The Impact of Digital–Green Synergy on Firm Innovation Resilience: Evidence from China

Abstract

Innovation is the core driving force behind high-quality development. This study uses a sample of Chinese A-share non-financial listed companies from 2011 to 2024. It empirically examines the impact of digital–green synergy on corporate innovation resilience. We find that digital–green synergy (DG) significantly enhances firm innovation resilience. The baseline regression coefficient is 0.031 (p < 0.01). This conclusion remains robust after addressing endogeneity and conducting various robustness checks. Mechanism tests show that digital–green synergy enhances innovation resilience by improving firms’ absorptive capacity, attracting capital market attention, and cultivating both resource and organizational synergy. Heterogeneity analyses reveal that the impact of this dual transformation depends on firms’ specific characteristics and their internal and external environments. This research provides micro-level evidence on the value-creation mechanisms of dual transformation synergy. The findings offer significant insights for supporting corporate innovation systems in navigating uncertainty and achieving high-quality, sustainable development.

1. Introduction

Against escalating global economic uncertainty, companies face intertwined risks. These include geopolitical frictions, restructuring of global industrial and supply chains, and frequent extreme climate events. These external disruptions have shifted from sporadic shocks to persistent, normalized challenges. Innovation remains the cornerstone of core competitive advantage. However, it requires high capital investment, elevated risk exposure, and long return cycles. This exposure makes innovation highly susceptible to interruption, stagnation, or even reversal when faced with disturbances. Therefore, the ability to sustain innovation continuity is crucial. Innovation resilience, the ability of enterprises to maintain stable innovation and quickly recover or upgrade after shocks, is essential for sustainable growth. Unlike static output indicators, innovation resilience offers a dynamic and more scientific means to assess the health and sustainability of innovation systems.

However, contemporary corporate innovation practices face a salient challenge. Large-scale investments in innovation are unlikely to consistently yield sustainable competitive advantages. At the micro-firm level, enhancing innovation resilience is constrained by a dual bottleneck. On the one hand, digital transformation enables decision-making optimization and iterative acceleration through big data, artificial intelligence, and other technologies. However, it may also trap firms in the “digital transformation paradox” [1,2]. Technology-related investments sometimes fail to achieve anticipated outcomes [2,3]. It may even reduce firms’ flexibility in coping with uncertainty due to path dependence [1,4,5]. On the other hand, green transformation aligns with the Chinese modernization concept of “harmonious coexistence between humanity and nature.” Yet, its strong positive externalities, high R&D costs, and long payback periods constrain internal motivation [6,7]. Digitalization and greening often advance in parallel in practice. They frequently fail to integrate effectively, so synergistic effects are absent. As a result, enterprises cannot systematically enhance their capacity to address complex challenges.

To address these practical problems, the Chinese government has issued policies such as the Implementation Guidelines for the Coordinated Transformation and Development of Digitalization and Greenization. These policies mark the evolution of digital–green synergy from an ideological initiative to a key implementation path for promoting high-quality development. However, academics still lack in-depth theoretical explanations of how digital and green factors work together to affect corporate innovation resilience. Key questions remain: How can digital technologies precisely empower green innovation? How does a green orientation guide digital investments to avoid aimlessness? What pathways do these factors use to enhance innovation resilience? Therefore, an in-depth investigation into how digital–green synergy influences corporate innovation resilience is essential. This not only aligns with national strategies but also has theoretical and practical value. Such research will help build an independent, controllable, and resilient national innovation system.

2. Literature Review

Corporate innovation resilience refers to an enterprise’s ability to maintain a stable innovation system and to recover. It also covers capability upgrades through dynamic adjustment in response to external shocks [8,9]. Studies define innovation resilience in terms of capability dimensions, behavioral perspectives, and dynamic processes. This includes survival, recovery, and breakthrough capabilities, and it emphasizes signal identification and collective action. Innovation resilience appears as a systematic attribute in the continuous process of “identification–response–recovery–renewal” [10,11]. Mainstream measurement methods include indicator-based assessments of innovation inputs and outputs. Other approaches quantify risk-resistance capability using survival analysis and assess dynamic adaptability with spatial econometrics, fsQCA and patent network analysis [11,12,13].

Existing studies analyze innovation resilience via external and internal factors. First, enterprise innovation resilience is shaped by external policies, financial policies, and market conditions. At the policy level, industrial and technology-finance policies boost risk resistance and innovation recovery through knowledge spillovers, credit support, and other channels [12,14,15]. At the market level, moderate competition helps build resilience against risks, but digitalization can create nonlinear risks such as “digital lock-in” [16]. Additionally, collaborative networks and industrial clusters form a composite ecosystem. These support enterprises in coping with shocks and improving innovation resilience. Second, internal factors such as resource base, dynamic capabilities, and strategic orientation affect resilience. The resource base is a static guarantee against shocks. It includes financial buffers and assets such as knowledge and data. Together, these resources enhance efficiency through policy spillovers and information integration [15,17]. Dynamic capability refers to a company’s ability to adapt and reorganize itself. This includes how it coordinates, combines, and changes. When linked to digital transformation and learning, dynamic capability helps companies respond quickly and recover from changes [12]. To leverage these strengths, firms should use resources wisely and build needed skills. They should also balance new and existing innovations based on their growth stage to stay strong [13,18].

Given the dual strategic contexts of “greenization and low-carbon development” and “building a Digital China”, digital and green transformations have become key corporate actions. These two trends jointly drive innovation shaped by broader influences. Research has summarized the main concepts and implications of digitalization, green transformation, and their synergy. Digitalization, in particular, means enterprises use technologies such as artificial intelligence and big data. By doing so, they systematically reconstruct processes, organizations, and business models. This approach improves efficiency, drives innovation, and reshapes competitive advantage [2,19,20]. Green transformation is the process by which enterprises use sustainable development principles to systematically reduce negative environmental impacts. Energy-saving technological innovation is central to this process, helping firms enhance resource efficiency through green technology research, energy conservation, emission reduction, and supporting management measures. This ultimately improves both economic and environmental performance [7,21]. The synergy between digitalization and greenization forms a dynamic, deeply integrated, closed-loop mechanism in which digitalization drives greening, and greening further drives digitalization [22,23]. This mechanism changes how companies innovate by using technology and strategic direction, offering solutions to build new productive forces. Research shows digital technology drives green innovation by making resource allocation and information more efficient [22]. The need for green innovation encourages digital investment, creating a positive cycle. This helps companies compete and perform better by speeding up innovation and improving operations [23,24]. However, most research focuses only on economic outcomes or broad trends, with little exploration of how digitalization and greenization jointly build innovation resilience.

Table 1. Literature summary analysis table.

Core Research Theme	Representative Literature	Main Research Methods	Core Findings & Research Gaps
Study on the Connotation, Measurement and Influencing Factors of Corporate Innovation Resilience	Oeij et al. (2016) [8]; Martin, R. (2019) [9]; Fey, S. (2020) [10]; Zhao, H.P. (2024) [12]; Huang X. (2025) [11] Chen, Y.L(2025) [14]	Theoretical definition & qualitative analysis, index system construction, panel regression, survival analysis, fsQCA, spatial econometrics	Core findings: Clarified the connotation of innovation resilience as a dynamic comprehensive capability and its multi-dimensional measurement paradigm. Verified the positive impact of the external environment and internal endowments on innovation resilience. Research gaps: Lack of analysis on the impact mechanism of digital–green synergy on innovation resilience. Insufficient exploration of heterogeneous effects across different contexts.
Research on Digital Transformation, Green Transformation and Their Synergy	Gomber, P. (2018) [19]; Vial, G. (2019) [20]; Tian H.F. (2023) [22]; Liang D. (2025) [23]; Li, G. (2025) [21]	Theoretical definition, text analysis, coupling coordination degree model, panel regression, mediation effect test	Core findings: Defined the connotation of mutual promotion and coupling of digital–green synergy. Confirmed its positive driving effect on corporate innovation and operational performance. Research gaps: Inadequate discussion on the transmission path of digital–green synergy affecting innovation resilience. Insufficient empirical evidence from emerging market contexts.

summarizes the above-mentioned literature.

Against this backdrop, this paper selects a sample of Chinese A-share non-financial listed firms for the period 2011 to 2024. It empirically examines the impact of digital–green synergy on corporate innovation resilience. This study also explores the underlying mechanisms and the heterogeneity of this relationship across different contexts. The potential marginal contributions are as follows. First, it enriches theoretical research on digitalization and greening. Unlike existing studies that mostly use text analysis of annual report word frequencies, event study methods [22,25], or single green patent indicators [23], this paper constructs a synergy index of digitalization and greenization at the micro-firm level. It employs a composite index method to measure digital transformation. An entropy-weighting approach helps build a multi-dimensional green development indicator system. This method increases the comprehensiveness and accuracy of measurement. Second, it reveals how digital–green synergy affects corporate innovation resilience. Mechanism tests show this synergy enhances resilience through four pathways: improving firms’ absorptive capacity, attracting capital market attention, optimizing resource synergies, and strengthening organizational synergies. These findings clarify how digital–green synergy empowers corporate innovation resilience. Third, digital–green synergy affects innovation resilience in different ways. Its positive impact varies among firms by political connections and scale, among industries by technological intensity and pollution, and across regions by digital infrastructure and public environmental awareness. These findings provide targeted policy insights.

3. Theoretical Analysis and Research Hypotheses

3.1. Digital–Green Synergy and Corporate Innovation Resilience

First, digital transformation lays the foundation for innovation resilience by improving firms’ efficiency in information processing and resource allocation. Specifically, digitalization enables more efficient internal information flow. It reduces information asymmetry and agency costs within organizations [26]. These changes enable firms to predict market changes more accurately and adjust innovation directions rapidly. The application of digital technologies also promotes optimal resource allocation. Driven by data flows, technologies, capital, and talent, sectors are concentrated in high-efficiency areas. This process helps correct resource misallocation effectively [3]. All these provide a solid basis for innovation activities. Second, green transformation enhances long-term adaptability by building a sustainable innovation system. Greening reduces operational costs and unlocks innovation resources through energy-saving technologies and process optimization. It also strengthens the sustainability of innovation by improving resource utilization efficiency, especially under resource constraints. More importantly, green transformation helps enterprises access government policy support and social recognition. This enables them to gain greater external assistance during innovation crises, thereby stabilizing innovation systems. Digitalization empowers green innovation by improving transparency and efficiency in information management. In return, the demand for sustainable development, driven by green transformation, fuels digital investment. This forms a virtuous circle of two-way promotion [27]. Such a synergistic mechanism can boost corporate innovation resilience by speeding up innovation cycles and optimizing efficiency. Based on the above analysis, this paper proposes the following hypothesis:

H1. 

The synergy between digitalization and greening significantly promotes corporate innovation resilience.

3.2. Mechanism for Digital–Green Synergy in Enhancing Corporate Innovation Resilience

This paper draws on dynamic capability theory, information asymmetry theory, resource orchestration theory, and complementarity theory. It constructs an analytical framework to explain how digital–green synergy enhances innovation resilience. The analysis focuses on four dimensions: absorptive capacity, capital market attention, organizational synergy, and resource synergy. This multi-dimensional approach covers the innovation value chain, from knowledge acquisition and capital allocation to internal collaboration and resource optimization. It provides a theoretical basis for revealing the pathways through which digitalization and greening together drive resilient enterprise development.

First, the dynamic capability theory highlights a firm’s ability to integrate, build, and reconfigure resources. These skills help firms adapt to rapid changes and gain sustainable advantages (Teece et al., 1997). Firms must respond dynamically to uncertainty and renew their capabilities. This involves learning, integrating, and reconfiguring their resource base. Absorptive capacity is a key capability. It enables firms to identify, assimilate, and apply external knowledge for innovation [28]. Absorptive capacity involves acquiring, assimilating, transforming, and using external knowledge [29]. Strong absorptive capacity helps enterprises integrate diversified resources through knowledge restructuring. It allows firms to respond rapidly to technological and market fluctuations. It also enhances stability in coping with external shocks by optimizing knowledge stock and distributed knowledge reserves [30]. Specifically, digital transformation deepens an enterprise’s connection with the external innovation ecosystem through digital tools. This greatly expands the breadth and depth of knowledge acquisition. It also enhances the efficiency of knowledge identification and assimilation [31]. Meanwhile, green transformation compels enterprises to upgrade their green human capital. Enterprises achieve this through environmental standards and sustainable development goals. As a result, they improve their ability to convert and apply knowledge in new areas, such as environmental protection [32]. The synergy of digitalization and greenization can improve how enterprises absorb and transform knowledge. This is done by boosting both potential and realized absorptive capacities [29]. Enterprises can then perceive innovation opportunities in uncertainty and maintain or restore innovation momentum. This leads to greater innovation resilience. Based on this analysis, this paper proposes Hypothesis 2:

H2. 

The synergy between digitalization and greenization can enhance enterprises’ innovation resilience by improving their absorptive capacity.

Second, from the perspective of capital market attention, this mainly reflects external governance forces such as institutional ownership, analyst coverage, and media attention. Based on information asymmetry theory (Akerlof, 1970), capital market attention alleviates information asymmetry between investors and firms. It does so mainly through the “information effect” and “monitoring effect.” This process optimizes the allocation of innovation resources and strengthens external governance. Digital transformation enhances positive market expectations by improving the efficiency of information processing and disclosure [25]. Green transformation boosts investor confidence by improving environmental performance and information disclosure [33]. The synergy between digitalization and greenization optimizes the structure and quality of information disclosure by enhancing transparency and reducing information screening costs [34]. This increased transparency attracts more attention from analysts and institutions, which in turn alleviates financing constraints and strengthens support for innovation [35]. As a result, the combined effect of digitalization and greenization enhances innovation resilience by increasing capital market attention, improving the financing environment and corporate governance, and supporting firms in maintaining or restoring innovation activities under shocks. Based on this analysis, this paper proposes Hypothesis 3:

H3. 

The synergy between digitalization and greenization enhances enterprises’ innovation resilience by increasing capital market attention.

Third, according to resource orchestration theory, organizational synergy is a dynamic process. Managers structure, bundle, and leverage resources systematically [36]. The core of synergy is to create value and gain a competitive advantage by integrating resources and optimizing processes. Digital transformation reduces cooperation costs. It blurs organizational boundaries through collaboration platforms and digital technologies. This enables more efficient integration of internal and external resources [24]. Green transformation drives change at the strategic level. It compels enterprises to establish mechanisms for strategic alignment. It also strengthens cross-value-chain collaboration [37]. The synergy between digitalization and greenization orchestrates key resources, including data, technology, and human resources. It improves operational and management efficiency by optimizing production and supply chain management [23]. This mechanism helps enterprises integrate scattered innovation resources into an organic whole. It also improves the allocation and response speed of innovation resources. By optimizing collaboration and breaking down silos, this process enhances the ability to handle uncertainty. Based on the above analysis, this paper proposes Hypothesis 4:

H4. 

The synergy between digitalization and greenization can improve enterprises’ innovation resilience by enhancing their organizational synergy capability.

Fourth, complementarity theory suggests that combining digitalization and greenization creates greater value than their individual contributions. This produces a significant synergistic effect (Milgrom and Roberts, 1990). Such complementarity is especially evident at the level of resource synergy. Digital technologies offer tools for precise monitoring and optimizing processes to meet green goals. Green transformation brings clear scenarios and a sustainable value orientation for applying these tools. The “technological symbiosis effect” from integration improves resource synergy within the firm. Digitalization empowers technology by enabling dynamic monitoring and optimization of critical resources, such as energy. This happens through information-processing and connectivity capabilities. The upgrade in human capital structure driven by digitalization further supports broader resource synergy [38]. In organizational restructuring, digital platforms break internal and external information barriers. They promote the visualization and efficient matching of resources. Green standards prompt enterprises to form cross-functional teams and optimize resource-allocation structures [23]. Integrating digital tools with green objectives helps enterprises identify resource gaps and mobilize resources efficiently. This approach maintains a sustainable, flexible resource supply in dynamic, complex environments. It provides a strong foundation for innovation and enhances firms’ innovation resilience in uncertainty. Based on this analysis, this paper proposes Hypothesis 5:

H5. 

The synergy between digitalization and greenization can improve enterprises’ innovation resilience by enhancing their resource synergy capability.

Figure 1 illustrates the influence mechanism of digital-green synergy on enterprises’ innovation resilience.

4. Research Design

4.1. Sample Selection and Data Sources

This paper uses China’s A-share non-financial listed firms from 2011 to 2024 as the sample. Data on digitalization, greenization, and financial characteristics are from CSMAR. Data on innovation resilience, media coverage, and other variables are from CNRDS. To ensure valid data and mitigate the impact of outliers, several steps were taken. First, PT or ST-labeled firms, those listed for under one year, delisted or suspended firms, and companies with abnormal data, such as an asset–liability ratio greater than 1, were excluded. Second, continuous variables were shrunk with parameters set at 1% and 99%. Third, samples with missing values in the main variables were excluded. After screening, 37,605 firm-year observations were obtained.

It should be noted that, although this study covers the period from 2011 to 2024, the sample size for 2024 is relatively small. This is mainly because some companies’ relevant data have not been fully updated, particularly when key indicators have not been released. As a result, the completeness of the 2024 data may be somewhat lacking. We have taken this into account in the analysis and made adjustments to the data processing.

4.2. Variable Description

4.2.1. Dependent Variable

LnIR is the dependent variable in this study, measured by the level of corporate innovation resilience. Unlike the comprehensive index method, the sensitivity index method uses fewer indicators. This approach reduces the risk of endogeneity caused by multicollinearity and limits differences from varying indicator selections or methods. Following Huang X.H. et al. (2025) [11] and Wang L. et al. (2024) [13], this study measures innovation capability by the number of patent grants awarded to a firm. Innovation resilience is shown by a firm’s ability to recover from shocks, measured as the relative change in patents granted. The calculation formula is as follows:

$${\mathit{IR}}_{\mathit{it}} = (\Delta Y_{\mathit{it}} - \Delta E_{\mathit{it}})/|\Delta E_{\mathit{it}}|$$

(1)

$$\Delta Y_{\mathit{it}}=(Y_{\mathit{it}}-Y_{\mathit{it}-1})$$

(2)

$$\Delta E_{\mathit{it}}=\left[\right(Y_{\mathit{rt}}-Y_{\mathit{rt}-1})/Y_{\mathit{rt}-1}]\ast Y_{\mathit{it}-1}$$

(3)

$$\mathit{Ln}{\mathit{IR}}_{\mathit{it}}=Ln ({\mathit{IR}}_{\mathit{it}}+100)$$

(4)

Among them, $Y_{\mathit{it}}$ and $Y_{\mathit{it}-1}$ represent the number of patents granted to firm i in year t and t − 1, respectively. $\Delta Y_{\mathit{it}}$ denotes the change in the number of patents granted for firm i. $Y_{\mathit{it}}$ and $Y_{\mathit{it}-1}$ represent the number of patents granted in region r where firm i is located in year t and t − 1, respectively. $\Delta E_{\mathit{it}}$ refers to the change in patents granted in region r where firm i is located. ${\mathit{IR}}_{\mathit{it}}$ measures the rebound degree of firm i’s innovation activities after being impacted by external environmental changes in year t. A larger value indicates stronger innovation resilience of the firm. To facilitate analysis, the natural logarithm is taken after adding 100 to the innovation resilience indicator in this paper. In the subsequent robustness tests, innovation resilience is recalculated using the number of patent applications to measure firms’ innovation capability.

4.2.2. Explanatory Variable

DG measures the coupling between enterprise digitalization and greenization. Following Wang S.H. & Guo Y.D. (2025) [39], this paper constructs separate indicator systems for digital transformation and green development, then calculates their coupling coordination degree as a measure of digital–green synergy:

$$C = {\displaystyle \frac{2\sqrt{U_1{U}_2}}{U_1 + U_2}}$$

(5)

$$\mathit{DG}=\sqrt{C \times T}, T={\mathsf{\alpha }}_1{U}_1+{\mathsf{\alpha }}_2{U}_2$$

(6)

In the above formulas, U1 and U2 represent the scores for enterprise digital transformation and green development, respectively. C denotes the coupling degree of the two subsystems, and T represents the comprehensive coordination index of the two subsystems, reflecting the overall synergistic effect between enterprise digitalization and greenization. This paper posits that both subsystems are equally crucial to the entire system’s operation and thus sets α1 = α2 = 0.5. DG refers to the degree of coupling between enterprise digitalization and greenization. Next, this paper constructs separate indicator systems for enterprise digital transformation and green development.

(1)

Construction of the Digital Transformation Indicator System

Following the method of Wang H. et al. (2025) [40], the degree of enterprise digitalization is measured using the enterprise digital transformation index from the CSMAR Enterprise Digital Transformation Database. The index consists of six dimensions: strategic leadership, technology-driven, organizational empowerment, environmental support, digital achievements, and digital application.

(2)

Construction of the Green Transformation Indicator System

Corporate green development is the coordinated advancement of economic profits, social value, and environmental benefits through green practices such as cost reduction, quality and efficiency improvement, increased social value, conservation of energy and resources, clean production, energy conservation, and emission reduction [41]. Based on the essence of green development and following the method of Li H.Y. et al. (2025) [42], this paper constructs a three-dimensional indicator system: corporate green system, social value, and green benefits. The entropy method determines the weight of each indicator, and a comprehensive green development score is calculated by linear weighting.

Table 2. Construction of the comprehensive index system for corporate green development.

Dimension Layer	Indicator Layer	Indicator Description	Indicator Attribute
Greenization System	Green Philosophy (X1)	The company discloses the company’s environmental protection philosophy, environmental policies, environmental management structure, circular economy development model, and green development: 1; otherwise: 0.	+
	Green Target (X2)	The company discloses the completion status of the company’s past environmental protection targets and future environmental protection targets: 1; otherwise: 0.	+
	Green Management System (X3)	If a company simultaneously implements the “Three Simultaneities” system, establishes an environmental management system, and sets up an environmental emergency response mechanism, it is assigned a value of 3. If any two of these are implemented, the value is 2. If one of these is implemented, the value is 1. Otherwise, the value is 0.	+
	Sustainability Information Disclosure (X4)	If relevant information is disclosed in the annual report, social responsibility report, and environmental report, the value is 3. If any two of these reports disclose the information, the value is 2. If only one report discloses the information, the value is 1. Otherwise, the value is 0.	+
Social Value	Basic EPS (X5)	(Net Profit for the Period − Preferred Stock Dividends)/Annual Weighted Average Total Share Capital	+
	Total Tax Paid (X6)	Total taxes and fees actually paid by the enterprise/Operating revenue	+
	Number of Employees (X7)	Total number of employees at the end of the year	+
	Total Employee Compensation (X8)	Cash paid to and on behalf of employees	+
Green Benefits	Waste Gas Emission Reduction and Treatment (X9)	0 = No description; 1 = Qualitative description; 2 = Quantitative description (monetary/numerical description)	+
	Wastewater Emission Reduction and Treatment (X10)	0 = No description; 1 = Qualitative description; 2 = Quantitative description (monetary/numerical description)	+
	Dust and Soot Treatment (X11)	0 = No description; 1 = Qualitative description; 2 = Quantitative description (monetary/numerical description)	+
	Environmental Rewards (X12)	The company discloses the formulation of environmental management systems, responsibilities, regulations and other management systems: 1; otherwise: 0.	+
	Environmental Compliance (X13)	0 if environmental violations, emergencies, and petition cases all occur; 1 if two of the three occur; 2 if only one occurs; 3 if none occur.	+
	ISO Certification (X14)	The company has passed ISO 14001 [43] certification: 1, otherwise: 0.	+

the symbol “+” indicates that the indicator has a positive effect on the level of green development.

provides detailed indicator descriptions.

The specific measurement method is as follows:

$$\mathit{GD}=\mathit{\sum }_{j=1}^p{w}_{\mathit{ij}}{U}_{\mathit{ij}} ,\mathit{\sum }_{j=1}^p{w}_{\mathit{ij}}=1$$

(7)

GD represents the comprehensive green development index for enterprises, p denotes the number of indicators in system i, and $w_{\mathit{ij}}$ denotes the weight of the j-th indicator in system i calculated using the entropy method. $U_{\mathit{ij}}$ is the value of the j-th indicator in system i. To avoid dimensionality issues, each indicator is standardized first. This paper employs normalization methods to linearly transform raw data, mapping results to the range [0, 1]. The standardization formulas for positive indicators and negative indicators are shown as follows:

$$U_{ij}={\displaystyle \frac{X_{ij}-min\left[X_j\right]}{\mathit{max}\left[X_j\right]-min\left[X_j\right]}}\ast 0.99+0.01$$

(8)

$$U_{ij}={\displaystyle \frac{{\mathit{max}\left[X_j\right]-X}_{ij}}{\mathit{max}\left[X_j\right]-min\left[X_j\right]}}\ast 0.99+0.01$$

(9)

Equation (8) is the standardization for positive indicators, and Equation (9) is the standardization for negative indicators. Where $X_{ij}$ is the original value of the j-th indicator for the i-th sample, min$\left[X_j\right]$ is the minimum value of the original data for the j-th indicator, and $\mathit{max}\left[X_j\right]$ is the maximum value of the original data for the j-th indicator. In Equation (1), $w_{ij}$ represents the indicator weight obtained by the entropy method. The specific calculation steps are as follows:

(1)

Calculate the information entropy of the indicator:

$$P_{\mathit{ij}}={\displaystyle \frac{U_{\mathit{ij}}}{{\sum }_{i=1}^n{U}_{\mathit{ij}}}}$$

(10)

(2)

Calculate the effect value of the J-th indicator:

$$E_j=-K\sum _{i=1}^m{P}_{\mathit{ij}}\mathit{ln}{P}_{\mathit{ij}}$$

(11)

(3)

Calculate indicator weights:

$${ W}_j={\displaystyle \frac{1-E_j}{{\sum }_{j=1}^p\left(1-E_j\right)}}$$

(12)

Figure 2 shows the schematic diagram of the DG calculation process.

4.2.3. Control Variables

To control for the influence of other factors on the study, this paper references the research of Liu B. and Liu Z. (2025) [44], focusing primarily on variables at the corporate financial and corporate governance levels. Specifically, at the corporate financial level, this study selected the following variables: Age (natural logarithm of listing duration), Fix (fixed assets/total assets), Liq (current assets/current liabilities), Lev (total liabilities/total assets), SG (sales growth rate), and ROE (net profit/average net assets). At the corporate governance level, this study selected the proportion of independent directors (IBDR, number of independent directors/total board members), equity concentration (top 1, shareholding ratio of the largest shareholder), dual role (dual, value 1 if the chairman and general manager roles are combined, otherwise 0), and equity balance (EB, sum of shareholding ratios of the 2nd to 10th largest shareholders/shareholding ratio of the largest shareholder).

4.2.4. Mechanism Variables

Based on the preceding theoretical analysis, the mechanism variables in this study are absorptive capacity, capital market attention, organizational synergy, and resource synergy. Specific measurements are as follows: ① Absorptive Capacity: Characterized by the proportion of employees with a bachelor’s degree or higher (HC). ② Capital Market Attention: Assessed jointly using institutional investor shareholding ratio (IH) and media coverage (Median). ③ Organizational Synergy: Measured by total asset turnover (ARO, operating revenue/average total assets) and operating cost ratio (OC, the ratio of the sum of administrative expenses and selling expenses to operating revenue). ④ Resource Synergy: Measured by enterprise investment efficiency and labor allocation efficiency.

4.3. Model Setting

To examine the impact of DG on LnIR, we construct the econometric model:

$${\mathit{LnIR}}_{\mathit{it}} = {\mathsf{\alpha }}_0 + {\mathsf{\alpha }}_1{\mathit{DG}}_{\mathit{it}} + c{X}_{\mathit{it}} + \sum \mathit{Year} + \sum \mathit{Ind} + {\mathsf{\epsilon }}_{\mathit{it}}$$

(13)

${\mathit{LnIR}}_{\mathit{it}}$ represents the innovation resilience level of firm i in year t, DG denotes the synergy level of digitalization and greenization, and $X_{it}$ stands for the control variables at the financial and corporate governance levels mentioned above. $\sum \mathit{Year}$ and $\sum \mathit{Ind}$ represent the year dummy variable and industry dummy variable, respectively, which are used to control for the impacts of year and industry factors, while ${\epsilon }_{it}$ is the random disturbance term. This study focuses on the coefficient ${\alpha }_1$: a positive value of ${\alpha }_1$ indicates that the synergy between digitalization and greenization can improve firms’ innovation resilience.

5. Empirical Analysis

5.1. Evolution Analysis of Digital–Green Synergy

This article first conducts a trend analysis of the dual transformation levels of the sample enterprises. As shown in

Table 3. Annual trend analysis of enterprise digital–green synergy.

Year	Digitalization	Greenization	Coupling Degree	Coupling Coordination Degree	Coordination Type
2011	0.318	0.119	0.663	0.375	Mild imbalance
2012	0.323	0.142	0.713	0.401	approaching imbalance
2013	0.341	0.146	0.722	0.413	approaching imbalance
2014	0.336	0.142	0.714	0.406	approaching imbalance
2015	0.367	0.145	0.710	0.420	approaching imbalance
2016	0.354	0.153	0.735	0.424	approaching imbalance
2017	0.364	0.163	0.760	0.440	approaching imbalance
2018	0.373	0.189	0.790	0.464	approaching imbalance
2019	0.379	0.202	0.801	0.476	approaching imbalance
2020	0.389	0.181	0.790	0.468	approaching imbalance
2021	0.383	0.238	0.848	0.507	Barely coordinated
2022	0.381	0.260	0.866	0.520	Barely coordinated
2023	0.386	0.312	0.889	0.549	Barely coordinated
2024	0.427	0.475	0.926	0.635	Primary coordination

, from 2011 to 2024, both digital and green transformation levels have steadily improved. Specifically, the digital transformation level increased from 0.318 to 0.427, and the green transformation level rose from 0.119 to 0.475, with significant growth. Since 2024, the development of both has begun to align more closely. The level of dual transformation collaboration has increased year by year. The coupling coordination degree rose from 0.375 (slight imbalance) in 2011 to 0.635 (primary coordination) in 2024. Notably, after 2021, the collaboration evolved from “barely coordinated” to the “primary coordination” stage. National strategies related to digitalization and greenization were proposed early on. However, enterprises initially focused on single-point development, with weak collaboration awareness. Since the concept of “dual transformation collaboration” was clarified in 2021, the collaboration level has improved. It is still constrained by factors such as insufficient institutional supply and lack of practical experience. There remains significant room for future improvement.

To assess the relationship between digital–green synergy (DG) and corporate innovation resilience (LnIR) from a trend perspective, this study plots a trend scatterplot in Figure 3. As shown in the figure, both variables exhibit an overall upward trend during the sample period. The movement of corporate innovation resilience and digital–green synergy is in the same direction, providing preliminary evidence of a positive correlation between the two.

5.2. Descriptive Statistics

Table 4. Descriptive statistics of variables.

Variable Type	Variable Symbol	Variable Declaration	Average	Standard Deviation	Minimum	Maximum	VIF
Explained Variable	LnIR	Innovation Resilience	4.611	0.111	4.078	5.182	--
explanatory variable	DG	Digital–Green Synergy	0.475	0.139	0.246	0.77	1.09
Control Variables	Age	Firm Age	2.201	0.779	0.693	3.401	1.35
	Fix	Fixed Asset Ratio	0.207	0.155	0.002	0.68	1.14
	Liq	Liquidity Ratio	2.47	2.502	0.33	17.766	1.86
	Lev	Leverage Ratio	0.423	0.202	0.051	0.899	1.95
	SG	Growth	0.331	0.87	0.701	5.942	1.06
	ROE	Return on Equity	0.044	0.154	0.908	0.311	1.12
	IBDR	Independent Director Ratio	0.377	0.054	0.333	0.571	1.02
	Top1	Ownership Concentration	0.336	0.148	0.084	0.742	2.29
	Dual	Duality	0.287	0.453	0	1	1.09
	EB	Equity Balancing	0.955	0.781	0.054	3.885	2.27

reports the descriptive statistics of the main variables. The mean value of the explanatory variable LnIR is 4.611, which aligns closely with existing studies [11]. The range suggests notable differences in innovation resilience among sample firms. The mean value of the core explanatory variable DG is only 0.475. This shows that, as a whole, sample enterprises remain in a state of “approaching imbalance.” Dual synergy needs improvement, and the degree of synergy varies across firms. All control variables fall within a reasonable range. For the fixed asset ratio (Fix), some firms show an asset-light trend. Duality (Dual) appears in 28.7% of sample firms. The variance inflation factor (VIF) for all variables is under 10. This indicates that no severe multicollinearity issues exist in the model.

5.3. Benchmark Regression Results Analysis

Table 5. Benchmark regression results.

Variable	(1)	(2)	(3)
	LnIR	LnIR	LnIR
DG	0.035 ***	0.031 ***	0.029 ***
	(6.59)	(5.92)	(5.57)
Age		0.000	0.001
		(0.10)	(1.40)
Fix		0.001	0.001
		(0.26)	(0.17)
Liq		0.000 **	0.000 **
		(2.26)	(2.12)
Lev		0.013 ***	0.013 ***
		(3.88)	(3.77)
SG		0.001	0.001
		(1.06)	(1.06)
ROE		0.011 ***	0.009 ***
		(3.41)	(2.80)
IBDR			0.016
			(1.32)
Top1			0.023 ***
			(3.89)
Dual			−0.001
			(−0.45)
EB			0.003 ***
			(3.20)
_cons	4.594 ***	4.589 ***	4.571 ***
	(2013.72)	(1408.83)	(638.91)
Ind FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
N	37,604	37,604	37,604
R2	0.004	0.005	0.005

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01. DG: Measures the coupling coordination degree between enterprise digitalization and greening by constructing indicator systems for each and evaluating the coordination of the two subsystems. LnIR: Represents the natural logarithm of enterprise innovation resilience, measured by the number of patent grants and the rate of change, which reflects recovery capacity after external shocks.

reports the baseline regression results. Column (1) shows the results of the benchmark regression excluding control variables; the DG coefficient was 0.035, significant at the 1% level. Column (2) incorporates the effects of corporate financial control variables; the DG coefficient decreased from 0.035 (p < 0.01) to 0.031 (p < 0.01), indicating that omitting relevant variables may lead to overestimation of the results. After further adding corporate governance variables, the results in Column (3) show that the coefficient of DG remains significantly positive. Thus, Hypothesis 1 was validated. In addition, some control variables, such as liquidity (Liq), profitability (ROE), ownership concentration (top 1), and internal control quality (EB), also exert significant positive effects, consistent with theoretical expectations.

5.4. Endogeneity Tests

5.4.1. Control for Fixed Effects

The baseline model only considers factors that do not vary across industries. However, the macroeconomic policies of the region where the firm is located may also influence the study. To avoid overlooking these unobservable factors, this paper further controls for provincial fixed effects (Prov FE). It also applies high-dimensional fixed effects, including industry-year and province-year fixed effects, in the baseline regression. As shown in Column (1) of

Table 6. Fixed effects regression results.

Variable	(1)	(2)	(3)	(4)
	LnIR	LnIR	LnIR	LnIR
DG	0.027 ***	0.026 ***	0.027 ***	0.026 ***
	(5.30)	(5.05)	(5.16)	(4.92)
Controls	Yes	Yes	Yes	Yes
Ind FE	Yes	Yes	Yes	Yes
Prov FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Ind $\times$ Year FE	No	Yes	No	Yes
Prov $\times$ Year FE	No	No	Yes	Yes
N	37,604	37,537	37,590	37,523
R2	0.005	0.001	0.005	0

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

, after controlling for industry, province, and year fixed effects simultaneously, the coefficient of DG remains significantly positive at the 1% level. Columns (2)–(4) respectively incorporate industry-year, province-year, and their joint interaction fixed effects. The DG coefficient stays stable and highly significant throughout. These results indicate that the core conclusion holds after controlling for potential unobservable factors at the regional, industry, and time-interaction levels. This further supports the robustness of the fact that digital–green synergy significantly promotes firm innovation resilience.

5.4.2. Dynamic Panel Model

To account for the dynamic continuity of innovation resilience and reduce endogeneity from omitting lagged terms, this study builds a dynamic panel model. It uses the system generalized method of moments (SYS-GMM) for estimation. Column (1) of Table 6 shows that the coefficient of the one-period lag of innovation resilience (L.LnIR) is significant. This supports the need for a dynamic model. The autocorrelation tests show first-order autocorrelation but no second-order autocorrelation in the disturbance term (AR(1) p = 0.021, AR(2) p = 0.131). The Hansen J test has a p-value of 0.247, indicating that the instrumental variables and the model specification are valid. After controlling for dynamic endogeneity, the DG coefficient is still significantly positive at the 5% level. This indicates that the core conclusion remains robust.

5.4.3. Instrumental Variables Method

To address endogeneity from omitted variables and reverse causality, this paper selects suitable instrumental variables for the endogenous explanatory variable and uses the two-stage least squares (2SLS) method. For the first instrument, following Liang D. et al. (2025) [23], we use the mean digital–green synergy of other firms in the same region and year (IV1). Peer effects and industry characteristics encourage firms in the same region to improve digitalization and greenization, satisfying the relevance condition. The digitalization–greenization synergy among other regional firms does not directly affect a specific firm’s innovation resilience, thereby meeting the exogeneity condition. For the second instrument, following Li Y. et al. (2025) [37], we use the mean value of DG for the same industry and year (IV2). This indicator reflects the industry’s overall level of digital–green integration. Firms in such industries tend to pursue digital and green transformation, which fulfills the relevance assumption. An industry’s digital–green synergy level is seldom influenced by a single firm, thereby satisfying the exogeneity assumption.

Table 7. SYS-GMM and IV-2SLS estimation results.

Variable	(1)	(2)	(3)
	SYS-GMM	IV-2SLS
	LnIR	LnIR	LnIR
L.LnIR	−0.446 **
	(−2.24)
DG	0.138 **	0.135 **	0.089 ***
	(2.06)	(2.01)	(2.84)
Controls	Yes	Yes	Yes
AR(1)	0.037
AR(2)	0.101
Hansen J	0.247
Underidentification test		190.539 ***	557.262 ***
Weak identification test		197.692	835.481
Firm FE	Yes
Ind FE	No	Yes	Yes
Year FE	Yes	Yes	Yes
N	32,130	37,590	37,604
R2		−0.013	−0.004

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

reports the instrumental variable tests. The regression results show that the instrumental variables selected in this paper pass both the underidentification and weak instrument tests. Meanwhile, the DG coefficients are all significantly positive, indicating that the baseline results hold after addressing endogeneity with the instrumental variable approach.

5.4.4. Propensity Score Matching

To reduce sample selection bias, this study uses the industry-level annual mean of digital–green synergy (DG) from the relevant literature as the grouping criterion. Samples above the mean are in the treatment group; those below are in the control group. We use control variables from the baseline model as covariates. Then, we perform propensity score matching (PSM) using the 1:1 nearest-neighbor method with a caliper radius of 0.01. Figure 4 shows that after matching, standardized biases for all covariates decrease significantly and approach zero. This result indicates that matching effectively reduces systematic differences between groups.

After passing the balance test, the baseline model is re-estimated using the PSM-matched sample. The results in

Table 8. Propensity score matching results.

Variable	(1)	(2)
	1:1 Nearest-Neighbor Matching with Replacement	1:1 Nearest-Neighbor Matching Without Replacement
DG	0.021 ***	0.022 ***
	(2.87)	(3.73)
Controls	Yes	Yes
Ind FE	Yes	Yes
Year FE	Yes	Yes
N	19,663	26,990
R2	0.002	0.002

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

show that the DG coefficient remains significantly positive at the 1% level across all matching methods, indicating that the core conclusion remains robust after controlling for sample selection bias.

5.5. Robustness Test

5.5.1. Replacement of Core Variables

First, we replace the measurement method of the dependent variable. Following Wang L. et al. (2024) [45], we remeasure it using industry-adjusted innovation resilience (IRm) and patent-based innovation resilience (IRA). In addition, to reduce sensitivity to the weighting method, we use an equal-weight approach to calculate the green development level and then recalculate the coupling coordination degree (DG1) for digital–green synergy. The results in Column 3 of

Table 9. Robustness tests by replacing core variables.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	IRm	IRA	LnIR	LnIR	IRm	IRA
DG	4.720 ***	0.044 ***
	(6.60)	(6.62)
DG1			0.072 ***
			(4.99)
Digt				−0.004	1.294	−0.010
				(−0.48)	(1.09)	(−0.89)
GD				−0.018	0.813	−0.021
				(−1.33)	(0.60)	(−1.46)
Digt × GD				0.110 ***	7.886 **	0.155 ***
				(2.90)	(2.32)	(3.69)
Controls	Yes	Yes		Yes	Yes	Yes
Ind FE	Yes	Yes		Yes	Yes	Yes
Year FE	Yes	Yes		Yes	Yes	Yes
N	37,604	37,034		37,604	37,604	37,034
R2	0.044	0.010		0.006	0.045	0.011

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

show that DG1’s coefficient remains significantly positive. Second, we replace the measurement method for the core explanatory variable. Coupling coordination (DG) is replaced by the interaction of digitalization and greenization (Digt × GD) for re-estimation. Finally, we change the measurements of both dependent and independent variables at once. The results in Table 9 show that, regardless of measurement method, the core variable coefficients (DG or Digt × GD) remain significantly positive at the 5% level or higher. This indicates that the baseline findings are robust and do not depend on the specific quantification of the core variables.

5.5.2. Exclusion of Exogenous Shocks

To stimulate the economy, the government may introduce loose fiscal policies. These policies often support innovative enterprises and enhance firm innovation resilience. To remove pandemic interference, we exclude samples from 2020 to 2022. As shown in the first column of

Table 10. Robustness tests excluding exogenous shocks.

Variable	(1)	(2)	(3)	(4)	(5)
	LnIR	LnIR	LnIR	LnIR	LnIR
DG	0.024 ***	0.026 ***	0.024 ***	0.029 ***	0.023 ***
	(4.45)	(3.83)	(4.34)	(3.73)	(4.38)
Controls	Yes	Yes	Yes	Yes	Yes
Ind FE	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
N	26,664	24,644	30,045	37,604	33,298
R2	0.005	0.003	0.002	0.005	0.01

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

, the DG coefficient remains significantly positive. This indicates that the baseline conclusion remains unchanged after the pandemic effect is excluded.

To ensure the reliability of the conclusions, this paper conducts a series of robustness tests by adjusting the sample scope and estimation methods (Table 10). First, to mitigate the impact of missing innovation data in non-manufacturing firms, we re-run the regression using only manufacturing firm samples. The core conclusions still hold (Column 2). Meanwhile, we control for the particularities of municipalities, such as resource endowment and development level, by excluding firms in four major municipalities, like Beijing and Shanghai. The coefficient on DG remains significantly positive at the 1% level (Column 3). In addition, we re-estimate the model by adjusting clustered standard errors from the firm to the industry level. The results remain robust (Column 4). Finally, to ensure stable, continuous observations, we keep only firms that have existed for at least 5 years. The core conclusions remain unchanged (Column 5).

5.6. Mechanism Testing

This study investigates whether digital–green synergy strengthens corporate innovation resilience. It may do so by enhancing absorptive capacity, increasing capital market attention, and producing organizational and resource synergies. To prevent multicollinearity from distorting results, we use the two-step mechanism test from Jiang (2022) [46]:

First, we build a model to examine how the main variable influences the mechanism variables. The model is as follows:

$${\mathit{Med}}_{\mathit{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathit{DG}}_{\mathit{it}}+c{X}_{\mathit{it}}+\sum \mathit{Year}+\sum \mathit{Ind} + {\mathsf{\epsilon }}_{\mathrm{it}}$$

(14)

Med denotes mechanism variables: firm absorptive capacity, capital market attention, organizational synergy, and resource synergy. Other variable definitions match those in the baseline regression.

Second, theoretical analysis to verify the impact of mechanism variables on the dependent variable. Drawing from the existing literature and practical economic conditions, this study analyzes how the mediating variables influence corporate innovation resilience. Detailed discussions follow in subsequent sections.

5.6.1. Mechanism Test of Absorptive Capacity

To test if digital–green synergy enhances innovation resilience by improving absorptive capacity, this study follows Li Y. et al. (2025) [37]. It uses the share of employees with a bachelor’s degree or above (HC) and the ratio of intangible assets (WXZC) to measure firms’ absorptive capacity in terms of human capital and knowledge base. Columns (2) and (4) of

Table 11. Mechanism tests: absorptive capacity and capital market attention.

Variable	(1)	(2)	(3)	(4)	(1)	(2)	(3)	(4)
	Absorptive Capacity				Capital Market Attention
	LnIR	HCS	LnIR	WXZC	LnIR	IH	LnIR	Median
DG	0.024 ***	0.024 ***	0.023 ***	0.005 **	0.023 ***	0.213 ***	0.023 ***	1.300 ***
	(4.42)	(2.87)	(4.56)	(2.54)	(4.38)	(11.86)	(4.14)	(13.44)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Ind FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	31,867	31,867	33,739	33,739	33,739	33,739	32,563	32,563
R2	0.005	0.415	0.005	0.254	0.005	0.491	0.005	0.320

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

show that, after controlling for relevant variables, the coefficients of DG are both significantly positive at the 5% level or higher. This indicates that DG boosts firms’ learning, absorption, and knowledge transformation capabilities. These findings support the mechanism inference of this paper. Digital–green synergy significantly improves firms’ absorption and internalization of external knowledge. It does so by building a cross-domain knowledge base and improving technology identification and integration. As a result, innovation resilience is enhanced. Thus, Hypothesis H2 is verified.

5.6.2. Mechanisms of Capital Market Attention

To test if digital–green synergy (DG) improves innovation resilience by attracting attention in the capital market, this paper uses the shareholding ratio of investment institutions (IH) and news media attention (Median) as measures. Table 10 shows that DG has a clear positive impact on IH and Median at the 1% level. This means that dual synergy increases market attention through sending clear signals. First, digital transformation makes it easier and faster to process and share information. This makes things more open and reduces gaps in what people know, which lowers investment uncertainty and makes the market more positive about future results. Second, green transformation strengthens corporate sustainability and investor confidence. It achieves this by improving environmental performance and disclosure. This is especially valuable for investors who prioritize social responsibility and sustainability. As a result, long-term competitiveness is enhanced. Third, DG optimizes information disclosure and reduces screening costs. It makes enterprise information easier for analysts and institutional investors to interpret. This broadens financing channels and improves the financing environment. Lastly, DG enhances the corporate image and boosts market confidence. It alleviates innovation constraints and offers external support, especially during market shocks. This support helps enterprises maintain innovation continuity and resilience. Thus, DG’s signaling effect attracts investor attention. It improves the financing environment and corporate governance and provides vital support for innovation during shocks. Ultimately, this verifies Hypothesis H3.

5.6.3. Mechanism Test of Organizational Synergy Effect

To examine whether digital–green synergy enhances innovation resilience by improving organizational synergy capacity, this study draws on Li W.L. et al. (2022) [47]. It uses total asset turnover (ATO) and operating cost ratio (OC) to measure corporate operational efficiency. The results in Column (1) of

Table 12. Mechanism tests: organizational synergy and resource synergy.

Variable	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	Organizational Synergy			Resource Synergy
	ATO	LnIR	OC	LnIR	Ineff	LnIR	Employ
DG	0.126 ***	0.023 ***	−0.067 ***	0.039 ***	−0.049 ***	0.023 ***	−0.110 ***
	(3.31)	(4.36)	(−6.53)	(4.03)	(−3.78)	(4.56)	(−5.42)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Ind FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	33,739	33,342	33,342	12,164	12,164	33,732	33,732
R2	0.354	0.005	0.375	0.004	0.031	0.005	0.123

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

show that dual synergy (DG) significantly improves total asset turnover. This indicates that DG enhances resource utilization efficiency by optimizing asset allocation and operational processes. The results in Column (3) show that DG significantly reduces the operating cost ratio. This suggests that dual coordination improves internal coordination efficiency by strictly controlling management and sales expenses. Based on resource orchestration theory, dual synergy integrates internal and external resources. It promotes technological innovation and facilitates green transformation. It also optimizes production processes and supply chain management. As a result, resource utilization efficiency and asset turnover improve. Resource orchestration theory emphasizes that effective integration of resources can reduce collaboration costs and management expenses. Dual synergy promotes cross-departmental collaboration and process optimization. It breaks down information silos and enhances organizational synergy and efficiency. This reduces operating costs and unnecessary expenditures. Resource orchestration theory emphasizes that, in the face of external shocks, enterprises can quickly reallocate resources. They coordinate actions through systematic integration and coordination capabilities. This enhances their ability to respond to innovation uncertainty. The mechanism of organizational synergy capability is empirically supported. The research Hypothesis H4 is confirmed.

5.6.4. Mechanism of Resource Synergy

Following Liang et al. (2025) [23], this study employs capital and labor allocation efficiency to measure overall resource allocation efficiency. Specifically, capital allocation efficiency is characterized by investment inefficiency (Ineff), which is calculated by the Richardson model:

$$\begin{gathered} {\mathit{Invest}}_{\mathit{it}}={\mathsf{\gamma }}_0+{\mathsf{\gamma }}_1{\mathit{Invest}}_{\mathit{it}-1}+{\mathsf{\gamma }}_2{\mathit{TQ}}_{\mathit{it}-1}+{\mathsf{\gamma }}_3{\mathit{Lev}}_{\mathit{it}-1}+{\mathsf{\gamma }}_4{\mathit{Cash}}_{\mathit{it}-1}+{\mathsf{\gamma }}_5{\mathit{Size}}_{\mathit{it}-1} \\ +{\mathsf{\gamma }}_6{\mathit{Ret}}_{\mathit{it}-1}+\sum \mathit{Ind}+\sum \mathit{Year}+{\mathsf{ϵ}}_{\mathit{it}} \end{gathered}$$

(15)

Invest represents a new corporate investment. Cash denotes net cash flow from operating activities divided by total assets; Ret denotes the annual stock return rate, accounting for reinvested cash dividends. Size is measured as the natural logarithm of total assets. Model (15) provides a reasonable investment level and residuals. A residual greater than zero indicates over-investment by the firm. This paper also uses employee deviation (Employ) to measure labor allocation efficiency. A higher deviation means lower labor allocation efficiency. First, Model (16) is adopted to estimate the normal number of employees (Employee) of the firm:

$${\mathit{Employee}}_{\mathit{it}}={\mathsf{\theta }}_0+{\mathsf{\theta }}_1{\mathit{Size}}_{\mathit{it}}+{\mathsf{\theta }}_2{\mathit{SG}}_{\mathit{it}-1}+{\mathsf{\theta }}_3{\mathit{Fixed}}_{\mathit{it}}+\sum \mathit{Ind}+\sum \mathit{Year} +{\mathsf{ϵ}}_{\mathit{it}}$$

(16)

$Employee$ is the natural logarithm of the number of employees. The fitted value from Model (16) represents the normal employment level of the firm, and the logarithm of the absolute difference between the actual value and the normal employment level is defined as the employee deviation ($Employ$).

The regression results in Columns (5) and (7) of Table 12 show that the coefficients of DG are both significantly negative at the 1% level. This indicates that digital–green synergy can effectively curb overinvestment, optimize labor allocation, and enhance resource integration within firms. Enhanced resource synergy enables firms to rapidly and accurately reallocate capital and human resources to key R&D sectors. This enables them to respond better to unexpected demands during the innovation process. As a result, firms can ensure continuity and adaptability in their innovation activities, thereby improving innovation resilience. Therefore, Hypothesis H5 is supported.

5.7. Heterogeneity Analysis

5.7.1. Firm Heterogeneity

In terms of political connections, this paper uses the political connection indicator to measure firms’ political–business relations. Following Hu et al. (2020) [48], a firm is defined as having political connections if its chairman or CEO formerly or currently serves as a deputy to the People’s Congress, a member of the Chinese People’s Political Consultative Conference, a local government official, or similar; otherwise, the firm is regarded as having no political connections. The subsample regression results in Columns (1) and (2) of

Table 13. Firm heterogeneity analysis results.

Variable	(1)	(2)	(3)	(4)
	Non-Politically Connected	Politically Connected	Large-Scale	Small-to-Medium-Sized
DG	0.019 ***	0.036 ***	0.028 ***	0.004
Controls	Yes	Yes	Yes	Yes
p-value	0.000		0.000
Ind FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
N	20,384	10,372	16,599	17,138
R2	0.005	0.003	0.008	−0.001

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

show that the regression coefficient of DG for politically connected firms (0.036) is significantly higher than that for non-politically connected firms (0.019). The empirical p-value from the Bootstrap intergroup difference test is 0.000, indicating that DG’s promoting effect on innovation resilience is stronger in firms with political connections. This may be because such firms can obtain more innovation resources and policy support through political networks, effectively realizing the improvement effect of digital–green synergy on innovation resilience.

Regarding firm size, this paper divides the sample into large and small-to-medium firms based on the industry-year median total asset level. The grouped regression results (Columns (3) and (4) of Table 13) show that the DG coefficient is 0.028 and significant at the 1% level for the large-firm sample, while the coefficient is insignificant for the small and medium-sized firm sample. The empirical p-value of the Bootstrap test for between-group differences is 0.000, further confirming that the effect differs significantly across firms of different sizes. This indicates that the improvement effect of digital–green synergy on innovation resilience is mainly concentrated in large firms. A likely reason is that large enterprises typically have more stable financing channels, stronger risk tolerance, and richer resources for innovation through trial and error. These factors enable them to better maintain the stability and resilience of the innovation system in promoting digital–green synergy, thereby achieving a more substantial enhancement of innovation resilience.

Politically connected enterprises access government resources, policy support, and information more easily. This helps them leverage digital–green synergy to boost innovation resilience. Large-scale enterprises integrate resources better and tolerate risk more effectively. These strengths help them face greater innovation risks and implement digital–green synergy, enhancing innovation resilience.

5.7.2. Industry Specific Heterogeneity

This paper uses the Catalogue of Strategic Emerging Industries, the 2012 Strategic Emerging Industry Classification (for Trial Implementation), and relevant documents of the Organization for Economic Co-operation and Development (OECD) to identify the industry codes of high-tech listed firms. A firm is classified as high-tech if it belongs to a high-tech sector; otherwise, it is traditional. We use these classifications to form subsamples for regression analysis. In

Table 14. Industry-specific heterogeneity analysis results.

Variable	(1)	(2)	(3)	(4)
	High-Tech Industry	Traditional Industry	Heavy Pollution	Non-Heavy Pollution
DG	0.025 ***	0.019 ***	0.008	0.028 ***
Controls	Yes	Yes	Yes	Yes
p-value	0.017		0.000
Ind FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
N	20,264	13,475	10,249	22,334
R2	0.004	0.011	0.004	0.006

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

, Columns (1) and (2), the DG coefficient in high-tech industries is 0.025 (significant at 1%), higher than 0.019 in traditional industries. The Bootstrap between-group difference test gives an empirical p-value of 0.017, showing that digital–green synergy more strongly promotes innovation resilience in high-tech industries. High-tech firms likely achieve this through stronger knowledge absorption and transformation capabilities, which enable them to respond quickly and accurately to market changes, understand customer demands, and improve products. As a result, they enhance their responsiveness and ability to make technological breakthroughs.

Regarding environmental sensitivity, according to the Guidelines for the Industry Classification of Listed Companies (Revised 2012) issued by the China Securities Regulatory Commission and the Guidelines for Environmental Information Disclosure of Listed Companies released by the Ministry of Ecology and Environment, and following the method of Zhou et al. (2021) [49], this paper classifies 19 industries as heavily polluting industries, including B06, B07, B08, B09, B10, C15, C17, C18, C19, C22, C25, C26, C27, C28, C29, C30, C31, C32, and D44. All other industries are defined as non-heavily polluting industries. The subsample regression results in Table 14 show that the coefficient of DG is 0.028 and significant at the 1% level in the non-heavily polluting industries subsample, while it is insignificant in the heavily polluting industries subsample. The empirical p-value of the between-group difference test is 0.000, further confirming a significant difference between the two groups. This indicates that the enhancing effect of digital–green synergy on innovation resilience is more pronounced in non-heavily polluting industries. A possible reason is that heavily polluting industries are often subject to stricter environmental regulations and higher pollution control costs, which crowd out resources for digital integration and coordinated transformation, thereby weakening the enabling effect of digital–green synergy on the innovation system.

5.7.3. Regional Heterogeneity

Regarding digital infrastructure, this paper classifies regions ranked in the top 10 of the 2024 China Information Infrastructure Competitiveness Index, released by the Internet Industry Research Institute of Tsinghua University, as high digital infrastructure regions. The regions listed below are classified as low digital infrastructure. Subsamples are used for regression analysis, as shown in

Table 15. Regional heterogeneity tests results.

Variable	(1)	(2)	(3)	(4)
	High Digital Infrastructure	Low Digital Infrastructure	High-Concern	Low-Concern
DG	0.032 ***	0.016 **	0.033 ***	0.023 ***
Controls	Yes	Yes	Yes	Yes
p-value	0.030		0.020
Ind FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
N	26,132	11,472	19,291	18,309
R2	0.006	0.000	0.007	0.001

t-statistics are reported in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

. The regression results (Columns (1) and (2)) show a higher coefficient of DG in high digital infrastructure regions (0.032) compared to low digital infrastructure regions (0.016). The empirical p-value of the Bootstrap between-group difference test is 0.030. This indicates that the digital–green synergy promotes innovation resilience more strongly in regions with well-developed digital infrastructure. Sound digital infrastructure primarily provides critical support for technology integration and application. This accelerates the synergy between digitalization and greenization of enterprises.

Following Tao et al. (2023) [50], this paper sums the daily average search volume for “environmental pollution” and “smog” at the prefecture-level city level from 2011 to 2024, adds 1, and takes the natural logarithm. Samples are split into high- and low-concern groups based on the median. Regression results in Table 15 show that the coefficient of DG in regions with high public concern is 0.033, higher than 0.023 in low-concern regions, with a between-group empirical p-value of 0.020. Thus, digital–green synergy’s impact on innovation resilience is stronger where public environmental concern is greater. This may be because greater concern leads to stricter regulation, pushing firms to embed digital technologies into green processes, deepening integration and boosting resilience.

6. Conclusions and Policy Implications

6.1. Conclusions

Faced with rising external uncertainty and increasing internal transformation pressure, firms must find ways to innovate and grow. Promoting synergy between digitalization and greenization is now a key strategy. This approach helps businesses build innovation resilience and achieve high-quality development. This paper uses data from China’s A-share non-financial listed firms between 2011 and 2024 to study the impact and mechanisms of this dual synergy on corporate innovation resilience. The main findings are as follows: First, the baseline regression results show that digital–green synergy (DG) significantly boosts corporate innovation resilience (LnIR). With all control variables included, a one-unit increase in DG leads to a 0.029-unit improvement in LnIR at the 1% significance level. This result stays robust after addressing endogeneity and performing multi-dimensional checks. Second, mechanism tests show that dual synergy promotes innovation resilience through four channels: enhancing absorptive capacity, increasing capital market attention, strengthening organizational synergy, and improving resource synergy. Third, the heterogeneity analysis shows that the effect of dual synergy on corporate innovation resilience is stronger in firms with political connections, larger size, and those operating in high-tech or lightly polluting industries. It is also more pronounced in regions with strong digital infrastructure or greater environmental focus. In conclusion, this study constructs a measurement system for dual synergy at the firm level. It also provides micro-level evidence that integrating digitalization and greenization is a key strategy to enhance corporate innovation resilience and achieve sustainable development in a complex environment.

6.2. Policy Implications

6.2.1. National-Level Policy Support for Digital-Green Synergy

The development of digital–green synergy is a critical path for enterprises to respond to external shocks and achieve innovative and sustainable development. At the national level, systematic arrangements can be put in place for “dual synergy.” First, specific guidelines for digital–green synergy should be formulated. These should clarify technological pathways, integration standards, and supporting policies. They should also address barriers to data interoperability and standard alignment. Second, investment in new infrastructure should be increased. Forward-looking deployment of 5G networks, industrial internet, big data centers, and other facilities is needed, while green and low-carbon requirements for enterprises should be strengthened. Finally, the fiscal, taxation, and financial support system should be improved. Policies such as industrial development funds, special green re-lending, and additional deductions for environmental protection R&D expenses should be adopted. These measures guide and incentivize enterprises to pursue “dual synergy” development.

6.2.2. Enterprise-Level Full-Process Integration of Digital-Green Synergy

To effectively enhance corporate innovation resilience, enterprises should deeply integrate digital–green synergy across the entire operational and management processes. They can establish a knowledge management system and hold regular technical exchanges to improve their ability to absorb and transform cutting-edge technologies. Enterprises should also improve their ESG information disclosure mechanisms, use green finance tools, and actively showcase their sustainable development value to the capital market. This will help broaden financing channels for innovation. Internally, companies should break down departmental barriers and form cross-functional teams. They can optimize workflows using digital platforms to boost operational efficiency. Additionally, enterprises should leverage intelligent technologies for precise monitoring and dynamic allocation of innovation resources, ensuring efficient focus on key innovation areas.

6.2.3. Differentiated and Targeted Support Strategies for Digital-Green Synergy

Implement differentiated and targeted support strategies. First, for small and medium-sized enterprises facing severe financing constraints, emphasis should be placed on providing targeted assistance such as credit guarantees and innovation vouchers. Large enterprises with abundant resources should be encouraged to act as chain leaders and drive coordinated transformation across the entire industrial chain. Second, high-tech industries should be supported in advancing cutting-edge technological integration and innovation, while traditional energy-intensive industries should be primarily promoted to pursue green technological transformation, supplemented by necessary transition financial support. Finally, policies should be tilted toward regions with weak digital infrastructure. Meanwhile, public oversight in regions with close environmental attention should be used to achieve balanced development of digital–green synergy and to enhance nationwide innovation resilience.