Sci-Tech finance efficiency promotes the construction of a modernized industrial system evidence from double machine learning

Abstract

Promoting a virtuous cycle among science, technology, finance, and industry is essential for advancing a modernized industrial system. Using provincial panel data from China spanning 2010–2023, this study applies a Double Machine Learning framework to investigate the causal impact of Sci-Tech finance efficiency (STFE) on the construction of a modernized industrial system(CMIS) and to uncover its internal mechanisms. The empirical results demonstrate that higher STFE significantly promotes industrial modernization by enhancing structural upgrading and innovation capacity. Mechanism analysis further reveals that STFE accelerates the transformation of scientific and technological achievements, strengthens the integration between digital technologies and the real economy, and optimizes the allocation of key production factors—including capital, talent, and technology. These mechanisms collectively foster the coordinated upgrading of industrial systems. Moreover, the heterogeneity analysis shows that the positive impact of STFE is more pronounced in regions with stronger economic foundations, higher degrees of marketization, and lower fiscal constraints, highlighting regional disparities in policy effectiveness. Overall, this study extends the theoretical understanding of the finance–technology–industry nexus under the DML framework and provides actionable insights for promoting regional coordination and differentiated policy design in the process of industrial modernization.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-35019-1.

Introduction

Amid the accelerating wave of the new global scientific and technological revolution and industrial transformation, technological innovation has become the primary engine of high-quality economic development and the modernization of industrial systems. At the Central Economic Work Conference in 2024, China elevated “promoting new quality productive forces through technological innovation and building a modernized industrial system” to a national strategic priority, highlighting the role of technological progress in optimizing industrial structures and enhancing their sophistication. Meanwhile, the resource-allocation function of the financial system in supporting technological innovation has become increasingly important. As a key bridge between the innovation chain and the capital chain, Sci-Tech finance helps channel financial resources toward innovation-intensive, green-oriented, and high-value-added sectors. The effectiveness of this process is captured by Sci-Tech finance efficiency, which reflects how efficiently financial resources are mobilized and allocated to support innovation and industrial upgrading. Accordingly, Sci-Tech finance efficiency constitutes an important driver of industrial upgrading and the construction of a modernized industrial system¹.

In recent years, the Chinese government has placed increasing emphasis on strengthening the Sci-Tech finance system and has issued a series of policy documents aimed at improving financial efficiency to support industrial restructuring and the modernization of the industrial framework. Despite this rapid expansion, several structural challenges remain, including inefficient resource allocation, mismatches between financial supply and innovation demand, and persistent regional disparities. These constraints may weaken the capacity of Sci-Tech finance to effectively serve industrial upgrading. Accordingly, enhancing STFE has become a pressing issue with both theoretical relevance and practical importance for advancing industrial modernization.

Although the evolution of Sci-Tech finance has provided crucial support for innovation-driven growth, the reconfiguration of the global industrial landscape and the increasing uncertainty of supply chains continue to pose complex challenges to China’s industrial system^2,3. Against this backdrop, building an internationally competitive and sustainable modernized industrial system is not only an intrinsic requirement for achieving high-quality development but also a strategic pathway to enhancing economic resilience and mitigating external shocks. Within this process, the efficiency of Sci-Tech finance plays a pivotal role. On the one hand, an efficient Sci-Tech finance system improves capital allocation, accelerates the industrialization of scientific and technological achievements, and promotes structural optimization and upgrading. On the other hand, an inefficient system can lead to capital stagnation and innovation bottlenecks, constraining the release of new-quality productive forces. Therefore, exploring the mechanisms through which STFE drives the modernization of industrial systems is essential for deepening our understanding of innovation-led development and advancing China’s transformation toward high-quality growth.

The potential contributions of this study may lie in the following aspects: (1) This study extends the literature on Sci-Tech finance and industrial modernization by shifting the focus from the scale or policy orientation of Sci-Tech finance to its efficiency dimension. Based on an input–output logic, we construct a DEA–BCC efficiency measure of STFE that evaluates how effectively finance and innovation related inputs are transformed into observable innovation outputs. (2) This study advances the measurement of industrial modernization by developing a seven-dimensional composite index of the CMIS, covering physical modernization, innovativeness, integration, greenness, openness, coordination, and institutional support. The index is constructed using the entropy-weight method, and its robustness and comparability are further validated using alternative evaluation approaches. (3) This study develops an efficiency-driven mechanism framework and provides causal evidence on how STFE promotes CMIS through three channels, namely the transformation of scientific and technological achievements, digital–real technology integration, and the optimization of factor allocation. Methodologically, we employ a Double Machine Learning framework to identify causal effects and mitigate confounding bias, complemented by robustness and endogeneity analyses, thereby improving the credibility of the empirical findings. Overall, this research advances the literature across three interrelated dimensions, namely indicator system construction, theoretical mechanism exploration, and methodological innovation, and provides new theoretical insights and empirical evidence for understanding how STFE drives industrial modernization.

The remainder of this paper is organized as follows. "Literature review" reviews the related literature on modernized industrial systems and the role of Sci-Tech finance, and identifies the research gaps. "Research hypothesis" develops the theoretical framework and proposes the testable hypotheses, accompanied by a conceptual diagram that summarizes the proposed mechanisms. "Model and variable" introduces the empirical strategy based on the Double Machine Learning framework, and describes the measurement of key variables and data sources. "Empirical analysis" presents the empirical results, including descriptive statistics, benchmark estimates, endogeneity analyses, robustness checks, and heterogeneity tests. "Conclusions and implications" concludes with the main findings and discusses policy implications, as well as limitations and directions for future research.

Literature review

The connotation and determinants of the modernized industrial system

Although the concept of a modernized industrial system (CMIS) is unique to China’s policy discourse, its theoretical underpinnings are deeply rooted in classical frameworks of industrial economics, development economics, and innovation theory⁴. Essentially, a modernized industrial system refers to an industrial structure that exhibits technological leadership, international competitiveness, and forward-looking adaptability to future industrial and technological developments, serving as the cornerstone of a modern economic system⁵. Its construction requires both the advancement of industrial foundations and the modernization of industrial chains, ensuring that industries evolve toward higher value-added, innovation-driven, and sustainable paradigms.

Recent studies have identified several key drivers shaping the evolution of modernized industrial systems. The digital industry has emerged as a critical enabler of industrial integration and structural transformation, providing an essential technological foundation for modernization⁶. Through the widespread application of digital technologies, cross-industry collaboration has been strengthened, and the efficiency of production, logistics, and innovation processes has been significantly improved. Moreover, extensive empirical evidence highlights that deeper integration between digital technologies and the real economy accelerates the intelligent upgrading of traditional industries, raises productivity, and enhances market competitiveness and resilience, thereby fostering new momentum for sustainable industrial modernization⁷. From the perspective of new quality productive forces, digital–real convergence enables the reconfiguration of traditional industries and the synergistic upgrading of entire value chains, representing a fundamental transformation in industrial organization and technological innovation⁸. Furthermore, the institutional and fiscal roles of government remain indispensable. Through policy guidance, fiscal support, and institutional innovation, governments provide the macro-level framework and systemic guarantees necessary for advancing modernization, particularly by aligning financial mechanisms with innovation incentives⁹. Collectively, these studies contribute to a deepening theoretical understanding of how technological progress, digital integration, and institutional capacity jointly underpin the CMIS, offering both conceptual insight and empirical evidence for the realization of innovation-driven and sustainable economic growth.

The enabling role of Sci-Tech finance in the modernized industrial system

Ensuring the autonomy and controllability of industrial chains has long been regarded as a primary objective of sustainable industrial development¹⁰. In this regard, Sci-Tech finance plays a crucial enabling role by optimizing the innovation ecosystem and supporting key enterprises in overcoming technological bottlenecks¹¹, thereby enhancing the self-sufficiency and security of modernized industrial systems. A well-functioning Sci-Tech financial system can channel capital efficiently toward frontier R&D activities, fostering breakthroughs in “choke-point” technologies and reducing systemic vulnerabilities in industrial chains.

From the perspective of neo-Schumpeterian growth theory, investment in technological R&D is a fundamental driver of technological progress and diffusion¹². Higher STFE significantly increases both the scale and effectiveness of R&D funding¹³, stimulating technological innovation and accelerating the shift toward intelligent and high-end industrial transformation. This process enhances the overall value of industrial chains¹⁴ and strengthens the capacity of industries to withstand external shocks by promoting innovation-driven structural upgrading and improving supply chain resilience^15,16. In addition, a more efficient Sci-Tech financial system can ease conventional financial frictions, including information asymmetry and high financing costs¹⁷, directing credit resources toward emerging industries aligned with the development of new quality productive forces¹⁸. This financial reallocation not only enhances the investment and production capacity of innovation-driven sectors but also accelerates disruptive innovation¹⁹, thereby fostering new competitive advantages within industrial systems. Moreover, improvements in STFE can enhance the circulation and utilization of information across different segments of the industrial chain, enabling manufacturing firms to upgrade along the global value chain. Overall, efficiency-oriented Sci-Tech finance strengthens both the technological depth and international competitiveness of modernized industrial systems and supports their role as a strategic pillar of sustainable economic modernization.

From an efficiency-and-allocation perspective, international evidence indicates that improvements in regional financial efficiency can facilitate resource accumulation and enhance allocation quality, thereby creating a more supportive financial environment for economic growth and industrial upgrading²⁰. Related research further shows that greater financial openness and financial efficiency can help dampen macroeconomic volatility²¹, suggesting that efficiency-enhancing financial mechanisms may stabilize the macro conditions under which industrial upgrading takes place. Moreover, financial innovation and deepening are found to improve capital-use efficiency and strengthen the provision of higher-quality financial services to the real economy²². Consistent with this view, enhanced financial service capacity can stimulate local innovation and entrepreneurship and support the emergence of new industries, thereby injecting new momentum into regional development²³. Taken together, these studies imply that a more efficient and resilient financial system can provide a stable and innovation-oriented foundation for the construction of a modernized industrial system.

Although the literature has examined the role of Sci-Tech finance in industrial upgrading, evidence remains limited on whether and how efficiency in finance technology matching contributes to the construction of a modernized industrial system. Existing studies more often rely on scale based measures of Sci-Tech finance and provide less direct identification of an efficiency channel. Moreover, the transmission pathways are not fully clarified, especially regarding whether Sci-Tech finance efficiency promotes industrial modernization through technology transformation and realization, digital real technology integration, and factor allocation optimization. These gaps motivate this study to focus on the efficiency dimension and to test the mechanism channels within a unified empirical framework.

Research hypothesis

Sci-Tech finance efficiency and the modernization of industrial systems

Enhancing STFE plays a pivotal role in optimizing resource allocation and fostering a synergistic industrial ecosystem. As a crucial bridge linking the innovation chain and the capital chain, the efficiency of Sci-Tech finance determines whether financial resources can flow swiftly and precisely toward innovation activities. An optimized regional Sci-Tech financial environment facilitates the effective integration of capital, talent, and technology, strengthens linkages across upstream and downstream industrial segments, and promote technology exchange and cooperation among enterprises²⁴. By accelerating technological upgrading and improving market responsiveness, efficient Sci-Tech finance provides a robust institutional and financial foundation for CMIS.

Moreover, higher STFE stimulates industrial innovation and the emergence of new-generation industries. Under an environment characterized by efficient financial allocation, firms face fewer capital constraints, enabling greater investment in R&D and faster technological transformation²⁵. As financial institutions enhance their service efficiency, capital flows increasingly toward strategic and future-oriented sectors such as artificial intelligence, renewable energy, and biopharmaceuticals. This two-way reinforcement between finance and innovation accelerates the upgrading of traditional industries while nurturing emerging sectors, leading to a more diversified, resilient, and innovation-driven industrial structure.

Finally, improving STFE also contributes to industrial internationalization and enhances global competitiveness. A high-performing Sci-Tech financial system not only supports domestic industrial coordination but also facilitates outward expansion by promoting cross-border integration of capital and technology²⁶. Through deeper coupling between financial capital and technological innovation, Chinese enterprises can move up the global value chain, strengthening their international competitiveness and expanding the external space for industrial development. Accordingly, we propose the following hypothesis:

Hypothesis H1

Improvements in Sci-Tech finance efficiency significantly promote the construction and advancement of a modernized industrial system.

Sci-Tech finance efficiency drives industrial modernization through technological achievement transformation

Enhancing STFE promotes the effective transformation of scientific and technological achievements, thereby strengthening the core competitiveness of the modernized industrial system. The major bottlenecks in this process typically lie in financing constraints, limited technological matching, and imperfect commercialization mechanisms. By alleviating financial constraints, Sci-Tech finance enables more effective risk sharing and value discovery throughout the process of technology transfer. Through the deep integration of finance and technology, efficient Sci-Tech finance directs capital toward emerging and high-tech industries, improves the financing conditions of technology-based small and medium-sized enterprises, and thus promotes the transformation and industrialization of technological innovation achievements²⁷. A high-efficiency Sci-Tech financial system also helps shorten the transformation cycle of scientific achievements. Rather than facing the traditional disconnection between research and market demand, capital can be swiftly allocated to research outcomes with strong application potential, accelerating their conversion into productive forces²⁸ and strengthening the alignment between technological supply and market demand²⁹.

Furthermore, the enhancement of STFE facilitates the cross-regional diffusion of innovation outcomes, mitigating spatial disparities in technological and financial resources. In China, the concentration of innovation and capital resources in eastern provinces often limits the nationwide dissemination of technological progress. An efficient Sci-Tech financial system can bridge this gap by breaking down regional barriers, accelerating interregional cooperation, and promoting the broader transfer and application of technological achievements. Such mechanisms foster an open, flexible, and adaptive industrial ecosystem³⁰, in which the integration of financial and technological elements drives the diffusion of innovation across regions. In summary, by strengthening the transformation and diffusion of scientific and technological achievements, STFE deepens the coupling between the innovation chain and the industrial chain, forming a virtuous cycle between technological innovation and industrial upgrading—a key strategic pathway toward constructing a modernized industrial system.

Hypothesis H2

Sci-Tech finance efficiency promotes the construction of a modernized industrial system by facilitating the transformation and diffusion of scientific and technological achievements.

Sci-Tech finance efficiency drives industrial modernization through digital–real technology integration

Enhancing STFE drives the deep integration of digital technologies with the real economy, injecting endogenous momentum into the CMIS. By leveraging digital empowerment, Sci-Tech finance establishes a sound ecosystem for digital–real integration³¹. Financial institutions increasingly employ emerging technologies such as big data, artificial intelligence, and blockchain to identify firms’ financing needs and risk profiles with greater precision. This enables more efficient matching of capital and technological resources, expands firms’ access to financing, information, and skilled labor, and enhances both the willingness and success rate of digital transformation among real-sector enterprises³².

Moreover, higher STFE fosters industrial digital transformation by accelerating funding flows to support digital upgrading initiatives. With efficient financial backing, enterprises can swiftly access resources for digital renovation, including smart supply chain management, blockchain-based traceability systems, and intelligent manufacturing applications³³. The interconnection among smart devices not only reduces information and coordination costs across production stages but also lays the foundation for the construction of smart factories, intelligent industrial chains, and data-driven supply networks. By mitigating investment uncertainty in the digitalization process, efficient Sci-Tech finance further promotes the deep fusion of the digital economy and the real economy, forming a collaborative innovation ecosystem characterized by continuous technological upgrading and cross-sectoral synergy. Consequently, by facilitating digital–real technological integration, STFE enables the modernized industrial system to exhibit greater intelligence, greenness, and integration, becoming a key endogenous driver of high-quality industrial development. Accordingly, we propose the following hypothesis:

Hypothesis H3

Sci-Tech finance efficiency promotes the structural optimization and capability upgrading of modernized industrial systems by fostering digital–real technology integration.

Sci-Tech finance efficiency reinforces industrial modernization through factor allocation optimization

Enhancing STFE promotes the agglomeration and efficient flow of key industrial factors—talent, capital, and technology—thereby providing a solid foundation for the CMIS. First, STFE facilitates the agglomeration and mobility of innovation talent. An efficient financial environment improves the allocation of scientific research resources and provides high-level researchers with better research conditions and competitive incentives, attracting top-tier talent to innovative enterprises and industrial clusters. Through its signaling effect, improved financial efficiency also shapes income expectations and strengthens the willingness of skilled professionals to migrate toward technology-intensive sectors, thereby cultivating an institutional and market environment that supports talent concentration and knowledge exchange³⁴.

Second, STFE drives the spatial concentration of venture capital. Empirical evidence suggests that venture capital in China exhibits pronounced spatial clustering patterns, influenced by technological, financial, and entrepreneurial environments³⁵. A highly efficient Sci-Tech financial system reduces uncertainty in capital allocation, enabling venture and private equity funds to channel resources more precisely toward strategic emerging industries and high-tech enterprises³⁶. This improves the matching efficiency between financial and industrial resources, strengthens the support capacity of industrial capital for innovation-driven sectors, and alleviates financing bottlenecks for new industries³⁷. Moreover, consistent with agglomeration externality theory, the clustering of venture capital generates positive financial spillover effects³⁸, enhancing capital effectiveness and reducing financing costs. Together, these mechanisms inject sustained financial momentum into the modernization of industrial systems.

Finally, improved STFE fosters the concentration of technological factors by accelerating the diffusion of innovations from R&D to industrial application³⁹. Through this process, emerging technologies permeate multiple stages of the industrial value chain, promoting breakthroughs and facilitating the entry of high-tech enterprises into innovation clusters⁴⁰. Overall, STFE enhances the coupling among talent, capital, and technology, optimizes the industrial resource endowment structure, and provides continuous endogenous power for building a modernized, innovation-driven industrial system. Accordingly, we propose the following hypothesis:

Hypothesis H4

Sci-Tech finance efficiency promotes the construction of modernized industrial systems by facilitating the optimal allocation and coordinated agglomeration of key production factors.

To further clarify the theoretical logic and the internal transmission channels, Fig. 1 provides a conceptual diagram of the mechanism framework linking STFE to the CMIS.

Fig. 1.

Theoretical framework of STFE driving the construction of modernized industrial system.

Model and variable

Model

Baseline regression model

Building on the theoretical analysis and hypotheses presented above, this study further examines the impact of STFE on the CMIS. Traditional policy evaluation studies often face limitations related to restrictive assumptions, sample selection bias, and data heterogeneity, which may lead to inconsistent or biased estimation results. To address these issues, this paper adopts the Double Machine Learning (DML) framework proposed by Chernozhukov et al. (2018)⁴¹, which effectively captures nonlinear relationships between variables and provides unbiased and asymptotically normal estimates of policy treatment effects even in finite samples.

Specifically, this study employs a partially linear DML model to estimate the causal effect of STFE on the modernization of the industrial system, as follows:

1

2

Where denotes the level of the CMIS in province at year ,and represents the STFE, The parameter measures the causal impact of STFE on industrial modernization. A significantly positive indicates that STFE promotes the construction of a modernized industrial system, whereas a significantly negative value implies a restraining effect. The vector represents a set of control variables that may simultaneously influence both and ,thereby acting as potential confounding covariates. The functions are unknown nonparametric components that allow for complex and potentially nonlinear relationships between the observed covariates and the outcome or treatment. These functions are estimated using advanced machine-learning algorithms, which allow flexible modeling of the data without imposing strict functional-form assumptions.denote error terms with zero conditional means, ensuring the identification of the treatment effect parameter under the orthogonality condition.

These nuisance functions are estimated using machine-learning algorithms that flexibly capture potentially nonlinear confounding patterns in the covariates, even when the set of controls is moderate. This specification therefore helps isolate the effect of STFE by combining flexible nuisance estimation with the orthogonalization step, yielding consistent causal estimates under the identifying assumptions.

We acknowledge that the modernization of industrial systems may also feed back into the development of the Sci-Tech finance system, which could introduce simultaneity or reverse causality. In this study, STFE is treated as an upstream financial–innovation capability that conditions industrial modernization, and the causal interpretation relies on the orthogonality condition in the DML framework after controlling for rich covariates and province and year fixed effects. To further alleviate endogeneity concerns, we additionally implement a partially linear DML instrumental-variable specification and conduct robustness checks using lagged STFE, both of which support the baseline conclusion.

Moderating effect model

To further explore the internal transmission mechanisms through which STFE influences the CMIS, this study employs a mediation effect model within the Double Machine Learning (DML) framework, as specified below.

3

Building upon the baseline model, the mechanism analysis proceeds in two stages.In the first stage, STFE is regressed on each mechanism variable to evaluate its direct influence on three key transmission channels: (1) Technology Transformation and Realization (TTR),(2) Digital–Real Technology Integration (DTI), and(3) Optimization of Factor Allocation (OFA).In the second stage, these mechanism variables are incorporated into the industrial modernization equation to assess their transmission effects between STFE and the CMIS. A significantly positive coefficient of STFE in the first stage and a significant coefficient of the mechanism variable in the second stage jointly confirm the existence of an effective mediating mechanism through which STFE promotes industrial modernization.

Variables

Dependent variable

The construction of modernized industrial systems (CMIS) is defined as the dependent variable in this study. CMIS represents the extent to which an industrial structure functions as a strategic pillar of a modernized economy and a key driver of national modernization. A well-developed modernized industrial system is expected to exhibit innovation leadership, security resilience, openness and competitiveness, and long-term sustainability. In essence, it simultaneously supports the modernization of the economic structure while embodying the attributes of innovation, coordination, and green transformation.

Following the analytical frameworks proposed by Wang(2023)⁴², a comprehensive evaluation index system is constructed to measure the level of CMIS development. The index comprises seven dimensions—physical modernization, innovativeness, integration, greenness, openness, coordination, and institutional support—which together capture the multi-faceted characteristics of industrial modernization, including structural upgrading, technological progress, digital–real integration, ecological transition, and institutional adaptability.

To quantify the level of CMIS development across regions, this study employs the entropy-weight method, which objectively assigns weights to each indicator based on the degree of information variation within the data. This approach minimizes subjective bias and enhances comparability across dimensions. The resulting composite index provides a systematic and data-driven measurement of the overall level of CMIS (see Table 1 for detailed indicators).

Table 1.

Comprehensive evaluation index system for the development of CMIS.

Primary indicator	Secondary indicator	Indicator description	Attribute
Physical modernization	Industrial development level	Growth rate of value added in the secondary industry (%)	+
	Industrial upgrading	Growth rate of fixed asset investment (%)	+
	Agricultural modernization	Value added per employee in the primary industry (10,000 CNY per person)	+
	Industrial modernization	Value added per employee in the secondary industry (10,000 CNY per person)	+
	Service sector modernization	Value added per employee in the tertiary industry (10,000 CNY per person)	+
Innovativeness	Innovation human capital investment	Full-time equivalent of R&D personnel (person-years)	+
	R&D investment intensity	R&D expenditure as a percentage of GDP (%)	+
	Innovation infrastructure	Number of higher education institutions	+
	Innovation output	Number of valid invention patents in large and medium-sized industrial enterprises	+
Integration	Digital industry scale	Total value of telecommunication services as a percentage of GDP (%)	+
	Information network infrastructure	Internet penetration rate (%)	+
	Communication digital infrastructure	Mobile telephone penetration rate (%)	+
	Digital labor input	Number of employees in the digital industry as a share of total employment (%)	+
	Digital capital investment	Growth rate of fixed asset investment in the digital industry (%)	+
Greenness	Energy conservation performance	Total electricity consumption per unit of GDP (kWh per 10,000 CNY)	–
	Water conservation performance	Total water supply per unit of GDP (10,000 m³ per 10,000 CNY)	–
	Pollutant discharge intensity (Wastewater)	Industrial wastewater discharge per unit of GDP (tons per 10,000 CNY)	–
	Pollutant discharge intensity (Waste Gas)	Industrial waste gas emissions per unit of GDP (10,000 m³ per 10,000 CNY)	–
	Solid waste recycling	Utilization rate of industrial solid waste (%)	+
	Urban greening level	Urban green coverage rate (%)	+
	Pollution control investment	Completed environmental protection investment as a share of total investment (%)	+
Openness	Foreign trade openness	Total import and export volume as a share of GDP (%)	+
	Foreign direct investment intensity	Inward FDI as a share of GDP (%)	+
	Foreign investment breadth	Number of newly established foreign-invested enterprises (units)	+
	International tourism intensity	Number of inbound international tourists (millions of persons)	+
Coordination	Economic development level	Per capita GDP (CNY/person)	+
	Industrial structure optimization	Value added of the secondary and tertiary industries as a share of GDP (%)	+
	Employment structure optimization	Employment in the secondary and tertiary industries as a share of total employment (%)	+
	Urban–rural income gap	Ratio of per capita disposable income of rural to urban residents (%)	+
	Urban residents’ income optimization	Engel coefficient of urban households (%)	–
	Rural residents’ income optimization	Engel coefficient of rural households (%)	–
Institutional support	Financial development scale	Value added of financial industry as a share of GDP (%)	+
	Financial efficiency	Ratio of outstanding loans to deposits (%)	+
	Human capital cultivation	Fiscal expenditure on education per capita (CNY/person)	+
	Public health investment	Fiscal expenditure on healthcare per capita (CNY/person)	+
	Social security and employment support	Fiscal expenditure on social security and employment per capita (CNY/person)	+
	Urban–rural infrastructure development	Fiscal expenditure on urban–rural community services per capita (CNY/person)	+

To further assess the robustness and comparability of the CMIS measure, alternative objective weighting schemes, including PCA and the CRITIC method, are employed, and the corresponding correlation and ranking-consistency results are reported in Supplementary Material.

Explanatory variables

Sci-Tech finance efficiency (STFE) is the core explanatory variable in this study and measures how efficiently the Sci-Tech finance system converts financial inputs and innovation-related human resources into observable innovation outputs. Conceptually, STFE reflects the quality of the finance–technology matching process by indicating how effectively capital and R&D labor are mobilized and transformed into scientific and technological outcomes. Empirically, STFE is estimated using the DEA–BCC model and is interpreted as a frontier-based relative technical efficiency score that measures the distance to the best-practice frontier under variable returns to scale. A higher STFE therefore signals a stronger supply-side financial–innovation capability, which is expected to facilitate technology commercialization and factor reallocation toward higher-value activities, thereby supporting industrial upgrading and the construction of a modernized industrial system. Within this process, governments and enterprises act as the dominant participants, while human talent functions as the essential productive element.

Following the analytical approach of Wang and Li (2022)⁴³,this study constructs a comprehensive evaluation framework for STFE in a frontier-based efficiency setting and estimates STFE using the Data Envelopment Analysis (DEA) Banker–Charnes–Cooper (BCC) model. The BCC model is particularly suitable for evaluating relative efficiency among decision-making units (DMUs) operating under variable returns to scale—an essential consideration for regional FinTech systems with heterogeneous innovation and financial capacities.

The input indicators represent the multi-source resource endowments supporting Sci-Tech finance and include: (1) Government capital investment, measured by fiscal expenditure on science and technology; (2) Enterprise capital investment, captured by the internal R&D expenditure and equipment investment of above-scale enterprises; (3) Labor input, represented by the number of R&D personnel across provinces.

The output indicators reflect the economic and social benefits generated by Sci-Tech finance–driven innovation activities and comprise: (1) Economic performance, represented by the turnover of the technology market; (2) Social performance, proxied by the number of published scientific papers and the number of authorized invention patents.

Together, these indicators comprehensively assess the extent to which financial and human resources are efficiently transformed into tangible innovation outputs. A higher STFE value signifies a more effective Sci-Tech financial system characterized by optimized capital allocation, faster technology commercialization, and closer operation to the best-practice frontier—thereby underscoring its crucial role in driving industrial modernization and high-quality economic development.

Mediating mechanism variables

To examine how STFE affects CMIS, we consider three mediation channels: technology transfer and commercialization, digital–real technology integration, and factor agglomeration. (1) Technology Transformation and Realization (TTR). This channel captures the extent to which scientific outputs are commercialized and converted into productive forces. TTR is proxied by the logarithm of the total transaction value in regional technology markets⁴⁴. A higher value indicates a more active technology market and stronger linkages between research outputs and industrial application, thereby facilitating industrial upgrading. (2) Digital–Real Technology Integration (DTI). This channel reflects the penetration of digital technologies into real-economy production networks. we construct a provincial DTI index based on patent classification codes and a technology–industry correspondence matrix, and use its logarithmic form in estimation⁴⁵. A higher DTI suggests deeper digital adoption in industrial processes and more efficient coordination across production activities, which supports intelligent and green upgrading. (3) Optimization of Factor Allocation (OFA). This channel captures whether STFE facilitates the reallocation and clustering of key production factors. Consistent with our empirical design, we operationalize this channel using three separate mediators. Capital agglomeration is measured by the logarithm of provincial venture capital investment⁴⁶. Talent agglomeration is proxied by the employment share in information transmission, computer services, software, scientific research, and technical services⁴⁷. Technological agglomeration is approximated using a composite index defined as: 0.5 × (government expenditure on science and technology / total fiscal expenditure) + 0.5 × (standardized per capita patent applications)⁴⁸. Higher values of these measures indicate stronger clustering and more efficient matching of capital, talent, and technology, which strengthens the endogenous momentum of industrial modernization.

Control variables

To address potential endogeneity and reduce omitted variable bias, six control variables are incorporated in this study following the approach of Zhang et al. (2025)⁹. These variables capture key socioeconomic and institutional characteristics that may influence the modernization of industrial systems.(1) Urbanization (Urban), measured as the share of the urban resident population in total population; (2) Economic development (Gdp), measured as the logarithm of per capita GDP; (3) Financial development (Finance), measured as the ratio of total deposits and loans of financial institutions to GDP; (4) Educational investment (Edu), measured as fiscal expenditure on education per capita; (5) Fiscal self-sufficiency (Fiscal), measured as the ratio of general public budget revenue to general public budget expenditure; and (6) Urban–rural consumption gap (Urrc), measured as the ratio of urban residents’ consumption level to rural residents’ consumption level. The definitions, measurements, and descriptive explanations of all variables used in this study are summarized in Table 2.

Table 2.

Variable definition.

Variable name	Variable labels	Specific explanation of variables
The modernized industrial systems	CMIS	Composite index of industrial modernization across seven dimensions: innovation, integration, greenness, openness, coordination, physical and institutional upgrading, measured by the entropy-weight method.
Sci-Tech finance efficiency	STFE	measured by the DEA-BCC method based on government, enterprise, and labor inputs and innovation output performance.
Technology transformation and realization	TTR	Log of total regional technology-market transactions
Digital–real technology integration	DTI	Index of digital–real industrial convergence based on patent–industry mapping data.
Optimization of factor allocation	OFA	Set of indicators capturing the agglomeration of capital, talent, and technology, reflecting the optimization of factor allocation.
Urbanization level	Urban	Ratio of urban to total population.
Economic development level	Gdp	Log of per-capita GDP.
Financial development level	Finance	Ratio of total deposits and loans to GDP
Educational investment intensity	Edu	Fiscal education expenditure per capita.
Fiscal self-sufficiency	Fiscal	Ratio of fiscal revenue to expenditure.
Urban–rural consumption gap	Urrc	Ratio of urban to rural consumption levels.

Data sources

Considering data availability and completeness, this study selects panel data from 31 provinces, municipalities, and autonomous regions in China covering the period 2010–2023. The data are primarily obtained from China Statistical Yearbook, China Science and Technology Yearbook, Statistical Communiqués on National Economic and Social Development, the CSMAR database, and various provincial statistical yearbooks.

To ensure data consistency and reliability, missing values in a few indicators are supplemented using the linear interpolation method, while all continuous variables are subjected to 1% winsorization at both tails to mitigate the influence of potential outliers on the estimation results.

Empirical analysis

Sample descriptive statistics

Table 3 reports the descriptive statistics of the main variables used in this study. The modernization level of the industrial system (CMIS) exhibits a mean value of 0.0856 and a standard deviation of 0.0371, suggesting moderate regional variation in industrial modernization across provinces. The average Sci-Tech finance efficiency (STFE) is 0.704, with a relatively large standard deviation (0.197), indicating notable heterogeneity in the efficiency of financial support for technological innovation. Among the control variables, the average urbanization rate (Urban) is 0.600, and the logarithm of per capita GDP (Gdp) averages 10.86, reflecting substantial economic disparities across regions. The mean financial development level (Finance) is 3.445, while educational investment (Edu) averages 2009, showing wide differences in human capital accumulation. Fiscal self-sufficiency (Fiscal) and the urban–rural consumption gap (Urrc) have mean values of 0.477 and 2.152, respectively, indicating moderate fiscal capacity but persistent regional inequality. Overall, the descriptive statistics display considerable cross-regional variability, providing a reliable foundation for subsequent regression analysis.

Table 3.

Descriptive statistics of variables.

Variable	N	Mean	p50	Sd	Min	Max
CMIS	420	0.0856	0.0723	0.0371	0.0506	0.250
STFE	420	0.704	0.704	0.197	0.288	1.042
Urban	420	0.600	0.588	0.123	0.363	0.893
Gdp	420	10.860	10.84	0.491	9.802	12.100
Finance	420	3.445	3.213	1.093	1.805	7.476
Edu	420	2009	1941	972.1	104.1	5193
Fiscal	420	0.477	0.440	0.200	0.0937	0.907
Urrc	420	2.152	2.094	0.378	1.533	3.249

Benchmark regression

This study employs a Double Machine Learning (DML) framework based on the Support Vector Machine (SVM) algorithm to estimate the impact of STFE on the construction of a modernized industrial system (CMIS). Following the standard DML procedure, we implement K-fold cross-fitting to obtain out-of-sample predictions of the nuisance functions and to mitigate overfitting. Specifically, alternative fold choices , , and are considered, and the estimated effect of STFE on CMIS remains positive and statistically significant across these specifications. A comparison of coefficient magnitudes and robust standard errors indicates that the cross-fitting delivers the most stable estimates. Therefore, we adopt as the baseline in the subsequent analysis, while reporting the results based on and as robustness checks. The regression results are presented in Table 4.

Table 4.

Empirical results of benchmark Regression.

Variable	(1)CMIS	(2)CMIS	(3)CMIS
STFE	0.0246*** (0.0095)	0.0248*** (0.0095)	0.0267*** (0.0097)
First-order term of control variable	YES	YES	YES
Quadratic term of control variable	NO	NO	YES
Time FE	NO	YES	YES
Province FE	NO	YES	YES
Observations	420	420	420

The values in parentheses represent robust standard errors.***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The same notation applies hereafter.

Columns (1)–(3) sequentially introduce the control variables, quadratic terms, and both province and year fixed effects. Across all specifications, the estimated coefficients of STFE remain positive and statistically significant at the 1% level, indicating that higher STFE effectively promotes the organic integration of capital and technology. This finding highlights STFE as a key driver in advancing the modernization of industrial systems and provides preliminary empirical support for Hypothesis H1.

Endogeneity issues

Following Chernozhukov et al. (2018)⁴¹, we strengthen the identification strategy by adopting a partially linear Double Machine Learning instrumental-variable framework to address potential endogeneity arising from omitted variables and reverse causality. This approach combines orthogonalization with flexible machine-learning nuisance estimation, allowing the causal effect of STFE to be identified while flexibly accounting for potentially nonlinear confounding relationships among the observed covariates.

IV method test

The first instrument (IV1) is constructed as the interaction between predetermined communication infrastructure and a time-varying digital-economy measure, namely, the number of fixed telephones in 1984 multiplied by the level of digital-economy development. Historically, early telecommunications infrastructure reduced communication costs and information frictions in financial transactions, thereby improving information transmission and risk screening and facilitating the subsequent evolution of finance–technology matching. Therefore, this interaction is expected to be strongly related to the efficiency of Sci-Tech finance. Regarding exogeneity, the identifying variation stems from the differential exposure generated by a predetermined historical endowment interacting with contemporary digital development, rather than from the historical endowment alone. In addition, our empirical specification conditions on rich time-varying covariates as well as province and year fixed effects, which absorb persistent cross-province differences and common macro shocks, making it less likely that the remaining variation is driven by historical industrialization levels per se.

Considering lagged effects

To further mitigate concerns about simultaneity and dynamic endogeneity, we additionally use the one-period lag of STFE as an alternative instrument (IV2). The lag structure captures path dependence in financial efficiency while ensuring temporal precedence relative to current industrial modernization. Estimation is implemented within the DML-IV procedure, where the nuisance components are learned using the Support Vector Machine algorithm to flexibly approximate nonlinear relationships without imposing restrictive functional-form assumptions. As reported in Table 5, the IV-based estimates remain positive and statistically significant, indicating that the baseline conclusion is robust after accounting for endogeneity concerns.

Table 5.

Instrumental variable regression Results.

Variable	(1)IV method	(2)Lagged STFE
STFE	0.5517*** (0.1272)	1.3982** (0.6207)
First-order term of control variable	YES	YES
Quadratic term of control variable	YES	YES
Time FE	YES	YES
Province FE	YES	YES
Observations	420	390

The empirical results, summarized in Table 5, show that both IV1 and IV2 yield positive and statistically significant coefficients at the 1% and 5% levels, respectively. These findings confirm that, after addressing potential endogeneity through the DML-IV procedure, the positive impact of STFE on the CMIS remains robust and statistically reliable. Notably, IV-based estimates should be interpreted as local average causal effects for the subset of provinces whose STFE is shifted by the instrument. When effects are heterogeneous, this local effect can be larger than the benchmark estimate, because identification relies on the provinces that respond most to the exogenous shift. This interpretation is consistent with the IV literature emphasizing that compliers may exhibit stronger marginal responses, which can lead to larger IV magnitudes without contradicting the baseline positive relationship.

Robustness tests

Adjustment of the research sample

Given that municipalities directly under the central government (Beijing, Shanghai, Tianjin, and Chongqing) possess distinctive characteristics in economic structure, resource allocation, and science–technology policy, their inclusion could bias the results. Therefore, these municipalities are excluded from the analysis. Moreover, the COVID-19 pandemic that erupted after 2020 imposed unprecedented shocks on regional economies and financial systems, potentially disturbing model estimations. To minimize pandemic-related effects, the sample period is restricted to 2013–2020.

Based on the adjusted sample, the regression results in Table 6, columns (1) and (2), show that the estimated coefficients of STFE remain positive and statistically significant at the 1% and 5% levels. This demonstrates that even after excluding municipalities and shortening the sample window, the positive linkage between STFE and the modernization of industrial systems remains robust. The results confirm that the study’s conclusions are not driven by sample selection or external shocks, but reflect a stable and generalizable empirical relationship.

Table 6.

Robustness Tests.

Variable	(1) Excluding special sample	(2) Adjusted sample period	(3) Changing sample split Ratio I	(4) Changing sample split ratio II
STFE	0.0252*** (0.0093)	0.0306** (0.0153)	0.0277***(0.0101)	0.0280***(0.0101)
First-order term of control variable	YES	YES	YES	YES
Quadratic term of control variable	YES	YES	YES	YES
Time FE	YES	YES	YES	YES
Province FE	YES	YES	YES	YES
Observations	364	240	420	420

Adjustment of sample split proportion

To examine the sensitivity of the main results to model parameter settings, the parameters of the Double Machine Learning (DML) framework were further adjusted. Specifically, the ratio of training to testing samples was modified from 1:2 to 1:4 and 1:6, and the estimation procedure was repeated 101 times. The median of the estimated coefficients was used as the final result to mitigate random variation.

As reported in Table 6, columns (3) and (4), the coefficients of STFE remain positive and statistically significant at the 5% level across all parameter configurations. This consistency confirms the robustness of the findings to alternative model specifications and validates that the positive effect of STFE on CMIS is stable and reliable.

Resetting the dual machine learning algorithm

To examine whether the benchmark results are sensitive to the choice of algorithm, the Support Vector Machine (SVM) used in the Double Machine Learning (DML) framework is replaced with a Random Forest (RF) algorithm. The model is then re-estimated accordingly. The results show that the coefficients of STFE remain positive and statistically significant at the 1% level. This finding further corroborates the robustness of the benchmark estimations, indicating that the positive impact of STFE on the CMIS is not driven by algorithm selection but reflects a consistent and stable empirical relationship.

Considering province-time interaction fixed effects

A remaining concern is that provinces may follow different long-run development trajectories, and such gradual, unobserved changes may simultaneously affect both STFE and industrial modernization. To absorb this source of time-varying heterogeneity, we augment the baseline specification by including province-specific linear time trends. The results indicate that the estimated effect of STFE on CMIS remains positive and statistically significant at the 1% level. This suggests that our findings are robust to differential long-run provincial trends and are unlikely to be driven by evolving historical industrialization paths.

Elimination of the influence of other policies

In recent years, various provinces in China have successively introduced strategic policies closely related to digitalization and innovation-driven development, such as the National Big Data Comprehensive Pilot Zones, Digital Economy Innovation Pilot Provinces, and Next-Generation Artificial Intelligence Innovation and Development Pilot Zones. The implementation of these policies may simultaneously affect both STFE and the CMIS, potentially generating policy-interaction effects that could bias the estimation results.

To mitigate the potential interference of such concurrent policy shocks, this study explicitly controls for these external policy variables—namely, whether a province has been designated as a National Big Data Pilot Zone, a Digital Economy Innovation Province, or an AI Innovation Pilot Zone. The regression results, presented in Table 7, column (3), demonstrate that the estimated coefficient of STFE remains positive and statistically significant at the 1% level, even after accounting for these exogenous policy effects. This finding provides strong evidence that the observed relationship between STFE and industrial modernization is not driven by overlapping policy influences, but rather reflects a genuine internal mechanism through which improved financial efficiency promotes industrial modernization.

Table 7.

Others robustness tests.

Variable	(1)Resetting algorithm	(2)Province-time interaction fixed	(3)Other policies	(4)Replacing dependent variable	(5)Replacing explanatory variable
STFE	0.0208***(0.0073)	0.0137** (0.0064)	0.0268***(0.0097)	0.5873** (0.2377)	0.0215***(0.0065)
First-order term of control variable	YES	YES	YES	YES	YES
Quadratic term of control variable	YES	YES	YES	YES	YES
Time FE	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES
Observations	420	420	420	420	420

Alternative dependent variable test

To verify the robustness of the findings with respect to the measurement of the dependent variable, this study reconstructs the CMIS index using the Principal Component Analysis (PCA) method and re-estimates the model accordingly. The regression results, reported in Table 7, column (4), show that the estimated coefficients of STFE remain positive and statistically significant at the 5% level, with no substantial change in magnitude or direction compared to the benchmark results.

This finding further confirms that the positive effect of STFE on industrial modernization is not driven by the specific measurement approach of CMIS, but rather reflects a robust and consistent empirical relationship between technological finance efficiency and the advancement of modern industrial systems.

Alternative explanatory variable test

To further verify the robustness of the main findings, this study replaces the original measurement of STFE with one derived from the EBM super-efficiency model, which accounts for data heterogeneity and the relative importance of inputs in assessing technical efficiency⁴⁹. Based on the recalculated STFE values, the regression is re-estimated, and the results are presented in Table 7, column (5).

The estimated coefficients of STFE remain positive and statistically significant at the 1% level, indicating that the benchmark conclusion is essentially unchanged. This result further confirms that the positive effect of STFE on CMIS is robust across different efficiency measurement methods, providing additional evidence of the stability and reliability of the empirical results.

Analysis of the impact mechanisms

To further elucidate the internal transmission channels of the benchmark relationship, this section investigates three potential mechanisms through which STFE facilitates industrial modernization: the transformation of scientific and technological achievements, the integration of digital and real technologies, and the optimization of factor allocation.

Facilitating the transformation of scientific and technological achievements

The first mechanism concerns the transformation of scientific and technological achievements, which serves as a vital bridge linking scientific research with industrial application and provides sustained momentum for industrial modernization. As shown in Table 8, column (1), STFE significantly promotes technology transfer, with the estimated coefficient being positive and statistically significant at the 1% level, thereby confirming Hypothesis H2. This result implies that higher financial efficiency strengthens the channels through which scientific achievements are commercialized, facilitating the deep integration of technological innovation and industrial innovation. By improving capital accessibility and reducing transaction costs in technology markets, Sci-Tech finance enhances the matching efficiency between R&D institutions and enterprises, enabling scientific outputs to evolve into tangible productive forces. This process not only accelerates industrial upgrading and structural optimization but also generates sustained innovation momentum that underpins the long-term competitiveness of the CMIS.

Table 8.

Results of mechanisms analysis.

Variable	(1) TTR	(2) DTI	(3) Capital agglomeration	(4) Talent agglomeration	(5) Technology agglomeration
STFE	1.4170***(0.4675)	1.0596** (0.5243)	1.3199*(0.7199)	3.0778***(0.6167)	0.3449***(0.1046)
First-order term of control variable	YES	YES	YES	YES	YES
Quadratic term of control variable	YES	YES	YES	YES	YES
Time FE	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES
Observations	420	420	420	420	420

Empowering digital–real technology integration

The integration of emerging digital technologies with traditional industries has become a defining direction for developing new forms of digital productivity within the modernized industrial system. As shown in Table 8, column (2), the regression results reveal that STFE significantly enhances digital–real integration, with the coefficient remaining positive and statistically significant at the 5% level, thereby confirming Hypothesis H3. This finding suggests that improvements in financial efficiency play a crucial role in promoting digital–industrial convergence by channeling resources into digital infrastructure, data platforms, and intelligent applications. Through this process, Sci-Tech finance strengthens the coupling between innovation and production, accelerates intelligent and green transformation, and injects new momentum into the upgrading of the CMIS.

Optimizing factor allocation

Technological innovation, modern finance, and human capital constitute the fundamental pillars of the real economy and represent indispensable foundations for the construction of a modernized industrial system⁵⁰. Building upon this theoretical premise, this study investigates how STFE contributes to industrial modernization by facilitating the coordinated agglomeration and optimal allocation of three critical production factors: capital, talent, and technology.

Capital agglomeration

As shown in Table 8, column (3), STFE significantly enhances capital agglomeration, with the coefficient positive and statistically significant at the 10% level. This finding suggests that an efficient Sci-Tech financial system mitigates uncertainty and transaction costs in capital allocation, allowing venture and entrepreneurial investment to flow more effectively toward strategic emerging industries and high-technology enterprises. The clustering of financial capital strengthens the financing capacity of innovation-oriented sectors and channels resources into innovation-driven domains, thereby providing robust financial support for industrial upgrading and structural optimization within the CMIS.

Talent agglomeration

As shown in Table 8, column (4), STFE exerts a significant and positive effect on talent agglomeration at the 1% statistical level. Improvements in financial efficiency create an enabling institutional and economic environment that attracts and retains high-quality human capital, particularly in technology-intensive and research-oriented sectors. Concentrated talent networks amplify knowledge spillovers, strengthen innovation diffusion, and foster cross-industry collaboration. These effects collectively enhance the dynamic capacity for innovation within regions, transforming financial support into human capital advantages that sustain technological progress and industrial competitiveness over time.

Technology agglomeration

The regression results, reported in Table 8, column (5), reveal that STFE significantly strengthens technology agglomeration by optimizing the financing and commercialization environment for innovation. Efficient Sci-Tech finance facilitates the continuous investment of funds into frontier R&D, accelerates the translation of research achievements into marketable products, and promotes the diffusion of new technologies across industries and regions. As a result, technological resources become more spatially concentrated and interconnected, forming an innovation ecosystem conducive to sustained technological upgrading. Through this channel, Sci-Tech finance not only enhances the regional density of innovation activities but also consolidates the technological foundations of industrial modernization.

While the coefficient magnitudes are not directly comparable across differently scaled mechanism variables, the results suggest heterogeneous strength across channels in terms of statistical salience. In particular, the association between STFE and talent agglomeration appears most pronounced, as reflected by its stronger statistical significance relative to capital and technology agglomeration. This pattern is consistent with the view that improvements in the efficiency of the finance–technology matching process can more rapidly attract and concentrate knowledge-intensive labor, which in turn enhances regional innovation capacity and supports sustained industrial upgrading.

Taken together, these results confirm that STFE serves as a key institutional driver for optimizing factor allocation. By promoting the coordinated clustering of capital, talent, and technology, STFE enhances innovation efficiency, strengthens endogenous growth momentum, and accelerates the transformation and upgrading of the CMIS.

Heterogeneity analysis

Heterogeneity by level of economic development

To further examine whether the impact of STFE on the CMIS varies across regions with different levels of economic development, provinces were divided into two groups based on the mean level of per capita GDP. The results, presented in Table 9, reveal a clear heterogeneity in the effect of STFE across regions. Specifically, in provinces with higher levels of economic development, the coefficient of STFE remains positive and statistically significant at the 1% level, indicating that financial efficiency strongly facilitates the advancement of industrial modernization. In contrast, in less-developed regions, the estimated coefficient is statistically insignificant, suggesting that the enabling effect of STFE on industrial modernization is weaker or even absent.

Table 9.

Results of heterogeneity Analysis.

Variable	(1) Higher economic development	(2) Lower economic development	(3) Higher marketization level	(4) Lower marketization level	(5) Higher fiscal pressure	(6) Lower Fiscal Pressure
STFE	0.0511***(0.0161)	-0.0033 (0.0074)	0.0610***(0.0177)	-0.0085(0.0081)	0.0293**(0.0137)	0.0338**(0.0140)
First-order term of control variable	YES	YES	YES	YES	YES	YES
Quadratic term of control variable	YES	YES	YES	YES	YES	YES
Time FE	YES	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES	YES
Observations	205	215	216	204	171	249

This divergence may stem from structural disparities in financial and innovation systems between the two groups of regions. Economically developed provinces typically possess more mature financial systems, stronger technological innovation capacities, and more efficient mechanisms for Sci-Tech finance operation and technology commercialization⁵¹. These institutional and structural advantages allow Sci-Tech finance to more effectively channel financial resources into high-value innovation activities, thereby amplifying its role in promoting industrial upgrading and structural optimization. Conversely, less-developed regions are often constrained by limited financial resources, narrow industrial bases, and weak innovation ecosystems, which hinder the effective transformation of improved financial efficiency into tangible industrial progress. Consequently, the positive role of STFE in fostering the modernization of industrial systems is predominantly concentrated in economically advanced areas.

Heterogeneity by degree of marketization

To further assess whether the effect of STFE on CMIS is influenced by the level of marketization, this study divides provinces into two groups based on the median value of the marketization index reported in the China Provincial Marketization Index Report. The results, presented in Table 9, reveal a distinct heterogeneity across regions. In provinces with higher degrees of marketization, the coefficient of STFE remains positive and statistically significant at the 1% level, indicating that a well-functioning market environment amplifies the positive impact of financial efficiency on industrial modernization. In contrast, in provinces with lower levels of marketization, the estimated coefficient is statistically insignificant, implying that the effectiveness of STFE in promoting industrial modernization is considerably constrained (Supplementary information).

The observed disparity can be attributed to differences in institutional environments and resource allocation mechanisms. Regions with higher marketization levels typically exhibit stronger property-rights protection, more flexible factor-mobility systems, and more competitive market structures, which together reduce institutional frictions in capital allocation and technology transfer. These institutional advantages enhance the efficiency with which financial resources are channeled into innovation-driven sectors, facilitating technological breakthroughs and industrial upgrading. Conversely, in regions with lower marketization degrees, underdeveloped financial systems and rigid institutional constraints hinder the transmission of financial efficiency into innovation outcomes, consistent with evidence that weak marketization dampens the impact of financial development on regional innovation efficiency⁵². Consequently, the capacity of STFE to drive industrial modernization is largely dependent on the maturity and openness of the regional market system.

Heterogeneity by fiscal pressure

To further examine the moderating effect of local fiscal conditions on the relationship between STFE and CMIS, this study use the ratio of the difference between local fiscal expenditure and revenue to total fiscal revenue as a proxy for fiscal pressure. Based on the mean value of this indicator, provinces were divided into two groups representing regions with high and low fiscal pressure.

The regression results, reported in Table 9, indicate that STFE exerts a significant positive effect on the CMIS in both groups, with coefficients statistically significant at the 5% level. However, a notable difference emerges in the magnitude of the coefficients: the positive impact of STFE is substantially stronger in regions with lower fiscal pressure. This finding suggests that fiscal conditions influence the extent to which financial efficiency can translate into industrial modernization outcomes.

The underlying mechanism may lie in the policy space and fiscal flexibility available to local governments. In regions with lighter fiscal pressure, governments typically possess greater capacity for public investment, industrial support, and financial innovation, which enables them to coordinate more effectively with Sci-Tech finance initiatives. This synergy amplifies the role of financial efficiency in promoting technological transformation and industrial upgrading. Conversely, in regions facing heavy fiscal burdens, limited fiscal capacity and rigid expenditure structures constrain policy support and reduce the complementarities between fiscal policy and financial efficiency. As a result, the enabling effect of STFE on industrial modernization is relatively weakened in fiscally constrained regions.

Conclusions and implications

Conclusions

Drawing on provincial panel data from China spanning 2010–2023, this study employs the Double Machine Learning (DML) framework to empirically examine the impact and mechanisms of STFE on CMIS. The main conclusions are as follows: First, STFE significantly promotes the development of CMIS, and this conclusion remains robust after a series of endogeneity and stability tests. Second, the mechanism analysis demonstrates that STFE drives high-quality industrial development primarily through three channels: (1) facilitating the transformation of scientific and technological achievements, (2) empowering digital–real technology integration, and (3) optimizing the agglomeration and allocation of key production factors, including talent, capital, and technology. Third, heterogeneity analysis reveals that the positive impact of STFE is more pronounced in regions with higher levels of economic development and marketization, as well as in those facing lower fiscal pressure, indicating that regional institutional and market environments play crucial moderating roles in shaping the effectiveness of Sci-Tech finance.

Recommendations

1. Improve the Sci-Tech finance system to enhance resource allocation efficiency. China should accelerate the improvement of institutional arrangements for financial support to technological innovation by developing a multi-tiered capital market and diversifying Sci-Tech financial instruments. Strengthening the integration of policy finance, long-term capital, and private investment can alleviate financing constraints, reduce information asymmetries, and facilitate the industrialization and commercialization of scientific achievements. This would accelerate the formation of new-quality productive forces and establish a more inclusive and efficient Sci-Tech financial ecosystem.

2. Deepen digital–real integration to cultivate new drivers of industrial growth. Financial institutions should leverage emerging technologies such as big data, artificial intelligence, and blockchain to enhance their capacity to identify and support innovation activities. By aligning capital flows with technological and industrial needs, they can promote the digital transformation of traditional sectors and accelerate the growth of strategic emerging industries. Moreover, improving supporting mechanisms—such as intellectual property pledge financing and technology insurance—can reduce the risks and uncertainties associated with enterprise digitalization, fostering a resilient and innovation-driven industrial ecosystem.

3. Foster factor agglomeration and promote regionally coordinated development through differentiated policies. STFE should serve as a catalyst for the spatial concentration of talent, capital, and technology, thereby facilitating the diffusion of innovation and the formation of regional industrial clusters. Policymakers should encourage advanced regions to further deepen financial innovation and enhance the marginal efficiency of Sci-Tech finance. Meanwhile, less-developed regions should expand public investment, strengthen policy-based financial support, and optimize their business environments to gradually narrow regional disparities. Coordinating factor mobility and regional policies can help establish a nationally integrated and synergistic Sci-Tech finance framework, injecting sustained momentum into the CMIS.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, R.H.; writing—original draft preparation and writing—review and editing, R.H., X.L.; writing—review and editing, X.L.; methodology, R.H., J.T.; data curation, S.W., C.L., Q.Z.; All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the 2025 Special Project for Research in Philosophy and Social Science in Shaanxi Province (Grant No. 2025YB0295), the Xi’an Social Science Planning Fund Project (Grant No. 25JX147), and the Xi’an International Studies University Research Project (Grant No. 25XWC05).

Data availability

Data will be made available through the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval and consent to participate

This study uses aggregated, publicly available province-level panel data (2010–2023) compiled from official statistical yearbooks and research databases. It does not involve human participants, human biological materials, or the collection, processing, or analysis of any personally identifiable information. Therefore, ethics approval and informed consent are not required. This determination is consistent with the Measures for the Ethical Review of Life Science and Medical Research Involving Humans (National Health Commission of the People’s Republic of China, 2023), which govern ethical review requirements for research involving human participants. In addition, the academic ethics review body of the School of Economics and Finance, Xi’an International Studies University, has confirmed that this study falls outside the scope of human-subject research and issued a formal waiver of ethics approval.