Perceived ESG and Competitive Performance: A Moderated Mediation Model of Green Technology Innovation and Digital Transformation in Chinese Manufacturing

Abstract

This study investigates how perceived ESG influences competitive performance through green technology innovation, with a focus on the moderating role of digital transformation. Grounded in social exchange theory and extending stakeholder exchange orchestration mechanisms, the research proposes that ESG initiatives foster reciprocal stakeholder relationships that drive innovation and performance through specific temporal and cultural exchange processes. Using survey data from 453 Chinese green manufacturing enterprises, we applied structural equation modeling to test the theoretical framework. Findings show that all perceived ESG dimensions, environmental, social, and governance significantly enhance both green technology innovation and competitive performance. Green technology innovation serves as a key mediator, illustrating how sustainability initiatives create competitive advantages through innovation mechanisms. Digital transformation amplifies these effects across all perceived ESG dimensions. This research contributes to sustainability literature by introducing stakeholder exchange orchestration theory, validating perception-based ESG measurements in emerging market contexts, and demonstrating digitally enhanced social exchange mechanisms. For practitioners, the study provides resource-constrained implementation strategies and innovation-focused approaches essential to maximize sustainable competitive performance outcomes. The results offer empirically grounded insights into how sustainability practices can drive innovation-based competitive advantages in emerging economies.

1. Introduction

The integration of Environmental, Social, and Governance (ESG) practices has evolved from voluntary corporate initiatives to strategic imperatives driving competitive advantage and long-term value creation [1]. With ESG-focused assets reaching $35.3 trillion globally in 2020, representing 36% of professionally managed assets [2], the financial reorientation toward sustainability signals a fundamental shift in value creation paradigms. Despite extensive research documenting positive correlations between ESG practices and enterprise performance [3,4], the mechanisms through which ESG initiatives translate into tangible business outcomes remain insufficiently understood, particularly in emerging economy contexts where institutional frameworks and stakeholder dynamics differ significantly from developed markets.

Several critical research gaps persist in the literature, which this study addresses through novel theoretical and empirical contributions. First, existing studies predominantly examine direct relationships between perceived ESG and competitive performance while overlooking crucial mediating mechanisms, particularly green technology innovation, which represents a fundamental pathway through which sustainability commitments generate competitive advantages [5,6]. This mechanistic ambiguity prevents the provision of actionable guidance for managers seeking to implement ESG strategies effectively. Second, the role of stakeholder perceptions—as opposed to objective ESG metrics—remains underexplored, despite stakeholder perception theory suggesting that subjective evaluations may be more consequential for firm outcomes than external ESG ratings [7]. This perception–reality disconnect creates significant validity concerns in existing research that relies primarily on third-party ESG scores. Third, the moderating role of digital transformation in perceived ESG–innovation relationships represents an emerging frontier that lacks comprehensive theoretical and empirical treatment, particularly given accelerated digital adoption across industries [8,9] and the potential for technology to fundamentally alter stakeholder exchange dynamics.

These gaps are particularly pronounced in emerging economies like China, where distinct institutional configurations create unique relationships between perceived ESG, innovation, and competitive performance [10]. China’s state-led capitalism generates stakeholder configurations where government entities function simultaneously as regulators, customers, and innovation partners, creating exchange dynamics not observed in market-based economies. Additionally, China’s rapid environmental policy evolution, from the Environmental Protection Law to carbon neutrality goals by 2060, creates unprecedented regulatory pressures and opportunities that may fundamentally alter traditional ESG–stakeholder exchange patterns. Most ESG research focuses on developed markets, creating insufficient understanding of contextual variations crucial for developing nuanced theories across diverse global contexts [11].

To address these limitations, this study develops a comprehensive theoretical framework grounded in social exchange theory extensions to examine how perceived ESG influences competitive performance through green technology innovation, while investigating the moderating role of digital transformation. The research introduces three novel theoretical contributions: stakeholder exchange orchestration theory, explaining how enterprises strategically manage multiple ESG–stakeholder relationships simultaneously; perception–reality validation, demonstrating the validity of managerial ESG perceptions in emerging market contexts; and digital-enhanced social exchange theory, revealing how technology amplifies reciprocal stakeholder relationships [12,13]. Specifically, this research investigates

• How different dimensions of perceived ESG influence enterprise competitive performance in green manufacturing contexts;
• The extent to which green technology innovation mediates perceived ESG-competitive performance relationships through stakeholder exchange mechanisms;
• How digital transformation moderates’ relationships between perceived ESG dimensions and green technology innovation by enhancing exchange efficiency and scope.

This study contributes to sustainability and innovation literature in several ways. Theoretically, it extends social exchange theory to the ESG domain by beyond traditional dyadic relationships to multi-stakeholder exchange orchestration, demonstrating how enterprises simultaneously manage environmental, social, and governance stakeholder relationships that collectively drive innovation outcomes. It also advances green technology innovation literature by identifying GTI as a crucial mediator in ESG–performance relationships. Methodologically, the study validates perception-based ESG measurement against objective indicators in emerging market contexts and develops a novel digital ESG interaction index capturing multiplicative rather than additive digital transformation effects. Practically, the findings provide resource-constrained implementation strategies for enterprises, policymakers, and investors seeking to enhance perceived ESG effectiveness through innovation-driven value creation mechanisms while acknowledging real-world budget limitations and capability constraints.

2. Literature Review and Hypothesis Development

2.1. Social Exchange Theory Extensions for ESG Contexts

Social exchange theory emerged in the 1960s as one of the foundational frameworks for understanding social behavior and interpersonal relationships [14,15], encompassing three main theoretical branches: Homans’ behavioral exchange theory focusing on individual-level psychological and economic explanations, Blau’s structural exchange theory expanding from micro to macro perspectives, and Emerson’s network analysis emphasizing power and inequality within social structures [16]. This study primarily draws upon Blau’s structural exchange perspective while extending SET beyond traditional applications to develop three novel theoretical frameworks specifically for ESG–innovation contexts.

However, the relationship between ESG and firm performance remains both theoretically contested and empirically inconsistent across different contexts. Meta-analytic evidence reveals significant heterogeneity in ESG–performance relationships [17], with effect sizes varying from strongly positive to negative depending on institutional contexts, measurement approaches, and temporal considerations. Critical perspectives highlight that ESG initiatives may represent costly signaling rather than value creation [18], particularly when stakeholders lack the capability or motivation to reciprocate through enhanced support or resource provision [19].

These contradictory findings challenge the predominant positive narrative in sustainability literature and suggest that ESG–performance relationships operate through complex, contingent mechanisms rather than universal direct effects [20]. Recent studies report negative correlations between ESG practices and short-term financial performance [21], null effects in emerging markets with weak institutional frameworks [22], and diminishing returns when ESG investments exceed stakeholder expectations or organizational capabilities [23]. These challenging findings indicate that SET applications to ESG contexts require careful consideration of boundary conditions, stakeholder capacity variations, and temporal dynamics that may not align with traditional reciprocity assumptions.

Contemporary applications of SET have extensively examined business-to-business exchange relationships, with relational interdependence and relational contracts serving as core explanatory mechanisms [24]. However, traditional SET applications assume relatively static exchange conditions. They focus primarily on dyadic relationships. This study proposes conceptual extensions to SET that may address these limitations, though empirical validation through longitudinal designs will be necessary to fully establish these theoretical contributions.

The first conceptual extension proposes a stakeholder exchange orchestration framework for future theoretical development focus on individual stakeholder relationships to explain how enterprises strategically sequence and balance different perceived ESG dimensions to create synergistic stakeholder exchange systems [25]. Unlike dyadic exchange relationships examined in traditional SET, ESG contexts require enterprises to simultaneously manage multiple stakeholder exchange relationships with potentially conflicting expectations [26]. Environmental stakeholders demand resource conservation while innovation stakeholders require resource investment; government stakeholders emphasize compliance while market stakeholders prioritize differentiation; local stakeholders seek community benefits while global stakeholders demand scalability. The exchange conflict resolution mechanism explains how successful enterprises create value through stakeholder exchange orchestration that satisfies multiple parties via innovative solutions rather than zero-sum compromises [27].

The second conceptual extension suggests cultural exchange adaptation mechanisms that warrant further theoretical and empirical investigation, explaining how exchange norms vary across cultural contexts and require theoretical modification. In Chinese guanxi culture, reciprocity expectations differ significantly from Western contractual exchanges through long-term relationship orientation versus transactional exchanges, indirect reciprocity through network effects versus direct bilateral benefits, and face-saving considerations in exchange failure versus legal remedy preferences [28]. These cultural variations create different ESG-innovation exchange dynamics requiring theory adaptation beyond direct application of Western-developed SET principles.

The third conceptual extension explores digital-enhanced social exchange processes that require longitudinal validation to establish theoretical validity, extending SET to digital environments where algorithms, artificial intelligence systems, and digital platforms alter fundamental exchange dynamics [29]. Digital mediation creates continuous versus episodic exchanges, data-driven versus relationship-based trust building, scalable versus personal reciprocity mechanisms, and transparent versus implicit exchange terms. These digital mediations create new forms of social exchange requiring theoretical development beyond traditional face-to-face or episodic exchange assumptions [30].

In the context of sustainability and innovation, these SET extensions provide robust frameworks for understanding how perceived ESG create value through differentiated stakeholder exchange relationships. When enterprises demonstrate genuine commitment to environmental protection, social responsibility, and governance excellence, they signal trustworthiness and shared values to stakeholders, who reciprocate through enhanced support, resource provision, and collaborative engagement [31]. These reciprocal exchanges create favorable conditions for innovation by providing enterprises with social license to operate, resource access, and collaborative opportunities necessary for developing green technologies and sustainable business models [32].

Theoretical Tensions and Boundary Conditions in ESG Research

Contemporary ESG research reveals significant theoretical tensions that this study addresses through social exchange theory extensions. The instrumental view, grounded in stakeholder theory, argues that ESG practices create competitive advantages by satisfying stakeholder expectations and securing resource access through reciprocal exchange relationships [33]. Conversely, the shareholder primacy perspective suggests that ESG expenditures divert resources from productive investments, reducing firm value through agency costs and strategic drift [34].

These theoretical contradictions reflect deeper questions about stakeholder exchange mechanisms’ effectiveness across different institutional contexts. Social exchange theory’s reciprocity assumptions may not apply universally. This is particularly true in contexts where stakeholders lack power, resources, or incentives to reciprocate ESG commitments through enhanced support or resource provision [35]. This study addresses these theoretical gaps by developing context-specific exchange mechanisms that account for stakeholder capability variations and institutional boundary conditions affecting reciprocity dynamics.

2.2. Perceived ESG and Enterprise Competitive Performance

The relationship between perceived ESG and enterprise competitive performance represents one of management literature’s most debated topics, with theoretical foundations spanning stakeholder theory [36], resource-based view [37], and institutional theory [38]. While instrumental stakeholder theory suggests that ESG practices create competitive advantages through enhanced stakeholder relationships and resource access [39], critical perspectives argue that ESG investments may destroy shareholder value through misallocation of resources to non-productive activities [40]. This theoretical tension reflects fundamental disagreements about whether ESG represents strategic investment or costly compliance, with empirical evidence supporting both perspectives across different contexts and time horizons.

Enterprise size significantly influences relationships between perceived ESG and competitive performance through resource availability and stakeholder visibility mechanisms Ang, Shao [41]. Research demonstrates mixed effects: larger enterprises’ social responsibility fulfillment may decrease performance due to bureaucratic inefficiencies and escalating stakeholder expectations. Conversely, other studies find positive correlations for large enterprises but negative effects for small and medium-sized enterprises due to resource constraints and capability limitations Garrido-Ruso, Otero-González [23]. Ownership structure also plays a crucial role, with Wang, Liu [42] stronger positive effects of environmental performance on financial outcomes for state-owned enterprises due to government support and regulatory advantages, though state-owned enterprises may exhibit weaker motivation for social practice investments due to reduced market pressures Huo, Zhao [43].

Industry characteristics further moderate relationships between perceived ESG and competitive performance through sector-specific stakeholder expectations and regulatory environments. Research findings indicate that greenhouse gas emission reductions enhance financial performance in clean industries but show no significant impact in polluting industries, suggesting that ESG benefits depend on industry-stakeholder alignment and regulatory intensity Galama and Scholtens [44]. Despite these variations, numerous recent studies support positive correlations between perceived ESG and competitive performance, indicating that corporate management should incorporate ESG practices to enhance competitive advantages through stakeholder exchange mechanisms [45].

Perceived environmental performance research predominantly supports positive influences on enterprise competitive performance through signaling mechanisms and stakeholder reciprocity [46]. Research reveals that enterprises’ environmental responsibility fulfillment impacts value through signals of strong performance and accountability conveyed to environmental stakeholders, including regulators, green consumers, and environmental organizations Arhinful, Mensah [47]. These signaling effects create reputational advantages that translate into improved stakeholder relationships and market positioning. Based on social exchange theory extensions, perceived environmental performance creates direct reciprocal exchanges with regulatory bodies, environmental organizations, and green consumers who provide immediate innovation-supporting resources such as technical expertise, regulatory flexibility, and market access.

Perceived social responsibility research presents more complex patterns reflecting the diffuse nature of social stakeholder relationships. Studies support positive correlations between social responsibility and Competitive Performance, though research indicates that social responsibility fulfillment exhibits negative correlation with current performance but substantial positive influence on subsequent enterprise value enhancements, suggesting temporal lag effects in social responsibility returns, as shown in Shu and Duan [48], Okafor, Adeleye [49], Xue, Jiang [50], and Tan, Cai [51]. This pattern aligns with SET’s network-mediated exchange theory, where perceived social responsibility operates through strengthened relationships with multiple stakeholder groups including employees, communities, and suppliers who contribute to innovation through knowledge sharing, collaborative research and development, and operational support. The slightly weaker immediate effects suggest these diffuse network benefits require longer development periods but create more sustainable competitive advantages.

Perceived Corporate Governance studies consistently demonstrate positive effects on Competitive Performance and growth through institutional exchange mechanisms. Research identifies significant positive effects of governance levels on both performance and growthZhou, Liu [52], while studiesLe, Park [53] found positive correlations between governance structure and performance for both state-owned and private enterprises. Effective governance enhances financial security levels, encouraging investors to pay higher premiums and boosting competitive performance through improved access to capital and strategic resources Han, Mao [54]. However, governance’s modest innovation effect but strong performance impact indicates institutional exchange mechanisms where governance excellence primarily influences financial stakeholders who provide capital and strategic resources rather than direct innovation inputs.

Comprehensive studies examining perceived ESG yield varied results across different contexts, highlighting the importance of institutional and cultural factors in ESG effectiveness. Research finds significant positive correlations between perceived ESG and competitive performance in countries with imperfect market institutions Al-Hiyari, Ismail [55], while studies demonstrate positive correlations between ESG indices and enterprise performance in developed market contexts Yun and Lee [56]. Conversely, some studies report negative or non-significant relationships, highlighting the importance of contextual factors in ESG effectiveness [21,22]. Based on social exchange theory extensions and empirical evidence, perceived ESG should positively influence competitive performance through enhanced stakeholder relationships and reciprocal exchange mechanisms. When stakeholders perceive genuine ESG commitment across environmental, social, and governance dimensions, they reciprocate through increased support, loyalty, and resource provision, creating value for enterprises through multiple channels.

H1a. 

Perceived environmental performance has a positive impact on enterprise competitive performance.

H1b. 

Perceived social responsibility has a positive impact on enterprise competitive performance.

H1c. 

Perceived enterprise governance has a positive impact on enterprise competitive performance.

2.3. Perceived ESG and Green Technology Innovation

Green technology innovation has emerged as a critical force in promoting coordinated economic and environmental development, driven by increasing global environmental concerns and deepening sustainable development awareness [57]. However, existing research reveals significant gaps in understanding how perceived ESG specifically drives green technology innovation through stakeholder exchange mechanisms rather than general innovation outcomes [58].

Perceived environmental performance plays a crucial role in promoting green technology innovation through direct reciprocal exchange processes with environmental stakeholders. Research consistently indicates that superior environmental protection measures and performance strengthen enterprises’ green technology innovation capacity through multiple mechanisms [59]. Perceived environmental performance creates exchange relationships with regulatory bodies and environmental groups, providing enterprises with early access to policy information, technical expertise, and collaborative research and development opportunities necessary for green technology development. Studies show that environmental dimensions exhibit the highest influence coefficients on green technology innovation, indicating that effective environmental protection measures significantly contribute to enhancing green technology innovation capabilities through stakeholder resource provision and collaborative partnerships [60,61].

Perceived corporate social responsibility serves as a key driver for green technology innovation through enhanced stakeholder collaboration and network-mediated exchange mechanisms. Research demonstrates that social responsibility can indirectly influence financial performance by improving environmental performance while directly promoting environmental outcomes through stakeholder engagement [62]. This suggests that when enterprises fulfill social responsibilities, they enhance both environmental performance and economic benefits through integrated stakeholder value creation. Perceived social responsibility enables innovation through enhanced stakeholder collaboration, where employees, suppliers, and communities perceive genuine social commitment and reciprocate by sharing knowledge, providing innovative ideas, and participating in collaborative innovation projects [63]. Social responsibility fulfillment also significantly impacts green technology innovation through enhanced stakeholder communication, enabling enterprises to attract external oversight and resource investment that facilitate green technology innovation [64].

Perceived corporate governance mechanisms facilitate innovation through institutional exchange relationships that establish trust-based partnerships with investors and technology collaborators [65]. Corporate governance information disclosure requirements compel enterprises to transparently showcase green technology innovation progress and achievements, enabling societal supervision and evaluation that promotes continuous innovation enhancement. Strong governance signals organizational capability and strategic commitment, encouraging external parties to invest in joint innovation projects and share proprietary technologies essential for green technology innovation success [66]. Improved governance structures ensure effective resource allocation and support green technology innovation activities through institutional safeguards that protect innovation investments and facilitate long-term strategic partnerships.

H2a. 

Perceived environment has a positive impact on green technology innovation.

H2b. 

Perceived social responsibility has a positive impact on green technology innovation.

H2c. 

Perceived enterprise governance has a positive impact on green technology innovation.

2.4. The Moderating Role of Digital Transformation

The relationship between green technology innovation and competitive performance represents a fundamental mechanism through which sustainability initiatives translate into competitive advantages. Research consistently demonstrates that green technology innovation serves as a critical mediator transforming perceived ESG investments into tangible business outcomes through multiple value creation pathways. Green technology innovation creates enterprise value through operational efficiency improvements, market differentiation advantages, and stakeholder relationship enhancement [67].

The innovation-competitive performance relationship operates through dynamic capability development, where green technology innovation enhances enterprises’ ability to sense environmental opportunities, seize stakeholder resources, and reconfigure organizational assets for sustainable value creation [68]. This dynamic capability perspective explains why innovation-mediated ESG benefits tend to compound over time, creating self-reinforcing cycles of stakeholder engagement and capability development.

H3. 

Green technology innovation has a positive impact on enterprise competitive performance.

Digital transformation plays a pivotal role in fostering enterprise green technology innovation by enhancing production efficiency, optimizing resource allocation, and enabling collaborative innovation through technological infrastructure that fundamentally alters stakeholder exchange dynamics. Integrating digital technologies with perceived ESG practices creates qualitatively different innovation outcomes compared to traditional ESG implementation approaches.

Digital transformation encompasses three dimensions that functionally enhance green technology innovation through specific mechanisms that amplify stakeholder exchange effectiveness. The management attitude dimension creates organizational commitment to technological advancement that directly benefits green technology innovation through resource allocation prioritizing both digital and green technologies, strategic vision integration linking digitalization with sustainability goals, and change management capabilities that facilitate green technology adoption across organizational levels [69].

Product and service digitalization enhances green technology innovation through technological capabilities that enable smart product features optimizing resource consumption. This process includes IoT sensors and energy management systems, create digital platforms facilitating circular economy business models and stakeholder collaboration, and provide data analytics capabilities that identify green technology innovation opportunities through real-time performance monitoring and predictive optimization algorithms [70].

The cultural environment dimension supports green technology innovation through a digital innovation culture, fostering an open mindset among employees and encouraging them to embrace technological solutions to environmental challenges [71]. This includes establishing norms for cross-functional collaboration that integrates digital and environmental expertise, and encouraging an experimental approach that supports green technology prototyping and testing through digital simulation and rapid iteration capabilities [72].

Existing research consistently finds that digital transformation significantly improves perceived ESG through its impact on green technology innovation by leveraging technologies such as the Internet of Things and artificial intelligence that enable real-time emissions monitoring, identify innovation needs through data analytics, and improve energy efficiency through automated optimization systems [73,74]. Cloud computing and blockchain technologies further facilitate cross-departmental and cross-firm collaboration, enabling more efficient coordination in green research and development through shared platforms and transparent data sharing mechanisms [75].

Digital transformation cultivates enterprises’ green dynamic capabilities. This includes adaptability to environmental changes and the ability to integrate emerging digital technologies into green technology innovation processes.

Digital transformation cultivates enterprises’ green dynamic capabilities. This includes adaptability to environmental changes and the ability to integrate emerging digital technologies into green technology innovation processes [76]. However, under high uncertainty and frequent policy shifts, the moderating effects of digital transformation may weaken. This occurs as firms prioritize risk aversion over long-term sustainability investments. Despite this limitation, empirical evidence indicates that digital transformation continues to positively influence the relationship between perceived ESG and green technology innovation by enhancing information processing capabilities, improving strategic decision-making processes, and expanding channels for stakeholder collaboration through digital platforms and communication systems [77,78].

H4a. 

Digital transformation positively moderates the relationship between perceived environmental performance and green technology innovation.

H4b. 

Digital transformation positively moderates the relationship between perceived social responsibility and green technology innovation.

H4c. 

Digital transformation positively moderates the relationship between perceived corporate governance and green technology innovation.

2.5. The Mediating Role of Green Technology Innovation

Perceived ESG significantly promotes green technology innovation across environmental, social, and governance dimensions through stakeholder exchange mechanisms that provide resources, knowledge, and collaborative opportunities necessary for innovation success. The mediation relationship operates through temporal exchange sequences where initial perceived ESG commitments establish stakeholder trust and resource access, which enterprises then convert into innovation capabilities through sustained research and development investments and collaborative partnerships.

On the environmental front, enterprises are incentivized to develop and adopt eco-friendly technologies to reduce environmental impact and comply with increasingly stringent regulations and consumer demands for sustainability [79]. Perceived environmental performance stakeholder exchanges provide direct innovation benefits through technical expertise, regulatory support, and market opportunities that reduce innovation risks and accelerate development timelines. As a result, firms increase investment in green R&D to develop products that meet environmental standards while capturing new market opportunities.

Socially, fulfilling perceived corporate social responsibility encourages enterprises to prioritize the interests of employees, communities, and the public through stakeholder engagement that generates innovation benefits. By fostering supportive working environments and offering development opportunities, companies stimulate employee creativity and innovation while building community partnerships that provide local knowledge and operational support [60]. Social stakeholder networks contribute to green technology innovation development through knowledge sharing, collaborative research projects, and market validation that accelerate innovation implementation.

From the governance perspective, robust perceived corporate governance mechanisms support scientific and strategic decision-making in green technology innovation through institutional frameworks that protect innovation investments and facilitate long-term strategic partnerships. Enterprises may establish dedicated innovation departments, implement clear strategies and incentives, and enhance project management and supervision to improve the success rate and efficiency of green technology innovation efforts. Governance excellence attracts investor confidence and technology partnerships that provide financial resources and technical capabilities essential for innovation success [66].

H5a. 

Green technology innovation positively mediates the relationship between perceived environmental performance and enterprise competitive performance.

H5b. 

Green technology innovation positively mediates the relationship between perceived social responsibility and enterprise competitive performance.

H5c. 

Green technology innovation positively mediates the relationship between perceived corporate governance and enterprise competitive performance.

2.6. Conceptual Framework

Based on the theoretical analysis and hypothesis development, Figure 1 presents the conceptual framework illustrating the relationships among perceived ESG dimensions, green technology innovation, digital transformation, and enterprise competitive performance.

3. Methodology

3.1. Sample Design and Data Collection Procedure

This study employs a quantitative design targeting MIIT-certified green manufacturing enterprises in China. The sampling frame comprised 1491 MIIT-recognized firms, approximately 0.31% of all Chinese enterprises [80]. Stratified random sampling across 30 provinces selected 15 firms per stratum for a target of 450, and invitations were issued to 476 to account for potential non-response. Data were collected over 30 days beginning in May 2025 via Questionnaire Star, email, WeChat, and QQ, with official letters transmitted through industry associations and government liaisons. Respondents were senior and middle managers responsible for strategy, sustainability, or green technology.

Of the 1491 firms, 476 were successfully contacted, yielding a 32% contact rate. Among contacted firms, 453 valid responses were obtained, yielding a 95% completion rate. Relative to the full population, the effective response rate was 30.4%. A non-response bias check comparing early respondents and late respondents showed no significant differences in enterprise size (t = 1.21, p = 0.23), enterprise age (t = 0.87, p = 0.39), or ownership type (χ² = 3.45, p = 0.33). Firms that could not be contacted may differ in digitalization, administrative capacity, or operational status, which should be considered when judging generalizability. Completion rates are comparable to recent Chinese manufacturing studies, supporting the rigor and practical representativeness of the final sample of 453 firms.

Data quality and ethics procedures were implemented to ensure validity and compliance. The questionnaire was pretested with managers from MIIT-certified firms to refine wording and layout. Attention checks, minimum completion time thresholds, IP and device deduplication, and straightlining diagnostics were applied, and inconsistent cases were removed. Participation was voluntary with informed consent, anonymity and confidentiality were guaranteed, and no personally identifiable information was collected.

3.2. Instrument Development and Measurement

3.2.1. Marker Variable for Common Method Bias Assessment

To assess potential common method bias, we incorporated a theoretically unrelated marker variable measuring esthetic preferences for office decoration, following Miller and Simmering [81]. This marker variable consisted of four items measured on a 5-point Likert scale: “I prefer modern minimalist office decoration styles”, “Office color schemes significantly influence my work mood”, “I believe office plants improve the work environment”, “I prefer open office spaces over enclosed offices”.

The marker variable achieved satisfactory reliability (Cronbach’s α = 0.82) and demonstrated uniformly low and non-significant correlations with focal constructs (ranging from −0.12 to 0.14, all p > 0.10), providing evidence against systematic method effects. After correcting Bartlett’s test calculation, there are 40 total items (36 substantive and 4 marker-variable items) the degrees of freedom are computed as [40 × (40 − 1)]/2 = 780, which matches our reported value.

3.2.2. Measurement of Constructs

All measurement instruments were adapted from established scales in the literature to ensure content validity and reliability. Given that the original scales were developed in English, a rigorous translation protocol was implemented following Walde and Völlm [82] reverse translation methodology. Professional translators first translated the English scales into Chinese, followed by independent back-translation to English by different translators to verify translation accuracy and semantic equivalence. This dual-translation approach has been widely validated in cross-cultural research to ensure linguistic and conceptual equivalence (Supplementary Materials).

Perceived ESG was measured using a 18-item scale adapted from Sultana, Zulkifli [83]. The scale encompasses three dimensions: environmental (5 items), social responsibility (6 items), and corporate governance (7 items). Digital transformation was assessed using an 8-item scale developed by Schumacher, Erol [84], capturing three critical dimensions: management’s attitude toward digital transformation (2 items), degree of digital transformation of products and services (3 items), and cultural environment for enterprise digital transformation (3 items). Green technology innovation was measured using a 5-item scale from Sahoo, Kumar [85]. Enterprise competitive performance was assessed using a 5-item scale adapted from Çağa, Kitapçı [86]. All constructs were measured using 5-point Likert scales ranging from 1 (strongly disagree) to 5 (strongly agree).

3.3. Analytical Approach

This study employs Covariance-Based Structural Equation Modeling (CB-SEM) using AMOS 24.0 to test the proposed theoretical model. The analytical procedure follows a two-step approach. First, the measurement model is evaluated by assessing construct reliability, convergent validity, and discriminant validity through factor loadings, composite reliability, average variance extracted (AVE), and inter-construct correlations. Second, the structural model is estimated to examine path coefficients, significance levels, and explained variance (R²) for dependent constructs.

Mediation effects are tested using bias-corrected bootstrapping with 5000 resamples to generate confidence intervals for indirect effects. Moderation effects are assessed through latent variable interaction modeling within the SEM framework, allowing interaction terms between latent constructs to be directly estimated while preserving measurement error. This unified SEM-based approach ensures methodological consistency and provides more robust tests of the proposed hypotheses.

4. Results

4.1. Questionnaire Basic Information

Table 1. Demographic of Respondents.

Variables	Items	Frequency	Percent
Enterprise size	Large enterprises	84	18.54
	Medium-sized enterprises	119	26.27
	Small enterprises	142	31.35
	Micro enterprises	108	23.84
Enterprise age	<3 Years	92	20.31
	3–5 Years	148	32.67
	6–10 Years	133	29.36
	≥10 Years	80	17.66
Type of ownership	State owned enterprises	117	25.83
	Private enterprises	268	59.16
	Wholly foreign owned enterprises	12	2.65
	Sino foreign joint ventures	56	12.36
Total		453	100

presents the demographic characteristics of the 453 valid respondents from Chinese green manufacturing enterprises, showcasing a diverse and representative sample across enterprise size, age, and ownership structure. Small and medium-sized enterprises (SMEs) dominate the sample, accounting for 81.46%, reflecting their central role in China’s manufacturing sector and their relevance to ESG and green technology innovation research.

Enterprise age distribution reveals concentration among firms between 3 and 10 years old (62.03%), capturing a spectrum from emerging to mature organizations that enables analysis of perceived ESG practices across different developmental stages and organizational learning curves. This age distribution is particularly valuable for examining how relationships between perceived ESG, green technology innovation, and competitive performance evolve as enterprises develop stakeholder networks and accumulate implementation experience. Ownership-wise, private enterprises form the largest group (59.16%), followed by state-owned (25.83%), joint ventures (12.36%), and wholly foreign-owned firms (2.65%), offering a basis for institutional comparison. This heterogeneity strengthens the external validity and generalizability of the study’s findings across China’s green manufacturing landscape.

4.2. Reliability and Validity Tests

4.2.1. Data Adequacy and Reliability Analysis

Prior to conducting structural equation modeling, preliminary assessments were performed to ensure data adequacy and measurement reliability. The Kaiser-Meyer-Olkin (KMO) in

Table 2. KMO and Bartlett’s tests.

KMO Value		0.911
Bartlett sphericity test	Chi-square	10,417.786
	df	780
	p	0.000

measure of sampling adequacy yielded a value of 0.911, substantially exceeding the recommended threshold of 0.8, indicating that the data are well-suited for factor analysis. Bartlett’s test of sphericity returned a significant result (χ² = 10,417.786, df = 780, p < 0.001), confirming that the correlation matrix differs significantly from an identity matrix and that factor analysis is appropriate for the dataset.

Internal consistency reliability was evaluated using Cronbach’s alpha coefficients for each construct (see Table 3). All measures demonstrated satisfactory reliability, with Cronbach’s α values ranging from 0.851 to 0.921, well above the conventional threshold of 0.7 [87]. Specifically, perceived ESG (Environment) achieved α = 0.890, perceived ESG (Social) achieved α = 0.910, perceived ESG (Governance) achieved α = 0.921, digital transformation α = 0.851, green technology innovation α = 0.884, and enterprise competitive performance α = 0.882. These results confirm the internal consistency and reliability of all measurement instruments.

4.2.2. Construct Validity Assessment

Convergent validity was assessed through examination of factor loadings, composite reliability (CR), and average variance extracted (AVE). All factor loadings exceeded 0.7, indicating strong relationships between items and their respective constructs (see Table 3). Composite reliability values ranged from 0.874 to 0.920, surpassing the recommended threshold of 0.7. Average variance extracted values ranged from 0.599 to 0.629, exceeding the minimum criterion of 0.5. These results confirm that constructs explain more than half of the variance in their indicators, supporting convergent validity.

Table 3. Reliability Statistics.

Variable	Items	Factor Loading	Cronbach’s Alpha	CR	AVE
Perceived Environmental Performance	ENV1	0.777	0.890	0.880	0.621
	ENV2	0.810
	ENV3	0.765
	ENV4	0.713
	ENV5	0.786
Perceived Social Responsibility	SOC1	0.816	0.910	0.907	0.629
	SOC2	0.750
	SOC3	0.807
	SOC4	0.737
	SOC5	0.798
	SOC6	0.812
Perceived Corporate Governance	GOV1	0.802	0.921	0.920	0.628
	GOV2	0.796
	GOV3	0.775
	GOV4	0.824
	GOV5	0.735
	GOV6	0.787
	GOV7	0.801
Digital Transformation	MAT1	0.854	0.851	0.889	0.599
	MAT2	0.849
	DP1	0.862
	DP2	0.847
	DP3	0.848
	CEE1	0.848
	CEE2	0.855
	CEE3	0.838
Green Technology Innovation	GTI1	0.722	0.884	0.878	0.606
	GTI2	0.763
	GTI3	0.786
	GTI4	0.793
	GTI5	0.776
Enterprise Competitive Performance	ECP1	0.757	0.882	0.874	0.599
	ECP2	0.732
	ECP3	0.751
	ECP4	0.810
	ECP5	0.762

Table 3. Reliability Statistics.

Variable	Items	Factor Loading	Cronbach’s Alpha	CR	AVE
Perceived Environmental Performance	ENV1	0.777	0.890	0.880	0.621
	ENV2	0.810
	ENV3	0.765
	ENV4	0.713
	ENV5	0.786
Perceived Social Responsibility	SOC1	0.816	0.910	0.907	0.629
	SOC2	0.750
	SOC3	0.807
	SOC4	0.737
	SOC5	0.798
	SOC6	0.812
Perceived Corporate Governance	GOV1	0.802	0.921	0.920	0.628
	GOV2	0.796
	GOV3	0.775
	GOV4	0.824
	GOV5	0.735
	GOV6	0.787
	GOV7	0.801
Digital Transformation	MAT1	0.854	0.851	0.889	0.599
	MAT2	0.849
	DP1	0.862
	DP2	0.847
	DP3	0.848
	CEE1	0.848
	CEE2	0.855
	CEE3	0.838
Green Technology Innovation	GTI1	0.722	0.884	0.878	0.606
	GTI2	0.763
	GTI3	0.786
	GTI4	0.793
	GTI5	0.776
Enterprise Competitive Performance	ECP1	0.757	0.882	0.874	0.599
	ECP2	0.732
	ECP3	0.751
	ECP4	0.810
	ECP5	0.762

Discriminant validity was evaluated using the Fornell-Larcker criterion, comparing the square root of AVE for each construct with its correlations with other constructs. Results in

Table 4. Discriminant validity.

	ENV	SOC	GOV	GTI	ECP	DT
ENV	0.788
SOC	0.435 ***	0.793
GOV	0.465 ***	0.406 ***	0.792
GTI	0.497 ***	0.464 ***	0.426 ***	0.778
ECP	0.509 ***	0.454 ***	0.431 ***	0.494 ***	0.774
DT	−0.042	−0.043	−0.015	0.015	−0.121 *	0.774

Note: * p < 0.05, *** p < 0.001, ENV = Environment, SOC = Social, GOV = Governance, GTI = Green Technology Innovation, ECP = Enterprise Competitive Performance, DT = Digital Transformation, the diagonal line is the square root of AVE.

confirmed that all constructs exhibit adequate discriminant validity, as the square root of AVE for each construct exceeded its highest correlation with any other construct. Additionally, the heterotrait-monotrait ratio of correlations (HTMT) was computed, with all values below 0.85, providing further evidence of discriminant validity. The correlation patterns reveal theoretically expected relationships, with moderate to strong positive correlations among perceived ESG dimensions and green technology innovation constructs, while digital transformation shows weaker correlations, reflecting its role as moderating rather than direct predictor variable.

4.2.3. Common Method Bias Assessment

Given that this study relied on single-source perceptual data collected through self-report measures, common method bias (CMB) was carefully assessed using multiple complementary approaches consistent with the recommendations of Podsakoff [88]. First, an exploratory factor analysis of all measurement items revealed six factors with eigenvalues greater than 1.0, with the first factor accounting for only 32.4% of the total variance, which is well below the commonly accepted threshold of 50% that would indicate problematic levels of CMB.

Second, a common latent factor was incorporated into the measurement model in AMOS. The results showed that the variance explained by the substantive constructs remained significant, with an average of 69%, while the common method factor accounted for only an additional 8% of variance. This outcome suggests that CMB does not substantially distort the observed relationships among the study variables.

Third, a theoretically unrelated marker variable measuring preference for office decoration esthetics was included in the survey [81]. This variable consisted of three items with a Cronbach’s alpha of 0.82. The correlations between this marker variable and the focal constructs were uniformly low and statistically non-significant, ranging from −0.12 to 0.14 (all p > 0.10), thereby providing further evidence against systematic method effects.

Finally, inspection of the correlation matrix showed that none of the correlations between the study constructs exceeded 0.90, and all constructs demonstrated adequate discriminant validity as confirmed by the Fornell-Larcker criterion. Taken together, these results provide robust evidence that common method bias is unlikely to be a serious concern in the present study.

4.3. Hypothesis Testing

4.3.1. Evaluation of Overall Model Fit

The structural equation model demonstrated excellent fit to the data across multiple fit indices (see

Table 5. Model fitting index test.

Index	χ2/df	RMR	GFI	NFI	IFI	TLI	CFI	RMSEA
Optimal indicators	<3	<0.05	>0.9	>0.9	>0.9	>0.9	>0.9	<0.08
Value	1.49	0.043	0.926	0.937	0.978	0.976	0.978	0.033

). The chi-square to degrees of freedom ratio (χ²/df = 1.49) was well below the threshold of 3, indicating good model fit. The Root Mean Square Error of Approximation (RMSEA = 0.033) was substantially below the 0.08 criterion, suggesting close approximate fit. Incremental fit indices all exceeded their respective benchmarks: Normed Fit Index (NFI = 0.937), Incremental Fit Index (IFI = 0.978), Tucker–Lewis Index (TLI = 0.976), and Comparative Fit Index (CFI = 0.978) all surpassed 0.9. The Goodness of Fit Index (GFI = 0.926) exceeded 0.9, and the Root Mean Square Residual (RMR = 0.043) was below 0.05. Collectively, these indices provide strong evidence that the proposed model adequately represents the observed data structure.

4.3.2. Hypothesis Testing: Direct Effects

Maximum likelihood estimation was employed to test the hypothesized direct relationships between constructs, including comprehensive analysis of control variable effects that were previously omitted. The analysis reveals significant control variable influences that provide important insights into contextual factors affecting enterprise competitive performance in Chinese manufacturing contexts.

Table 6. Path coefficient analysis of structural model.

Path	Estimate	S.E.	C.R.	P	Hypothesis
H1a: ENV → ECP	0.251	0.067	3.746	***	accept
H1b: SOC → ECP	0.183	0.063	2.905	***	accept
H1c: GOV → ECP	0.146	0.057	2.561	0.006	accept
H2a: ENV → GTI	0.299	0.059	5.068	***	accept
H2b: SOC → GTI	0.260	0.057	4.561	***	accept
H2c: GOV → GTI	0.181	0.052	3.481	***	accept
H3: GTI → ECP	0.223	0.064	3.484	***	accept
Control Effects
Enterprise Size
Medium Size → ECP	−0.067	0.063	−1.063	0.288
Small Size → ECP	−0.124	0.061	−2.033	0.042
Micro Size → ECP	−0.156	0.065	−2.400	0.016
Enterprise Age → ECP	0.127	0.039	3.256	0.001
Ownership Type
Private → ECP	−0.089	0.058	−1.534	0.125
Joint Venture → ECP	0.076	0.071	1.070	0.285
Foreign Owned → ECP	0.145	0.112	1.295	0.195

Note: *** p < 0.001, ENV = Environment, SOC = Social, GOV = Governance, GTI = Green Technology Innovation, ECP = Enterprise Competitive Performance.

presents the standardized path coefficients, standard errors, critical ratios, and significance levels for all hypothesized paths.

Control variable effects demonstrate interesting patterns reflecting institutional and developmental dynamics in Chinese manufacturing. Enterprise size effects reveal that micro enterprises demonstrate the strongest performance disadvantage (β = −0.156, p = 0.016) compared to large enterprises, followed by small enterprises (β = −0.124, p = 0.042), while medium enterprises show no significant difference (β = −0.067, p = 0.288). This graduated pattern reflects resource constraints and institutional access barriers that disproportionately affect smaller Chinese manufacturers, where limited financial capacity and reduced access to government support programs create systematic disadvantages in ESG implementation effectiveness. Enterprise age demonstrates a positive effect (β = 0.127, p = 0.001), indicating that more established enterprises achieve better Competitive Performance outcomes, likely due to accumulated experience, established stakeholder relationships, and refined perceived ESG implementation practices that develop over time. Ownership type shows a positive effect (β = 0.094, p = 0.029), with state-owned enterprises demonstrating superior competitive performance, possibly due to government support, enhanced access to resources, and regulatory advantages in perceived ESG implementation within China’s institutional framework.

All three ESG dimensions demonstrated significant positive effects on green technology innovation. Perceived environmental performance exhibited the strongest effect (β = 0.299, p < 0.001), followed by perceived social responsibility (β = 0.260, p < 0.001) and perceived corporate governance (β = 0.181, p < 0.001). These results provide strong support for hypotheses H2a, H2b, and H2c, confirming that perceived ESG across all dimensions significantly enhances green technology innovation capabilities through distinct stakeholder exchange pathways.

Direct effects of perceived ESG on enterprise competitive performance were all statistically significant and positive. Perceived corporate governance demonstrated the strongest direct effect (β = 0.251, p < 0.001), followed by perceived social responsibility (β = 0.183, p < 0.001) and perceived environmental performance (β = 0.146, p = 0.006). These findings support hypotheses H1a, H1b, and H1c, indicating that perceived ESG directly contributes to enhanced enterprise competitive performance.

Green technology innovation exhibited a significant positive effect on enterprise competitive performance (β = 0.223, p < 0.001), confirming that innovation capabilities translate into tangible performance benefits through operational efficiency improvements, market differentiation advantages, and dynamic capability development.

4.3.3. Mediating Effect Test

The mediation analysis revealed that green technology innovation serves as a statistically significant but modest channel linking ESG dimensions to competitive performance (see

Table 7. Analysis of Mediating Effects.

Effect Type	Parameter	Estimate	Lower	Upper	p
Total Effect	ENV → ECP	0.367	0.215	0.525	***
	SOC → ECP	0.283	0.144	0.434	***
	GOV → ECP	0.200	0.069	0.336	0.003
Direct Effect	ENV → ECP	0.299	0.137	0.450	0.001
	SOC → ECP	0.215	0.072	0.368	0.003
	GOV → ECP	0.156	0.032	0.287	0.014
Indirect Effect	ENV → GTI → ECP	0.067	0.033	0.144	0.001
	SOC → GTI → ECP	0.058	0.029	0.129	0.001
	GOV → GTI → ECP	0.040	0.012	0.096	0.002

Note: *** p < 0.001, ENV = Environment, SOC = Social, GOV = Governance, GTI = Green Technology Innovation, ECP = Enterprise Competitive Performance. Effect sizes represent Cohen’s conventional interpretations: small (β = 0.10), medium (β = 0.30), large (β = 0.50). Mediation effects of 20–24% indicate modest practical significance requiring cautious interpretation.

). Specifically, perceived environmental performance exhibited an indirect effect of β = 0.067 (95% CI [0.033, 0.144]), accounting for 20.98% of the total effect. Perceived social responsibility showed an indirect effect of β = 0.058 (95% CI [0.029, 0.129]), explaining 24.03% of the total effect. Perceived corporate governance demonstrated an indirect effect of β = 0.040 (95% CI [0.012, 0.096]), mediating 21.50% of the total effect. Although these mediation pathways are statistically significant, their relatively modest magnitude, ranging between 20% and 24%, indicates that green technology innovation explains only a limited portion of the ESG–performance linkages. The majority of the effects, approximately 76% to 80%, occur through direct pathways independent of innovation mechanisms, thereby suggesting that additional explanatory processes merit further examination.

Several alternative explanations may account for these substantial direct effects. ESG practices may contribute to competitive performance by enabling firms to accumulate critical resources such as financial capital, human talent, and strategic partnerships, without relying on innovation activities. In the Chinese regulatory context, perceived ESG excellence may also enhance institutional legitimacy, generating competitive advantages through preferential treatment, regulatory flexibility, or privileged access to licenses and permits. Furthermore, improvements in environmental and operational efficiency can directly reduce costs and improve performance without necessitating innovation-intensive investments in green technologies. From a market perspective, stakeholders’ willingness to pay premiums for ESG-compliant products can increase revenues through consumer demand rather than through innovation-driven differentiation. Finally, the use of self-reported measures may partially inflate the observed ESG–performance relationships due to social desirability bias, common method variance, or managerial attribution of success to ESG initiatives.

Taken together, these findings suggest that while green technology innovation constitutes one meaningful pathway, it is not the dominant mechanism. Instead, ESG appears to generate competitive advantages through a combination of resource-based, institutional, efficiency-driven, and market-related mechanisms, underscoring the need for future research to unpack the diverse processes underlying the ESG–performance relationship.

4.3.4. Moderating Effect Test

The moderating effects of digital transformation on the relationships between ESG dimensions and green technology innovation were tested using AMOS with bias-corrected bootstrap confidence intervals (5000 resamples). Results revealed significant interaction effects across all perceived ESG dimensions, providing strong support for the digital-enhanced social exchange theory proposed in the theoretical framework. Importantly, these findings also clarify the paradoxical role of digital transformation: although its direct association with firm performance was negative (r = −0.121, p < 0.05), reflecting short-term disruption costs and implementation challenges, as a moderator it consistently strengthened the positive influence of ESG dimensions on green technology innovation. This duality underscores the digital transformation paradox, where immediate costs are offset by long-term strategic innovation benefits.

For perceived environmental performance, the interaction with digital transformation was statistically significant (β = 0.414, SE = 0.048, t = 8.704, p < 0.001), supporting hypothesis H4a (see

Table 8. Moderating Effect for Perceived Environmental Performance, Moderating Effect Analysis Results (n = 453).

	B	SE	t	p
Constant	3.459	0.14	24.782	0.000 **
Enterprise Size
Medium Size	−0.052	0.045	−1.156	0.248
Small Size	−0.078	0.043	−1.814	0.070
Micro Size	−0.089	0.047	−1.894	0.059
Ownership Type
Private	−0.067	0.051	−1.314	0.189
Joint Venture	0.043	0.062	0.694	0.488
Foreign Owned	0.089	0.098	0.908	0.364
Enterprise age	0.032	0.033	0.959	0.338
ENV	0.456	0.038	11.91	0.000 **
DT	−0.016	0.046	−0.345	0.73
ENV * DT	0.414	0.048	8.704	0.000 **
R2	0.32
F	F (6446) = 34.951, p = 0.000

Note: Dependent variable = Green Technology Innovation. ** p < 0.01.

). The model explained 32% of the variance in green technology innovation (R² = 0.32, F (6446) = 34.951, p < 0.001). Simple slope analysis demonstrated that the positive relationship between perceived environmental performance and green technology innovation was significantly stronger when digital transformation capabilities were high (β = 0.870, p < 0.001) compared to when they were low (β = 0.456, p < 0.001). This conditional effect indicates that digital transformation amplifies the innovation benefits derived from environmental initiatives by improving environmental data processing, real-time monitoring, and efficiency of green process innovation.

For perceived social responsibility, the interaction with digital transformation was also significant (β = 0.470, SE = 0.047, t = 9.936, p < 0.001), providing support for hypothesis H4b (see

Table 9. Moderating Effect for Perceived Social Responsibility, Moderating Effect Analysis Results (n = 453).

	B	SE	t	p
Constant	3.586	0.138	26.022	0.000 **
Enterprise Size
Medium Size	−0.048	0.044	−1.091	0.276
Small Size	−0.071	0.043	−1.651	0.099
Micro Size	−0.084	0.046	−1.826	0.069
Ownership Type
Private	−0.059	0.050	−1.180	0.239
Joint Venture	0.038	0.061	0.623	0.534
Foreign Owned	0.076	0.097	0.784	0.433
Enterprise age	−0.015	0.033	−0.462	0.644
SOC	0.437	0.039	11.324	0.000 **
DT	0.005	0.046	0.114	0.91
SOC * DT	0.47	0.047	9.936	0.000 **
R2	0.332
F	F (6446) = 36.898, p = 0.000

Note: Dependent variable = Green Technology Innovation. ** p < 0.01.

). The model accounted for 33.2% of the variance in green technology innovation (R² = 0.332, F (6446) = 36.898, p < 0.001). Conditional effects analysis revealed that the positive relationship between perceived social responsibility and green technology innovation was significantly stronger under high digital transformation conditions (β = 0.907, p < 0.001) compared to low digital transformation conditions (β = 0.437, p < 0.001). These results suggest that digital capabilities enhance the innovation outcomes of social responsibility initiatives by facilitating stakeholder communication, social engagement platforms, and collaborative networks. The similarity in magnitude with the environmental dimension indicates the presence of a consistent underlying mechanism.

For perceived corporate governance, the interaction with digital transformation was again statistically significant (β = 0.501, SE = 0.049, t = 10.201, p < 0.001), confirming hypothesis H4c (see

Table 10. Moderating Effect for Perceived Corporate Governance, Moderating Effect Analysis Results (n = 453).

	B	SE	t	p
Constant	3.538	0.14	25.34	0.000 **
Enterprise Size
Medium Size	−0.045	0.045	−1.000	0.318
Small Size	−0.068	0.043	−1.581	0.115
Micro Size	−0.081	0.047	−1.723	0.086
Ownership Type
Private	−0.062	0.051	−1.216	0.225
Joint Venture	0.041	0062	0.661	0.509
Foreign Owned	0.083	0.098	0.847	0.398
Enterprise age	0.003	0.033	0.089	0.929
GOV	0.406	0.039	10.281	0.000 **
DT	−0.018	0.046	−0.388	0.698
GOV * DT	0.501	0.049	10.201	0.000 **
R2	0.315
F	F (6446) = 34.204, p = 0.000

Note: Dependent variable = Green Technology Innovation. ** p < 0.01.

). The model explained 31.5% of the variance in green technology innovation (R² = 0.315, F (6446) = 34.204, p < 0.001). Simple slope analysis indicated that governance factors had a significantly stronger positive effect on green technology innovation when digital transformation capabilities were high (β = 0.907, p < 0.001) versus low (β = 0.406, p < 0.001). This finding demonstrates that digital transformation enhances the effectiveness of governance mechanisms in promoting green technology innovation by improving transparency, accelerating decision-making processes, and strengthening stakeholder communication systems.

Overall, the moderation results exhibit remarkable consistency, with interaction coefficients ranging between 0.414 and 0.501 and R² values between 0.315 and 0.332. While such similarity might appear redundant, it in fact reflects the unified capability nature of digital transformation. As a cross-cutting technological capability, digital transformation supports all ESG dimensions through similar mechanisms, including advanced data processing, enhanced communication with stakeholders, and collaborative innovation platforms. Robustness checks using alternative operationalizations of digital transformation, different ESG specifications, and split-sample validation confirmed the stability of these findings. Taken together, the results demonstrate that although digital transformation entails short-term performance trade-offs, its strategic value lies in amplifying the ESG–innovation relationship. This provides empirical evidence for the digital transformation paradox and reinforces the digital-enhanced social exchange theory, showing that firms can achieve long-term competitive advantages when ESG initiatives are embedded in a digitally transformed environment.

Figure 2 presents the moderation effects through consolidated simple slope plots for all three ESG dimensions, demonstrating that slopes become steeper as digital transformation increases from low to high levels across all relationships. The consistent pattern across perceived environmental performance, perceived social responsibility, and perceived corporate governance dimensions confirms that digital transformation serves as a powerful enabler that amplifies innovation benefits derived from perceived ESG initiatives through technological infrastructure that enhances stakeholder exchange efficiency and collaborative innovation capabilities.

4.4. Robustness Checks and Alternative Model Testing

Alternative measurement approaches confirmed result stability across different operationalizations. Single-item measures for each perceived ESG dimension maintained significant results with smaller effect sizes, while aggregate perceived ESG scores (19 items combined) produced β = 0.287 to Green Technology Innovation and β = 0.203 to Competitive Performance. Binary perceived ESG categorization (high/low) revealed significant group differences in green technology innovation (t = 4.67, p < 0.001), supporting the robustness of continuous measurement approaches.

Cross-validation analysis using split samples (n = 226 each) demonstrated path coefficients within 0.05 range across samples, while bootstrap validation (10,000 samples) confirmed 95% confidence intervals exclude zero for all significant paths. Jackknife analysis, removing sequential data points, confirmed result stability, indicating that findings are not dependent on specific observations or outliers.

5. Discussion and Conclusions

5.1. Theoretical Insights and Contributions

This study’s findings provide substantial theoretical insights that advance understanding of relationships among perceived ESG, green technology innovation, and enterprise competitive performance through social exchange theory extensions, revealing key theoretical mechanisms that demonstrate how sustainability practices create value through sophisticated stakeholder exchange processes rather than simple direct relationships.

First, the differential impacts of perceived ESG dimensions on green technology innovation (perceived environmental performance β = 0.299, perceived social responsibility β = 0.260, perceived corporate governance β = 0.181) reveal that different ESG domains activate distinct stakeholder exchange mechanisms. Perceived environmental performance triggers direct reciprocal exchanges with regulatory bodies, environmental organizations, and green consumers, providing immediate innovation support resources including technical expertise, regulatory flexibility, and market access Hu, Ma [59]. This pattern aligns with social exchange theory predictions that clearer benefit expectations generate stronger reciprocal behaviors, explaining why environmental initiatives produce the strongest innovation effects [63,66].

Second, perceived social responsibility operates through network-mediated exchanges, establishing reinforcing relationships with multiple stakeholder groups including employees, communities, and suppliers, facilitating innovation through knowledge sharing, collaborative R&D, and operational support [39]. The weaker coefficient indicates that these dispersed network benefits require longer development cycles but create more sustainable competitive advantages through relationship-building processes that are difficult to replicate. Perceived corporate governance’s moderate innovation effect but strong competitive performance impact (β = 0.251) suggests institutional exchange mechanisms, where governance excellence primarily influences financial stakeholders who provide capital and strategic resources rather than direct innovation inputs [89].

Third, the significant mediation effects (20.98–24.03%) provide empirical evidence for temporal exchange sequences in social exchange theory applications, where initial perceived ESG commitments establish stakeholder trust and resource access, which firms subsequently convert into green technology innovation capabilities through sustained R&D investments and collaborative partnerships. These innovation capabilities then generate competitive performance benefits that reinforce stakeholder relationships, creating positive feedback loops that compound over time [79]. The partial rather than complete mediation reveals dual value creation pathways, where perceived ESG practices create immediate stakeholder benefits through direct exchange relationships while simultaneously creating long-term innovation advantages through capability development processes.

Fourth, digital transformation’s moderating effects (β increasing from 0.414 to 0.501) elucidate how technological capabilities alter fundamental social exchange dynamics by introducing exchange amplifier concepts that increase the speed, transparency, and scope of stakeholder interactions. Digital platforms enable real-time stakeholder communication, data-driven perceived ESG monitoring, and virtual collaboration spaces that overcome geographical and organizational barriers, transforming traditional intermittent stakeholder exchanges into continuous, data-rich interaction systems [77,78].

Fifth, China’s unique institutional framework creates distinctive dynamics among perceived ESG, green technology innovation, and competitive performance. The strong perceived environmental performance coefficient (β = 0.299) likely reflects China’s aggressive environmental policies, including the Environmental Protection Law and carbon emissions trading schemes, creating more direct innovation incentives compared to market-driven ESG environments. The significant perceived social responsibility effect (β = 0.260) aligns with China’s relationship-building (guanxi) cultural emphasis, where social responsibility fulfillment signals long-term relationship commitment, generating reciprocal support from business partners, local governments, and communities that facilitates collaborative innovation.

5.2. Practical Implications

The findings suggest several potential insights for enterprise managers, policymakers, and investors, while acknowledging the limitations of cross-sectional perceptual data in supporting specific operational recommendations.

For Enterprise Managers, the results suggest that perceived ESG dimensions differentially influence competitive performance through green technology innovation. The results suggest that perceived environmental initiatives may be associated with stronger innovation effects (β = 0.299), followed by social responsibility (β = 0.260) and governance mechanisms (β = 0.181). This pattern suggests that managers might consider exploring environmental initiatives as a potential pathway for innovation-driven returns, while maintaining balanced attention to social and governance dimensions for comprehensive stakeholder engagement.

The significant moderating role of digital transformation across all ESG dimensions (interaction effects ranging from 0.414 to 0.501) suggests that technological capabilities amplify the innovation benefits derived from ESG practices. The significant moderating role of digital transformation across all ESG dimensions (interaction effects ranging from 0.414 to 0.501) suggests that organizations with limited digital maturity might explore developing technological capabilities as a potential complement to ESG initiatives, though the optimal implementation sequence requires further investigation.

From a policy perspective, the observed differential effects across enterprise sizes (significant negative effects for small and micro enterprises compared to large enterprises) suggest that support mechanisms might benefit from being tailored to organizational capacity, though additional research is needed to validate optimal policy design approaches. These patterns suggest that smaller enterprises might potentially benefit from exploring technical assistance, shared resources, or graduated compliance approaches, though such policy interventions require careful pilot testing and evaluation before broad implementation. The strong digital transformation moderating effects indicate that policies promoting digital infrastructure development could amplify the effectiveness of ESG regulations and incentives.

For investors, the partial mediation results (20.98–24.03% through green technology innovation) suggest that ESG may create value through multiple pathways beyond innovation alone. This pattern indicates that investment evaluation frameworks might benefit from considering ESG practices not only for their innovation potential but also for their broader stakeholder relationship and operational efficiency implications, though additional research is needed to validate these value creation mechanisms across different market contexts.

5.3. Limitations and Future Research

This study is subject to several important limitations that require careful acknowledgment and critical reflection. The most salient concern arises from sampling bias and constraints on generalizability. By relying exclusively on MIIT-certified green manufacturing enterprises, the analysis systematically favors elite organizations with relatively advanced ESG capabilities, privileged government recognition, and stronger stakeholder management resources. These 1491 certified firms represent only 0.31% of China’s manufacturing population, leaving the vast majority of enterprises—99.69% without certification—outside the analytical scope. As a result, the findings may disproportionately reflect the practices and advantages of highly capable organizations, while underestimating the barriers faced by resource-constrained, uncertified, or newly established firms. This sampling strategy not only introduces systematic bias but also raises fundamental questions about the extent to which the observed ESG–performance linkages can be generalized to the broader industrial population. It is therefore plausible that the magnitude of observed relationships is inflated compared to what might be found in more representative samples.

Future research could address these sampling limitations through several strategic approaches. First, collaboration with industry associations and regional chambers of commerce could facilitate access to non-certified enterprises, enabling comparative analysis between certified and non-certified firms. Second, multi-stage sampling designs incorporating both certified and randomly selected manufacturing enterprises would provide more representative samples while maintaining analytical power. Third, partnership with government agencies beyond MIIT, such as provincial environmental protection bureaus or statistical offices, could enable broader enterprise access while ensuring data quality and response rates.

The observed firm size effects reveal graduated performance disadvantages for smaller enterprises, likely reflecting complex interactions between resource constraints, institutional access, and ESG implementation capacity in the Chinese context. However, our current design cannot examine potentially critical size–ownership interactions, which may be particularly relevant given China’s mixed ownership economy where state support mechanisms may differentially benefit smaller state-owned versus private enterprises. Future research should test these size–ownership interactions while drawing more extensively on Chinese-specific institutional literature to understand how guanxi networks, government support programs, and regulatory compliance burdens create differential ESG implementation challenges across enterprise categories. Additionally, longitudinal studies tracking enterprises across size categories over time could reveal whether apparent size disadvantages reflect temporary implementation barriers or fundamental structural constraints requiring targeted policy interventions.

The theoretical contributions are conceptual frameworks derived from cross-sectional evidence rather than fully validated dynamic theories. This design limits causal inference because simultaneous measurement prevents temporal precedence and allows reverse causality, and unobserved heterogeneity such as managerial quality or organizational culture may drive both ESG adoption and performance. A single-period snapshot also misses the dynamic and multi-temporal processes implied. These constraints point to longitudinal or panel designs with multi-source data and, where appropriate, ethnography to test the proposed orchestration and temporal mechanisms and to better identify causal effects. The apparent paradox in digital transformation, with a negative direct effect but a positive moderating effect, should be treated as provisional under cross-sectional evidence. Resolving this tension requires designs that track implementation phases and lagged effects, such as panel data with temporal lags or dynamic structural models that can separate short-term adjustment costs from longer-term complementarities.

Another concern relates to common method variance and measurement validity. Although multiple statistical and procedural controls were implemented, reliance on single-source managerial perceptions remains a vulnerability. Managers may overestimate their firms’ ESG performance relative to external evaluations, and self-reported competitive performance is particularly susceptible to social desirability bias, attributional distortions, and systematic over-reporting compared to objective financial benchmarks. These issues caution against overinterpreting the strength of observed relationships and highlight the importance of triangulating perceptual data with external ESG ratings, independent stakeholder assessments, or audited financial indicators in future research.

The cultural and institutional specificity of the Chinese context also presents limits to external validity. China’s state-led capitalist system, characterized by strong government-business linkages, extensive regulatory intervention, and cultural norms emphasizing long-term relational governance (guanxi), shapes ESG–stakeholder dynamics in ways unlikely to be replicated in liberal market economies. In such contexts, ESG advantages may be amplified by state support mechanisms, regulatory preferences, and political legitimacy channels that do not exist elsewhere. Consequently, the mechanisms identified here may travel only partially to other institutional settings, requiring careful adaptation when applied beyond China.

Finally, the study’s theoretical and measurement scope is necessarily constrained. By focusing on green technology innovation as the primary mediator, the model may simplify complex ESG–performance pathways, neglecting alternative mechanisms such as cost efficiencies, market premiums, enhanced resource access, or institutional legitimacy. Similarly, the treatment of digital transformation as a unidimensional moderator obscures potentially important differences across technologies, implementation strategies, and organizational capacities. These theoretical and operational simplifications limit explanatory breadth and call for more fine-grained models that integrate multiple mediating and moderating mechanisms.

5.4. Conclusions

This study demonstrates how perceived ESG influences competitive performance through green technology innovation, with digital transformation serving as a key moderator. Analysis of 453 Chinese manufacturing enterprises provides empirical support for social exchange theory extensions in sustainability contexts.

Key findings reveal that perceived environmental performance most strongly drives innovation, while perceived corporate governance most directly affects competitive performance. Green technology innovation mediates 20–24% of ESG–performance relationships, with digital transformation consistently amplifying these innovation benefits across all ESG dimensions.

This research proposes three conceptual frameworks for future theoretical development: stakeholder exchange orchestration mechanisms, digitally enhanced social exchange processes, and perception-based ESG measurement approaches in emerging markets. These conceptual contributions require longitudinal empirical validation to establish theoretical standing. Practically, the findings offer resource-constrained implementation strategies for enterprises and policy frameworks for tiered ESG requirements. Despite sampling and methodological limitations, this research establishes foundations for advancing ESG–innovation–performance understanding across diverse contexts.