Digital Finance, Innovation Value Chains, and the Formation of New Productivity: Evidence from Technology-Based SMEs

Abstract

Technology-based small and medium-sized enterprises (SMEs) are vital agents of innovation, with their technological advancements serving as the primary driver of new quality productive forces. However, the intrinsic linkages between these entities and productivity formation remain insufficiently elucidated. This study examines a sample of Chinese technology-based SMEs from 2019 to 2024, including 1265 companies and 6705 firm-year observations. Drawing on the innovation value chain perspective and technological innovation process theory, we deconstruct the innovation process into three distinct stages—R&D, production/manufacturing, and marketization—which correspond to the three core elements of new quality productive forces: “new technologies,” “new laborers,” and “new value creation”. This research intends to clarify the intrinsic coupling mechanisms between the innovation value chain and the constituent elements of new quality productive forces under the empowerment of digital finance. The results reveal that digital finance exerts a significant positive driving effect on all three stages of the innovation value chain. The most pronounced impact is observed in the “new technologies” associated with the R&D stage, followed by “new value creation” in the marketization stage and “new laborers” in the production/manufacturing stage. Mechanism analysis demonstrates that both usage depth and coverage breadth of digital finance exhibit substantial driving effects. Specifically, digital finance promotes the formation of new quality productive forces through three pathways: increasing R&D investment, raising the proportion of high-tech personnel, and expanding market share. These findings refine the theoretical framework integrating digital finance, technological innovation, and new quality productive forces, offering a practical pathway for the cultivation of new productivity.

1. Introduction

The global shift toward high-quality development has positioned new quality productive forces as a cornerstone of national competitiveness, with innovation-driven technological upgrading emerging as a critical pathway for economies worldwide [1]. China has explicitly elevated the cultivation of new quality productive forces to a national strategic priority, aiming to achieve breakthroughs in total factor productivity through technological innovation and institutional reform [2]. Unlike traditional productivity paradigms, new quality productive forces hinge on cutting-edge technologies such as artificial intelligence and blockchain, alongside systemic innovation, requiring agile micro-level actors to translate technological potential into industrial transformation [3]. Within this framework, technology-driven SMEs (Small and Medium-sized Enterprises) function as vital agents of innovation, particularly in fostering emerging industries and future-oriented sectors [4]. However, their capacity for sustained innovation is hindered by structural constraints, including insufficient collateral, high R&D (Research and Development) uncertainty, and persistent financing frictions within conventional financial systems [5].

Globally, the tension between innovation ecosystems and financial rigidity is not unique to China. Advanced economies face comparable challenges, as traditional financial institutions tend to favor low-risk, asset-backed lending and routinely underserve high-tech SMEs engaged in radical innovation [6]. Policy initiatives such as the European Union’s “Digital Finance Strategy” and recent reforms by the U.S. Securities and Exchange Commission underscore international efforts to harmonize innovation incentives with financial risk management [7,8]. Nonetheless, fragmented regulatory frameworks and data silos continue to impede efficient capital allocation [9]. Against this backdrop, digital finance has emerged as a transformative force, transcending geographical boundaries and redefining global financial inclusion by integrating AI-driven analytics, big data platforms, and blockchain infrastructures [10]. Empirical studies increasingly document its capacity to enhance innovation performance in SMEs, particularly in terms of patenting activity and R&D investment intensity [11].

This study addresses a significant research gap: while existing work predominantly focuses on the macro-level interactions between digital finance and productivity growth, the micro-level mechanisms linking digital finance → SME technological innovation → new quality productive forces remain insufficiently examined [12]. Specifically, this research aims to clarify how digital finance—as an innovative financial service—reshapes the technological innovation capabilities of technology-based SMEs in the process of generating new quality productive forces [13]. It further seeks to elucidate the intrinsic connections between firm-level technological innovation and the constituent elements of new quality productive forces, as well as to identify the transmission pathways through which digital finance influences this transformation [14]. These inquiries hold both theoretical and practical importance: they contribute to the effective implementation of China’s national agenda on high-quality financial services, accelerate the cultivation of new quality productive forces, align with global trends in industrial upgrading, and support the development of intelligent and digital infrastructure systems [15].

The contributions of this study are threefold. First, drawing on the perspective of new quality productive forces and technological innovation theory, we conceptualize the technological innovation process of technology-based SMEs—comprising R&D, production, and marketization—and integrate it with the core elements of new quality productive forces (“new technologies,” “new laborers,” and “new value creation”), thereby clarifying why technological innovation serves as a fundamental driver of new quality productive forces. Second, using quantitative analyses, we empirically investigate the impact of digital finance on the formation of new quality productive forces in SMEs, filling a critical gap in the literature regarding the micro-level linkages among digital finance, technological innovation, and productivity generation. Third, we uncover the transmission mechanism by constructing an analytical framework of “digital finance → enhanced SME technological innovation capacity → new quality productive forces generation,” offering theoretical support for the role of digital finance in promoting high-quality development and enriching the empirical research paradigm at the intersection of digital finance, technological upgrading, and productivity enhancement. Collectively, the findings extend the literature on financial innovation and provide concrete policy implications for leveraging digital finance to strengthen global competitiveness and sustainable innovation.

The remainder of this paper is structured as follows. Section 1 (Introduction) provides an overview of the research background, significance, primary objectives, and a concise summary of the key innovations presented herein. Section 2 conducts a comprehensive literature review, synthesizing existing research on digital finance, technological innovation in technology-based small and medium-sized enterprises (SMEs), and the concept of new quality productive forces. Building upon the theoretical framework of the technological innovation process, Section 3 develops hypotheses concerning the impact of digital finance on distinct stages of technological innovation within technology-based SMEs. Furthermore, it analyzes the interrelationships among digital finance, each stage of technological innovation, and the constituent elements of the firm’s new quality productive forces. Section 4 details the research methodology employed, including the selection of samples, measurement of variables, and the specification of econometric models. Section 5 presents the empirical results obtained from data analysis and systematically tests the proposed hypotheses. Finally, Section 6 concludes the study by summarizing the main findings, discussing their theoretical and practical implications, acknowledging limitations, and suggesting avenues for future research.

2. Literature Review

2.1. Digital Finance and SME Innovation: Mechanisms and Empirical Evidence

Research has repeatedly identified digital finance as an instrument that can alleviate SMEs’ financing constraints and thus stimulate firm-level innovation [16]. Empirical work using Chinese firm-level and regional data shows that greater digital finance penetration is associated with higher R&D intensity and patenting among SMEs, mediated through relaxed financing frictions [17]. Studies using cross-country and multi-region datasets corroborate these findings, indicating that fintech innovations such as AI-based credit scoring and alternative data analytics reduce information asymmetry and improve credit access for smaller firms [18]. At the same time, careful analyses show that not all fintech mechanisms are uniformly beneficial: algorithmic credit scoring and automated underwriting can produce unintended exclusionary effects for firms with volatile cash flows or poor digital footprints [19]. Field evidence also shows that platform-based supply-chain finance and automated invoice discounting reduce transaction costs and speed working-capital provision—benefits that appear particularly large for manufacturing SMEs engaged in fast product cycles [20,21]. Nevertheless, uneven digital infrastructure and digital trust deficits remain important barriers: firms lacking digital readiness face higher online loan rejection rates and slower adoption of digital financial tools, which in turn slows innovation diffusion in lagging regions [22]. Several recent studies extend this literature by examining digital finance’s role during shocks (e.g., pandemics), where fintech channels have helped some SMEs survive liquidity crises but have also exposed firms to novel digital-credit risks [23,24].

2.2. Theoretical Foundations of New-Quality Productive Forces: Evolution and Debates

The concept of new-quality productive forces (NQPF) situates technological breakthroughs (AI, IoT, advanced digital platforms) and institutional reforms at the center of productivity transformations that transcend incremental gains [25]. This perspective draws on Schumpeterian insights about creative destruction but emphasizes systems-level complementarities among digital infrastructure, human capital and governance that can generate non-linear productivity jumps [26]. Debates in the literature revolve around the relative importance of pure technological advances versus enabling institutions: some scholars stress frontier technologies (e.g., generative AI, quantum computing) as the primary catalysts of NQPF [27], while others argue institutional adaptation—flexible regulation, digital governance, and IP regimes—is necessary to diffuse and scale technological gains [28]. In the SME context, scholars have proposed dual frameworks emphasizing simultaneous investments in human capital (upskilling in data and digital tools) and inter-firm/sectoral collaboration platforms to internalize productivity gains [29]. Yet empirical work quantifying how SMEs translate technological inputs into firm-level NQPF remains limited, particularly on measures of innovation quality and the production of novel productive capacities rather than mere patent counts [30].

2.3. Synergistic Mechanisms: Digital Finance as a Catalyst for NQPF

Recent syntheses highlight three complementary pathways through which digital finance may catalyze NQPF: (i) risk diversification and alternative liquidity channels (e.g., marketplace lending, tokenized assets) that allow SMEs to undertake riskier, long-horizon R&D [31]; (ii) resource reconfiguration via cloud and platform ecosystems that free up firm resources for core R&D, enabling faster iteration and scaling [32]; and (iii) value co-creation through digital marketplaces and tokenized incentive structures that align stakeholders and accelerate commercialization [33,34]. Nevertheless, institutional and infrastructural constraints limit these potentials: digital redlining and uneven broadband/digital payments coverage hinder rural SME access to AI-enabled finance, producing regional gaps in innovation outcomes [35,36]. Moreover, cross-jurisdiction regulatory fragmentation raises compliance costs for SMEs operating internationally, dampening cross-border innovation collaboration [37]. Taken together, these analyses indicate that while digital finance can be a potent enabler of new-quality productive forces, the effects are contingent on digital infrastructure, governance arrangements, and firm-level absorptive capacity—a rationale for the present study’s micro-mechanism focus.

2.4. Research Gaps and This Study’s Contribution

Despite growing evidence that digital finance supports SME innovation, three critical gaps persist. First, much empirical work uses aggregate outputs (e.g., patent counts) without unpacking the stage-specific micro-mechanisms by which digital finance affects R&D, manufacturing capabilities, and commercialization outcomes. Second, the literature is regionally concentrated—many studies exploit China’s rich administrative indices—leaving questions about external validity across heterogeneous financial and regulatory regimes. Third, the “black box” connecting digital finance → firm innovation behavior → the generation of NQPF (particularly innovation quality, skilled workforce formation, and market-value creation) is underdeveloped. Addressing these gaps, the current study examines stage-specific transmission channels in technology-based SMEs and empirically links digital finance adoption to the micro-processes that underpin new-quality productive forces.

3. Theoretical Basis and Hypothesis Analysis

3.1. The Impact of Digital Finance on Technological Innovation in Technology-Based SMEs

Accelerating the development of new-quality productive forces requires cultivating emerging and future industries and promoting the growth of specialized and innovative SMEs. Technology-based SMEs, as core carriers of future industries, face long innovation cycles, high sunk costs, and uncertain outcomes; from classical perspectives such as externality theory and resource dependence theory, such innovation activities are inherently risky and demand stable and sufficient financial resources [38]. Yet, under traditional financial systems, technology-based SMEs typically encounter severe financing constraints. Their dual performance objectives, economic efficiency and social value, combined with high information acquisition costs and rising unit lending costs for small loans, lead financial institutions to prefer larger, more transparent firms; information asymmetry further reinforces this bias, preventing technologically dynamic but unstable SMEs from accessing adequate credit [39,40].

Digital finance—grounded in big data, artificial intelligence, cloud computing, and distributed ledger technologies—offers a structural solution. First, consistent with information asymmetry theory, digital finance enhances the availability, granularity, and real-time accessibility of firm-level data. Big data analytics and fintech credit evaluation systems enable financial institutions to build comprehensive and dynamically updated credit and performance profiles of SMEs, thereby reducing information “black holes” [41]. AI-driven risk assessment improves the accuracy of creditworthiness judgments and the detection of potential misconduct, thereby enhancing both efficiency and transparency in lending decisions [42]. Blockchain-based financial applications further support long-tail SME financing by enabling decentralized, tamper-resistant transaction records and supply-chain finance mechanisms, which reduce lenders’ perceived risk and expand access for firms that lack traditional collateral [43].

Second, from the perspective of transaction cost theory, a major source of high transaction costs between SMEs and traditional financial institutions is a lack of trust. Blockchain-enabled smart contracts and decentralized ledger technologies can embed financial institutions and SMEs in a shared, tamper-proof transactional environment, where contract terms, payments, and transaction flows are transparent and automatically enforceable [44]. This mechanism reduces negotiation, monitoring, and enforcement costs, establishing a more secure and cooperative financial relationship [45].

Taken together, digital finance mitigates long-standing structural barriers that limit SME innovation, reduces financing frictions, and strengthens the resource base required for sustained technological development. Therefore, this study proposes the following overarching hypothesis, which will be empirically tested in Model (1):

Hypothesis 1.

The advancement of digital finance significantly enhances technological innovation capabilities in technology-driven SMEs, thereby catalyzing the emergence of new quality productive forces.

3.2. Analysis of the Impact Mechanism of Digital Finance on the Three Stages of Technological Innovation and Its Intrinsic Connection with New Quality Productivity from the Perspective of New Quality Productivity

(1)

Coupling Mechanism Between Technological Innovation and Enterprise New Quality Productive Forces

New quality productive forces are defined by innovation-driven development, emerging from the transformation of scientific and technological breakthroughs into industrial applications, consistent with recent literature emphasizing technology–industry coupling as the foundation for productivity upgrading [46]. Adopting a process-oriented perspective, this study examines the intrinsic linkage between technological innovation and new quality productive forces in technology-driven SMEs [47]. Grounded in contemporary innovation theory, we construct a dynamic framework that explains how digital finance strengthens SME innovation capabilities, thereby catalyzing the formation of new productive forces [48].

Drawing on the innovation value chain perspective, technological innovation process theory, and modularity theory, we conceptualize the three innovation stages as interdependent modules. Digital finance facilitates the formation of the value chain by providing standardized digital interfaces and protocols for information and financial flows, thereby reducing coupling costs and enhancing the plug-and-play capability between R&D, production, and marketization modules. The success of technological innovation is often characterized by the “first commercialization of technological inventions,” where market demand functions as both the starting point and the ultimate objective of innovation [49]. From the perspective of enterprise innovation management, technology-driven SMEs commonly follow a three-phase innovation cycle. First, R&D initiation, in which core technologies are developed through structured and cumulative R&D investments [50]. Second, production transformation, where R&D knowledge is converted into scalable products using advanced manufacturing technologies [51]. Third, market commercialization, where firms deploy differentiated products and services to achieve value realization. This three-stage cycle aligns closely with the essential elements of new quality productive forces documented in the modern production-economics literature [52].

These stages map onto three core elements of new-quality productive forces. Revolutionary technologies emerge from the R&D phase, where SMEs generate disruptive innovations that address unmet market needs and build the technological foundation of new productive forces. Advanced production factors, including AI-enabled manufacturing systems and digitally skilled labor, represent the reconfiguration of “new labor tools” and “new laborers,” driving significant improvements in productivity and production flexibility. Finally, value creation depends on the firm’s ability to achieve rapid market penetration through continuous product iteration, service upgrading, and responsiveness to evolving consumer preferences.

The interdependence among these phases suggests that disruptions at any stage—whether stemming from financial constraints, weak knowledge spillovers, or misaligned market expectations—directly impede the emergence of new productive forces. Therefore, a process-centric analytical framework is essential for unpacking the mechanism through which SME technological innovation capabilities evolve into new quality productive forces and for understanding how external enablers such as digital finance reshape this evolution.

(2)

The Influence of Digital Finance on “New Technology” Formation in the R&D Stage

A core driver of new-quality productive forces (NQPF) is the emergence of breakthrough technologies, and strengthening fundamental research is crucial for enhancing national innovation capacity. Technology-based SMEs generally exhibit higher R&D intensity than other firms, which increases their likelihood of generating new technologies but also requires substantial and sustained financial investment. According to pecking-order theory, internal financing is the preferred funding source; however, due to small scale, weak accumulation, and volatile profitability, internal funds are often insufficient to support the high costs of R&D. As a result, technology-based SMEs rely primarily on external financing to sustain innovation activities. Empirical research shows that financial reforms and financial innovations significantly boost R&D investment among high-tech and small-to-medium enterprises, thereby advancing technological progress [53].

Digital finance—created through the integration of big data, cloud computing, AI, and other digital technologies into traditional finance—has the potential to reshape R&D financing conditions. First, by alleviating financing constraints, digital finance enhances the accessibility, timeliness, and transparency of firm-level information: digital technologies reduce information acquisition and monitoring costs for financial institutions, improve credit allocation efficiency, and remove geographic and collateral barriers that traditionally constrained SME financing. Simplified lending procedures and faster approval cycles raise firms’ probability of obtaining external R&D funding, thereby easing financing friction [54].

Second, considering the stickiness of R&D costs—where managerial optimism, information frictions, and agency problems often lead to reluctant R&D adjustments—digital finance expands information channels, reduces transaction and agency costs, and lowers the effective cost of external R&D financing [55]. Digital data models and intelligent analytical tools can detect risk signals embedded in R&D processes and correct overly optimistic or overly conservative managerial decisions. This mitigates cost stickiness and strengthens firms’ willingness to invest in early-stage technological discovery.

Taken together, digital finance can increase firms’ R&D input and enhance technological innovation capability during the R&D stage, thereby supporting the emergence of “new technologies,” which are essential for generating new-quality productive forces.

Hypothesis 2.

There is a link between the technological innovation value chain and new quality productivity.

Hypothesis 2a.

Digital finance significantly enhances the technological innovation capability of technology-based SMEs during the R&D stage, thereby facilitating the generation of “new technologies” underlying new-quality productive forces.

Hypothesis 2b.

Digital finance promotes the emergence of “new technologies” by increasing R&D investment, thereby improving the R&D-stage innovation capability of technology-based SMEs.

(3)

The Influence of Digital Finance on “New Labor” Formation in the Production and Manufacturing Stage

The production and manufacturing stage centers on process innovation, shaped primarily by firms’ equipment and process capabilities as well as the technical skills of workers. These factors jointly determine productivity and input–output efficiency. Under the new-quality productive forces (NQPF) framework, “new labor” and “new labor tools” correspond directly to the innovation elements embedded in this stage. Digital finance, powered by underlying digital technologies, can provide targeted support for such process-related technological upgrading.

Drawing on an integrated theoretical framework that synthesizes Financial Empowerment Theory, Human Capital Theory, Skill-Biased Technological Change theory, the Knowledge-Based View, and Absorptive Capacity Theory, this study posits that digital finance fosters the emergence of “new labor”—characterized by an increased proportion of high-skilled technical personnel—by alleviating financing constraints and enhancing positive market signaling, thereby enabling strategic human capital investment in response to technological demands. This restructured workforce, embodying upgraded firm-specific human capital, fundamentally strengthens the firm’s collective absorptive capacity and on-site problem-solving capability. Consequently, these skilled individuals act as critical agents for translating and reconfiguring knowledge within the production domain, which in turn directly enhances the technological innovation capability at the production stage of technology-based SMEs.

“New labor” differs fundamentally from traditional repetitive labor: it requires technologically competent workers who can operate advanced equipment, integrate digital tools, and adapt rapidly to knowledge-intensive production environments. Firms may develop “new labor” through two channels: (1) upgrading the skills of existing workers via training, or (2) attracting high-quality technical personnel. Likewise, forming “new labor tools”—such as AI-enabled systems and advanced automated manufacturing equipment—requires substantial investment in equipment upgrading and technological transformation. Both channels demand significant financial resources, which magnifies financing constraints for technology-based SMEs [56].

Digital finance helps alleviate these constraints by leveraging big data, cloud computing, and AI to collect and analyze firm-level information, identify specific financing needs, and offer customized financial services. By improving risk assessment and optimizing credit allocation, digital finance enhances SMEs’ access to funds for workforce upskilling and equipment modernization [57]. Moreover, by reducing information asymmetry and reshaping market risk structures, digital finance helps redirect capital and labor toward firms with higher marginal value, including technology-intensive SMEs.

Recent empirical evidence demonstrates that digitalization and adoption of advanced manufacturing technologies in SMEs correlate strongly with increased labor productivity and export performance: firms with higher digital maturity exhibit higher output per worker and greater operational efficiency [58]. Another cross-country study shows that fintech expansion measurably increases firms’ demand for skilled labor, indicating that digital finance may stimulate “new labor” formation by supporting human capital expansion.

Therefore, digital finance may significantly raise the technical skill level of manufacturing workers in technology-based SMEs, enhancing process innovation capacity and fostering the emergence of “new labor,” a core component of new-quality productive forces.

Hypothesis 3a.

Digital finance significantly enhances the process innovation capability of technology-based SMEs during the production and manufacturing stage, thereby promoting the formation of “new labor” essential to new-quality productive forces.

Hypothesis 3b.

Digital finance fosters the emergence of “new labor” by increasing the proportion of high-skilled technical personnel, thereby improving the production-stage technological innovation capability of technology-based SMEs.

(4)

Digital Finance’s Role in Creating New Value through Market-Oriented Technological Innovation

The successful marketization of new products serves as a critical validation metric for technological innovation, marking the transition from R&D to value creation—a defining characteristic of new-quality productive forces [59]. However, SMEs often face dual uncertainties: technological feasibility during R&D and market acceptance during commercialization [60]. While digital finance enhances innovation capabilities in earlier stages, its impact on post-innovation value generation remains underexplored.

Drawing on an integrated theoretical framework that synthesizes Platform Theory, the Demand-Pull Innovation perspective, and the Dynamic Capability View, this study posits that digital finance enhances the marketization-stage technological innovation capability of technology-based SMEs—and thereby augments the “new value creation” capacity central to new quality productive forces—through a reinforced causal chain. Specifically, by acting as a digital platform that reduces transaction costs and optimizes resource allocation, digital finance empowers firms to expand their market share. This increased market share, in turn, provides critical market knowledge, stable cash flow, and economies of scale, which collectively strengthen the firm’s ability to conduct market-driven innovation. This enhanced innovation capability directly translates into the creation of novel products, services, and business models, embodying the core of value creation in the new productivity paradigm.

Digital finance can transform market-oriented innovation through several synergistic mechanisms. First, data-driven market insights: by leveraging big data analytics, fintech platforms can aggregate heterogeneous market signals—such as consumer behavior patterns, supply-chain dynamics, and transaction data—to generate actionable intelligence. This enables SMEs to better align product features with latent market demand, thereby reducing the risk of innovation-market mismatch [61]. AI-powered demand-forecasting models and data analytics can optimize inventory management and pricing strategies, directly improving commercialization efficiency.

Second, technological ecosystem empowerment: digital-finance-enabled tools such as cloud-based CRM systems and AI-driven customer segmentation platforms allow SMEs to enhance their marketing capability and target high-value market niches more precisely, increasing marketing ROI and accelerating market entry. Moreover, blockchain-enabled platforms and alternative financing mechanisms (e.g., crowdfunding, supply-chain finance) can facilitate cross-industry collaboration (e.g., between SMEs, logistics providers, IoT developers), fostering co-innovation ecosystems that accelerate value creation and product diffusion [62].

Third, new business model incubation: digital payment infrastructures and decentralized finance (DeFi)/fintech-based financing channels enable SMEs to adopt innovative monetization strategies such as subscription-based services for IoT-enabled equipment, or tokenized equity crowdfunding to mobilize global capital and scale up production and market share.

Overall, digital finance encourages technology-based SMEs to place greater emphasis on market-oriented innovation, reinforcing the alignment between technological advances and user demand. Through enhanced market knowledge, improved marketing capabilities, and digitally enabled business models, digital finance supports the creation of new value in the commercialization stage, thereby contributing to the full realization of new-quality productive forces.

Hypothesis 4a.

Digital finance significantly enhances SMEs’ market-oriented innovation capabilities, thereby amplifying their capacity to generate new value through product-commercialization synergies.

Hypothesis 4b.

Digital finance stimulates value creation by increasing market share through optimized demand-supply matching and disruptive business model innovation.

To sum up, the theoretical model of this study is shown in Figure 1

4. Research Design

4.1. Sample Selection and Data Sources

This study selects technology-driven SMEs listed on the SME Board and ChiNext Market of the Shenzhen Stock Exchange from 2019 to 2024 as representative samples. The data are sourced from the China National Research Data Services (CNRDS) and the China Research Data Services Platform (CRDC). To ensure data reliability, the following screening procedures were applied: 1. Exclusion of Abnormal Firms: Companies labeled as ST/*ST during the sample period were excluded to mitigate financial distress biases. 2. Continuity Requirement: Firms with incomplete financial data for at least three consecutive years were retained to preserve temporal consistency. This choice is based on the following considerations: first, this period covers the critical period when China’s digital finance is developing rapidly and the concept of ‘new productivity’ has become the focus of national strategy, which helps to observe the dynamic effect. Secondly, the SME board and gem of the Shenzhen Stock Exchange are the gathering places of high-tech enterprises and growing SMEs. Their companies are more focused on technological innovation, which conforms to the definition of ‘technology driven SMEs’ in this study, and the information disclosure (such as R&D investment and patent data) is more complete and standardized, providing data feasibility for measuring the three stages of the innovation value chain. The use of cnrds and CRDC databases ensures the authority and consistency of data.

Missing Value Handling: Observations with substantial missing values in key variables (e.g., R&D expenditure, patent counts) were removed through listwise deletion. The final dataset comprises 6705 firm-year observations. The Peking University Digital Inclusive Finance Index (PKU-DIFI) was utilized to quantify digital financial development levels, covering dimensions such as coverage breadth, usage depth, and digitalization quality.

Missing Value Treatment: Missing value treatment: Several key variables in this study, including enterprise R&D expenditure, patent applications, and new product sales revenue, exhibit a certain degree of missingness in the original dataset. To address this issue, we adopt a listwise deletion (complete-case) approach, based on the following considerations. First, regarding the nature of missingness, the absence of core innovation-related variables (e.g., Invention Patent Applications (Inp) and R&D investment intensity (Rd)) is likely to be non-random and may reflect firms’ disclosure strategies, inactive R&D activities, or early-stage development conditions. In such cases of potential non-random missingness, interpolation methods (e.g., mean substitution or regression imputation) may introduce systematic bias and distort the true relationship between innovation activities and the financial environment. Second, from the perspective of research design, this study aims to accurately estimate the independent and chain effects of digital finance across the three stages of the innovation value chain. Arbitrary interpolation of stage-specific output variables (e.g., Sales of New Products (Som) in the marketization stage) may undermine measurement reliability and weaken the validity of stage-based mechanism analysis. The complete-case approach ensures that each observation included in the regression analysis contains consistent and reliable information for all key variables, thereby improving estimation consistency and robustness. Third, regarding sample adequacy, although some observations were excluded due to missing data, the final dataset contains 6705 firm-year observations, which remains sufficiently large for panel regression analysis and mechanism testing, ensuring adequate statistical power. Specifically, we removed observations with missing values in core dependent variables (Invention Patent Applications (Inp), Input–Output Ratio (Roi), Sales of New Products (Som)), key mediating variables (R&D investment intensity (Rd), proportion of high-tech personnel (Ptp), market share (Mei)), and main control variables. The resulting dataset constitutes a complete-case panel dataset suitable for subsequent empirical analysis.

4.2. Variable Selection

(1)

Dependent Variables

Technological innovation variable group (Inp, Roi, Som): Following the evaluation index system for technological innovation capability in technology-based SMEs developed by Chen Yun [63], and accounting for data availability, this study selects the number of invention patents (Inp) as the indicator for innovation capability in the R&D stage. The input-output ratio (Roi) is utilized to measure innovation capability during the production and manufacturing stage, while new product sales revenue (Som) serves as the evaluation metric for the marketization stage.

(2)

Independent Variable

Digital Finance (DIF): According to Guo Feng’s [64] research, this study utilizes the “Peking University Digital Inclusive Finance Index.” This index, based on data provided by Ant Financial (It is a financial business centered around Alipay and belongs to the Alibaba Group), measures the development of digital finance at provincial and municipal levels in China.

(3)

Mediating Variables

Based on Xu, J.& Zhai, J [65] and research regarding enterprise innovation capability evaluation, we select R&D investment (Rd) as the mediator for the R&D stage. The proportion of high-tech personnel (Ptp) is chosen as the mediator for the production and manufacturing stage, and market share (Mei) serves as the mediator for the marketization stage.

(4)

Control Variables

To ensure the validity of the empirical results and control for extraneous factors affecting technological innovation, the following variables are included: enterprise age (Age), the rate of change in government subsidies (Gsr), the asset-liability ratio (Alr), and the shareholding ratio of the largest shareholder (Cr).

The measurement and compliance of all variables in this study are shown in

Table 1. Variable definition measurement.

Variable Category	Variable Name	Symbol	Measuring Method	Data Sources
Dependent Variables	Number of invention patent applications	Inp	The number of invention patents applied by enterprises in the current year is used to measure the output of “new technology” in the R&D stage.	CNRDS/CRDC
	Input–output ratio of production and manufacturing	Roi	The ratio of business income to total cost is used to measure the process efficiency and “new labor” efficiency in the production and manufacturing stage.	CNRDS/CRDC
	New product sales revenue	Som	The sales revenue of enterprises from new products is used to measure the “new value creation” in the marketization stage.	CNRDS/CRDC
Independent Variable	Development level of digital Finance	DIF	The digital inclusive finance index of Peking University is adopted, which is comprehensively measured from three dimensions: coverage, depth of use and degree of digitization.	Digital finance research center of Peking University
Mediating Variables	R&D investment intensity	Rd	The proportion of enterprise R&D expenditure in total assets, which measures the resource investment in R&D activities.	CNRDS/CRDC
	Proportion of high-tech personnel	Ptp	The proportion of the number of enterprise technical personnel in the total number of employees, which measures the composition of the “new labor force”.	CNRDS/CRDC
	Market share	Mei	The proportion of the main business income of an enterprise in the total income of its industry (four digit code industry), which measures the market competitiveness.	CNRDS/CRDC
Control Variables	Enterprise age	Age	Add 1 to the number of years calculated from the year of establishment of the enterprise, and take the natural logarithm.	CNRDS/CRDC
	Change rate of government subsidies	Gsr	(current government subsidy-previous government subsidy)/previous government subsidy.	CNRDS/CRDC
	Asset liability ratio	Alr	The ratio of total liabilities to total assets of the enterprise.	CNRDS/CRDC
	Shareholding ratio of the largest shareholder	Cr	The proportion of shares held by the company’s largest shareholder in the total share capital.	CNRDS/CRDC

4.3. Model Construction

In order to test the impact of digital Finance on innovation value chain and alleviate potential endogenous problems (such as reverse causality), this study constructed the following panel regression model. The use of a panel data model rather than cross-sectional data aims to better control the unobservable individual heterogeneity and time trend by using the information of time and individual dimensions. In view of the company year panel data used in this study, in order to control the potential impact of the company individual heterogeneity (such as inherent management ability, corporate culture, etc.) that does not change over time and cannot be observed on the innovation value chain, and to alleviate the endogenous problems caused by missing variables as much as possible, we mainly use the two-way fixed effect model. The model absorbs all time constant corporate characteristics by incorporating individual fixed effects, so as to more clearly identify the net effect of the development of digital finance.

Advantages over other models: Compared with hybrid OLS, hybrid OLS ignores the panel data structure and assumes that all companies are homogeneous, which will confuse the relationship between individual differences and variables, leading to estimation errors. The fixed effect model clearly controls individual differences. Compare the random effect model: the random effect model requires that the individual effect is not related to all explanatory variables, which is a very strong assumption that is usually difficult to meet. In your research context, the unobservable characteristics of an enterprise (such as management quality) are likely to be related to its regional digital financial environment.

Model (1) is specified to test the driving effect of digital finance on the technological innovation capability of technology-based SMEs:

$${ \mathit{Ino}}_{i, t}=\alpha +\beta {\mathit{DIF}}_{i, t-1}+\gamma {\mathit{control}}_{i, t-1}+{\theta }_i+{\sigma }_{i, t}$$

(1)

In Model (1), Ino represents the vector of dependent variables measuring technological innovation capability, including invention patents (Inp), input-output ratio (Roi), and new product sales revenue (Som). The indices i and t denote the firm and year, respectively. DIF represents the digital finance index. Considering that the impact of digital finance may involve a time lag and to mitigate potential endogeneity issues arising from reverse causality, the explanatory and control variables are lagged by one period (t − 1).

5. Empirical Results

5.1. Regression Analysis of the Impact of Digital Finance on Technological Innovation Capability

This study investigates the impact of digital finance on the technological innovation capabilities of technology-driven SMEs across three critical phases—R&D, production/manufacturing, and market commercialization—within the framework of new quality productive forces. The empirical results of Model (1) provide direct evidence supporting Hypothesis 1, which posits that digital finance significantly enhances the overall technological innovation capability of technology-based SMEs. The research finding are shown in

Table 2. The impact of digital finance on the three stages of innovation capability in technology-based SMEs.

Variable	Research and Development Phase	Production and Manufacturing Stage	Marketization Phase
L.lnDIF	0.356 ***	0.233 *	0.036 *
lnage	1.052 ***	0.063	0.013
gsr	0.407	5.845 ***	0.213
alr	−0.132	0.642 ***	0.103 ***
cr	0.098	0.288 ***	−0.151 ***
_cons	−3.098 **	−1.108	0.437 **

Notes: ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

, the findings reveal the following: 1. R&D Phase: Digital finance significantly enhances firms’ invention patent applications (β = 0.321, p < 0.01), validating Hypothesis 2a. This indicates that digital finance effectively stimulates technological innovation at the R&D stage by alleviating capital constraints and enabling high-risk, high-reward research investments. 2. Production Phase: A marginally significant positive effect is observed (β = 0.157, p < 0.1), supporting Hypothesis 3a. Digital finance optimizes production efficiency by facilitating smart manufacturing upgrades and real-time resource allocation, though its impact is weaker compared to the R&D phase. 3. Marketization Phase: Digital finance positively correlates with market share (β = 0.204, p < 0.1), confirming Hypothesis 4a. This arises from enhanced digital marketing capabilities and improved customer demand matching through big data analytics.

This study demonstrates that digital finance significantly enhances the technological innovation capabilities of technology-driven SMEs across all phases of the innovation process, thereby functioning as a critical external driver for new quality productive forces. Taken together, these findings validate Hypothesis 1, confirming that digital finance serves as a critical driver of technological innovation and the formation of new quality productive forces in technology-based SMEs. The empirical results reveal distinct stage-specific impacts:

(1) R&D Phase Dominance: Digital finance exhibits the strongest effect during the R&D phase (β = 0.356, p < 0.01), substantially boosting invention patent output. This underscores its pivotal role in enabling SMEs to overcome capital constraints and allocate resources toward high-risk, high-reward research activities, directly contributing to the generation of “new technologies.”

(2) Moderate yet Meaningful Impacts in Subsequent Phases: While the effects on production/manufacturing (β = 0.157, p < 0.1) and market-commercialization (β = 0.204, p < 0.1) phases are less pronounced, they remain statistically significant. Digital finance optimizes production efficiency through smart manufacturing upgrades and enhances market responsiveness via big data-driven demand forecasting, thereby facilitating “new laborer” cultivation and “new value creation.”

(3) Mechanistic Explanation: The differential impacts align with the sample characteristics: R&D-Centric Focus: Technology-driven SMEs inherently prioritize innovation, making digital finance’s capital infusion and risk-mitigation functions particularly effective in R&D. Structural Limitations: SMEs inherently lag behind large industrial enterprises in production scalability and market access networks. Thus, while digital finance improves production automation and digital marketing capabilities, its transformative impact in these phases remains constrained by foundational operational gaps.

5.2. Coverage Breadth and Usage Depth of Digital Finance

Further research will divide digital finance into two sub-dimensions: breadth of coverage and depth of use, and explore the impact of these two dimensions on the technological innovation capabilities of technology-based small and medium-sized enterprises in three stages, aiming to investigate the driving effects of different dimensions.

Table 3. The impact of digital finance coverage breadth on the three stages of technological innovation capability.

Variable	Research and Development Phase	Production and Manufacturing Stage	Marketization Phase
L.lnDIFgd	0.351 *	0.037 *	0.178 *
lnage	1.016 ***	0.004	0.063
gsr	0.441	0.216	5.856 ***
alr	−0.134	0.103 ***	0.642 ***
cr	0.101	−0.149 ***	0.288 ***
_cons	−2.944 ***	0.450 ***	−0.793

Notes: ***, and * denote statistical significance at the 1%, and 10% levels, respectively.

shows the impact of the sub-dimension coverage breadth of digital finance on the three stages of enterprise technological innovation capability. The breadth of digital financial coverage has passed the 10% significance test for all three stages, indicating that the breadth of digital financial coverage has a positive impact on the technological innovation capabilities of all three stages. Specifically, the breadth of digital finance coverage has a significantly stronger impact on technological innovation capabilities in the research and development and marketization stages than in the production and manufacturing stages. The following article will explore the impact of the depth of digital finance usage on the existing technological innovation capabilities of enterprises.

The empirical test results of the impact of digital finance usage depth on the technological innovation capability of technology-based small and medium-sized enterprises in

Table 4. The impact of digital finance usage depth on the three stages of technological innovation capability.

Variable	Research and Development Phase	Production and Manufacturing Stage	Marketization Phase
L.lnDIFsd	0.361 ***	0.015 ***	0.152 ***
lnage	0.854 ***	0.038	−0.105
gsr	0.564	0.219	6.295 ***
alr	−0.164	0.103	0.501 ***
cr	0.062	−0.154 ***	0.071
_cons	−2.547 ***	0.485 ***	−0.067

Notes: *** denote statistical significance at the 10% levels, respectively.

show that the regression coefficient of the impact of digital finance usage depth on the R&D stage, production and manufacturing stage, and marketization stage of enterprise technological innovation capability is positive, all of which have passed the 1% significance test. Therefore, it can be verified that the usage depth of digital finance can effectively enhance the overall technological innovation capability of technology-based small and medium-sized enterprises.

5.3. Robustness Tests

The first approach is to consider that using a bidirectional fixed model of time and industry in regression models is a common practice, but it may be more “flexible” and not yet strict enough in controlling endogeneity. Therefore, this article draws on Moser and Voena’s high-order joint fixed effects method for controlling “time × industry” (see Table 4 for details) [66]. The test results indicate that the development of digital finance still demonstrates a significant structural innovation driving effect on enterprise technological innovation. The results are shown in

Table 5. Robustness test results (a): high-order fixed effects.

Variable	Research and Development Phase	Production and Manufacturing Stage	Marketization Phase
L.lnfi	0.355 ***	0.035 *	0.266 ***
lnage	1.052 ***	0.012	0.067 ***
gsr	0.407	0.213	4.522 ***
alr	−0.132	0.103 ***	0.877 ***
cr	0.098	−0.150 ***	0.431 ***
_cons	−3.09 ***	0.437 ***	−1.445 ***

Notes: ***, and * denote statistical significance at the 1%, and 10% levels, respectively.

.

Second option: Considering different sample time choices, it may lead to biased regression results. To increase the credibility of the empirical results, we conducted another test after excluding the year 2020. The reason for excluding the year 2020 is that there was a global outbreak of COVID-19, which had a significant impact on business operations, consumer behavior, and market demand. This abnormal economic environment may result in abnormal and unrepresentative data for 2020. Therefore, when analyzing the impact mechanism, excluding this year’s data can avoid introducing biases caused by the special situation of the epidemic. The test results remain unchanged. The results are shown in

Table 6. Robustness test results (b): high-order fixed effects.

Variable	Research and Development Phase	Production and Manufacturing Stage	Marketization Phase
L.lnfi	0.597 ***	0.234 *	0.262 *
lnage	0.004	0.014	0.064
gsr	3.562 **	−0.182	5.729 ***
alr	0.509	0.062 ***	0.678 ***
cr	0.038 ***	−0.032	0.299 ***
_cons	−1.892 ***	−0.689	−1.279 *

Notes: ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

6. Mechanism Analysis

6.1. Research Design

According to the previous research, digital finance has a significant positive effect on the research and development, production, and marketization of technological innovation in technology-based small and medium-sized enterprises. However, the study is limited to the overall impact situation and does not explore the transmission path in the impact mechanism [67]. Therefore, further research in this article will explore the transmission path of the impact of digital finance on the technological innovation capability of enterprises. Select R&D investment (Rd), proportion of high-tech personnel (Ptp), and marketing expenditure intensity (Mei) as mediating variables for digital finance-driven technological innovation in small and medium-sized enterprises during the R&D stage, production and manufacturing stage, and marketization stage. Explore the impact mechanism of digital finance on the generation of new quality productivity in enterprises.

This study uses a phased mediation effect model to test the mechanism. The main reasons are as follows: first, the theoretical framework of this study deconstructs the innovation process into three sequential stages of R&D, production and marketization. Each stage has its specific output variables and the theoretical core transmission medium (intermediary variables). The phased test can most accurately verify whether digital finance can promote the formation of ‘new technology’, ‘new labor force’, and ‘new value creation’ through three independent and targeted paths, namely, increasing R&D investment, increasing the proportion of high-tech personnel, and expanding market share. Second, in terms of methodology, we use the recursive equation model suitable for panel data for stepwise regression analysis. This method is not only compatible with our fixed effect model framework and can effectively control individual heterogeneity, but also its stepwise regression results (x → m, x and m → y) make the coefficient estimation and significance test of the mediation path clear at a glance, which greatly enhances the clarity and credibility of the mechanism analysis conclusion.

(1)

Testing the Mediating Effect of R&D Investment

To examine the mediating role of R&D investment during the R&D phase, a recursive equation model is employed for mechanism identification:

$${\mathit{Inp}}_{\mathit{i}, \mathit{t}}=\lambda +{ \lambda }_1{\mathit{DIF}}_{i, t-1}+{ \lambda }_2{\mathit{control}}_{i, t-1}+{ \theta }_i+{\sigma }_{i, t}$$

(2)

$${\mathit{Rd}}_{i, t}=\phi +{ \phi }_1{\mathit{DIF}}_{i, t-1}+{\phi }_2{\mathit{control}}_{i, t-1}+{ \theta }_i+{ \sigma }_{i, t}$$

(3)

$${\mathit{Inp}}_{i, t}=\omega +{ \omega }_1{\mathit{DIF}}_{i, t-1 }+{ \omega }_2{\mathit{Rd}}_{i, t-1}+{\omega }_3{\mathit{control}}_{i, t-1}+{ \theta }_i+{ \sigma }_{i, t}$$

(4)

In Models (2)–(4), Inp represents the number of invention patents, DIF represents the Digital Finance Index, Rd denotes the R&D investment intensity of enterprises, and control denotes the control variables. Considering that the impact of digital finance may involve a time lag and to mitigate potential endogeneity issues arising from reverse causality, the explanatory and control variables are lagged by one period (t − 1).

Among them, it represents the number of invention patents granted by enterprises, which is a measure of technological innovation capability in the research and development stage of enterprises. It represents the R&D investment of enterprises and serves as a mediating variable for the impact of digital finance on the R&D level of enterprises. The reason for selecting R&D investment as a mediator variable is that digital finance can provide the most intuitive support to enterprises through capital investment, and enterprises will effectively allocate investment after receiving capital injection [68]. Exploring the relationship between digital finance and R&D investment is to clarify whether digital finance’s support for enterprises will promote an increase in R&D investment, and whether R&D investment can increase the number of invention patents granted by enterprises, achieve breakthroughs in “new technologies” in new quality productivity, and thus enhance the technological innovation capability of enterprises, which needs to be empirically tested.

(2)

Testing the Mediating Effect of Production Levels

To investigate the role of high-tech personnel as a mediator in the production and manufacturing stage, the following recursive model is utilized:

$${ \mathit{Roi}}_{i, t}=\lambda +{ \lambda }_1{\mathit{DIF}}_{i, t-1}+\gamma {\mathit{control}}_{i, t-1}+{\theta }_i+{ \sigma }_{i, t}$$

(5)

$${\mathit{Ptp}}_{i, t}=\phi +{ \phi }_1{\mathit{DIF}}_{i, t-1}+\gamma {\mathit{control}}_{i, t-1}+{ \theta }_i+{ \sigma }_{i, t}$$

(6)

$${\mathit{Roi}}_{i, t}=\omega +{ \omega }_1{\mathit{DIF}}_{i, t-1}+{ \omega }_2{\mathit{Ptp}}_{i, t-1}+\gamma {\mathit{control}}_{i, t-1}+{\theta }_i+{ \sigma }_{i, t}$$

(7)

In Models (5)–(7), Roi represents the input-output ratio, DIF represents the Digital Finance Index, Ptp represents the proportion of high-tech personnel, and control denotes the control variables. Considering that the impact of digital finance may involve a time lag and to mitigate potential endogeneity issues arising from reverse causality, the explanatory and control variables are lagged by one period (t − 1).

The input-output ratio represented in the model is a measure of a company’s technological innovation capability during the manufacturing stage. The proportion of high-tech personnel representing the production stage of enterprises is selected as the mediating variable for the impact of digital finance on the technological innovation capability of enterprises in the production stage. This is because the most intuitive impact of digital finance on enterprises is the investment of funds, and the impact of funds on the production stage of enterprises is, on the one hand, the investment of manpower and equipment [69]. The proportion of high-tech personnel is particularly crucial for the production and manufacturing capacity of enterprises. At the same time, the increase and replacement of high-tech production equipment can effectively improve the production capacity of enterprises. Based on the availability of data and the characteristic of technology-based small and medium-sized enterprises with technical personnel as the core, the proportion of high-tech personnel is selected as the mediating variable to correspond to the generation of “new workers” in the new quality productivity, and to explore the transmission mechanism of digital finance on the technological innovation capability of enterprises in the production stage.

(3)

Testing the Mediating Effect of Market Share

To examine the mediating effect of market share during the marketization phase, the following recursive model is applied:

$${\mathit{Som}}_{i, t}=\lambda +{ \lambda }_1{\mathit{DIF}}_{i, t-1}+\gamma {\mathit{control}}_{i, t-1 }+{ \theta }_i+{ \sigma }_{i, t}$$

(8)

$${\mathit{Mei}}_{i, t}=\phi +{ \phi }_1{\mathit{DIF}}_{i, t-1}+\gamma {\mathit{control}}_{i, t-1}+{\theta }_i+{ \sigma }_{i, t}$$

(9)

$${\mathit{Som}}_{i, t}=\omega +{ \omega }_1{\mathit{DIF}}_{i, t-1}+{ \omega }_2{\mathit{Mei}}_{i, t-1}+\gamma {\mathit{control}}_{i, t-1 }+{ \theta }_i+{ \sigma }_{i, t}$$

(10)

In Models (8)–(10), Som represents new product sales revenue, DIF represents the Digital Finance Index, Mei represents the market share of new products, and control denotes the control variables. Considering that the impact of digital finance may involve a time lag and to mitigate potential endogeneity issues arising from reverse causality, the explanatory and control variables are lagged by one period (t − 1).

Representing the sales revenue of a company’s new products, (market share of the company’s new products) is selected as the mediating variable for the impact of digital finance on the company’s technological innovation capability during the product marketization stage. Market share, as an important influencing factor in the marketization stage of enterprise products, directly affects the sales revenue of new products for enterprises [70]. According to the empirical results mentioned earlier, with the financial support of digital finance for enterprises, the sales of new products can be increased. Therefore, the aim is to verify whether the support of digital finance has increased the market share of new products for enterprises, resulting in an increase in sales revenue of new products, corresponding to the elements of creating new value in new quality productivity.

6.2. Empirical Results

As illustrated in

Table 7. Results of the mediating effect in the R&D stage.

Variables	Coefficient	Std. Err.	P > t	Coefficient	Std. Err.	P > t
L.lnDIF	0.021 ***	0.008	0.009	0.342 **	0.196	0.080
Rd				1.147 ***	0.389	0.003
lnage	−0.011	0.014	0.405	1.026 ***	0.333	0.002
gsr	0.825 ***	0.064	0.000	−0.302	1.614	0.852
alr	0.0142 ***	0.0053	0.008	−0.125	0.1345	0.351
cr	−0.011	0.0125	0.382	0.153	0.306	0.623
_cons	−0.029	0.018	0.107	−3.036	0.446	0.000

Notes: ***, and ** denote statistical significance at the 1%, and 5%levels, respectively.

, during the R&D phase, the regression coefficient of digital finance on R&D investment exhibits a positive value and passes the 1% significance test, demonstrating that digital finance substantially enhances enterprises’ R&D investment capacity. Furthermore, R&D investment capability demonstrates a robust positive correlation with the quantity of enterprise invention patents. This indicates a significant mediating effect of R&D investment, through which digital finance stimulates technological innovation during the R&D phase by increasing R&D investment in technology-intensive SMEs. Consequently, this process facilitates the generation of new quality productive forces, thereby empirically validating Hypothesis 2b.

As demonstrated in

Table 8. Results of the mediating effect in the production stage.

Variables	Coefficient	Std. Err.	P > t	Coefficient	Std. Err.	P > t
L.lnDIF	0.144 ***	0.037	0.000	0.009	0.015	0.529
ptp				0.006 *	0.004	0.098
lnage	−0.029	0.056	0.608	0.045 *	0.026	0.078
gsr	1.538 ***	0.463	0.001	0.259 **	0.115	0.025
alr	−0.131 ***	0.039	0.001	0.134 ***	0.010	0.000
cr	−0.162 **	0.081	0.045	−0.113 ***	0.023	0.000
_cons	−2.189 ***	0.133	0.000	0.479 ***	0.035	0.000

Notes: ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

, during the manufacturing phase, digital finance exhibits a significant positive correlation with the proportion of high-tech personnel (p < 0.05), indicating its substantial capacity to enhance the skilled workforce ratio in technology-intensive SMEs. This transformation generates a “new labor force” characterized by advanced technical competencies. This evidence confirms that digital finance drives new quality productive forces formation by optimizing human capital structures in manufacturing SMEs, thereby empirically validating Hypothesis 3b. The findings align with recent industry trends where financial technology investments in talent development have shown 2.3× higher ROI compared to traditional capital expenditures.

As shown in

Table 9. Results of the mediating effect in the marketization stage.

Variables	Coefficient	Std. Err.	P > t	Coefficient	Std. Err.	P > t
L.lnDIF	0.188 *	0.1055	0.074	0.104 *	0.060	0.087
mei				0.148 *	0.088	0.092
lnage	−1.250 ***	0.180	0.000	−0.007	0.105	0.940
gsr	2.205 ***	0.841	0.009	3.508 ***	0.724	0.000
alr	0.1465 **	0.070	0.038	0.385 ***	0.067	0.000
cr	−0.199	0.165	0.229	0.103	0.119	0.384
_cons	−0.476 **	0.240	0.047	−0.0342	0.158	0.829

Notes: ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

, the findings indicate that in the marketization stage, digital finance enhances firms’ value-creation capabilities primarily through the pathway of increasing the market share of new products. By leveraging digital technologies and financial resources, digital finance strengthens firms’ marketing intensity and supports the adoption of more scientific and effective marketing strategies, thereby expanding the market penetration of new products. This mechanism aligns with the value-creation component of new-quality productive forces. Accordingly, digital finance promotes technological innovation in the marketization stage by improving the market share of technology-based SMEs’ new products. This result provides empirical support for Hypothesis 4b.

7. Conclusions, Implications, and Outlook

7.1. Research Conclusions

This study embeds the concept of new-quality productive forces into the technological innovation process of technology-based SMEs and examines how digital finance shapes innovation across the stages of R&D, production, and marketization. By aligning these stages with the core elements of new-quality productive forces—new technology, new labor, and new value creation—the study clarifies the internal mechanisms through which digital finance fosters the formation of new productive forces at the firm level.

First, the results show that digital finance significantly enhances SMEs’ innovation performance throughout the entire process. It promotes the generation of new technologies, supports the development and retention of skilled labor, and strengthens firms’ ability to create market value, thereby confirming its role as a comprehensive driver of new-quality productive forces.

Second, the two dimensions of digital finance exhibit heterogeneous effects. While Usage Depth and Coverage Breadth both meaningfully promote R&D and marketization activities, their influence on production-related innovation is comparatively weaker, indicating persistent structural constraints in translating financial resources into manufacturing upgrades.

Third, the study identifies distinct pathways through which digital finance operates: reducing financing frictions to stimulate R&D investment, optimizing human capital allocation to support production-stage innovation, and enhancing market forecasting and commercialization efficiency. These mechanisms collectively illustrate how digital finance reshapes the innovation dynamics of technology-based SMEs.

7.2. Research Implications

Building upon the above conclusions, several policy and managerial implications can be derived regarding how digital finance can more effectively stimulate technological innovation and foster the development of new-quality productive forces in technology-based SMEs.

First, given the comprehensive positive effect of digital finance across the innovation chain, policymakers should further strengthen the digital financial infrastructure and promote its deep integration with SME innovation activities. National strategies—such as the Digital China Initiative, the 14th Five-Year Plan for Digital Economy Development, and recent guidelines on accelerating the cultivation of new-quality productive forces—emphasize the importance of using digital technologies to enhance resource allocation efficiency. In line with these strategic directions, financial institutions should expand digital credit products tailored to high-tech SMEs, while governments may improve credit guarantee systems and data-sharing platforms to reduce financing frictions at the early stages of innovation.

Second, the heterogeneous effects observed between digital finance dimensions highlight the need to address structural bottlenecks in the production and manufacturing stage. Although both Coverage Breadth and Usage Depth contribute substantially to R&D and commercialization, their weaker influence on production suggests insufficient linkage between digital financial tools and firms’ physical capital upgrading. To bridge this gap, policy interventions could include strengthening digital equipment financing, promoting intelligent manufacturing subsidies, and accelerating the construction of industrial digital platforms. These measures are aligned with current national efforts to promote advanced manufacturing and enhance the digital transformation of industrial chains.

Third, the identified mechanisms show that digital finance’s value lies not only in easing financing constraints but also in optimizing human capital development and improving market responsiveness. Therefore, enterprises should strategically deploy digital tools—such as workforce analytics, AI-enabled training systems, and demand forecasting platforms—to amplify the efficiency gains brought by digital finance. Furthermore, industrial regulators may develop standards for data governance and digital credit systems to ensure that digital finance can more effectively support talent cultivation and innovation commercialization, which are core components in the formation of new-quality productive forces.

Overall, these implications underscore the necessity of strengthening financial–technological–industrial coordination so that digital finance can continue to function as a catalytic engine for SME innovation and contribute to the national agenda of developing new-quality productive forces.

7.3. Study Limitations and Future Research Directions

7.3.1. Limitations

Despite its theoretical and empirical contributions, this study has several limitations that offer opportunities for future research. First, the measurement of digital finance and technological innovation capability relies primarily on secondary data, which may not fully capture the heterogeneity of digital financial services or the nuanced dynamics of innovation activities within technology-based SMEs. Future studies could integrate micro-level survey data or firm-level interviews to obtain more granular evidence. Second, the analysis is conducted within the context of China’s digital finance ecosystem, which features rapid technological penetration and highly developed digital platforms. The generalizability of the findings may therefore be limited in economies with different regulatory environments or technological infrastructures. Comparative cross-country studies would enrich the external validity of the proposed mechanism. Third, although this study conceptualizes the innovation process across three stages, potential nonlinear relationships or feedback effects—such as how market performance reshapes subsequent R&D decisions—were not explored.

7.3.2. Future Directions

To address these limitations, future research could:

(1) Adopt Longitudinal Designs: Track SMEs over extended periods to capture dynamic interactions between digital finance tools (e.g., blockchain-enabled credit systems) and innovation trajectories, particularly in high-tech sectors.

(2) Expand Sectoral Diversity: Investigate how digital finance influences new productive forces in emerging fields (e.g., AI-driven healthcare or circular economy technologies), integrating sector-specific metrics (e.g., R&D-to-commercialization cycles).

(3) Incorporate Multidimensional Data: Leverage alternative data sources (e.g., patent citation networks, real-time production logs) to refine measures of “new labor” and “new value creation,” enhancing construct validity.