structures are reshaped in ways that promote closer coordination among related industries and innovation actors. In this context, industrial collaborative agglomeration represents a spatial form of economic organization in which related industries, firms, and supporting institutions are geographically concentrated and functionally interconnected. Existing literature suggests that such agglomeration enhances knowledge exchange, improves resource allocation efficiency, and accelerates innovation through proximity, specialization, and network effects [25,26]. Beyond productivity gains, industrial collaborative agglomeration can strengthen regional innovation ecosystems by fostering denser linkages among firms, research institutions, and universities [27,28].

The development of NQPFs is closely associated with deeper forms of industrial coordination and integration. Regions characterized by strong NQPFs typically exhibit higher levels of technological sophistication, digital connectivity, and industrial upgrading, which favor the emergence of collaborative industrial clusters rather than isolated sectoral expansion [29]. Compared with fragmented industrial growth, such clusters are more likely to generate cross-sector complementarities, joint innovation platforms, and shared infrastructures, thereby creating a regional environment in which knowledge production, technology diffusion, and organizational learning become more continuous and mutually reinforcing.

This transformation has important implications for higher education quality. On the one hand, industrial collaborative agglomeration embeds universities more deeply within regional production and innovation networks. When universities are located in regions where firms, industries, and supporting institutions interact intensively, they gain greater access to external knowledge sources, research partnerships, practical training platforms, and technology transfer opportunities. This closer integration not only strengthens university–industry cooperation, but also improves the relevance of teaching content, the applicability of research activities, and the responsiveness of talent cultivation to socioeconomic needs [5]. On the other hand, collaborative agglomeration reshapes regional demand for advanced human capital. As industrial linkages become more sophisticated and innovation activities more interconnected, universities face stronger incentives to adjust disciplinary structures, promote inter-disciplinary programs, strengthen applied research capacity, and cultivate talents better aligned with emerging industries and innovation-oriented development [30].

More importantly, the effect of industrial collaborative agglomeration on higher education is not limited to isolated cooperation projects; it also operates through broader changes in the regional resource environment. A more collaborative industrial structure tends to attract innovation resources, policy support, and high-level talent, while also lowering coordination costs among universities, firms, and research institutions [31,32]. Under these conditions, universities are better positioned to improve research productivity, expand students’ experiential learning opportunities, and enhance their overall institutional capacity. From a sustainability perspective, such a configuration supports the long-term integration of education, industry, and innovation, allowing higher education institutions to participate more effectively in regional development while simultaneously benefiting from the spillovers generated by industrial upgrading.

Therefore, industrial collaborative agglomeration should be understood as a key transmission channel through which NQPFs influence the high-quality development of higher education. By reshaping the structure of regional innovation networks, the allocation of external resources, and the demand for talent and knowledge, collaborative agglomeration translates productivity transformation into tangible improvements in teaching, research, and talent cultivation. Accordingly, the following hypothesis is proposed:

H2. 

Industrial collaborative agglomeration mediates the positive effect of NQPFs on the high-quality development of higher education.

2.3. The Mediating Effect of Digital Infrastructure

The development of NQPFs is also closely associated with the upgrading of regional digital environments. In this process, digital infrastructure—encompassing broadband networks, data platforms, cloud computing facilities, and intelligent systems—provides an essential foundation for knowledge production, organizational coordination, and educational innovation in the digital era. A growing body of research suggests that well-developed digital infrastructure enhances the flexibility, accessibility, and effectiveness of higher education by enabling blended learning models, expanding access to digital resources, and supporting data-driven teaching and learning processes [33,34,35]. Beyond teaching, digital infrastructure also plays a vital role in research activities by facilitating access to scientific databases, computational resources, and collaborative research tools.

The development of NQPFs is inherently intertwined with improvements in digital infrastructure. As productivity increasingly relies on data-driven technologies, platform coordination, and intelligent systems, regions characterized by stronger NQPFs tend to invest more heavily in digital connectivity and related technological facilities [36]. Such investments not only support industrial upgrading and innovation activities, but also reshape the broader technological environment in which universities operate. Compared with regions with weaker digital foundations, digitally advanced regions provide more favorable conditions for higher education institutions to adopt digital technologies, improve governance efficiency, and integrate more effectively into wider knowledge networks.

From the perspective of higher education quality, digital infrastructure affects multiple dimensions through several interrelated channels. First, it lowers barriers to accessing high-quality educational resources and reduces constraints imposed by geography and institutional disparities, thereby improving the accessibility and inclusiveness of higher education [37,38]. This enables universities to enrich teaching content, diversify instructional methods, and expand students’ access to high-level learning resources. Second, digital infrastructure enhances the efficiency, scale, and continuity of research collaboration by supporting data sharing, remote cooperation, and access to advanced computational tools, allowing universities to participate more actively in national and international knowledge networks [39]. In this sense, digital infrastructure not only improves research productivity but also strengthens the openness and connectivity of higher education systems.

Third, digital infrastructure supports the digital transformation of university governance and institutional operations. As teaching, research, and administrative processes become increasingly data-intensive, stronger digital foundations improve information processing, resource coordination, and organizational responsiveness within higher education institutions. This is particularly important in periods of rapid technological and social change, as digital infrastructure enhances institutional resilience and enables universities to adapt more effectively to external shocks and evolving development demands [40]. From a sustainability perspective, such improvements help create more inclusive, flexible, and durable educational systems capable of supporting long-term human capital development.

Importantly, digital infrastructure should therefore be understood not merely as a technical condition, but as a crucial transmission channel through which NQPFs influence the high-quality development of higher education. By strengthening the techno-logical environment, expanding access to knowledge resources, improving research collaboration, and enhancing institutional adaptability, digital infrastructure translates regional productivity transformation into tangible improvements in teaching quality, research capacity, and educational sustainability. Based on the above analysis, this study proposes the following hypothesis:

H3. 

Digital infrastructure mediates the positive effect of NQPFs on the high-quality development of higher education.

2.4. The Spatial Spillover Effects of NQPFs on the High-Quality Development of Higher Education

The impacts of NQPFs are unlikely to be confined within administrative boundaries. Both productivity transformation and higher education development are embedded in spatially interconnected systems characterized by the mobility of knowledge, talent, capital, and technology [41]. A growing literature on regional development emphasizes that innovation-driven growth processes often generate significant spatial spillover effects, whereby advancements in one region influence economic and institutional outcomes in neighboring areas through diffusion and network interactions [42].

The development of NQPFs, which relies heavily on intangible factors such as digital technologies, innovation capabilities, and knowledge-intensive activities, is particularly conducive to cross-regional spillovers. Unlike traditional factor inputs that are geographically bounded, the core elements of NQPFs—data, algorithms, technological know-how, and skilled labor—can circulate across regions through collaborative networks, labor mobility, and digital platforms [43]. As a result, regions with strong NQPFs may shape the development trajectories of surrounding areas, even in the absence of direct policy coordination.

Higher education systems are especially sensitive to such spatial dynamics [44]. Universities participate in interregional research collaborations, academic exchange programs, and shared innovation platforms that transcend local boundaries [45]. Technological breakthroughs and industrial upgrading driven by NQPFs in leading regions can diffuse to neighboring regions through joint research projects, cross-regional funding schemes, and the spillover of managerial and pedagogical practices [46]. These processes enable universities in adjacent regions to access advanced knowledge, upgrade research agendas, and improve teaching quality without replicating the full scale of local innovation inputs.

From a sustainability perspective, spatial spillovers play a crucial role in reducing regional disparities and promoting coordinated development of higher education systems. Positive externalities generated by NQPFs can help mitigate uneven distribution of educational resources by enabling lagging regions to benefit from nearby innovation hubs [47]. This mechanism supports a more balanced and resilient higher education landscape, which is essential for sustaining long-term human capital development in an integrated economy.

Taken together, these arguments suggest that NQPFs not only enhance higher education quality within a given region but also exert significant positive spillover effects on neighboring regions through interregional knowledge diffusion and factor mobility. Accordingly, this study proposes the following hypothesis:

H4. 

NQPFs exert significant spatial spillover effects on the high-quality development of higher education.

3. Methods

This study adopts a quantitative empirical approach to examine the impact of NQPFs on the high-quality development of higher education. The analysis is based on a longitudinal panel dataset covering 30 Chinese provinces from 2015 to 2022, using secondary data from official national databases. The research design is non-experimental and relies on econometric techniques. Specifically, the entropy weight method is used to construct multidimensional indices, while the empirical framework incorporates mediation models to test potential mechanisms and a Spatial Durbin Model (SDM) to capture regional spillover effects.

3.1. Model Setting

To investigate the impact of NQPFs on the high-quality development of higher education and to test the proposed research hypotheses, this study constructs a series of econometric models based on panel data. First, a baseline fixed-effects model is established to estimate the direct effect of NQPFs on the high-quality development of higher education:

$$ED{U}_{it}=\alpha +{\beta }_{1}NQP{F}_{it}+\gamma X_{it}+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(1)

where ${EDU}_{it}$ represents the level of high-quality development of higher education in province i at time t. ${NQPF}_{it}$ represents new quality productive forces, which serve as the core explanatory variable. $X_{it}$ is a vector of control variables that may influence higher education development. ${\mu }_i$ and ${\lambda }_i$ denote province and year fixed effects, respectively, and ${\epsilon }_{it}$ is the error term. The coefficient ${\beta }_1$ captures the direct impact of NQPFs on the high-quality development of higher education and provides the primary test for H1.

Second, to explore the underlying mechanisms through which NQPFs affect the high-quality development of higher education, mediation models are constructed by incorporating potential mediating variables. Specifically, industrial collaborative agglomeration (ICA) and digital infrastructure (DI) are introduced as mediators to test H2 and H3, respectively. The mediation effect is tested using the following equations:

$$M_{it}=\alpha +{\beta }_{1}NQP{F}_{it}+\gamma X_{it}+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(2)

$$ED{U}_{it}=\alpha +{\beta }_{1}NQP{F}_{it}+{\beta }_2{M}_{it}+\gamma X_{it}+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(3)

where $M_{it}$ represents the mediating variable (either ICA or DI). If both ${\beta }_1$ and ${\beta }_2$ are statistically significant, it indicates the presence of a mediation effect. The significance of the indirect effect is further verified using the Sobel test.

Finally, considering that regional economic development and innovation activities often exhibit spatial interactions, this study further employs a Spatial Durbin Model (SDM) to investigate potential spatial spillover effects and test H4. The SDM specification is expressed as follows:

$$ED{U}_{it}=\rho (WEDU{)}_{it}+{\beta }_{1}NQP{F}_{it}+\theta (WNQPF{)}_{it}+\gamma X_{it}+\delta (WX{)}_{it}+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(4)

where $W$ is the spatial weight matrix based on provincial adjacency, $WEDU$ is the spatial lag of the dependent variable, and $WNQPF$ and $WX$ denote the spatially lagged terms of the explanatory variable and the control variables, respectively. $\rho$ captures the spatial dependence of higher education quality, while $\theta$ and $\delta$ reflect the spillover effects of NQPFs and the control variables.

As highlighted by LeSage and Pace [48], the estimated coefficients in the Spatial Durbin Model cannot be directly interpreted as marginal effects because of spatial feedback effects. Therefore, this study further decomposes the impacts of NQPFs into direct and indirect effects using the partial derivative approach. The matrix of partial derivatives of the expected value of higher education quality with respect to the $k$-th explanatory variable is defined as follows:

$$\begin{gathered} \left[\begin{array}{ccc}{\displaystyle \frac{\partial E\left({EDU}_t\right)}{\partial x_{1t}}},& \dots & ,{\displaystyle \frac{\partial E\left({EDU}_t\right)}{\partial x_{nt}}}\end{array}\right]= \left[\begin{array}{cccc}{\displaystyle \frac{\partial E\left({EDU}_{1t}\right)}{\partial x_{1t}}}& {\displaystyle \frac{\partial E\left({EDU}_{1t}\right)}{\partial x_{2t}}}& \dots & {\displaystyle \frac{\partial E\left({EDU}_{1t}\right)}{\partial x_{nt}}}\\ {\displaystyle \frac{\partial E\left({EDU}_{2t}\right)}{\partial x_{1t}}}& {\displaystyle \frac{\partial E\left({EDU}_{2t}\right)}{\partial x_{2t}}}& \dots & {\displaystyle \frac{\partial E\left({EDU}_{2t}\right)}{\partial x_{nt}}}\\ ⋮& ⋮& \ddots & ⋮\\ {\displaystyle \frac{\partial E\left({EDU}_{nt}\right)}{\partial x_{1t}}}& {\displaystyle \frac{\partial E\left({EDU}_{nt}\right)}{\partial x_{2t}}}& \dots & {\displaystyle \frac{\partial E\left({EDU}_{nt}\right)}{\partial x_{nt}}}\end{array}\right] \\ ={\left(I_n-\rho W_n\right)}^{-1}\left[\begin{array}{cccc}{\beta }_k& w_{12}{\theta }_k& \dots & w_{1n}{\theta }_k\\ w_{21}{\theta }_k& {\beta }_k& \dots & w_{2n}{\theta }_k\\ ⋮& ⋮& \ddots & ⋮\\ w_{n1}{\theta }_k& w_{n2}{\theta }_k& \dots & {\beta }_k\end{array}\right] \end{gathered}$$

(5)

Based on this matrix, the average of the diagonal elements represents the direct effect, while the average of the off-diagonal elements represents the indirect (spillover) effect, and the total effect is defined as the sum of the direct and indirect effects.

3.2. Data and Variables

This study employs panel data from 30 provinces in mainland China covering the period from 2015 to 2022. The final sample was determined based on data availability and consistency. Specifically, Hong Kong, Macau, and Taiwan were excluded because their statistical systems and indicator definitions are not fully comparable with those of mainland provinces. Tibet was excluded due to substantial missing values in several key indicators over the sample period. The data are sourced from authoritative publications, including the China Statistical Yearbook, Tertiary Industry Statistical Yearbook, China High-tech Industry Statistical Yearbook, China Education Statistical Yearbook, China Population and Employment Statistical Yearbook, and the Compilation of Science and Technology Statistics in Higher Education Institutions. Additional data were obtained from the official websites of the Ministry of Education and provincial governments. For the provinces included in the final sample, a small number of intermittently missing observations were imputed using linear interpolation based on adjacent years within each province.

3.2.1. Dependent Variable

The high-quality development of higher education is a multidimensional and systematic process. Drawing on established frameworks that conceptualize higher education quality in terms of performance outcomes, institutional conditions, and global engagement [49,50,51], this study constructs an index system capturing the core functions and development patterns of higher education across three primary dimensions: output performance, teaching environment, and international engagement. In this framework, the index is designed to reflect not only the substantive outcomes of higher education but also the institutional conditions and support capacity that sustain its long-term improvement. The specific indicators and their definitions used in the construction of the high-quality development of higher education index are summarized in

Table 1. Indicator system for measuring the level of high-quality development in higher education.

Primary Dimension	Secondary Dimension	Indicator Description	Direction	Weight
Output Performance	Talent Output	Number of graduate degrees awarded by universities/institutions	+	0.0527
		Number of undergraduate and junior college degrees awarded	+	0.0288
	Research Output	Number of intellectual property rights and patents granted to universities	+	0.0577
		Number of natural science and technology achievements awarded at or above the provincial level	+	0.0388
		Number of scientific research papers published by universities	+	0.0503
	Research Application	Number of R&D applications and science & technology service projects in universities	+	0.0767
		Contract value of technology transfers by universities	+	0.1344
		Proportion of employed individuals with higher education	+	0.0373
Teaching Environment	Campus Environment	Proportion of green areas and sports facilities on campus	+	0.0083
		Proportion of multimedia classrooms to total classrooms	+	0.0154
	Facilities & Equipment	Total value of fixed assets in higher education	+	0.0324
		Proportion of teaching and research equipment in fixed assets	+	0.0183
	Faculty Resources	Number of full-time teachers in higher education institutions	+	0.0294
		Student–faculty ratio in higher education	−	0.0068
		Proportion of full-time teachers with senior professional titles at main campuses	+	0.0223
		Proportion of full-time teachers with doctoral degrees at main campuses	+	0.0364
International Exchange	Academic Exchange	Number of international academic conferences hosted by universities	+	0.0965
		Number of academic papers presented at hosted international conferences	+	0.1040
	Scientific Collaboration	Number of international research exchange visits initiated by universities	+	0.0648
		Number of international research exchange visits received by universities	+	0.0886

Note: “+” denotes a positive indicator, and “−” denotes a negative indicator.

.

Specifically, output performance reflects the key roles of higher education in talent cultivation, scientific research, and social service, and is measured by indicators such as the number of postgraduate and undergraduate degrees awarded, the number of patents granted and science and technology awards received, the volume of research publications, the number of Research and Development (R&D) service projects, the value of technology transfer contracts, and the proportion of employed individuals with higher education. Teaching environment captures the institutional and infrastructural foundations supporting higher education development, assessed by indicators including the share of campus area devoted to green space and sports facilities, the proportion of multimedia classrooms, the total value of fixed assets and the share allocated to research and teaching equipment, the number of full-time faculty, the student–faculty ratio, and the proportions of senior-ranked and doctoral-holding faculty. International engagement evaluates the global orientation and external linkages of higher education institutions, based on metrics such as the number of international academic conferences hosted, the number of papers presented, and the volume of inbound and outbound international research collaborations.

To synthesize the above indicators into a composite measure, this study employs the entropy weight method. All indicators are first normalized to eliminate differences in scale, with positive and negative indicators standardized separately to ensure directional consistency. Based on the normalized values, the proportion of each indicator is calculated, and its information entropy is derived according to the variation across provinces and years. The corresponding entropy redundancy is then used to determine indicator weights, such that indicators with greater variation are assigned larger weights. The weighted indicator values are subsequently aggregated to obtain the composite index of the high-quality development of higher education for each province and year. The indicator directions and entropy-based weights are reported in Table 1.

3.2.2. Independent Variable

NQPFs refer to a new form of advanced productivity that departs from traditional growth models by emphasizing innovation, digital transformation, and ecological sustainability. Rooted in new development philosophy, this concept is characterized by breakthroughs in technology, innovative factor allocation, and deep industrial upgrading. Drawing on and synthesizing existing studies [10,15], this study constructs a composite NQPF index based on four dimensions: innovation capacity, factor allocation efficiency, industrial transformation performance, and green low-carbon development. The detailed indicator system and variable definitions used to measure the development level of NQPFs are reported in

Table 2. Indicator system for measuring the NQPFs.

Primary Dimension	Secondary Dimension	Indicator Description	Direction	Weight
Innovation-Driven Capacity	Intensity of Innovation Input	Full-time equivalent of R&D personnel	+	0.0527
		R&D expenditure as a percentage of Gross Domestic Product (GDP)	+	0.0287
	Innovation Output	Number of invention patents granted per capita	+	0.0537
		Number of valid high-tech invention patents per capita	+	0.0848
	Innovation Economic Benefits	Sales revenue of new products in high-tech industries	+	0.0978
		Total profit of high-tech industries	+	0.0701
Efficiency of Innovative Factor Allocation	Labor Allocation Efficiency	Labor misallocation index	−	0.0060
	Capital Allocation Efficiency	Capital misallocation index	−	0.0110
	Technology Allocation Efficiency	Transaction volume in the technology market	+	0.0833
	TFP	TFP estimated using DEA-Malmquist method	+	0.0124
Effectiveness of Industrial Transformation	Industrial Digitization	E-commerce sales volume	+	0.0646
		Number of enterprises with websites	+	0.0466
	Digital Industrialization	Number of legal entities in information transmission, software, and IT services	+	0.0480
		Software business revenue	+	0.0896
	Industrial Structure Optimization	Industrial structure sophistication index	+	0.0089
		Industrial structure rationalization index	+	0.1123
Green and Low-Carbon Development	Green Technology	Ratio of green invention patent applications to total patent applications	+	0.0170

Note: “+” denotes a positive indicator, and “−” denotes a negative indicator.

.

Specifically, innovation capacity is reflected in indicators such as R&D personnel and expenditures, the number of invention patents granted, and the output and profitability of high-tech products. Factor allocation efficiency captures the reorganization of labor, capital, and technology, as reflected by misallocation indices, technology market transactions, and TFP levels. Industrial transformation performance is assessed by indicators of digitalization and industrial upgrading, including e-commerce sales, enterprise internet presence, the size of the information service sector, software industry revenue, and measures of industrial advancement and rationalization. Green low-carbon development is measured by the share of green patent applications, energy intensity, and the proportion of government spending on environmental protection.

Consistent with the construction of the high-quality development of higher education index, the above indicators are aggregated into a composite measure of NQPFs using the same entropy-based weighting approach, ensuring methodological consistency across key variables while avoiding subjective weight assignment.

3.2.3. Mediating Variables

Two mediating variables are introduced to examine the mechanisms through which NQPFs affect the high-quality development of higher education. Industrial collaborative agglomeration (ICA) captures the coordinated spatial concentration of related industries within a region. Following Zhang et al. [52], this study measures ICA based on the co-agglomeration of manufacturing and producer service industries, which exhibit strong complementarities in regional production systems. The degree of industrial co-agglomeration is estimated using the location quotient of these two sectors based on employment data, reflecting their spatial coupling and synergy intensity.

Digital infrastructure (DI) provides the essential technological foundation for both NQPF development and the digital transformation of higher education. Drawing on existing studies and considering data availability, regional digital infrastructure is measured using three indicators: broadband access ports, fiber-optic cable length, and the number of mobile communication base stations [53,54,55,56].

3.2.4. Control Variables

To control for potential confounding influences on higher education development, several control variables are included in the model. Economic development (PGDP) is measured by GDP per capita, as stronger economies are more likely to support educational investment [57]. Urbanization (URB) is measured by the share of urban population in total population, as higher urbanization often correlates with increased demand for skilled labor and expanded educational resources [58]. Government support (GOV) is proxied by the share of education expenditure in local government budgets, capturing regional differences in policy attention and funding allocation to higher education [59]. Finally, the level of openness (OPEN) is measured by the ratio of foreign direct investment to GDP, as regions with greater openness may have better access to global knowledge and research resources [57].

Table 3. Descriptive statistics of variables.

Variable Type	Variable	Observations	Mean	Std. Dev.	Min	Max
Dependent Variable	EDU	240	0.188	0.126	0.032	0.572
Independent Variable	NQPFs	240	0.127	0.108	0.027	0.610
Mediating Variables	ICA	240	2.591	0.457	1.804	3.847
	DI	240	0.308	0.176	0.039	0.851
Control Variables	PGDP	240	6.823	3.185	2.617	19.031
	URB	240	0.627	0.107	0.429	0.893
	GOV	240	0.160	0.026	0.103	0.217
	OPEN	240	0.017	0.013	0.000	0.080

reports the descriptive statistics for all key variables used in this study.

4. Results

4.1. Analysis of NQPFs and the High-Quality Development of Higher Education

From 2015 to 2022, the development levels of NQPFs and high-quality higher education in China both exhibited steady growth. At the national level, the average NQPF index increased from 0.0903 to 0.1709, with an average annual growth rate of 9.53%, while higher education quality rose from 0.1682 to 0.2096, growing at 3.19% annually.

Regionally, eastern provinces consistently outperformed the central and western regions in both indicators, with the latter two lagging behind the national averages. Intra-regional disparities were most pronounced in the east for NQPFs and in the west for higher education quality, while the central region showed the least variation in both dimensions.

At the provincial level, Guangdong led in NQPF development, with a value of 0.6103 in 2022, while Qinghai and Gansu ranked lowest. For higher education quality, Beijing, Jiangsu, Shanghai, and Guangdong were among the top performers, whereas several western provinces significantly lagged. These spatial patterns reveal persistent regional imbalances that are broadly consistent across both dimensions. Detailed annual and provincial data are provided in the appendix for further reference.

4.2. Benchmark Regression Analysis

In the benchmark model, both fixed effects and random effects models are employed to estimate the impact of the NQPFs on the high-quality development of higher education. The estimation results are presented in columns (1) and (2) of

Table 4. Results of benchmark regression.

	(1)	(2)
	Fixed Effects	Random Effects
Variables	EDU	EDU
NQPFs	0.250 ***	0.228 ***
	(3.06)	(2.73)
PGDP	−0.005 **	−0.004
	(−2.13)	(−1.57)
URB	0.572 ***	0.447 ***
	(3.40)	(3.56)
GOV	0.224	0.256
	(1.14)	(1.61)
OPEN	0.166	0.201
	(0.44)	(0.90)
Constant	−0.197 **	−0.141 **
	(−1.92)	(−2.00)
Province Fixed Effects	YES	NO
Year Fixed Effects	YES	NO
N	240	240
R2	0.957	0.582

Note: t-statistics for estimated coefficients are in parentheses; *** p < 0.01, ** p < 0.05.

. The regression results show that the estimated coefficients of the core explanatory variable are significantly positive at the 1% level under both models, indicating that the NQPFs significantly promote the high-quality development of higher education. Thus, H1 is supported.

4.3. Mechanism Analysis

Based on the theoretical framework, mediation models were used to examine mechanisms, with results shown in

Table 5. Results of the mediation effect test.

	(1)	(2)	(3)	(4)
Variables	ICA	EDU	DI	EDU
NQPFs	0.464 **	0.233 ***	0.480 ***	0.243 ***
	(2.19)	(4.39)	(5.14)	(6.81)
ICA		0.035 **
		(2.01)
DI				0.398 ***
				(5.43)
Control Variables	YES	YES	YES	YES
Province Fixed Effects	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES
Sobel test		0.076 ***		0.222 ***
		(2.859)		(5.076)
N	240	240	240	240
R2	0.9776	0.9822	0.9708	0.9831

Note: t-statistics for estimated coefficients are in parentheses; z-statistics are reported in parentheses for the Sobel test; *** p < 0.01, ** p < 0.05.

. Columns (1)–(2) indicate that NQPFs significantly promote industrial collaborative agglomeration (coefficient = 0.464, p < 0.05), which in turn positively influences higher education quality (coefficient = 0.035, p < 0.05). The Sobel test confirms the mediating role of ICA, with an estimated indirect effect of 0.076 (z = 2.859, p < 0.01). These findings provide empirical support for H2.

Columns (3)–(4) reveal a strong mediation via digital infrastructure, with NQPFs positively impacting DI (coefficient = 0.480, p < 0.01) and DI significantly enhancing EDU (coefficient = 0.398, p < 0.01). The Sobel test likewise supports DI’s mediating role (indirect effect = 0.222, z = 5.076, p < 0.01). Therefore, H3 is also supported.

4.4. Analysis of Spatial Spillover Effects

Using a binary contiguity spatial weight matrix, this study tests the spatial autocorrelation of NQPFs and high-quality development of higher education via the global Moran’s I index (

Table 6. Global Moran’s I statistics.

Year	2015	2016	2017	2018	2019	2020	2021	2022
$NQPF$	0.2755 ***	0.2683 ***	0.2385 **	0.1717 *	0.2088 **	0.2064 **	0.2012 **	0.2213 **
	(2.681)	(2.625)	(2.383)	(1.813)	(2.141)	(2.131)	(2.484)	(2.097)
$EDU$	0.2226 **	0.1981 **	0.1663 *	0.2098 **	0.186 *	0.2217 **	0.2233 **	0.2381 **
	(2.201)	(1.991)	(1.703)	(2.053)	(1.858)	(2.125)	(2.177)	(2.272)

Note: z-statistics for the Global Moran’s I test are in parentheses; *** p < 0.01, ** p < 0.05, and * p < 0.1.

). Results show significant spatial autocorrelation for both variables, supporting the use of spatial econometric models.

Based on the LM, Robust LM, Likelihood Ratio, and Hausman tests reported in

Table 7. Results of spatial model specification tests.

Test Type	Test Statistic
LM-lag	123.482 ***
Robust LM-lag	97.482 ***
LM-error	26.000 ***
Robust LM-error	0.000
Hausman	102.09 ***
LR test for province fixed effects	51.13 ***
LR test for year fixed effects	635.79 ***
LR test (SDM vs. SAR)	12.01 **
LR test (SDM vs. SEM)	9.83 *

Note: *** p < 0.01, ** p < 0.05, and * p < 0.1.

, significant spatial dependence is detected in the data, which justifies the use of spatial econometric models. The Robust LM-lag statistic remains statistically significant, whereas the Robust LM-error statistic is not statistically significant, suggesting that spatial lag dependence may be more relevant in this sample. In addition, the LR tests comparing the SDM with the SAR and SEM specifications reject model simplification, indicating that the SDM cannot be reduced to either alternative model. The Hausman test further supports the fixed-effects specification, and the LR tests for province and year effects indicate that the two-way fixed-effects SDM is the most appropriate specification for the analysis.

According to Equation (5), the effect of NQPFs on higher education quality is decomposed into direct and indirect (spillover) effects.

Table 8. Decomposition of spatial effects.

Variables	Direct Effect		Indirect Effect		Total Effect
	Coef.	$\mathit{z}$	Coef.	$\mathit{z}$	Coef.	$\mathit{z}$
NQPFs	0.264 ***	5.20	0.172 *	1.88	0.436 ***	3.85
PGDP	−0.006 ***	−3.94	0.003	0.98	−0.003	−0.81
URB	0.657 ***	3.83	−0.046	−0.25	0.611 ***	3.41
GOV	0.108	0.59	−0.436	−1.32	−0.327	−0.87
OPEN	0.129	0.504	0.048	0.13	0.176	0.47

Note: z-statistics are reported for the estimated spatial effects; *** p < 0.01, and * p < 0.1.

shows a significant direct effect of 0.264 (p < 0.01) and an indirect effect of 0.172 (p < 0.10), resulting in a total effect of 0.436 (p < 0.01). These results indicate that NQPFs not only promote the high-quality development of higher education within the local province but also generate positive spillover effects on neighboring provinces. Therefore, the estimated spillover pattern is more consistent with a positive radiation effect than with a siphoning effect, thereby supporting H4.

4.5. Endogeneity Test

Potential endogeneity between NQPFs and higher education quality may arise from reverse causality and omitted variables. To alleviate this concern, this study employs the one-period lag of NQPFs as an instrumental variable and estimates a two-stage least squares (2SLS) model (

Table 9. Results of endogeneity testing.

	First Stage	Second Stage
Variables	NQPFs	EDU
NQPFs		0.183 ***
		(2.84)
NQPF_1	0.856 ***
	(17.67)
Control Variables	YES	YES
Province Fixed Effects	YES	YES
Year Fixed Effects	YES	YES
Anderson canon. corr. LM statistic	136.235 ***
Cragg-Donald Wald F statistic	312.120
N	240	240

Note: t-statistics for estimated coefficients are in parentheses; *** p < 0.01.

). The use of the lagged term is based on the assumption that NQPFs exhibit temporal persistence, so that the past level is strongly correlated with the current level, while the lagged value is less likely to be jointly determined by contemporaneous shocks to higher education quality after controlling for province fixed effects, time fixed effects, and other covariates.

The first stage shows a strong, significant correlation between the IV and NQPFs (p < 0.01). In the second stage, NQPFs continue to show a positive and significant effect on EDU (coefficient = 0.183, p < 0.01). Notably, the 2SLS coefficient is smaller than the benchmark estimate, suggesting that the baseline model may have overestimated the magnitude of the effect due to endogeneity, such as reverse causality or omitted regional factors that jointly influence technological upgrading and higher education development. Nevertheless, the coefficient remains statistically significant after instrumenting for NQPFs, indicating that the positive relationship is unlikely to be driven solely by endogenous bias and lending stronger support to a causal interpretation. Diagnostic tests further confirm model identification (Anderson LM test p = 0.000) and instrument strength (Cragg-Donald F = 312.12 > critical value 16.38). Therefore, while the baseline model may overstate the size of the effect, the IV results continue to support a significant positive role of NQPFs in promoting the high-quality development of higher education.

4.6. Robustness Test

To further verify the robustness of the baseline findings, three additional checks were conducted (

Table 10. Results of robustness test.

	(1)	(2)	(3)
Variable	Replacing Explanatory Variable	Winsorized Data	Alternative Spatial Weight Matrix
NQPFs	4.889 ***	0.255 ***
	(10.850)	(3.19)
Direct Effect			0.279 ***
			(5.64)
Indirect Effect			0.127 *
			(1.92)
Total Effect			0.406 ***
			(4.37)
Control Variables	YES	YES	YES
Province Fixed Effects	YES	YES	YES
Year Fixed Effects	YES	YES	YES
N	240	240	240
R2	0.9045		0.4846

Note: t-values are reported in parentheses for Columns (1) and (2), while z-values are reported in parentheses for Column (3); *** p < 0.01, and * p < 0.1.

). First, the core explanatory variable was replaced with a PCA-based index of NQPFs to reduce the possibility that the results depend on a specific measurement approach. The estimated coefficient remained positive and statistically significant, indicating that the main findings are not sensitive to alternative index construction methods. Second, a two-sided 1% winsorization was applied to all continuous variables to alleviate the influence of extreme observations, and the positive effect of NQPFs on the high-quality development of higher education remained robust. Third, the spatial weight matrix was replaced with a nested economic–geographic matrix that incorporates both geographic distance and interprovincial economic differences. The direct and indirect effects remained significantly positive, suggesting that the identified spatial spillover effect is not driven by a particular matrix specification alone. Overall, these results provide consistent evidence that the baseline conclusions are stable under alternative variable construction, data treatment, and spatial dependence assumptions.

5. Discussion

5.1. Theoretical Implications

This study contributes to the literature by extending the analytical scope of NQPF research to the field of higher education. Existing studies have mainly examined NQPFs in relation to economic growth, industrial upgrading, and environmental performance, while their implications for higher education development have received much less attention [14,15,16]. By showing that NQPFs significantly promote the high-quality development of higher education, this study suggests that productivity transformation should be understood not only as an economic process but also as an important force shaping educational upgrading.

First, the findings broaden the understanding of the relationship between regional productivity transformation and higher education development. Higher education is often discussed primarily as a provider of human capital and knowledge for innovation-driven growth. However, the results of this study indicate that it is also shaped by the wider regional context in which technological upgrading, digital transformation, and green development unfold [17,18]. In this sense, higher education is not merely a supporting condition for regional development, but part of a co-evolving system in which changes in productive forces reshape the environment for teaching, research, and external engagement [7,12,15].

Second, this study refines the understanding of how NQPFs affect higher education by identifying industrial collaborative agglomeration as an important transmission channel. Existing literature has shown that collaborative agglomeration enhances knowledge spillovers, factor matching, and innovation efficiency through spatial proximity and inter-industry coordination [25,26], while also strengthening linkages among firms, universities, and research institutions [27,28]. Building on this literature, the present study shows that industrial collaborative agglomeration is not only an outcome of productivity transformation but also a mechanism through which NQPFs are translated into higher education upgrading. This finding enriches existing discussions by highlighting the institutional and network-based pathways that connect regional industrial restructuring with improvements in higher education quality.

Third, the mediating role of digital infrastructure adds to the literature on the digital transformation of higher education. Prior studies have mainly emphasized the role of digital infrastructure in improving resource accessibility, supporting teaching innovation, and facilitating research collaboration [33,34,35]. The present study extends this discussion by showing that digital infrastructure also functions as a regional-level channel linking NQPFs to higher education quality [36]. This suggests that the contribution of digitalization to higher education should be understood not only in terms of internal institutional applications, such as online platforms or smart campus initiatives, but also in terms of broader regional conditions that support knowledge production, connectivity, and resource integration. In this respect, digital infrastructure serves as a structural enabler of higher education modernization.

Finally, the identified spatial spillover effects enrich the literature on the regional dynamics of higher education development. Previous studies have emphasized that productivity transformation and innovation-driven growth are inherently spatial, involving knowledge diffusion, talent mobility, and cross-regional externalities [17,41,42,43]. Higher education systems are similarly embedded in spatial networks through interregional collaboration, academic exchange, and shared innovation platforms [44,45,46]. The findings of this study support this perspective by showing that NQPFs generate not only local effects but also positive spillover effects on neighboring regions. This result points to a more relational understanding of higher education quality improvement under evolving regional productivity regimes.

5.2. Practical Implications

The findings of this study also provide several practical implications for policymakers, regional planners, and university administrators. Since NQPFs significantly promote the high-quality development of higher education, efforts to improve higher education quality should be embedded more closely in broader regional strategies of technological innovation, industrial upgrading, digital transformation, and green development. This means that higher education policy should not be formulated in isolation, but coordinated with the wider transformation of regional productive forces.

First, greater attention should be paid to policy coordination between NQPF development and higher education reform. As NQPFs reshape regional demand for knowledge, skills, and innovation capacity, higher education institutions need to respond more actively to these structural changes. This requires closer alignment between university development strategies and regional economic priorities, particularly in talent cultivation, discipline adjustment, research orientation, and social service functions. By linking higher education planning more closely with the development of advanced productive forces, policymakers can better support the mutual reinforcement between educational upgrading and regional transformation.

Second, the mediating role of industrial collaborative agglomeration suggests that higher education development can be promoted by strengthening the integration of universities into regional innovation and industrial systems. Rather than viewing universities solely as educational organizations, policy should further emphasize their role as key actors in collaborative innovation networks. Governments can support this process by encouraging stronger linkages among universities, enterprises, research institutes, and local industries, thereby creating more favorable conditions for knowledge transfer, applied research, and talent matching. In this way, industrial collaborative agglomeration can contribute not only to economic upgrading but also to the improvement of higher education quality.

Third, the results highlight the importance of continued investment in digital infrastructure. Since digital infrastructure functions as an important channel through which NQPFs enhance higher education quality, improving regional digital connectivity and technological support systems can generate broad benefits for higher education institutions. This includes not only better access to digital resources and platforms, but also stronger support for teaching innovation, research collaboration, and the sharing of educational opportunities across regions. Digital infrastructure should therefore be treated as a foundational condition for higher education modernization rather than merely a technical supplement.

Finally, the positive spatial spillover effects imply that higher education development should be approached from a regional perspective. Because the effects of NQPFs extend beyond local boundaries, policies confined to individual provinces may not fully capture the benefits of regional interaction. Cross-regional cooperation in innovation platforms, digital infrastructure, academic exchange, and educational resource sharing should therefore be further encouraged. A more coordinated regional policy framework would help amplify the broader benefits of NQPFs and support more balanced improvements in higher education quality across regions.

5.3. Limitations and Directions for Future Research

Despite these contributions, several limitations should be acknowledged. First, the analysis relies on provincial-level panel data, which may obscure important heterogeneity across cities or individual higher education institutions. Micro-level data could provide more detailed insights into how productivity transformation influences university behavior, research performance, or educational outcomes. Second, although the study employs mediation models and spatial econometric techniques to explore potential mechanisms, measurement challenges remain when constructing composite indices for complex concepts such as NQPFs and the high-quality development of higher education. Alternative measurement approaches or additional indicators could further improve the robustness and precision of the analysis. Third, while the empirical design identifies statistically significant relationships and potential mechanisms, the observational nature of the data limits the extent to which strong causal claims can be made. Future studies may consider quasi-experimental approaches or natural experiments to further strengthen causal inference.

Future research could extend this work in several directions. First, micro-level datasets, such as university-level or city-level data, could be used to examine how productivity transformation influences institutional strategies, research collaboration patterns, or student outcomes. Second, future studies may explore additional mechanisms linking productivity transformation and higher education development, including human capital mobility, research funding allocation, or the diffusion of digital technologies within universities. Third, comparative studies across countries could help determine whether the relationships identified in this study are specific to China’s institutional context or represent broader patterns associated with innovation-driven development. Such efforts would contribute to a deeper understanding of how higher education systems evolve within rapidly changing technological and economic environments and would further clarify the role of productivity transformation in shaping educational development in the knowledge economy.

6. Conclusions

Utilizing mediation effect models and the Spatial Durbin Model, the analysis yields several salient findings. First, both NQPFs and the high-quality development of higher education exhibit substantial regional imbalances, reflecting uneven development across provinces. Second, NQPFs exert a statistically significant and robust positive effect on the high-quality development of higher education, a relationship that remains stable across multiple model specifications and robustness tests. Third, the mechanisms through which NQPFs enhance the high-quality development of higher education operate primarily through two channels: industrial collaborative agglomeration and digital infrastructure. These mechanisms highlight the importance of structural and technological conditions in translating productivity transformation into educational outcomes. Finally, the influence of NQPFs extends beyond individual regions, generating significant spatial spillover effects. This suggests that advancements in one region can produce positive externalities for neighboring areas, implying that regional coordination may strengthen overall higher education development.