1. Introduction

Amid the confluence of technological revolution and industrial progression, new quality productivity represents an advanced developmental model propelled by scientific and technological advancements. Unlike traditional growth patterns, enterprises—serving as key actors—must coordinate innovative resource allocation, facilitate industry-wide upgrading, and lead technological breakthroughs [1]. Powered by artificial intelligence and cloud computing, digital transformation reshapes organizational structures, operational processes, and innovation mechanisms through the capitalization of data, thereby modernizing business models [2]. As a result, it becomes a crucial enabler for the evolution of new quality productivity.

Current research encompasses both theoretical and empirical aspects. Theoretically, multi-perspective frameworks outline the routes through which digital transformation influences new quality productivity. For example, resource-based viewpoints [3] reveal that executives' digital proficiency positively moderates this association; internal control theory suggests that digitization improves system automation, creating settings conducive to productivity [4]; other studies advance a "drivers-elements-structure" framework to elucidate developmental processes [5]. Empirically, studies on Chinese listed companies pinpoint avenues—such as cost savings, efficiency improvements, risk management [6], innovation amplification, and human capital refinement [7]—via which digital transformation raises total factor productivity.

Nonetheless, earlier studies have placed excessive emphasis on macro-efficiency while paying insufficient attention to micro-level mechanisms. This research adds value in three respects: Firstly, it goes beyond qualitative approaches by measuring new quality productivity using entropy based on dual-factor theory. Secondly, through a multi-angle examination, it explores how digitalization boosts new quality productivity with the mediating effects of green innovation and supply chain effectiveness. Thirdly, it delves into variations across financing constraints, ownership forms, and R&D intensity.

2. Theoretical analysis and research hypotheses

2.1. The direct impact mechanism of enterprise digital transformation on new quality productivity

Digital transformation strengthens new quality productivity through three primary channels. In terms of production instruments, Marx highlighted in Capital that tools objectively reflect the developmental phase of productive forces. Embedding AI and IoT technologies into industrial apparatus converts conventional equipment into intelligent digital twins with self-governing decision-making abilities [8], greatly improving operational accuracy and system reliability. The essential “algorithms+data+computing power” structure allows comprehensive systemic enhancements of traditional productivity setups. Regarding labor objects, companies build unified data platforms to integrate cross-functional information covering R&D, production, and supply chain activities. This shift turns physical assets into interactive digital representations that correct inefficiencies in resource distribution, moving production methods toward data-focused models that represent radical improvements in factor allocation. Lastly, concerning labor skills, as digital tools increasingly handle routine operations, the need for sophisticated digital expertise rises, considerably raising the quality of human capital [9] and strengthening R&D innovation potential. This progression in labor supplies ongoing impetus for the growth of new quality productivity.

Accordingly, this paper proposes Hypothesis 1: Enterprise digital transformation promotes the development of new-quality productive forces.

2.2. The mediating role of green technological innovation

Green technology innovation involves strategic R&D efforts aimed at reducing environmental impacts while maximizing ecological gains through technology use, accomplishing synergistic economic and environmental results. Digital transformation directly meets the substantial data needs of green innovation projects while raising their technical level [10]. Sustained productivity growth fundamentally depends on advances in green technology, establishing green innovation as the key conduit through which digital transformation propels new quality productivity—a defining trait being the thorough merging of technological advance with environmental sustainability. From a resource-based theory standpoint, data resources gathered during digital transformation form unique strategic assets for green innovation, marked by strong non-replicable technical obstacles. These exclusive resources allow firms to improve energy-saving technologies and circular economy methods. When applied, green innovation updates technological foundations while also boosting market reception for eco-friendly goods, thus speeding up sustainable productivity expansion. From an organizational information processing perspective, digital transformation greatly extends information-handling capabilities [11]. Businesses use digital tracking systems to monitor real-time changes in market desire for sustainable products, encouraging ongoing commitment to ecological innovation. Moreover, blockchain-supported information exchange along value chains enables entry to advanced green technologies [12], significantly lowering resource waste and innovation expenses due to information imbalances.

Accordingly, this paper proposes Hypothesis 2: Enterprise digital transformation promotes the development of new-quality productive forces through green technological innovation.

2.3. The mediating role of supply chain efficiency

Supply chain efficiency measures a firm's ability to optimize resource use, lower stock levels, shorten procurement periods, and reduce operational unpredictability [13]. Digital transformation boosts efficiency via linked technologies like IoT sensors and distributed ledger systems, allowing real-time data harmonization across supply network points while curbing bullwhip effect distortions. This technological merger improves alignment between demand and supply, cuts logistics expenses, and boosts system reactivity. Digital approaches methodically break down information barriers typical of conventional linear supply chains [14], supplying accurate operational insights for node enterprises to swiftly adjust to changing market situations—thus improving efficiency in sourcing, manufacturing, logistics, and distribution operations. These efficiency gains then stimulate new quality productivity development. Rooted in endogenous growth theory, refined supply chain management lowers transaction costs and raises total factor productivity measures. Adaptable digital supply networks reduce innovation trial costs while optimizing innovation environments. Excess capacity freed through efficiency improvements can be tactically redirected toward high-value tasks, together propelling advanced productivity development.

On this basis, this paper proposes Hypothesis 3: Corporate digital transformation advances the development of new-quality productivity by enhancing supply chain efficiency.

3. Model construction

3.1. Sample selection and data sources

The data used in this study come from the Guotai An (CSMAR) Listed Company Database for the period 2015–2022. The data were processed as follows: (1) Excluding samples of firms with abnormal statuses like ST and *ST, and removing companies in the financial sector; (2) Removing samples with serious variable omissions; (3) Adopting a winsorization approach to trim extreme values, ensuring all variables lie within the 1% to 99% range to prevent outlier-induced estimation errors. After processing, 15,462 observations remained.

3.2. Variable setup

3.2.1. Dependent variable

The dependent variable is corporate new-quality productivity (NP). Building on prior scholarship [15] and integrating the two-factor theory of productivity, an indicator system for new-quality productivity was developed across three aspects: labor force, labor objects, and labor instruments. The entropy method was applied to compute firm-level new-quality productivity indicators.

3.2.2. Core explanatory variable

The central independent variable is enterprise digital transformation (DIG). In line with earlier work [16], annual reports of listed companies served as the main text source. Analysis concentrated on five technological fields: artificial intelligence, big data analytics, cloud computing, blockchain technology, and integrated uses of digital technology. Python was utilized to perform text processing, which involved: first, retrieving feature data from the annual reports; second, detecting and counting keyword frequencies across the five categories. After frequency tallying, raw data were normalized. A natural log transformation was then used to create a composite digital transformation index, mainly to counter data skewness and thus improve model stability and result interpretability.

3.2.3. Control variables

To ensure precision and rigor, this paper draws on the approach of [17], selecting firm size (Size), leverage ratio (Lev), board size (Board), firm age (ListAge), and CEO duality (Dual) as control variables.

3.2.4. Mediating variables

Green technology innovation (Green) and supply chain efficiency (Stock_day) are chosen as mediators. In accordance with [18], Green is measured as ln(number of green patent applications + 1). Following [19], supply chain efficiency is gauged by inventory turnover days, computed as ln(365 / inventory turnover ratio).

3.3. Model design

To test Research Hypothesis 1, this paper constructs the following benchmark regression model:

$N{P}_{i,t}={\alpha }_0+{\alpha }_1{\mathrm{D}\mathrm{I}\mathrm{G}}_{i,t}+\sum {\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{i,t}+\sum EndDate+\sum ind+{\epsilon }_{i,t}$(1)

 Here, NP denotes the level of new-quality productivity; DIG represents digital transformation intensity; Control is a set of control variables; EndDate and ind stand for year and industry fixed effects; ε is the error term; i and t indicate firm and year.

To test research hypotheses 2 and 3, we refer to Chinese scholars' [20]research on mediating effects and construct the following mechanism model:

$M{V}_{i,t}={\beta }_0+{\beta }_1{DIG}_{i,t}+\sum {Control}_{i,t}+\sum EndDate+\sum ind+{\epsilon }_{i,t}$(2)

MV represents the mediating variable. In regression, Green and Stock_day are used as dependent variables to examine the links between digital transformation and green innovation, and digital transformation and supply chain efficiency, respectively.

4. Empirical analysis

4.1. Benchmark regression

The table below displays the baseline regression outcomes. The results demonstrate that the coefficient of DIG is positive and significant at the 1% level, implying that corporate digital transformation notably facilitates the improvement of new-quality productivity. Column (1) shows results absent control variables, where the estimated effect of digital transformation on new-quality productivity is 0.00194, significant at 1%; Column (2) includes control variables, and the DIG coefficient remains positively significant at the 1% level. Thus, Hypothesis 1 receives initial support.

Standard errors in parentheses^* p < 0.1, ^** p < 0.05, ^*** p < 0.01

4.2. Robustness test

4.2.1. Replacing the dependent variable

In corporate economics, productivity levels are closely tied to total factor productivity, which captures the technical efficiency of new-form productivity. Hence, we substitute new-type productivity with total factor productivity (TFP_LP) as the dependent variable. As shown in Table 2 Column (1), after swapping the dependent variable, the regression coefficient for DIG is 0.359, remaining positive and significant at the 1% level, corroborating the baseline results.

4.2.2. Adding control variables

Drawing on [21], we include return on equity (Roe) and sales growth (Growth) to counteract omitted variable bias. Column (2) shows that the coefficient of digital transformation on new-quality productivity is 0.00513, significantly positive at the 1% level, aligning with baseline conclusions.

4.2.3. Removing abnormal years and changing the sample period

To minimize the influence of external shocks, particularly the COVID-19 pandemic’s severe effect on business operations post-2020, data from 2020 to 2022 are omitted. Column (3) indicates that the digital transformation coefficient is 0.00618, significantly positive at the 1% level, further affirming the baseline findings.

Standard errors in parentheses^* p < 0.1, ^** p < 0.05, ^*** p < 0.01

4.3. Mechanism analysis

4.3.1. Green technology innovation

Column (1) of Table 3 shows that the DIG coefficient is 0.637, significant at the 1% level, indicating that digital transformation stimulates green technology innovation. Column (2) shows that the DIG coefficient is 0.00503 and the Green coefficient is 0.000334, both significant at 1%, confirming that green innovation mediates the relationship between digital transformation and new productivity. Hypothesis 2 is supported.

4.3.2. Supply chain efficiency

Column (3) reveals that the DIG coefficient is 2.624, significant at 1%, implying that digital transformation enhances supply chain efficiency. Column (4) shows that the DIG coefficient is 0.00480 and the Stock_day coefficient is 0.000166, both significant at 1%, indicating that supply chain efficiency serves as a mediator. Hypothesis 3 is validated.

Standard errors in parentheses^* p < 0.1, ^** p < 0.05, ^*** p < 0.01

4.4. Heterogeneity analysis

4.4.1. Financing constraints

Financing constraints describe the difficulty firms face in obtaining external funding at reasonable rates due to information asymmetry and high transaction costs [22]. Using the SA index as in [23], we split the sample into high and low financing constraint groups based on the median. Columns (1) and (2) of Table 4 show that the DIG coefficient is larger for firms with low constraints, suggesting that digital transformation demands sustained investment in data systems and talent, which cash-strapped firms struggle to afford, thus hindering the transformation’s benefits.

4.4.2. Property rights nature

Columns (3) and (4) reveal that non-state enterprises exhibit a stronger and statistically significant impact on new productivity at the 1% level compared to state-owned enterprises. This may be because non-state firms face fiercer market competition, view digital transformation as vital for survival, and have fewer social duties and more flexible structures, enabling quicker digitization.

4.4.3. R&D investment

Using the median ratio of R&D spending to operating revenue, we categorize firms into high and low R&D groups [24]. Columns (5) and (6) demonstrate that high-R&D firms have a larger DIG coefficient, indicating that greater R&D investment amplifies the contribution of digital technology to new-quality productivity.

Standard errors in parentheses^* p < 0.1, ^** p < 0.05, ^*** p < 0.01

5. Conclusions

This empirical study, using panel data from Chinese A-share listed companies (2015–2022), clarifies the inherent mechanisms by which digital transformation enhances new quality productivity. The results robustly show that corporate digital transformation has a significant positive impact on new quality productivity. This relationship is partially mediated through green technology innovation and supply chain efficiency. Heterogeneity tests further disclose that the effect is more pronounced among firms with lower financing constraints, non-state-owned enterprises, and those with higher R&D intensity.

Based on these results, we propose the following recommendations:

For businesses, digital transformation should be embedded as a strategic priority for developing new quality productivity. Investments should focus on smart equipment and integrated digital platforms. Integration of green innovation with digital infrastructure is essential—e.g., through real-time environmental monitoring and algorithms for low-carbon production. Supply chain digitization should be optimized using IoT for real-time synchronization of inventory, logistics, and production, enhancing overall efficiency.

For policymakers, support for corporate digital transformation should be strengthened through improved digital infrastructure (e.g., 5G, Industrial Internet) to lower access costs, and fiscal measures such as subsidies and tax incentives. Policies should be tailored based on heterogeneity: a dedicated digital transformation fund could offer low-interest loans for green R&D to ease financial pressure; a public service platform could provide technical support and training, especially for R&D-weakened firms; firms of different ownership should be guided toward transformation paths fitting their goals while boosting social responsibility. Ultimately, leveraging the synergy between green innovation and digitization will provide lasting impetus for enterprise and social development.

While this study confirms mediation through green innovation and supply chain efficiency, other paths—such organizational restructuring and human capital improvement—remain unexamined. Moreover, measuring digital transformation via keyword frequency in annual reports, though empirically useful, may not fully capture implicit digital practices. Future research should develop integrated theoretical frameworks and refined metrics to better guide policy and strategy.