How Can Digital Transformation Drive a Green Future?—Intermediary Mechanisms for Supply Chain Innovation: Evidence from Chinese A-Share Listed Companies

Abstract

Against the backdrop of stricter global carbon emission policies, corporate green transition performance has become a key driver for advancing sustainable development. Based on data from A-share listed companies in China from 2015 to 2022, this study empirically examines the mechanisms by which digital transformation impacts corporate green transformation performance. The findings reveal that (1) Digital transformation significantly promotes corporate green transformation, with supply chain innovation serving as a critical mediating factor; (2) The environmental awareness of senior executives and the strategic proactiveness of enterprises exert a significant moderating effect on this relationship. Enhanced environmental awareness among executives drives enterprises to leverage digital tools for green transformation; conversely, excessive strategic proactiveness exerts a constraining influenc; (3) Heterogeneity analysis indicates that firm-specific characteristics, industry attributes, and regional disparities produce differentiated effects. State-owned enterprises, benefiting from their policy support and resource advantages, are more likely to advance green innovation through enterprise digital transformation. Non-high-tech industries tend to optimize production processes, control pollution, and improve operational efficiency through digitalization. Moreover, in regions with stringent environmental regulations, the positive impact of digitalization on both innovation performance and environmental outcomes becomes particularly pronounced. This study enriches theoretical understanding of the integration between digitalization and greening, and by uncovering the pivotal role of supply chain innovation provides practical guidance and policy insights for enterprises advancing sustainable development.

1. Introduction

Currently, the world collectively faces the severe challenge of overlapping climate change and resource scarcity. As the world’s largest developing country, with the dual carbon goals of “carbon peak by 2030 and carbon neutrality by 2060”, it has made commitments to the international community and elevated “digital-green synergy” to a national strategy through the “Overall Plan for Digital China Construction”, forming a transformation paradigm driven by digital technology and guided by green and low-carbon development [1]. This “dual carbon” vision not only demonstrates China’s commitment to global sustainable development but also charts a roadmap and timeline for domestic enterprises to undergo green transformation. Against this backdrop, the Chinese government has clearly recognized that digital technology, as a key engine of the new round of technological revolution and industrial transformation, can inject new momentum and possibilities into green development [2]. In 2023, the “Overall Plan for Building a Digital China” was officially released, elevating “digital-green synergistic transformation” to a national strategic priority for the first time. Amid the urgent demand for green transformation and the surging wave of digitalization, enterprises—as micro-level drivers of the digital economy—urgently need to leverage digital technologies to build differentiated, customized transformation pathways. This will propel the deep integration and coordinated advancement of digitalization and greening [3,4]. This topic has become a crucial direction requiring in-depth exploration within China’s digital economy strategy.

Green transition performance (EGT) refers to the process of achieving sustainable development by incorporating environmentally friendly practices and technologies into economic and social activities, thereby reducing resource consumption, lowering carbon emissions, and enhancing ecological efficiency [5]. In recent years, the international community has attached considerable importance to the concept and practice of green transformation. Empirical evidence from various nations also robustly confirms that green transformation is a necessary pathway to address climate change [6], resource depletion [7], and to achieve stable economic growth [8]. The growing importance of green transition has reached a consensus in recent academic discourse, with emerging literature shifting focus to pathways for achieving sustainable transformation. Gea-Bermúdez [9] highlight the role of sector coupling in facilitating the green transition within the North-Central European energy system by 2050. Hossain et al. [10] empirically investigate the relationship between financial digitalization and green innovation using data from 15 advanced and emerging economies. Ma and Zhu [11] demonstrate that the digital economy directly drives high-quality green development, while Mahmood et al. [12] identify environmental regulations as significant catalysts for green growth through OECD data. Despite increasing attention to the supportive role of digital technologies in sustainability, a scholarly consensus remains elusive regarding the mechanisms through which digitalization influences green transition pathways. Particularly underexplored is how corporate digital transformation, after reshaping supply chain dynamics, subsequently reconfigures organizational trajectories toward a green transition—a critical research gap that requires systematic investigation.

Digital transformation (DCG) represents a fundamental restructuring of organizations and economies, propelled by advances in artificial intelligence, IoT, cloud computing, and big data [13,14]. At its heart, this shift focuses on harnessing digital tools to facilitate the seamless flow and integration of information, improve how resources are allocated, and boost operational performance—all aimed at fostering competitive advantage and sustainable development [15]. However, it extends beyond mere technological adoption; it fundamentally reconfigures organizational architectures, workflows, and revenue models [16,17]. Digital technologies, exemplified by cloud computing, big data analytics, the Internet of Things (IoT), and artificial intelligence (AI), are fundamentally reshaping corporate operational paradigms and exerting profound influences on production decision-making and organizational performance [18]. From an economic performance perspective, enterprises leverage digital technologies such as big data analytics and AI to conduct in-depth analyses of consumer behavior, enabling the delivery of personalized services and fostering enhanced customer loyalty. Li et al. highlights that digital overhaul reconstitutes operational models and generates positive spillovers on economic outcomes. Similarly, Guo et al. [19] corroborate the economic benefits of digitalization, Li et al. [20,21] arguing that it enhances operational cost efficiency, optimizes asset turnover ratios, and improves administrative productivity, thereby elevating overall economic performance. In terms of environmental performance, Rachinger et al. [22] posit that digital tools revolutionize product lifecycle management by enabling end-to-end tracking and optimization across design, production, usage, and recycling phases. Pan et al. [23] further investigate the synergistic effects of dual digital-green innovation in driving corporate sustainability transitions. Fernando et al. [24] find that eco-innovation enhances sustainable performance, while service innovation capabilities achieve corporate differentiation by emphasizing value creation, ultimately benefiting the enterprise. While extant literature has explored the multidimensional impacts of digital transformation, scholarly attention to the nexus between digital technologies and corporate environmental governance remains nascent. Existing studies predominantly focus on singular aspects such as emission reduction or pollution control, often neglecting a comprehensive framework integrating pollution mitigation processes and green performance metrics. This oversight obscures the intrinsic mechanisms through which digital technologies reconfigure organizational pathways toward green transition. Addressing this gap, this study employs a dual analytical lens of green innovation and green performance to systematically examine the effects and pathways of digital technologies in corporate green transitions. By synthesizing pollution governance processes with green performance outcomes, we clarify the underlying logic of how digital tools promote sustainable practices, providing an important extension to the existing literature.

How can enterprises leverage digital transformation to drive green transformation? From a theoretical perspective, first, according to production function theory, a firm’s output, under given technological conditions, depends on the input of production factors, including labor and capital [25]. However, the law of diminishing marginal returns emphasizes that as resources are gradually depleted, the additional output generated by an incremental unit of input declines. Digital technologies help reduce resource consumption, not only generating cost-saving effects but also enhancing resource allocation efficiency, which is critical for achieving corporate energy-saving goals. Furthermore, transaction cost theory highlights that due to factors such as information asymmetry and negotiation costs, coordination costs within firms and across supply chains remain high [26]. Digital reformation, through the integration of intelligent information systems and the restructuring of automated processes, lowers contract enforcement costs and interdepartmental coordination costs in environmental management [27]. A digital monitoring system covering the entire production lifecycle effectively mitigates information asymmetry and, through the effect of declining marginal costs, reduces transaction frictions in green technology adoption. Finally, innovation diffusion theory reveals that digital transformation as a technological carrier accelerates the absorption and iteration of green innovation technologies. By leveraging tools such as digital twins and the Internet of Things, firms can overcome the spatial and temporal constraints of traditional environmental governance, fostering a “digital-green” collaborative innovation paradigm. This tripartite mechanism collectively constitutes the endogenous driving force behind corporate green transition [28].

The reason for selecting China as the subject of this paper is based on the following considerations: First, the Chinese government has recently accelerated its efforts to promote green and digital transformation by setting targets such as “carbon peaking and carbon neutrality” and outlining clear decarbonization strategies in the “14th Five-Year Plan.” Strong policy support and mandatory regulations have spurred enterprises to pursue both green and digital transformations. As a result, China has gained a significant advantage in green innovation and technology applications, making it an exemplary case for study [29,30]. Second, as the world’s second largest economy, China features a highly complex industrial structure that spans a diverse range of sectors—from energy-intensive manufacturing to high-tech industries [31]. Studying China thus provides a comprehensive perspective for analyzing how digital transformation affects green transformation performance across different industrial environments, offering valuable insights for other countries. Lastly, China occupies a critical position in the global supply chain as the world’s largest manufacturing hub and a key node [32]. Consequently, the digital and green transformation efforts of Chinese enterprises directly impact the sustainability and carbon emissions of the global supply chain, with worldwide implications. Research into the synergistic effects of digital and green transformation in China can offer useful lessons for other nations, particularly in the area of cross-national cooperation on carbon reduction and technological innovation.

Our research faces the following main challenges: First, both digital transformation and green transformation are complex, multidimensional phenomena [33]. Our first challenge is to develop a robust and practical set of indicators. To address this, we attempt to construct comprehensive indices for both transformations. Specifically, for digital transformation, we focus on dimensions such as the application of digital technologies and the organization’s digital strategy, and we employ text analysis—using word frequency counts—to gauge its extent [34]. In contrast, to capture enterprise green transformation, we measure both environmental performance and green innovation performance. Second, there is an endogeneity challenge in estimating the impact of digital transformation on enterprise green transformation performance. Reverse causality, omitted variables, and self-selection bias may exist between these two processes, making it difficult to identify clear causal relationships. To mitigate these issues, this paper utilizes fixed effects models, instrumental variable techniques, and the Heckman test in order to more accurately identify the causal effects.

Using a sample of Chinese A-share listed companies from 2015 to 2022, this study empirically demonstrates that corporate digital transformation effectively enhances green performance. This result is robust to multiple tests and remains stable. The analysis further examines the mediating role of supply chain innovation and the moderating effects of executives’ environmental awareness and corporate strategic aggressiveness [35].

The primary contributions of this paper are mainly in the following aspects: First, it expands the research on the influencing factors of enterprises’ green transformation performance. Different from the existing literature that focuses on executive characteristics [36], environmental regulation [12], and technological innovation [37], we focus on digital transformation that is of strategic significance to enterprises. We also establish the connection between digital transformation and the performance of enterprises’ green transformation, effectively promoting the research progress of related literature. Second, our research deepens our understanding of the impact of digital transformation on environmental performance. Although past literature has been highly focused on the impact of enterprise digital transformation on environmental performance [23], the specific mechanism by which digital transformation affects the green transformation of enterprises remains unclear. Building upon existing research, this study employs Resource-Based View (RBV) to reveal how digital resources underpin green transformation, applies Transaction Cost Theory to explain how supply chain innovation reduces coordination costs in transformation, and utilizes Innovation Diffusion Theory to elucidate the cross-entity transmission mechanism of digital practices. This approach overcomes the limitations of single-theory explanations for the “resources–collaboration–diffusion” chain and fills a theoretical gap in understanding the interaction between digital and green transformation within supply chain contexts. Third, the evidence based on Chinese listed companies provides practical references for countries to balance economic growth and green transformation, especially having significant reference value for those facing similar challenges.

The remainder of the paper is organized as follows: Section 2 provides theoretical analysis and hypothesis development; Section 3 presents the research design, describing the model, variables, and data; Section 4 presents the empirical results and analysis; Section 5 provides further exploration, including analysis of mechanism effects and moderating effects; and Section 6 provides conclusions and policy recommendations.

2. Theoretical Framework and Hypotheses

Against the backdrop of rapid digital and green transformation, clarifying the interplay between these two processes holds significant theoretical value and practical relevance. This study constructs an integrated analytical framework to explore how digital transformation drives corporate green transformation through supply chain innovation. It further examines the moderating roles of executive characteristics and strategic orientation in this process. The paper proposes a series of differentiated research hypotheses to systematically reveal the intrinsic pathways through which digitalization enables green transformation. Based on the above research process, this paper proposes the research framework as shown in Figure 1.

2.1. Digital Transformation and Green Transformation

Corporate digital transformation has been widely recognized for its positive effects in enhancing resource allocation efficiency, reducing transaction costs, and driving technological innovation [37], all of which create favorable conditions for the green transition. The Resource-Based View (RBV) posits that a firm’s competitive advantage stems from its unique resources and capabilities [38]. Digital technologies—such as the Internet of Things (IoT) and artificial intelligence (AI)—strengthen corporate IT infrastructure and data analytics capabilities, thereby forming distinctive digital resources. This resource advantage directly translates into enhanced resource allocation efficiency, providing the material foundation for green transformation [39]. For instance, through the Internet of Things (IoT) and artificial intelligence (AI), smart manufacturing and precision agriculture can be realized, significantly enhancing the efficiency of water, energy, and material utilization, directly converting resource optimization into green benefits. This resource capability built by digital technologies is precisely what provides the foundation for the practical application of transaction cost theory. The integration of information systems and process automation brought by digital transformation not only embodies resource capabilities but also reduces transaction frictions in green production by lowering monitoring costs and mitigating information asymmetry (e.g., real-time tracking of production processes) [39]. This aspect bridges the “capability output” of the resource-based view, transforming static resource advantages into dynamic institutional safeguards, propelling green transformation from “technically feasible” to “economically viable.” When enterprises possess both digital resource capabilities (RBV) and achieve reduced transaction costs (transaction cost theory), digital transformation itself—as a form of technological innovation—accelerates the diffusion and application of green technologies. As innovation diffusion theory underscores, the spread and adoption of technological innovations are pivotal drivers of industrial transformation [40]. Sharing economy platforms through credit systems and algorithmic matching propels the large-scale development of circular business models. In summary, starting from the resource-based view establishes the “capability foundation” where digital transformation drives green transformation. Transaction cost theory serves as the “conversion bridge,” while innovation diffusion theory forms the “amplification engine.” This reveals how digitalization as a vehicle for technological innovation accelerates the penetration of green technologies from isolated applications to full-chain integration. Based on this, a closed-loop system of “resource support → mechanism safeguards → scaled diffusion” is formed. Accordingly, the following hypothesis is proposed:

H1. 

Digital transformation enhances corporate green transition performance.

2.2. Digital Transformation, Supply Chain Innovation, and Green Transformation

Digital transformation drives supply chain innovation by deeply restructuring internal processes and external collaboration networks. Based on the resource-based view (RBV) and dynamic capabilities theory, firms maximize value by optimizing existing resources while integrating, building, and reconfiguring resources in dynamic environments to adapt to external changes. Digital transformation enables firms to efficiently integrate internal data resources, optimize information systems, and enhance intelligent decision-making capabilities [41]. This not only strengthens real-time monitoring and high-efficiency coordination across supply chain processes but also generates a compounding effect in process reengineering and internal innovation, enhancing organizational responsiveness and flexibility. Transaction cost economics highlights the impact of transaction costs on corporate decision-making, arguing that reducing information asymmetry and transaction costs enhances economic efficiency [42]. By improving information transparency and coordination efficiency, digital transformation lowers communication and coordination costs among suppliers, distributors, and customers [43], thereby strengthening cross-organizational synergy. Changhong Group has established a closed-loop green supply chain through a digital management system. This system enables real-time monitoring of supplier compliance while facilitating the effective deployment of energy-saving and emission reduction technologies. These outcomes illustrate how digital transformation fosters multidimensional enhancement in supply chain efficiency, resilience, and sustainability through resource reconfiguration, process refinement, and ecosystem-wide collaboration. This enhances corporate resilience in response to evolving market and policy environments, driving continuous innovation and sustainable development across the supply chain network.

Under the premise that supply chain sustainability serves as a contextual condition, the role of supply chain innovation in driving green strategic practices may be even more significant. From an aggregate perspective, supply chain innovation integrates resources across the entire value chain [44], facilitating coordinated operations in production, logistics, and warehousing. Through synergy effects, it promotes information sharing across the value chain, reduces energy consumption and waste emissions, and alleviates environmental burdens.

Regarding marginal effects, each process optimization or introduction of a new technology typically leads to a leap in production efficiency [45]. In line with technology diffusion theory, the generation and application of new technologies yield positive outcomes for corporate green transition. For instance, the Internet of Things (IoT) and big data technologies enable real-time monitoring and precise regulation of production processes, where even incremental improvements contribute to reduced energy waste and pollution emissions. Consequently, supply chain innovation not only optimizes resource allocation at the macro level but also continuously enhances green efficiency at a granular level, jointly advancing corporate green transition. Based on this, the following hypothesis is proposed:

H2. 

Supply chain innovation mediates the relationship between digital transformation and green transition performance.

2.3. The Moderating Effect of Enterprise Digital Transformation on Green Transformation Performance: The Moderating Effect of Executives’ Green Perception and Corporate Strategic Activism

Upper echelons theory posits that executives’ cognition, values, and experiences shape corporate strategic direction [36]. When executives possess a heightened level of green cognition, they prioritize environmental protection, resource conservation, and green innovation in the digital transformation process. This prioritization not only guides firms to emphasize the application and promotion of green technologies during digital transformation but also fosters a “guidance effect,” channeling corporate resources toward green innovation. Consequently, digital technologies are better leveraged to support the green transition.

Drawing on institutional theory [46] and stakeholder theory [47], executive green cognition facilitates firms’ proactive responses to government environmental policies and market-driven green demands, thereby enhancing external legitimacy and attracting green investments to improve environmental performance. This helps firms better align with societal expectations regarding environmental protection, bolstering corporate image and reputation. Such an “external spillover effect” strengthens relationships with governments, investors, and consumers while expanding access to green investment opportunities, ultimately enhancing green transition performance.

From the perspectives of the technology–organization–environment (TOE) framework and dynamic capabilities theory, executive green cognition enables firms to swiftly adjust organizational structures and resource allocation in response to external environmental changes, accelerating the transformation of digitalization outcomes into green achievements. In sum, executive green cognition positively moderates the relationship between digital transformation and green transition by influencing strategic decision-making and resource allocation, thereby fostering the synergistic enhancement of economic and environmental benefits [48].

Overly aggressive strategies, however, often lead top management teams to prioritize short-term market expansion and financial gains at the expense of long-term environmental protection and green investments [49]. According to the upper echelons theory, executives’ decisions directly affect resource allocation. When strategy becomes excessively aggressive, firms may prioritize allocating limited digital resources to high-risk, high-return projects rather than green technological innovation and energy conservation initiatives, thereby weakening the positive impact of digital transformation on green performance.

Moreover, an excessively aggressive strategic orientation can cause resource fragmentation and instability, undermining the firm’s ability to develop a unique, sustainable competitive advantage that aligns digital transformation with green transition [50]. Dynamic capabilities theory suggests that in rapidly evolving market and technological environments, overly aggressive strategic behavior may impair firms’ sensitivity to external environmental regulations and stakeholders’ green demands, making it difficult to timely adjust and optimize green strategies during transformation. Additionally, aggressive strategies may trigger regulatory scrutiny and negative public opinion [51], obstructing firms from generating green value through digital transformation and ultimately weakening the impact of digital transformation on green transition performance. Based on this, the following hypothesis is proposed:

H3a. 

Executive green cognition moderates the core relationship between digital transformation and green transition performance.

H3b. 

Corporate strategic aggressiveness moderates the core relationship between digital transformation and green transition performance.

3. Data Description and Model Specification

To empirically examine the relationship between digital transformation and corporate green transformation, this study utilizes firm-level panel data encompassing key indicators of digital transformation and green transformation, along with control variables reflecting firm characteristics, industry attributes, and regional institutional contexts. Building upon the theoretical framework, we establish an econometric model that accounts for the direct impact of digital transformation on green transformation.

3.1. Data Collection

In 2015, the adoption of the Paris Agreement marked a significant global commitment to carbon reduction and mitigating climate change. China subsequently introduced a series of related initiatives, including the “dual carbon goals” and the “Green Digital Guidelines.” This study uses data from A-share listed companies on the Shanghai and Shenzhen stock exchanges from 2015 to 2022 as the research sample. To ensure the integrity of the sample data and the reliability of the research conclusions, the following screening criteria were applied to the sample companies: ① Exclude ST (Special Treatment Stock) and ST* companies; ② Exclude companies with missing key data; ③ Exclude financial and insurance companies. The green patent application data come from the China Research Data Service Platform (CNRDS), digital transformation data and green executive awareness from company annual reports, environmental performance from the ENV database in CSMAR, and the remaining data from the Wind Database (WIND). This paper uses Excel to organize the relevant data and ultimately employs Stata 17.0 statistical software for data analysis and statistics.

3.2. Measures

3.2.1. Dependent Variable

The dependent variable in this study is green transition performance (EGT), which is measured across two dimensions: green transition quality and environmental performance. To assess green transition quality, the study employs the number of independently filed green invention patents (Inva) per year. To mitigate the issue of right-skewed distribution, the number of green patent applications is increased by one and then log-transformed [52].

For environmental performance (EP), this study follows prior research by utilizing the ENV database from CSMAR and adopting a comprehensive scoring approach. The environmental performance index is constructed based on the following components [53]:

(1)

Whether the firm incorporates an environmental protection philosophy;

(2)

Whether the firm sets environmental protection goals;

(3)

Whether the firm implements environmental management systems;

(4)

Whether the firm conducts environmental training programs;

(5)

Whether the firm engages in specific environmental protection initiatives;

(6)

Whether the firm has an emergency response mechanism for environmental incidents.

(7)

Whether the firm adheres to the “Three Simultaneities” system;

(8)

Whether the firm has received environmental awards or honors;

(9)

Whether the firm has obtained ISO 14001 certification [54].

Each criterion is assigned a score of 1 if met and 0 if unmet. The total score is then used as a proxy for the firm’s environmental performance. As shown in Figure 2, to enhance the clarity of the methodological design, we provide a flowchart to illustrate the measurement process of the dependent variable green transition (EGT). As shown in Figure 2, EGT is achieved through two dimensions: green transition quality (Inva) and environmental performance (EP).

3.2.2. Independent Variable

The independent variable in this study is digital transformation (DCG). Following the approach of [39] and others, this study measures the degree of digital transformation by calculating the total frequency of digital transformation-related terms in corporate annual reports of listed firms. Specifically, the study identifies 76 characteristic keywords across five major categories: artificial intelligence, blockchain technology, cloud computing, big data, and digital technology applications. The sum of occurrences of these keywords is then log-transformed (natural logarithm of the total frequency plus one) to mitigate skewness. A higher value indicates a greater degree of digital transformation within the firm. As shown in Figure 3, we provide a flowchart to illustrate digital transformation.

3.2.3. Control Variables

Considering the numerous factors influencing this study, control variables are introduced to mitigate potential confounding effects. Following the approach of Zhang et al. [55] and other scholars, this study selects control variables at both the firm level and the macro level. At the firm level, the following control variables are included: Firm size (Size), Leverage ratio (Lev), Return on equity (Roe), and Dual leadership structure (Dual). At the macro level, the study incorporates Industrial structure (Struc), Economic development level (lngdp), and Information technology development level (ITlev). Additionally, year (year) and firm (id) fixed effects are introduced as dummy variables to control for unobserved heterogeneity and their potential impact on corporate innovation performance. The control variables are shown in Figure 4 and

Table 1. Control variables.

Variable Name	Symbol	Definition
Firm Size	Size	Natural logarithm of total assets
Leverage Ratio	Lev	Total liabilities/Total assets
Return on Equity	Roe	Return on equity
Dual Leadership	Dual	Whether the chairman also serves as the CEO
Industrial Structure	Struc	Tertiary industry output/Secondary industry output
Economic Development Level	lngdp	Per capita GDP
Informatization Level	ITlev	Total postal and telecommunication business volume/Regional GDP

.

3.3. Model

To examine the relationship between corporate digital transformation and green transition performance, the following baseline regression model is constructed:

$$EG{T}_{idt}={\beta }_0+{\beta }_{1}DC{G}_{idt}+{\beta }_2{X}_{idt}+{\gamma }_i+{\mu }_d+{\delta }_t+{\epsilon }_{idt}$$

(1)

The dependent variable represents the performance of green transformation, which is reflected in the green innovation performance (lnva) and environmental performance (EP) of enterprises. The core independent variable DCG represents digital transformation. The set of control variables includes firm size (Size), leverage ratio (Lev), return on equity (Roe), dual leadership structure (Dual), industrial structure (Struc), economic development level (lngdp), and IT development level (ITlev). Firm fixed effects account for unobserved heterogeneity, industry fixed effects control for industry-specific variations, and year fixed effects address time-related shocks. The random error term c captures unexplained variations.

3.4. Descriptive Statistics

This study conducts a descriptive statistical analysis of the sample data, including the mean, standard deviation, minimum, maximum, and median values. The detailed results of the descriptive statistics are presented in

Table 2. Descriptive Statistics.

Variables	Sample Volume	Mean	Sd	Min	Max	P50
DCG	18,566	1.604	1.377	0.000	5.030	1.386
Inva	18,566	0.257	0.665	0.000	6.328	0.000
EP	18,566	1.952	2.064	0.000	8.000	1.000
Size	18,566	22.273	1.280	17.641	28.293	22.133
Lev	18,566	0.412	0.195	0.008	1.238	0.403
ROE	18,566	0.074	0.691	−85.647	2.379	0.079
Dual	18,566	0.297	0.457	0.000	1.000	0.000
Struc	18,566	1.003	0.272	0.518	3.231	0.956
Ingdp	18,566	45,155.057	31,616.738	2759	143,466	39,418.04
ITlev	18,566	0.064	0.082	0.015	2.520	0.053

.

The descriptive statistical results reveal the fundamental characteristics and distribution of each variable, providing a foundation for subsequent empirical analysis. The mean value of corporate digital transformation (DCG) is 1.604, with a standard deviation of 1.377, indicating substantial variation among firms—some have achieved a high degree of digital transformation, while others remain in the early stages. The mean value of corporate green innovation investment (Inva) is only 0.257, with a median of 0, suggesting that most firms allocate limited resources to green innovation, potentially due to financial constraints, technological limitations, or insufficient policy incentives. The mean value of corporate environmental performance (EP) is 1.952, with a maximum value of 8, demonstrating that while some firms have attained high environmental performance, there is still significant room for overall improvement.

4. Results of Estimation

This section presents the empirical results of our analysis. We apply a benchmark regression to test Hypothesis 1, which examines the direct impact of digital transformation on corporate green transformation. To address potential endogeneity issues, we conduct a series of endogeneity tests and robustness checks to ensure the reliability of our findings. Subsequently, we extend the analysis through heterogeneity tests, examining how equity structure, industry characteristics, and regional regulatory environments shape this relationship. Collectively, these findings provide comprehensive evidence on the mechanisms and boundary conditions through which digital transformation facilitates corporate green transformation.

4.1. Baseline Regression Outcomes

To verify the validity of Hypothesis 1, namely the impact of corporate digital transformation on green innovation, an empirical test was conducted. The regression result presented in

Table 3. Baseline regressions.

	(1)	(2)	(3)	(4)
Variables	Inva	Inva	EP	EP
DCG	0.013 ***	0.009 *	0.080 ***	0.051 ***
	(0.005)	(0.005)	(0.019)	(0.019)
Size		0.058 ***		0.448 ***
		(0.013)		(0.042)
Lev		−0.011		−0.259 *
		(0.038)		(0.145)
ROE		0.001		0.013
		(0.001)		(0.009)
Dual		−0.013		−0.063
		(0.011)		(0.040)
Struc		−0.006		0.020
		(0.013)		(0.044)
Ingdp		−0.000		0.000 **
		(0.000)		(0.000)
ITlev		0.068 **		−0.095
		(0.034)		(0.083)
constant	0.237 ***	−1.042 ***	1.829 ***	−8.025 ***
	(0.008)	(0.287)	(0.030)	(0.918)
individual fixed effect	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES
Year fixed effects	YES	YES	YES	YES
R2	0.700	0.701	0.669	0.674
N	18,498	18,498	18,498	18,498

“*** p < 0.01”, “** p < 0.05”, “* p < 0.10”.

reveal a significant positive effect of digital transformation (DCG) on both green innovation investment (Inva) and environmental performance (EP).

Regression results indicate that digital transformation significantly promotes both green innovation investment and environmental performance in enterprises. Regarding green innovation investment, each unit increase in digital transformation levels leads to an average rise of approximately 1.3% in corporate green innovation spending (Model 1). This effect persists at around 0.9% even after controlling for additional variables (Model 2). Regarding environmental performance, the marginal effect of digital transformation is even more pronounced: a one-unit increase yields an approximately 8.0% improvement in environmental performance (Model 3). Even under more stringent model specifications, an improvement of about 5.1% persists (Model 4). These findings underscore the positive role of digital transformation in driving corporate green transition and sustainable development [56,57]. This improvement may stem from adopting digital technologies that enhance environmental monitoring, improve resource efficiency, and promote green production practices, thereby advancing corporate sustainability [58]. In summary, these findings provide empirical support for Hypothesis 1. More importantly, the management and policy value of digital transformation as a lever for promoting sustainable development is clarified, providing clear guidance for all stakeholders. Enterprises should adopt a “digitalization for greening” approach, utilizing digital tools to collect real-time energy consumption data during production processes and tailor corresponding energy-saving solutions. Governments should create a favorable environment for corporate transformation by refining incentive policies, establishing public service platforms, and combining strengthened oversight with scientific guidance. Through concerted efforts across multiple sectors, a win–win outcome of economic development and ecological conservation will ultimately be achieved.

4.2. Endogeneity Test

4.2.1. Instrumental Variables Test

The conclusions of this study may be affected by endogeneity issues, with one critical source being potential reverse causality. The advancement of digital transformation can optimize a firm’s information environment, thereby enhancing the level of green innovation. However, firms with higher green innovation capabilities may simultaneously be more inclined to increase their investment in digital transformation to strengthen their competitive advantage [57], thus giving rise to reverse causality concerns. To mitigate this endogeneity issue, this study employs the instrumental variable (IV) approach. Specifically, the average level of digital transformation among other firms in the same region, industry, and year is selected as the instrumental variable [59]. This variable is highly correlated with an individual firm’s degree of digital transformation, satisfying the relevance requirement. Meanwhile, as a single firm’s green transition decisions do not directly affect the overall level of digitalization within the industry, the instrumental variable also meets the exogeneity assumption. Furthermore, the first-stage regression results indicate that both the F-statistic and the minimum eigenvalue statistic of the instrumental variable far exceed 16.38, demonstrating that the choice of the instrumental variable is valid and that weak instrument concerns are not present. Based on this, a regression analysis using the instrumental variable approach (2SLS) is conducted, with the second-stage regression results, as shown in

Table 4. Instrumental variables test.

	(1)	(2)	(3)
Variables	DCG	lnva	EP
IV	0.384 ***	0.038 ***	0.017 **
	(32.721)	(5.105)	(2.424)
constant	1.048 ***	−1.396 ***	−1.135 ***
	(51.996)	(−8.970)	(−7.641)
control variable	YES	YES	YES
individual fixed effect	YES	YES	YES
industry fixed effect	YES	YES	YES
Year fixed effects	YES	YES	YES
R2	0.047	0.008	0.006
N	25,266	24,542	24,542

“*** p < 0.01”, “** p < 0.05”.

, revealing that the positive effect of digital transformation on corporate green transition remains significant. This finding further corroborates the robustness of the study’s core conclusion, indicating that digital transformation indeed serves as an effective driver of corporate green transition. The instrumental variable regression results indicate that digital transformation significantly promotes corporate green innovation investment and environmental performance. After controlling for firm characteristics, industry, and year fixed effects, each additional unit of digital transformation increases green innovation investment by approximately 3.8% on average and improves environmental performance by about 1.7%. This finding suggests that digital transformation not only enhances firms’ green innovation capabilities but also contributes to better environmental outcomes. The IV coefficient of 0.384* indicates that the selected instrumental variables possess strong explanatory power for digital transformation, providing a reliable basis for causal analysis.

4.2.2. Heckman Two-Stage Test

The estimation results of the Heckman two-stage model reveal the impact of digital transformation (DCG) on firms’ green innovation investment (Inva) and environmental performance (EP) [60], as shown in

Table 5. Heckman two-stage test.

	(1)	(2)	(3)
Variables	Indicator Variable	lnva	EP
DCG		0.0673 ***	−0.0512 ***
		(9.9250)	(−3.2549)
IV	0.8693 ***
	(30.0318)
imr		0.1473 ***	−0.0033
		(3.8098)	(−0.0300)
Constant	−5.2532 ***	−3.8263 ***	−14.4014 ***
	(−17.9096)	(−17.2609)	(−34.2304)
control variable	YES	YES	YES
individual fixed effect	YES	YES	YES
industry fixed effect	YES	YES	YES
Year fixed effects	YES	YES	YES
R2		0.182	0.344
N	18,463	13,562	13,562

“*** p < 0.01”.

. In the first stage, this study treats firms’ digital transformation adoption as the dependent variable and conducts a Probit regression using the average digital transformation level of firms in the same region, industry, and year as an instrumental variable. The inverse Mills ratio (IMR) is then calculated. In the second stage, the IMR is introduced into the original model (1) to re-estimate the effect of digital transformation on green innovation efficiency. Columns (2) and (3) of the table report the regression results of the Heckman two-step method. For the dependent variable Inva, the IMR is significant at the 1% level, indicating the presence of sample selection bias in this study. Additionally, the coefficient of digital transformation is 0.0641 and remains significantly positive at the 1% level, consistent with the baseline regression results. For the dependent variable EP, the IMR is not significant, suggesting that sample selection bias is not a concern in this case.

4.3. Robustness Check

4.3.1. Exclusion of Special Samples

To further validate the robustness of the research conclusions, this study conducted exclusion tests on specific samples and years. The regression results are shown in

Table 6. Exclusion of special samples.

	Excluding Special City Samples		Exclusion of Special Year Samples
	(1)	(2)	(3)	(4)
Variables	lnva	EP	lnva	EP
DCG	0.014 ***	0.052 ***	0.009 *	0.054 ***
	(2.826)	(3.076)	(1.793)	(3.384)
constant	−0.837 ***	−7.917 ***	−0.995 ***	−7.534 ***
	(−3.672)	(−10.446)	(−4.842)	(−11.155)
control variable	YES	YES	YES	YES
individual fixed effect	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES
Year fixed effects	YES	YES	YES	YES
R2	0.750	0.724	0.746	0.726
N	16,021	16,021	17,466	17,466

“*** p < 0.01”, “* p < 0.10”.

. First, considering the profound impact of the COVID-19 outbreak at the end of 2019 on corporate operations and market environments [61], this study excluded 2020 data to eliminate potential exogenous shocks triggered by the pandemic. After re-running the regression analysis within the adjusted sample, the positive impact of digital transformation on green transition performance remained robust. Each unit increase in digital transformation levels was associated with an approximate 1.4% to 5.2% increase in corporate green investment. This indicates that the research findings hold even when accounting for the effects of this exceptional year.

Additionally, to control for potential interference from regional economic development levels, this study further excluded samples from highly developed regions such as Beijing, Shanghai, Guangzhou, and Shenzhen, thereby mitigating biases arising from regional economic heterogeneity. After removing these cities, the regression results still demonstrated a significant positive impact of digital transformation on corporate green transition, with an increase of approximately 0.9% to 5.4% in the regression excluding the special year sample. This further strengthened the robustness and generalizability of the research findings.

4.3.2. Incorporation of Multidimensional Fixed Effects

To further verify the robustness of the research conclusions, this study added an additional dimension of fixed effects in the baseline regression to control for potential bias of omitted variables. Specifically, on the basis of the existing fixed effects of individuals, industries and years, the interactive fixed effects of time cities and industrial cities were introduced. The “time-city effect” refers to the fixed effect that controls the unobserved shocks of a specific city over time, while the “industrial-city effect” refers to the changes within industries between cities. We use this to more accurately identify the impact of digital transformation on the performance of enterprises’ green transformation. As shown in

Table 7. Incorporation of multidimensional fixed effects.

	Time × City Fixed Effects		Industry × City Fixed Effect
	(1)	(2)	(3)	(4)
Variables	lnva	EP	lnva	EP
DCG	0.010 *	0.057 ***	0.010 *	0.052 ***
	(1.733)	(2.841)	(1.776)	(2.595)
constant	−1.125 ***	−7.141 ***	−1.193 ***	−8.605 ***
	(−3.516)	(−7.233)	(−3.572)	(−8.503)
control variable	YES	YES	YES	YES
individual fixed effect	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES
year fixed effects	YES	YES	YES	YES
R2	0.775	0.764	0.754	0.733
N	17,434	17,434	18,418	18,418

“*** p < 0.01”, “* p < 0.10”.

, the positive impact of digital transformation on green transition performance remains significant, even after controlling for multidimensional fixed effects, including city-time and industry-city factors. Specifically, a one-unit increase in digital transformation is associated with approximately 1% to 5.7% higher investment in green innovation, demonstrating a meaningful marginal effect. This finding suggests that even under more stringent control conditions digital transformation remains a critical driver of corporate green transition. It highlights the practical relevance of promoting digital tools to enhance sustainable practices and strengthens the reliability of the study’s conclusions and their policy implications.

4.3.3. Replacement of the Measurement Model

To further test the robustness of the research conclusions, this study employs alternative econometric models to ensure that the results are not dependent on a specific model specification. In addition to the fixed-effects model used in the baseline regression, the bidirectional clustering robustness and the negative binomial model are also applied as robustness checks. As shown in

Table 8. Replacement of the measurement model.

	Bidirectional Clustering Robustness		Negative Binomial Model
	(1)	(2)	(3)	(4)
Variables	lnva	EP	lnva	EP
DCG	0.010 ***	0.046 **	0.279 ***	−0.093 ***
	(3.669)	(2.609)	(23.556)	(−13.299)
constant	−1.138 ***	−11.418 ***	−8.370 ***	−6.749 ***
	(−6.463)	(−4.644)	(−26.301)	(−36.533)
control variable	YES	YES	YES	YES
individual fixed effect	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES
year fixed effects	YES	YES	YES	YES
N	18,566	18,566	18,566	18,566

“*** p < 0.01”, “** p < 0.05”.

, regardless of the econometric model employed, the positive impact of digital transformation on green transition performance remains significant and directionally consistent. This indicates that the research conclusions are not driven by model selection but rather exhibit strong robustness. Furthermore, this reinforces the notion that digital transformation plays a stable role in promoting corporate green development, as the relationship holds across different model specifications, providing strong support for the reliability of the study’s findings.

4.3.4. Other Robustness Tests

To further validate the robustness of the research conclusions, this study conducts multiple robustness tests, including replacing core variables, eliminating the lagged effect of listed companies’ annual report data, and adjusting clustering methods to ensure the reliability and consistency of the results. As shown in

Table 9. Other robustness tests.

	Replacement of Core Variables		The Effect of Lagging Data From Annual Reports of Listed Companies Is Excluded		Adjustment of the Clustering Method
	(1)	(2)	(3)	(4)	(5)	(6)
Variables	lnva	EP	lnva	EP	lnva	EP
DCG(DCG_L)	0.015 ***	0.035 **	0.009 *	0.070 ***	0.009 *	0.052 *
	(2.720)	(2.021)	(1.662)	(4.106)	(2.051)	(2.314)
constant	−0.938 ***	−8.883 ***	−0.675 ***	−6.141 ***	−1.046 **	−8.065 ***
	(−3.511)	(−10.151)	(−2.796)	(−7.800)	(−3.161)	(−4.744)
control variable	YES	YES	YES	YES	YES	YES
individual fixed effect	YES	YES	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES	YES	YES
year fixed effects	YES	YES	YES	YES	YES	YES
R2	0.776	0.746	0.769	0.738	0.749	0.726
N	15,705	15,705	17,085	17,085	18,498	18,498

“*** p < 0.01”, “** p < 0.05”, “* p < 0.10”.

, regardless of whether the core variables are replaced, the lagged data effects are removed, or the clustering method is adjusted, the regression results consistently demonstrate the stable positive effect of digital transformation on green transition performance. This indicates that the research conclusions exhibit strong robustness, remaining unaffected by variable selection, data lag, or error clustering methods, thereby providing more solid empirical support for theoretical inferences and policy formulation.

4.4. Heterogeneity Test

This study conducts a heterogeneity analysis at the firm level, industry level, and regional level to examine how the impact of digital transformation on green transition performance varies across different types of firms.

4.4.1. Heterogeneity Analysis at the Firm Level

(1)

Different ownership structures

Firms with varying ownership structures exhibit differences in their emphasis on and engagement in green innovation activities. To further examine the differential impact of digitalization on green innovation performance across ownership types, this study classifies the sample firms into state-owned enterprises (SOEs) and non-state-owned enterprises (non-SOEs) and conducts a regression analysis. The objective is to explore how digital transformation influences the green transition performance of SOEs and non-SOEs differently.

Table 10. Heterogeneity analysis at the firm level.

	SOEs	Non-SOEs	SOEs	Non-SOEs
	(1)	(2)	(3)	(4)
Variables	Inva	Inva	EP	Ep
DCG	0.015 *	0.007	0.051	0.042 **
	(1.724)	(1.225)	(1.627)	(2.469)
constant	0.271	−1.294 ***	−6.225 ***	−7.938 ***
	(0.600)	(−5.561)	(−3.804)	(−11.090)
control variable	YES	YES	YES	Yes
individual fixed effect	YES	YES	YES	Yes
industry fixed effect	YES	YES	YES	Yes
year fixed effects	YES	YES	YES	Yes
R2	0.783	0.721	0.723	0.721
N	5942	13,705	5942	13,705

“*** p < 0.01”, “** p < 0.05”, “* p < 0.10”.

results indicate that in the dimension of green innovation (Inva), digital transformation significantly promotes green innovation in SOEs (β = 0.015, p < 0.10), whereas its effect on non-SOEs is less pronounced (β = 0.007, p > 0.10). At the environmental performance (EP) level, digital transformation does not exhibit a significant effect on SOEs (β = 0.051, p > 0.10), while it demonstrates a significantly positive impact on non-SOEs (β = 0.042, p < 0.05).

This result indicates that there is a logical divergence between the two types of enterprises: “policy-driven innovation manifeficialization” and “market-driven behavior materialization”. The actions of state-owned enterprises should give priority to responding to government policy orientations (such as the “dual carbon” goals and green development strategies) and demonstrate their ability to implement policies through explicit achievements. Green patents are the core hard indicator to prove the policy implementation ability and are linked to assessment and subsidies. Digital transformation provides state-owned enterprises with technical tools to increase the number of green patents, such as integrating R&D resources and accelerating technological iteration. This forms the transformation path of “digitalization–green patent”. The survival of non-state-owned enterprises depends on cost control and reputation accumulation in market competition. The environmental performance reflected in the ENV database of CSMAR (such as the reduction in wastewater and waste gas and the operational efficiency of environmental protection facilities) is directly related to the market reputation of enterprises. Digital tools provide it with precise means to optimize environmental behavior, such as monitoring emissions and tracing inputs. Digitizing actual environmental behaviors leads to more significant improvements in environmental performance.

4.4.2. Analysis of the Heterogeneity of the Industry in Which the Company Operates

(1)

Whether it is a high-tech industry.

The impact of digital transformation (DCG) on green innovation performance (Inva) and environmental performance (EP) varies between high-tech and non-high-tech industries. As shown in

Table 11. Whether it is a high-tech industry.

	High-Tech Industries	Non-High-Tech Industries	High-Tech Industries	Non-High-Tech Industries
	(1)	(2)	(3)	(4)
Variables	Inva	Inva	EP	EP
DCG	0.013 *	0.001	0.023	0.085 ***
	(1.905)	(0.193)	(1.123)	(3.426)
constant	−1.498 ***	−0.495 *	−7.301 ***	−9.443 ***
	(−4.789)	(−1.722)	(−8.025)	(−7.930)
control variable	YES	YES	YES	YES
individual fixed effect	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES
year fixed effects	YES	YES	YES	YES
R2	0.766	0.624	0.728	0.734
N	11,365	7074	11,365	7074

“*** p < 0.01”, “* p < 0.10”.

,, in terms of green innovation (Inva), digital transformation has a significantly positive effect in high-tech industries (β = 0.013, p < 0.10), whereas this effect is not significant in non-high-tech industries (β = 0.001, p > 0.10). At the environmental performance (EP) level, digital transformation does not exhibit a significant effect in high-tech industries (β = 0.023, p > 0.10), but it has a pronounced impact in non-high-tech industries (β = 0.085, p < 0.01).

High-tech industries are knowledge-intensive sectors whose production and innovation rely on cutting-edge technologies. Digital tools such as big data analytics and digital twins can directly support green technology R&D processes—for instance, by simulating and optimizing the performance of new energy materials or using machine learning to identify optimal pathways for energy-saving and emission reduction technologies. Simultaneously, high-tech industries possess stronger technological absorption capabilities, enabling them to rapidly convert digital resources into green innovation outcomes (e.g., green patents). Therefore, the adoption of digitally enabled strategies exerts a more pronounced effect on enhancing green innovation performance (Inva) [62]. In contrast, non-high-tech industries (such as traditional manufacturing, textiles, and heavy industry) are predominantly “capital/labor-intensive” with lower technological barriers. Their production processes involve significant visible environmental issues, including high energy consumption and direct discharge of wastewater and exhaust gases. For these industries, digital transformation tends to focus on addressing “practical environmental issues at the production end.” Digital tools can directly optimize production processes—for example, reducing energy waste through real-time monitoring of consumption data or minimizing pollutant exceedances caused by human error via automated equipment—thereby enhancing corporate environmental performance.

(2)

Industry Pollution Intensity.

Given the differences in industry characteristics, firms exhibit varying needs and conditions for engaging in green innovation activities. Accordingly, the sample firms are divided into heavily polluting and non-heavily polluting enterprises to further examine the differential impact of digital transformation on green innovation performance across industries. A regression analysis is conducted based on Model (1), with results presented in

Table 12. Industry pollution intensity.

	Heavily Polluting Industries	Non-Heavily Polluting Industries	Heavily Polluting Industries	Non-Heavily Polluting Industries
	(1)	(2)	(3)	(4)
Variables	Inva	Inva	EP	EP
DCG	0.017 *	0.007	0.044	0.057 ***
	(1.645)	(1.182)	(1.182)	(3.335)
constant	−1.116 **	−1.051 ***	−8.518 ***	−7.886 ***
	(−2.422)	(−4.122)	(−5.135)	(−9.939)
control variable	YES	YES	YES	YES
individual fixed effect	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES
year fixed effects	YES	YES	YES	YES
R2	0.646	0.770	0.737	0.703
N	4385	14,076	4385	14,076

“*** p < 0.01”, “** p < 0.05”, “* p < 0.10”.

. Columns (1) and (3) display the regression results for heavily polluting industries, while Columns (2) and (4) present those for non-heavily polluting industries. The findings indicate that in terms of green innovation investment (Inva), digital transformation has a significantly positive effect on heavily polluting industries (β = 0.017, p < 0.10), whereas its impact on non-heavily polluting industries is not significant (β = 0.007, p > 0.10). Regarding environmental performance (EP), digital transformation does not exhibit a significant effect on heavily polluting industries (β = 0.044, p > 0.10), but it has a significant and positive effect on non-heavily polluting industries (β = 0.057, p < 0.01).

4.4.3. Analysis of Regional Heterogeneity of Firms

(1)

Intensity of different environmental regulations.

Government-led environmental regulations, whether incentive-based or mandatory, play a crucial role in curbing ecological degradation and compelling firms toward green transformation [63]. It is reasonable to expect that the driving force of digital transformation (DCG) on green transition efforts varies with the stringency of regional environmental regulations. The regression results in

Table 13. Intensity of different environmental regulations.

	Low Environmental Regulations	High Environmental Regulations	Low Environmental Regulations	High Environmental Regulations
	(1)	(2)	(3)	(4)
Variables	Inva	Inva	EP	EP
DCG	0.004	0.017 **	0.037	0.083 ***
	(0.451)	(2.027)	(1.347)	(3.384)
constant	−2.149 ***	−1.010 **	−7.202 ***	−9.031 ***
	(−5.605)	(−2.549)	(−5.565)	(−7.879)
control variable	YES	YES	YES	YES
individual fixed effect	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES
year fixed effects	YES	YES	YES	YES
R2	0.800	0.767	0.762	0.768
N	8260	7149	7149	8260

“*** p < 0.01”, “** p < 0.05”.

show that under low environmental regulation conditions, the impact of DCG on green innovation and environmental performance is not significant. Under low environmental regulatory conditions, the effect of DCG on green innovation and environmental performance is not significant. However, under high environmental regulatory conditions, the positive impact of DCG on innovation investment and environmental performance is significantly enhanced. Notably, the effect of DCG on environmental performance under stringent environmental regulations is strongly positive (coefficient = 0.083, p < 0.01), underscoring the critical role of environmental regulations in fostering both digital and green transformation.

(2)

Different geographic locations.

The impact of digitalization of regional industries on green innovation performance (Inva) and environmental performance (EP) shows significant regional heterogeneity. Specifically, as can be seen in

Table 14. Different geographic locations.

	Western	Central and Eastern	Western	Central and Eastern
	(1)	(2)	(3)	(4)
Variables	Inva	Inva	EP	EP
DCG	0.032 **	0.006	0.017	0.057 ***
	(2.514)	(1.133)	(0.372)	(3.453)
constant	−0.773	−1.116 ***	−7.874 ***	−8.088 ***
	(−1.356)	(−4.624)	(−3.798)	(−10.597)
control variable	YES	YES	YES	YES
individual fixed effect	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES
year fixed effects	YES	YES	YES	YES
R2	0.668	0.758	0.739	0.728
N	2488	16,005	2488	16,005

“*** p < 0.01”, “** p < 0.05”.

, in the western region, DCG has a significantly positive effect on innovation investment (p < 0.05), suggesting that the integration of digital and green technologies may be driving innovation in this area. In contrast, in the central and eastern regions, DCG does not significantly influence innovation investment but has a strong and positive effect on environmental performance (p < 0.01). This divergence may reflect the higher maturity of digital technology applications and resource management in the central and eastern regions, allowing DCG to have a more direct impact on environmental improvement. These regional differences underscore the need for differentiated policy approaches.

5. Further Analysis

This section examines underlying mechanisms by testing the mediating role of supply chain innovation (Hypothesis 2) and the moderating role of managerial environmental awareness and strategic proactiveness (Hypothesis 3), thereby deepening our understanding of how and under what conditions digital transformation drives corporate green transformation.

5.1. Mechanism Analysis

The preceding analysis examined how customer digital transformation drives the green transition of supplier firms. This study adopts the two-step method proposed by Baron and Kenny (1986) to test the mediating mechanism [64]. This study further introduces supply chain innovation as a mediating variable and employs the two-step method to test the mediation mechanism. In the two-step method, the first stage involves regressing the mediating variable M on the independent variable X to obtain the predicted value of M; the second stage regresses the dependent variable Y on the predicted value of M, with the resulting coefficient representing the estimated causal effect of M on Y.

To identify the causal mediation effect, this method must satisfy three key assumptions: (1) Relevance—there must be a strong correlation between X and M; (2) Exogeneity—X must be random or quasi-random; and (3) Exclusion restriction—the effect of X on Y should be fully mediated through M. In the empirical test, relevance can be evaluated using weak instrument diagnostics. In the context of this study, both the exogeneity and exclusion restriction conditions are reasonably satisfied. First, corporate digital transformation is typically driven by macro-level policies, industry trends, technological shifts, and long-term strategic planning, rather than by a firm’s current green transition performance or level of supply chain innovation, thereby suggesting strong theoretical exogeneity. Second, the empirical model incorporates a comprehensive set of control variables—including firm size, industry characteristics, financial indicators, and regional attributes—to account for potential confounding pathways and enhance the validity of mediation identification.

This method consists of two main steps: examining whether the independent variable (X) significantly affects the mediating variable (M) and whether the mediating variable (M) significantly affects the dependent variable (Y) in order to determine whether a mediation effect exists.

The first step is to test whether the independent variable (X) significantly influences the mediating variable (M). The regression equation for this test is constructed as follows:

$$SSC={\beta }_0+{\beta }_{1}DC{G}_{idt}+{\beta }_2{X}_{idt}+{\gamma }_i+{\mu }_d+{\delta }_t+{\epsilon }_{idt}$$

(2)

From the first stage, the predicted value of the mediating variable, SSC fitted, is obtained. In the second stage, a regression is conducted with the dependent variable (Y) on the predicted value of SSC. The regression coefficient of SSC fitted on Y represents the causal estimate of M on Y. The following regression equation is constructed:

$$DC{G}_{idt}+{\beta }_0+{\beta }_{1}SSCfitted+{\beta }_2{X}_{idt}+{\mu }_d+{\delta }_t+{\epsilon }_{idt}$$

(3)

As shown in

Table 15. Mechanism analysis.

	(1)	(2)
Variables	SSC	Inva
DCG	−0.408 ***	−0.063 *
	(−3.109)	(−1.657)
constant	69.994 ***	12.068 *
	(9.746)	(1.848)
control variable	YES	YES
individual fixed effect	YES	YES
industry fixed effect	YES	YES
year fixed effects	YES	YES
R2	0.016	0.009
N	18,571	20,158

“*** p < 0.01”, “* p < 0.10”.

, supply chain innovation plays a mediating role in digital and green transformation performance. Since corporate environmental performance is measured by internal corporate initiatives, which supply chain innovation does not affect, this study uses only corporate green innovation performance for validation.

5.2. Analysis of Moderating Effects

This study further explores the moderating roles of executives’ green perception (Egp) and firms’ strategic aggressiveness (ST) in the relationship between digital transformation (DCG) and corporate green transition. The regression analysis in

Table 16. Analysis of moderating effects.

	Executive Green Perceptions		Corporate Strategy Aggressiveness
	(1)	(2)	(3)	(4)
Variables	Inva	EP	Inva	EP
DCG	0.050 ***	0.038 **	0.005	0.063 ***
	(16.422)	(2.325)	(0.803)	(3.175)
Egp	0.005 ***	0.062 ***
	(6.569)	(15.241)
ST			−0.003	−0.010 *
			(−1.565)	(−1.754)
DCG × Egp	0.002 ***	0.006 **
	(2.860)	(2.568)
DCG × ST			−0.004 ***	−0.010 ***
			(−3.633)	(−2.975)
constant	−1.044 ***	−8.261 ***	−0.787 ***	−9.498 ***
	(−4.505)	(−11.081)	(−2.631)	(−9.580)
control variable	YES	YES	YES	YES
individual fixed effect	YES	YES	YES	YES
industry fixed effect	YES	YES	YES	YES
year fixed effects	YES	YES	YES	YES
R2	0.750	0.731	0.778	0.745
N	17,639	17,639	12,945	12,945

“*** p < 0.01”, “** p < 0.05”, “* p < 0.10”.

shows that executives’ green perception has a significant positive moderating effect on the impact of digital transformation on green innovation investment (Inva) and environmental performance (EP) (DCG × Egp: β = 0.002, p < 0.01; β = 0.006, p < 0.05). This indicates that when executives possess a stronger green perception, the positive effect of digital transformation on corporate green innovation and environmental performance becomes more pronounced. A heightened environmental awareness among executives encourages firms to proactively integrate digital technologies into green innovation while enhancing the execution of environmental strategies, thereby accelerating the application of digital technologies in environmental management and sustainable development.

In contrast, strategic aggressiveness exhibits a significant negative moderating effect on the impact of digital transformation on green innovation and environmental performance (DCG × ST: β = −0.004, p < 0.01; β = −0.010, p < 0.01). This finding suggests that when firms adopt a more aggressive strategic orientation, the positive influence of digital transformation on green innovation is weakened, and its effect on environmental performance is significantly reduced. One possible explanation is that firms with an aggressive strategic posture prioritize short-term performance objectives, allocating resources primarily toward market expansion and profit maximization, thereby diminishing investments in digital technologies for green innovation and environmental sustainability.

Based on the above research results, our H2 is proven.

6. Conclusions

This section summarizes the key findings of this study, highlighting how digital transformation facilitates corporate green transition through supply chain innovation and under varying contextual conditions. Based on these results, we provide policy implications aimed at promoting sustainable digital and green development. At the same time, we acknowledge the study’s limitations in data coverage and methodological scope and outline future research directions to further explore alternative mechanisms.

6.1. Deliberations

As environmental issues intensify, the question of how to promote green transition has become a focal point of global attention. In addition to enhancing the economic performance of enterprises, digital technologies also present new opportunities for improving their environmental performance. The baseline findings of this study demonstrate that the digital transformation of enterprises significantly enhances green transition performance, aligning with the conclusions of Hart et al. [65], who stated that the widespread adoption of digital technologies and resource optimization contributes to reducing energy consumption and pollution emissions, thereby improving environmental performance. The research by Wang et al. [66] further supports this conclusion, analyzing how the digital economy through intelligent and digital upgrades promotes high-quality development in the energy sector, ultimately reducing carbon emissions. Additionally, Lyu et al. [67] find that digital transformation, by optimizing energy utilization efficiency, helps alleviate energy poverty and improve environmental performance.

However, some literature presents contrasting viewpoints. Zhong et al. [68] argue that digital transformation may exhibit a “data curse” effect or a reversed U-shaped relationship. Zhong et al. [68] examine the impact of information and communication technologies (ICT) on carbon emissions reduction, suggesting that in certain instances, the widespread adoption of ICT may have a limited effect on carbon reduction due to technological substitution effects, and in some high-energy-consumption industries, it could even lead to an increase in carbon emissions. The study by Moyer and Hughes [69] also posits that the extensive application of ICT in certain contexts may result in higher carbon emissions, particularly regarding the energy consumption of data centers and network infrastructure. These divergent conclusions may arise from differences in industry contexts, data samples, and technological applications across the various studies.

Our mechanism analysis reveals that supply chain innovation mediates the relationship between digital transformation and green transformation, supporting the theory of Ketchen and Hult [70]: digital technologies drive green transformation by optimizing supply chain coordination, reducing transaction costs, and promoting information sharing. However, Büyüközkan and Göçer [71] contend that while digitalized supply chains enhance efficiency, their impact on green transformation remains uncertain, particularly when firms fail to effectively align digital technologies with environmental objectives, potentially leading to dispersed benefits or limited short-term improvements in environmental performance. Teece [72] emphasizes that digital supply chain innovation primarily focuses on economic benefits, and its contribution to environmental gains in the short term may be constrained by industry characteristics and technological maturity. Consequently, in some instances, digital transformation may not directly drive green transformation. Moreover, our research also identifies that executives’ green awareness and the aggressiveness of corporate strategy play a moderating role in the “digital transformation–green transformation performance” relationship, consistent with the findings of Qader et al. [73]. This study indicates that Industry 4.0 technologies enhance the resilience and performance of supply chains, with executives’ green awareness and strategic aggressiveness playing a pivotal role in this process. Green awareness helps firms better leverage digital technologies to advance green transformation.

6.2. Key Findings

This study focuses on China, aiming to identify the impact and mechanisms of digital transformation on corporate green development. Using data from A-share listed companies in the Shanghai and Shenzhen stock exchanges from 2015 to 2022, the key findings are as follows:

First, digital transformation significantly enhances corporate green transition performance. This conclusion remains robust after a series of robustness checks. However, the impact of digital transformation on green transition varies across firms, industries, and regional characteristics. In terms of ownership structure, digital transformation exerts different effects on green innovation and environmental performance in state-owned and non-state-owned enterprises. At the industry level, its influence differs depending on whether a firm operates in a high-tech or heavily polluting sector. Regionally, variations in environmental regulation intensity and geographical location also lead to differential effects on green innovation and environmental performance. Second, digital transformation stimulates supply chain innovation, which in turn facilitates corporate green transition by enhancing supply chain efficiency and innovation capacity. Supply chain innovation thus serves as a mediating mechanism between digital transformation and green transition. Third, executives’ green cognition and firms’ strategic aggressiveness moderate the core relationship between digital transformation and green transition performance.

6.3. Practical Applications

In addition to expanding relevant theoretical research, the conclusions of our study offer significant practical insights. First, it is crucial to accelerate the advancement of digital technologies and digital transformation. Consequently, governments should refine the digital governance framework and promote the development of the digital economy, assisting enterprises in enhancing their digital infrastructure and bridging the digital divide. Simultaneously, enterprises should increase their investment in digital transformation, utilizing advanced technologies to optimize supply chain management and green production processes, thereby improving operational efficiency and environmental outcomes. Such integrated measures will not only drive domestic enterprises to achieve breakthroughs in green transformation and supply chain innovation but also provide a Chinese solution for the inclusive and sustainable development of the global digital economy. These insights are especially valuable for developing countries with relatively scarce resources, offering important lessons for their development.

Second, it is essential to enhance the development and deployment of digital technologies in the public environmental consciousness, promoting the widespread use of green digital tools to raise public awareness of environmental protection and provide data-driven decision support for governments and enterprises in environmental governance. By strengthening environmental awareness and stimulating innovation in green, low-carbon, and circular economies, governments and enterprises can collaboratively advance the deep application of digital technologies in areas such as environmental monitoring and ecological restoration, thereby providing robust support for global sustainable development and economic recovery.

Finally, enterprises should adopt digital tools to build a unified green-supply-chain platform that embeds environmental goals at every stage. By installing IoT sensors at supplier sites to monitor energy use and emissions, applying big data analytics to identify high-impact processes, and guiding targeted upgrades, firms can continuously refine production. Artificial intelligence can then optimize transport routes, favoring electric vehicles or consolidated shipments to cut carbon output, while blockchain traces raw materials and waste to close the loop from green production through low-carbon logistics to circular recycling. Studies show that sharing supply chain data dissolves information silos, accelerates joint environmental R&D, reduces transaction costs, and sharpens green innovation through accurate, end-to-end environmental feedback. The outcome is a synchronized green upgrade across the entire value chain that maximizes both precision and impact.

6.4. Limitations

At the empirical level, this study relies exclusively on data from Chinese A-share listed firms between 2015 and 2022. The relatively narrow temporal window, combined with the restriction to domestic listed companies, inevitably constrains the generalizability of the findings. Future inquiries may broaden the evidentiary base by incorporating cross-national and cross-market comparisons, integrating data from both advanced economies and emerging markets within a unified analytical framework to test the robustness and contextual dependence of the results.

The operationalization of digital transformation also remains contested, particularly with respect to measurement. Here, digitalization is proxied by indicators constructed through text-frequency analysis. While methodologically tractable, this approach may fail to capture the full complexity of an enterprise’s digital reality. Building on the work of Yoo and colleagues, future research could incorporate dimensions such as digital platforms, distributed innovation, and portfolio innovation, thereby offering a more nuanced and precise assessment of digital transformation within organizational contexts [74].

The research design itself would benefit from greater refinement and longitudinal depth. Digital transformation tends to advance green transition through protracted processes of technological assimilation and organizational restructuring, effects that often materialize with considerable delay. Extending the temporal horizon—tracking data over a decade or more—combined with dynamic panel models or event-study approaches could illuminate these lagged mechanisms and reveal their evolutionary trajectories. Moreover, embedding both firm-level and industry-level data within a multi-level analytical framework would enrich explanatory power. Linking micro-level factors, such as digital investment intensity and green R&D capacity, with macro-level attributes, such as industry-wide digital maturity and environmental regulatory regimes, can uncover how institutional contexts condition corporate behavior. Such cross-level perspective promises deeper insights into the interplay between structural environments and organizational practice.