Does Digital Transformation Improve Manufacturing ESG Performance: Evidence from China

Abstract

In the Industry 4.0 era, marked by rapid digital technological breakthroughs, the adoption of environmental, social, and corporate governance (ESG) is crucial for improving corporate management capabilities and promoting sustainable corporate development. We analyze data from 769 A-share listed companies in China’s manufacturing sector from 2011 to 2023 to examine the impact and transmission mechanism of digital transformation on ESG performance in the manufacturing industry. The findings demonstrate that digital transformation significantly improves manufacturing ESG performance. The results of the mechanism study demonstrate that digital transformation can enhance the ESG performance of the manufacturing sector through three channels: strengthening organizational resilience, promoting technological innovation dynamics, and increasing the green total factor productivity of enterprises. The heterogeneity test results indicate that the influence of digital transformation on ESG is more significant in state-owned firms, where a more a lenient policy environment and moderate market competitiveness promote improved ESG performance via digital transformation.

1. Introduction

The global challenges of climate change, resource depletion, and socio-economic instability have intensified alongside rapid population growth, urbanization, and industrialization [1,2,3]. In response, international efforts such as the Sixth United Nations Environment Assembly and the 29th Conference of the Parties of the United Nations Framework Convention on Climate Change in 2024 have emphasized regulatory coordination, pollution control, and carbon market mechanisms to promote sustainable development. However, the effectiveness of these measures is often undermined by persistent geopolitical tensions and structural market imbalances. Against this backdrop, enterprises play a pivotal role as both economic engines and agents of sustainability. Particularly in China, the manufacturing sector lies at the heart of industrial modernization and global competitiveness [4,5]. While the country has made significant progress in industrial system development, challenges remain in independent innovation, digital integration, and efficient resource use [6,7]. Transitioning to a sustainable manufacturing paradigm thus requires not only technological upgrades but also greater attention to social responsibility and long-term strategic planning.

ESG is a contemporary framework rooted in sustainable operational management, designed to integrate environmental, social, and governance considerations into corporate investment decisions and business practices. The ESG framework necessitates that corporations actively assume responsibilities in environmental stewardship, social equity, and governance, while maintaining profitability. It aims to establish a novel set of corporate value measurement standards, enabling the quantification and comparison of corporate practices and performance in fulfilling social responsibilities to foster sustainable corporate advancement [8,9,10]. As capital markets mature and company models evolve, CSR investing behaviors have been integrated into business activities. In the current intricate and unstable economic landscape, corporate performance regarding social responsibility, environmental sustainability, and governance has evolved beyond mere business metrics, garnering significant interest from investors and society at large [11,12]. As ESG considerations become increasingly important in investor decision-making and public scrutiny, improving ESG performance is vital for both firm-level sustainability and macroeconomic progress.

In the Industry 4.0 era, sophisticated digital technologies such as artificial intelligence, blockchain, big data, and Internet technology are incessantly merging with company operations, that are reshaping business models and governance structures [13,14,15]. The 2024 Chinese Enterprise Digital Transformation Index indicates that a greater number of Chinese enterprises intend to augment their investments in digitalization and persist in advancing digital transformation. The majority of experts con-tend that digital transformation enhances an organization’s information control capabilities, improves the quality of corporate information disclosure, and facilitates the proactive strengthening of internal governance [16]. Conversely, the advancement of digital technology also brings higher expectations for sustainable production and increased regulatory pressures. However, existing empirical findings on the relationship between digital transformation and ESG performance remain mixed. While some studies support the view that digital tools can improve environmental monitoring and stakeholder engagement, others point out that digital transformation alone does not necessarily lead to substantial improvements in ESG outcomes, especially when lacking strategic alignment or sufficient institutional support [17,18]. This ambiguity suggests that the ESG implications of digital transformation should not be taken for granted and warrant further systematic investigation. Clarifying the underlying mechanisms between digital transformation and ESG performance has therefore become particularly important.

While previous studies have explored the link between digital transformation and sustainability, few have systematically examined the mechanisms through which digitalization affects corporate ESG performance. Existing research often isolates individual ESG dimensions or lacks large-scale empirical data, particularly in the Chinese manufacturing context. As a result, the internal transmission pathways between digital transformation and ESG outcomes remain poorly understood. To fill this gap, this study investigates how digital transformation influences the ESG performance of Chinese manufacturing firms using panel data from A-share listed companies between 2011 and 2023. Compared to the existing literature, this study makes several key contributions, outlined as follows:

(1)

It not only established a unified framework integrating environmental, social, and governance dimensions, but it also conducted explorations in each dimension to assess the impact of digitalization on the ESG of manufacturing enterprises.

(2)

It identifies three main transmission mechanisms, through which digitalization enhances ESG performance.

(3)

It explores the heterogeneous effects of digital transformation across different ownership structures, policy environments, and industry competition levels.

The rest of the paper is arranged as follows: Section 2 describes the theoretical analysis. Section 3 outlines the models and variables. Section 4 details the empirical results. Section 5 provides a detailed discussion of the most noteworthy and intriguing findings. Section 6 presents the conclusions, policy implications, and limitations of the study.

2. Theoretical Analysis

2.1. DT and ESG Performance in the Manufacturing Industry

Digital technology is widely recognized as a powerful enabler of the real economy and is increasingly viewed as a critical lever for improving the ESG performance of the industrial sector. However, current research presents mixed conclusions on whether digital transformation consistently delivers positive ESG outcomes across different contexts. While digital finance (DF), a key component of digital development, was initially regarded as a vehicle for ESG enhancement, it primarily emphasizes short-term economic gains by leveraging virtual economy tools to stimulate real industrial growth. This model, though effective in catalyzing rapid expansion, often exacerbates structural distortions such as industrial hollowing, supply chain fragility, and the neglect of environmental and technological upgrading challenges in sectors like traditional manufacturing [19,20]. In contrast, digital transformation, which is distinct from digital finance, prioritizes technological enablement and organizational adaptation. It emphasizes the acquisition and integration of digital capabilities to reconfigure resource allocation, production models, and corporate governance paradigms [21]. While the generalizability of its ESG benefits remains debated, evidence from China’s manufacturing sector increasingly supports the view that digital transformation contributes positively to ESG performance.

Specifically, digital transformation in Chinese manufacturing fosters ESG improvements through multiple mechanisms. Technologically, the embedding of digital technologies into production processes enhances operational efficiency, facilitates the integration of renewable energy, and boosts system responsiveness [22]. From a capability perspective, it cultivates firms’ abilities to collect, analyze, and manage digital assets, enabling real-time environmental monitoring and transparent disclosure that builds trust among stakeholders [23]. Institutionally, digitalization disrupts legacy governance models and drives organizational innovation by strengthening labor protections, formalizing stakeholder responsibilities, and improving social sustainability outcomes [24,25]. From a knowledge standpoint, it breaks through cognitive and organizational boundaries, enabling flexible decision-making and dynamic adaptation to evolving ESG-related market expectations.

Based on the above analysis, the following hypotheses are proposed:

Hypothesis 1.

Digital transformation helps improve corporate ESG performance.

2.2. The Mechanism of Organizational Resilience

Organizational resilience refers to a firm’s dynamic capacity to endure, adjust, and recover in challenging environments. In unfavorable external environments, firms depend on robust organizational resilience to endure external risks, formulate adaptable business strategies, and maintain supply chain stability [26]. Organizational resilience enhances ESG performance through several mechanisms. At the environmental level, it improves the manufacturing industry’s capacity for strategic adaptation and execution. Firms exhibit greater resilience to environmental changes and climate risks by implementing clear contingency plans and maintaining resource reserves [27]. Executives’ cognitive abilities, risk perception, and leadership are critical determinants of organizational resilience. These factors not only protect employee safety during crises but also enhance the organization’s civic image and sense of responsibility [28]. At the governance level, the establishment of a transparent decision-making mechanism and risk management system encourages enterprise management to identify long-term value orientation during operations and enhance the enterprise’s capacity for sustainable development [29].

The development of organizational resilience depends on a combination of internal and external factors influencing the enterprise. The merger and elimination of enterprise departments necessitate the redistribution of resources and rights. Due to organizational inertia and the presence of vested interest groups, the adjustment of organizational structure is frequently challenging to implement effectively. The emergence of digital technologies disrupts corporate cognition and influences the organizational structure of manufacturing firms, compelling them to integrate, learn, and apply diverse and fragmented new knowledge. Consequently, digital knowledge serves as the foundation for developing a new type of organizational structure [30,31]. Furthermore, digital transformation can indirectly improve company management by increasing responsiveness and recovery capabilities during crises, influencing long-term competitiveness and growth potential, as well as enhancing resource efficiency and risk management [32].

Based on the above analysis, the following hypotheses are proposed:

Hypothesis 2.

Digital transformation improves corporate ESG performance by enhancing organizational resilience.

2.3. The Mechanism of Technological Innovation

Technological innovation serves as a crucial factor for organizations seeking to enhance their ESG performance. Technological innovation aids enterprises in optimizing resource allocation, enhancing corporate visibility and social impact, improving governance transparency, and addressing ESG standards and requirements, thereby establishing a competitive advantage in long-term competition. The ongoing advancement and refinement of technology compel enterprises to phase out high-pollution and low-efficiency production sectors, thereby lowering production costs and facilitating the greening of production processes [33]. Secondly, enterprises create inclusive technologies via technological innovation to offer services to small and medium-sized businesses and households, thereby contributing to the reduction in social inequality and enhancement of social inclusion [34]. Moreover, technological innovation addresses practical issues affecting societal functioning, enhances the working conditions of frontline workers, and mitigates risks in high-risk environments. Technological innovation enhances corporate transparency and compliance, facilitates the tracing of supply chain sources, and increases the effectiveness of corporate digital governance [35].

Digital transformation and technological innovation exhibit a reciprocal relationship: digital transformation depends on technological innovation for foundational support, while technological innovation achieves value realization through digital transformation. Together, they form the essential engine for sustainable enterprise development [36,37]. Digital technology establishes the basis for the reconstruction of enterprise production processes, while digital transformation facilitates greater public access to advanced technology. The growing public demand for smart products and the rise of smart manufacturing necessitate that enterprises enhance their R&D investment to bolster their innovation capabilities. The ongoing advancement of digital transformation has established a beneficial cycle within the manufacturing sector characterized by “technology-application-re-innovation”.

Based on the above analysis, the following hypotheses are proposed:

Hypothesis 3.

Digital transformation improves corporate ESG performance through enhanced innovation capabilities.

2.4. The Mechanism of Green Total Factor Productivity

There is a consensus among researchers regarding the significant synergistic relationship between green total factor productivity (GTFP) and firms’ ESG performance. The GTFP enhances ESG performance by facilitating cost savings, providing policy incentives, ensuring market responsiveness, and improving risk resilience [20,38]. Initially, companies decrease production costs and enhance resource utilization efficiency by increasing GTFP. Firms exhibiting higher GTFP are more inclined to receive government subsidies and carbon quota surpluses, and they are likely to enhance their ESG performance through policy-driven mechanisms. Third, enterprises exhibiting higher GTFP are more inclined to create green and efficient products that align with the preferences of consumers and investors, thereby advancing low-carbon manufacturing. Furthermore, enhancing GTFP to improve resilience against fluctuations in raw material prices and environmental regulations can assist firms in mitigating the crises associated with ESG risk exposure.

In contrast to total factor productivity (TFP), GTFP indicates the efficiency of resource factor development and utilization, highlighting the importance of minimizing resource consumption and environmental pollution while enhancing efficiency. Digital technology is defined by its cleanliness and efficiency, aligning well with the notion of green total factor productivity [39]. IoT technology facilitates real-time data collection on production energy consumption, pollution emissions, and equipment operation, thereby optimizing resource allocation. Blockchain technology enables enterprises to trace the source of suppliers’ materials, ensuring compliance with environmental protection standards. AI technology supports the predictive maintenance of production processes and optimizes transportation routes through algorithms. Furthermore, digital transformation enhances green GTFP by promoting participation in total product lifecycle management and clean technology integration [40,41].

Based on the above analysis, the following hypotheses are proposed:

Hypothesis 4.

Digital transformation enhances ESG performance by boosting GTFP and thus ESG performance.

The technical route and research hypothesis of this paper are shown in Figure 1.

3. Research Design

3.1. Models

In order to verify the above research hypotheses, this paper refers to Yang and Chen [9,25] and constructs specific models as follows:

$$ES{G}_{it}={\mu }_0+{\mu }_{1}D{T}_{it}+{\mu }_{2}Con{s}_{it}+Year+Ind+Place+{\epsilon }_{it}$$

(1)

It utilizes existing scholarly research to analyze the causal relationship between digital transformation and the mechanism variables. The model for this mechanism test is represented in Equation (2), expressed as follows:

$$M_{it}={\theta }_0+{\theta }_{1}D{T}_{it}+{\theta }_{2}Con{s}_{it}+Year+Ind+Place+{\epsilon }_{it}$$

(2)

In this model, ESG_it indicates the ESG performance of firm i in year t, DT_it refers to the level of digital transformation, M_it is the mechanism variables, and Cons_it is a set of control variables; Year, Ind, and Place denote year, industry, and region fixed effects; and ε_it is a random error term.

3.2. Variables

3.2.1. Explained Variable

ESG (ESG). Currently, academic studies on corporate ESG performance mostly refer to the ESG indicator scores of corporate research organizations. In this paper, to mitigate scoring discrepancies among different rating agencies and conduct a comprehensive assessment of ESG performance in the manufacturing sector, we select the ESG scores published by Hexun.com, Bloomberg, WIND, Huazheng, and Allied Wave with reference to the research results of Avramov [42]. Among them, the data type of Hexun.com and Bloomberg databases is direct scoring, while the ESG scoring scale of WIND, WIND, and Allied Wave are nine grades, respectively. First, the ESG rating scale of WIND, Huazheng, and Allied Wave is assigned as 1–9 using the assignment method, and then the ESG scores of these five institutions are standardized separately, and the average score of the standardized data is taken as the values. Due to data availability, the scores for the environmental (E), social (S), and governance (G) sub-dimensions were obtained from the Huazheng ESG index.

3.2.2. Explanatory Variable

Digital transformation (DT). Regarding the measurement of the level of enterprise digital transformation has not been fully unified, some scholars use the regional digital finance index to measure the level of digital transformation, but the digital finance score mainly refers to the local user’s Alipay data, from which it is difficult to assess the situation of the enterprise; there are scholars who use indicators such as the percentage of intangible assets of the enterprise and the density of the installation of industrial robots to react to the degree of digital transformation, but due to the size of the manufacturing industry and the specific characteristics of the industry, it is difficult to accurately measure the level of digital transformation of different enterprises. For listed companies, the annual report of enterprises not only discloses the financial information and operational status but also contains the enterprise’s judgment on the future development direction as well as deployment, so it has an important reference value for enterprise decision-making and strategy, so this paper refers to the research of Wu [43] and adopts the text analysis method to measure the index of digital transformation of the manufacturing industry. The specific operation is to convert the annual statements disclosed by A-share listed companies in the manufacturing industry into text format, and to count the frequency of relevant keywords based on the four dimensions of digital technology application, intelligent manufacturing, Internet business model, and modern information system. Figure 2 shows the keywords of the digital transformation index of manufacturing industry.

3.2.3. Mechanism Variables

The mechanism variables mainly include three dimensions: organizational resilience, technological innovation, and green total factor productivity.

Organizational resilience (OR). Previous studies have primarily measured organizational resilience using two approaches: a firm’s long-term operational performance, and the degree of loss and recovery following specific external shocks. As this study focuses on the long-term resilience of enterprises, the first approach is deemed more appropriate. Drawing on the method proposed [44,45,46], we construct an organizational resilience index based on two dimensions: rebound resilience and surpassing resilience. Rebound resilience refers to a firm’s ability to recover to its original state or functional level after encountering external shocks. It is evaluated using the following four indicators: the quick ratio, stockpiled slack resources, non-stockpiled slack resources, and return on equity. Surpassing resilience, on the other hand, captures the firm’s capacity not only to recover but also to grow stronger in the aftermath of disruptions, representing its potential for growth and transformation. This dimension is assessed using the year-on-year growth rates of total assets, operating revenue, and net profit. All indicators are standardized, and the average values of the standardized indicators are calculated to obtain a composite score of organizational resilience.

Technological innovation (TI). At present, a variety of indicators are used to evaluate firms’ technological innovation capabilities, including the ratio of R&D expenditure to total expenditure, the proportion of R&D personnel to total employees, and the output of research achievements. Given that the manufacturing sector is a key arena for technological innovation competition, this study adopts the natural logarithm of (1 + the number of granted patents) as a proxy for a firm’s technological innovation performance [33].

Green total factor productivity (GTFP). GTFP is commonly measured using the super-efficiency SBM model [39]. In this study, input variables include the total number of employees as a proxy for labor input, the net value of fixed assets to represent capital input, and energy consumption converted into standard coal equivalent as energy input. The desirable output is measured by the firm’s operating revenue, while undesirable outputs are captured by emissions of sulfur dioxide, nitrogen oxides, and particulate matter from industrial exhaust.

3.2.4. Control Variables

The control variables that may affect the ESG performance of manufacturing companies are selected, including firm size (Size), net fixed assets (Fixed), debt-to-asset ratio (Asset), marginal profit margin (Profit), earnings per share (Roa), CEO duality (Dual), board size (Board), ownership concentration (Top5), and firm age (Age), and the variables are specifically defined in

Table 1. Variables and their definitions.

Types	Variables	Notations	Definitions
Explained variable	ESG	ESG	The standardized mean ESG score across the five institutions
Explanatory variable	Digital transformation	DT	The word frequency of relevant texts in the annual reports
Mechanism variables	Organizational resilience	OR	The standardized mean values of the indicators under the rebound and surpassing dimensions
	Technological innovation	TI	Logarithm of (the number of granted patents + 1)
	Green total factor productivity	GTFP	The calculation of the super-efficient SBM model
Control variables	Firm size	Size	Logarithm of total business assets
	Fixed assets	Fixed	Logarithm of fixed assets of the enterprise
	Debt-to-asset ratio	Asset	Total liabilities/total business assets
	Marginal profit margin	Profit	(Sales revenue − variable costs)/sales revenue
	Earnings per share	Roa	(Net income − preferred stock dividends)/average shares outstanding
	CEO duality	Dual	If the chairman and general manager are the same person take 1; alternatively, take 0
	Board size	Board	Total number of board members at the end of the year + 1 to take natural logarithms
	Ownership concentration	Top5	Natural logarithm of the percentage of shares held by the top five shareholders
	Firm age	Age	Logarithm of years of business establishment

.

3.3. Descriptive Statistics of Variables

This study selects A-share listed manufacturing firms in China from 2011 to 2023 as the research sample. The year 2011 marks the beginning of China’s 12th Five-Year Plan, capturing the initial phase of the country’s digital transformation. Moreover, the selected period encompasses key stages in the evolution of China’s digitalization policies and the development of the ESG framework. Enterprise data are mainly from databases such as Cathay Pacific (CSMAR), China Research Data Service Platform (CNRDS), and SkyEye, and some data are from companies’ annual reports and official websites. This paper excludes the samples according to the following criteria: (1) the occurrence of ST, * ST, and PT enterprises; (2) enterprises with more missing data samples; and (3) suspended enterprises. It finally obtains 769 eligible manufacturing enterprises. In order to minimize the impact of extreme values on firms, this paper shrinks all continuous variables by 1% up and down, and

Table 2. Descriptive statistics of variables.

Variable	Observation	Mean	Standard Deviation	Min	Max
ESG	9997	0.343	0.245	0.010	0.940
DT	9997	0.010	0.014	0.000	0.084
Size	9997	22.405	1.225	20.061	25.888
Fixed	9997	20.799	1.421	17.539	24.609
Asset	9997	0.425	0.196	0.057	0.910
Profit	9997	1.007	0.064	0.776	1.351
Roa	9997	0.401	0.673	−1.261	3.621
Dual	9997	0.752	0.432	0.000	1.000
Board	9997	2.262	0.159	1.792	2.708
Top5	9997	0.494	0.148	0.184	0.852
Age	9997	2.532	0.603	0.693	3.367

reports the descriptive statistics of the variables.

Table 2 presents the mean value of ESG as 0.343, with a standard deviation of 0.245. The minimum and maximum values are 0.01 and 0.94, respectively. These statistics suggest significant variability in ESG performance across different enterprises, highlighting the currently low ESG performance of Chinese manufacturing enterprises. The mean and standard deviation of DT are 0.01 and 0.014, respectively, while the minimum and maximum values are 0 and 0.084. These statistics suggest that the manufacturing transformation faces challenges, including difficulties and a low success rate in digital transformation.

4. Analysis of Empirical Results

4.1. Benchmark Regression

Table 3. Benchmark regression results.

Variable	ESG			E	S	G
	(1)	(2)	(3)	(4)	(5)	(6)
DT	1.806 *** (10.22)	1.534 *** (8.57)	1.271 *** (6.49)	39.855 *** (7.20)	32.120 *** (4.92)	13.542 *** (2.70)
Size		0.04 *** (8.52)	0.045 *** (8.94)	0.989 *** (6.90)	1.437 *** (8.50)	1.276 *** (9.82)
Fixed		0.023 *** (6.11)	0.02 *** (4.8)	0.838 *** (7.23)	0.233 * (1.70)	0.068 (0.65)
Asset		−0.235 *** (−17.08)	−0.227 *** (−15.52)	−0.568 (−1.38)	−1.772 *** (−3.54)	−12.040 *** (−32.16)
Profit		−0.04 (−1.11)	−0.054 (−1.45)	−1.609 (−1.54)	−5.292 *** (−4.29)	−1.331 (−1.40)
Roa		0.032 *** (8.04)	0.028 *** (6.95)	0.244 ** (2.13)	1.124 *** (8.32)	0.850 *** (8.19)
Dual		0.015 *** (2.74)	0.017 *** (3.11)	0.644 *** (4.16)	0.037 (0.20)	0.544 *** (3.88)
Board		0.02 (1.33)	0.026 * (1.68)	1.348 *** (3.10)	1.779 *** (3.47)	−1.542 *** (−3.91)
Top5		−0.003 (−0.2)	−0.008 (−0.47)	−0.137 (−0.28)	−3.545 *** (−6.07)	0.631 (1.41)
Age		−0.05 *** (−11.26)	−0.033 *** (−5.95)	0.236 (1.52)	−2.799 *** (−15.30)	0.455 *** (3.24)
Constant		−0.844 *** (−12.29)	−0.933 *** (−12.88)	19.442 *** (9.66)	46.449 *** (19.56)	56.722 *** (31.06)
Year/Ind/Place	NO	NO	YES	YES	YES	YES
Observations	9997	9997	9997	9997	9997	9997
Adj-R2	0.0102	0.1266	0.1657	0.2339	0.2157	0.2270

Note: Standard errors in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01.

presents the results of the benchmark regression of digital transformation affecting ESG performance of manufacturing firms. It is found that DT passes the test with a confidence level of 1% and the estimated coefficient is significantly positive regardless of whether control variables are added or not. The above results indicate that digital transformation helps manufacturing enterprises to improve the performance of corporate participation in environmental governance and fulfillment of social governance responsibilities, playing an important role in improving the ESG performance of manufacturing enterprises, and Hypothesis 1 is verified. An analysis of the individual E, S, and G dimensions reveals that the coefficients of digital transformation are all positive and statistically significant at the 1% level, indicating that digital transformation exerts a consistently positive effect across all three ESG sub-dimensions.

At the level of control variables, the estimated coefficients of Size and Fixed are significantly positive, indicating that larger firms usually have better ESG performance, and large firms are subject to more social pressures, but they also have better resource endowment, and it is easier for them to transform ESG into competitive advantages by virtue of their capital, technology, and supply chain control. The estimated coefficients of Asset are significantly negative, indicating that a high degree of indebtedness is not conducive to corporate sustainability; Roa is significantly positive at the 1% level, indicating that the more profitable a firm is, the more it tends to improve its ESG performance. The estimated coefficient of Roa is significantly positive at the 1% level, indicating that the more profitable a firm is, the more likely it is to improve its ESG performance. Meanwhile, the estimated coefficient of Dual is also significantly positive, which suggests that firms with more centralized management power may be more inclined to consider corporate sustainability and therefore contribute to ESG performance.

4.2. Robustness Tests

4.2.1. Endogeneity Tests

This paper employs the instrumental variable method and the Heckman two-stage model to address potential reverse causality in the benchmark regression analysis and to mitigate the endogeneity problem affecting the results.

First, instrumental variable method. To mitigate the endogeneity issue arising from the reverse causality of “better ESG performance leading to a higher stage of digital transformation”, we utilize the average level of digital transformation of other firms in the same industry, region, and year as the instrumental variable, following the approach of Feng [47], and we test this using the two-stage least squares (2SLS) method. The results of the two-stage least squares (2SLS) test are presented in columns (1) and (2) of

Table 4. Endogeneity test results.

Variable	Instrumental Variable		Heckman Two-Stage		PSM
	(1)	(2)	(3)	(4)	(5)
	Stage 1 DT	Stage 2 ESG	Stage 1 DT_H	Stage 2 ESG
DT		2.268 *** (3.41)		1.148 *** (5.67)	0.959 *** (2.93)
IV_DT	0.007 *** (29.98)		0.125 ** (2.46)
IMR				0.015 ** (2.52)
Constant	−0.036 *** (−9.62)	−0.818 *** (−10.99)	−6.853 *** (−14.30)	−0.928 *** (−12.40)	−0.992 *** (−6.71)
Controls	Yes	Yes	Yes	Yes	Yes
Year/Ind/Place	Yes	Yes	Yes	Yes	Yes
Observations	9997	9997	9997	9997	3983
R2/Adj-R2	0.2069	0.1147	0.2687	0.1629	0.2517

Note: Standard errors in parentheses; ** p < 0.05, *** p < 0.01.

. The regression results indicate that the estimated coefficient of digital transformation level on the ESG performance of manufacturing firms remains significantly positive, suggesting that the findings of this paper hold significance even when addressing the endogeneity issue arising from reverse causality. The F-value of the first-stage regression is 254.71, significantly exceeding the threshold of 10, indicating the absence of a weak instrumental variable issue.

Second, Heckman two-stage estimation procedure. To mitigate the impact of sample selection bias on regression outcomes, this study employs the Heckman two-stage model. Initially, a dummy variable representing the median level of digital transformation in manufacturing enterprises (DT_H) is constructed. If the digitalization level of enterprise i in year t meets or exceeds the median, it is assigned a value of 1; otherwise, it is assigned a value of 0. Subsequently, DT_H is incorporated as an explanatory variable alongside the control and instrumental variables in model (1) to conduct Probit regression for estimating the inverse Mills ratio (IMR). In the second stage, the IMR is included in the regression of model (1), with results presented in columns (3) and (4) of Table 4. The results indicate that the coefficient of IMR is significant, suggesting the presence of a sample selection problem. Additionally, the DT coefficient from the re-regression is significantly positive, demonstrating that the findings of this paper remain robust when addressing the potential endogeneity issue arising from reverse causality.

Third, propensity score matching method (PSM). Firms with different levels of digital transformation may also have significant differences in other aspects, and these differences may affect firms’ ESG performance and cause sample selection bias. To address the impact of such differences on the estimation results, this paper adopts the propensity score matching method (PSM) to mitigate this endogeneity problem. First, the samples are grouped using the median annual digital transformation as the critical condition; second, the propensity score is matched between the experimental group and the control group by the kernel matching method using the nine control variables in this study as the covariates to ensure that there is no significant difference between the matched samples in terms of the basic characteristics except for the digital transformation; and lastly, the main effect is re-tested by using the matched samples, and the results are shown in column (5) of Table 4. The coefficients of digital transformation are significantly positive at the 1% level, indicating that there is still a facilitating effect of digital transformation on manufacturing ESG performance after mitigating the endogeneity problem of self-selection bias, and Hypothesis 1 still holds.

4.2.2. Replacement of Explanatory Variables

The robustness test is performed by substituting the primary explanatory variables with the subsequent indicators as follows: The level of digital transformation is assessed by the ratio of intangible assets to total assets of manufacturing firms for the year, as indicated in column (1) of

Table 5. Results of robustness tests.

Variable	Replacement of Digital Transformation Measurement			Clustering Standard Error	Excluding Municipalities	Reduced Timeframe
	(1)	(2)	(3)	(4)	(5)	(6)
DT	0.016 *** (6.77)	1.169 *** (5.88)	0.003 *** (10.23)	1.271 *** (3.58)	1.208 *** (5.83)	1.252 *** (5.74)
Constant	0.045 *** (8.76)	4.817 *** (11.29)	−0.948 *** (−13.18)	−0.933 *** (−5.07)	0.966 *** (−11.96)	0.961 *** (−10.89)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Year/Ind/Place	Yes	Yes	Yes	Yes	Yes	Yes
Observations	9997	9997	9997	9997	8605	6151
Adj-R2	0.1661	0.9541	0.1610	0.1657	0.1591	0.1925

Note: Standard errors in parentheses; *** p < 0.01.

. A one-to-one correspondence exists between all firms and their respective cities throughout the study period, utilizing the digital finance index of prefecture-level cities published by Peking University to assess the level of digital transformation. The results are presented in column (2) of Table 5. The CSMAR database evaluates the digital transformation of listed companies from multiple dimensions, including strategic leadership, technological advancement, organizational empowerment, and the outcomes and applications of digitalization. The results are presented in column (3) of Table 5. The results indicate that the estimation coefficient of DT remains significantly positive when the measurement standard of digital transformation is altered, suggesting that the benchmark regression results are robust.

4.2.3. Modification of the Clustering Standard Error

To enhance the stability of the regression results, it is important to consider the significant impact of regional policy and economic environments on the manufacturing industry, which may lead to serial correlation in the error terms at the region-industry level. This paper utilizes the two-way clustering method [47] to cluster standard errors at the “region-industry” level, with results presented in column (4) of Table 5. The estimated coefficients of DT remain significantly positive, suggesting that clustering the standard errors to the region-industry level indicates an increase in the degree of digital transformation positively contributes to the ESG performance of the manufacturing industry. The benchmark regression results demonstrate robustness.

4.2.4. Adjust the Sample Size

First of all, the strategic position of China’s municipalities is somewhat special, considering the impact of policy, economic, and other unobservable factors, this study censors the samples of such regions, and the regression results are shown in column (5) of Table 5. The results show that without considering the sample of municipalities, digital transformation still has a significant positive contribution to the ESG performance of the manufacturing industry, which verifies the robustness of the benchmark regression results in this paper. Second, the development and application of digital technology will change significantly over a longer time horizon, so only the 2015–2023 sample is retained for re-estimation, and the results are shown in column (6) of Table 5. The results show that the driving effect of digital transformation on the ESG performance of manufacturing firms remains significant when considering the effect of the time factor.

4.3. Mechanism Tests

4.3.1. Mechanism Effects of Organizational Resilience

Column (1) of

Table 6. Mechanism tests.

Variable	OR	TI	GTFP
	(1)	(2)	(3)
DT	2.840 *** (11.92)	13.292 *** (10.81)	0.170 *** (4.34)
Constant	−0.470 *** (−5.14)	−14.106 *** (−31.53)	−0.422 *** (−29.6)
Controls	Yes	Yes	Yes
Year/Ind/Place	Yes	Yes	Yes
Observations	9997	9997	9997
Adj-R2	0.1737	0.4646	0.5274

Note: Standard errors in parentheses; *** p < 0.01.

presents the regression results. The results indicate a significantly positive regression coefficient for digital transformation on organizational toughness, suggesting that digital transformation enhances the organizational toughness of manufacturing enterprises. This improvement is reflected in the organization’s resistance, adaptation, and recovery capabilities when confronted with complex external environments. Digital transformation, within the framework of adjusting and restructuring enterprise management, can effectively mediate conflicts of interest among various departments by enhancing organizational resilience. This improvement subsequently bolsters the enterprise’s strategic determination and execution capabilities, leading to enhanced ESG performance in the manufacturing sector, thereby validating Hypothesis 2.

4.3.2. Mechanism Effects of Technological Innovation

Column (2) of Table 6 presents the regression results. The results indicate a significantly positive regression coefficient for digital transformation on technological innovation. This suggests that as enterprises pursue digitization and intelligence, their independent innovation capacity is strengthened, thereby enhancing their sustainable development capability. The results indicate that enterprises can enhance the ESG performance of the manufacturing industry by improving their technological innovation capabilities, thereby validating Hypothesis 3.

4.3.3. Mechanism Effects of GTFP

Column (3) of Table 6 presents the regression results. The results indicate a significantly positive regression coefficient for digital transformation on green total factor productivity. This suggests that digital transformation substantially enhances the green total factor productivity of enterprises, leading to improved resource utilization efficiency and pollution management capabilities in manufacturing enterprises through the application of digital technology. The manufacturing industry improves its green total factor productivity via digital transformation, subsequently enhancing its ESG performance, thereby validating Hypothesis 4.

4.4. Heterogeneity Tests

The prior analysis indicates that digital transformation can improve manufacturing ESG performance via the following three mechanisms: bolstering organizational resilience, fostering innovation, and augmenting green total factor productivity. Factors such as the regional development plan and business environment significantly influence the digital transformation process and the management’s attitude towards ESG within the enterprise. This paper systematically analyzes the differentiated impacts of digital transformation on the ESG performance of manufacturing enterprises, considering the three dimensions of property rights, policy environment, and market competition.

4.4.1. Nature of Property Rights

The impact of digital transformation on ESG performance may vary across firms with different ownership structures. Utilizing the nature of enterprise property rights as a distinguishing criterion and incorporating it into model (1), the findings presented in columns (1) and (2) of

Table 7. Results of heterogeneity tests.

Variable	Nature of Property Rights		Policy Environment		Industry Competition
	(1)	(2)	(3)	(4)	(5)	(6)
	State-Owned	Non-State-Owned	Strict	Lenient	Intense	Lenient
DT	1.43 *** (4.04)	1.208 *** (5.07)	1.322 *** (5.86)	0.86 ** (2.22)	1.101 *** (4.16)	1.561 *** (5.24)
Constant	−0.835 *** (−7.48)	−0.887 *** (−8.42)	−1.165 *** (−13.29)	−0.51 *** (−3.92)	−0.942 *** (−9.49)	−0.981 *** (−8.17)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Year/Ind/Place	Yes	Yes	Yes	Yes	Yes	Yes
Observations	4373	5624	6408	3589	6048	3949
Adj-R2	0.2	0.19	0.173	0.176	0.159	0.207

Note: Standard errors in parentheses; ** p < 0.05, *** p < 0.01.

demonstrate that the DT for both state-owned and non-state-owned enterprises achieves significance at the 1% confidence level, with the regression coefficients for state-owned enterprises exceeding those of non-state-owned enterprises. This suggests that state-owned enterprises possess notable advantages related to policy rigidity constraints, resource endowment, and governance standardization, resulting in a more pronounced enhancement effect of digital transformation on environmental, social, and governance (ESG) factors.

4.4.2. Policy Environment

The influence of the policy environment on corporate ESG is multifaceted. Mandatory legal constraints directly encourage enterprises to enhance their environmental performance. Conversely, moderate policies can encourage enterprises to invest in environmental and social initiatives, leading to a reduction in violations and a decrease in administrative penalties through improved environmental and social outcomes. According to Zhang and Chen [48], the policy environmental strength (ER) of prefecture-level cities is quantified by the ratio of the frequency of environmental terms in sentences related to manufacturing enterprises to the total word count of the entire government work report. Secondly, the average value of environmental regulation across all enterprises (ER_A) serves as the criterion for division. If ERit ≥ ER_t for a specific enterprise’s city, it indicates a strict policy environment in that region. Conversely, if ERit < ER_t, it suggests a lenient policy environment in the same region. The heterogeneity test for model (1) is performed according to the specified division criteria, with results presented in columns (3) and (4) of Table 7. The findings indicate that the estimated coefficient of DT is 1.322 in a loose policy environment, achieving significance at the 1% confidence level. In contrast, the estimated coefficient of DT is 0.86 in a strict policy environment, with significance at the 5% confidence level. The findings suggest that in a relatively permissive policy context, manufacturing firms exhibit greater dynamism, and the impact of manufacturing digital transformation on enterprise ESG is more pronounced.

4.4.3. Industry Competition

Market competition, as a significant external factor, can influence the investment strategies and behaviors of enterprises to a certain extent. This paper employs the Herfindahl Index (HHI) to assess a company’s market share, drawing on research by Amini [49]. A higher HHI indicates a less competitive environment. Utilizing the average value of the Herfindahl Index (HHI_H) for all manufacturing firms as a criterion for division, a firm is considered to be in an intensely competitive environment if HHIit ≥ HHI_Ht, whereas a firm is deemed to be in a relatively lenient competitive environment if HHIit < HHI_Ht. The sample is regressed in groups, with results presented in columns (5) and (6) of Table 7. The findings suggest that the influence of enterprise digital transformation on ESG is more significant in conditions of moderate market competition. This environment enables enterprises to prioritize social responsibility alongside corporate profitability, facilitating long-term planning and adjustments to production processes and organizational structures through digital technology.

5. Discussion

Based on the research findings, we focus our discussion on two main aspects:

5.1. The Role of Digital Transformation in Enhancing ESG Performance in Manufacturing

From the empirical results, it is evident that digital transformation in China’s manufacturing sector exerts a positive influence on ESG performance, both in the aggregate and across individual dimensions. In the environmental dimension, digital transformation enables real-time monitoring and dynamic optimization of energy consumption in production processes through technologies such as the Internet of Things, artificial intelligence, and digital twins. In the social dimension, blockchain and supply chain management systems enhance end-to-end traceability, ensuring supplier compliance with labor standards. Smart factories reduce the demand for high-risk labor positions through automation and sensor technologies, while digital training platforms help improve employees’ green skills and awareness [50]. In the governance dimension, big data analytics integrate environmental, social, and governance risks to support dynamic strategic adjustments by top management, thereby mitigating ESG-related operational risks.

The reason digital transformation can effectively enhance ESG performance lies in its ability to manage large-scale environmental and social data at low cost, shifting ESG practice from passive compliance to proactive optimization. Government subsidies also incentivize the synergistic advancement of digitalization and green transformation. Meanwhile, firms with higher ESG ratings often enjoy better access to financing, creating a virtuous cycle [5,51]. Ultimately, digital technologies are inherently suited to the sustainable management of complex systems [52,53]. In summary, digital transformation in China’s manufacturing industry systematically improves ESG performance through technological empowerment, data integration, and governance restructuring.

5.2. Causes of Heterogeneous Impacts

The results of the heterogeneity analysis reveal that the impact of digital transformation on ESG performance is more pronounced among state-owned enterprises. In these firms, a more relaxed legal framework and moderate market competition have facilitated ESG improvements through digital transformation. This may be primarily attributed to SOEs’ greater ability to access policy resources, enabled by government credit endorsement and lower financing costs, which allows them to invest more readily in long-term ESG initiatives [54]. In contrast, private enterprises often face constraints from short-term profit pressures, which hamper their capacity to establish sustainable business models.

A lenient policy environment, such as fiscal subsidies, tax incentives, and low-cost financing, directly reduce the cost of digital transformation and foster more effective ESG market incentives. Owing to the high financial burdens and significant pollution levels inherent in manufacturing firms, overly stringent regulatory regimes can suppress technological upgrading by reallocating scarce resources toward regulatory compliance rather than innovation [55]. Consequently, manufacturers may prioritize end-of-pipe treatments over upstream technological innovation, aiming only to meet minimum compliance requirements, thereby lacking motivation for continuous ESG improvement and ultimately demonstrating suboptimal ESG performance.

Furthermore, moderate market competition encourages investment in environmental technologies, accelerating the pace of technological iteration. Such competition pressures management to incorporate ESG considerations into core strategies, enhancing operational transparency and credibility, strengthening governance to mitigate legal risks, avoiding short-termism, and boosting investor confidence [56]. In contrast, cutthroat price wars tend to erode firms’ capacity to invest in ESG initiatives.

6. Conclusions and Limitations

6.1. Conclusions and Implications

Based on the data of Chinese listed manufacturing companies from 2011 to 2023, these findings reveal manufacturing enterprises significantly improve their corporate ESG performance through digital transformation, and this conclusion still holds after a series of robustness tests and endogeneity tests. Through digital transformation, the organizational resilience, innovation capability, and green total factor productivity of manufacturing firms are all enhanced, and corporate ESG performance is improved through these three pathways. Additionally, the driving effect of digital transformation on enterprise ESG performance is more significant when the nature of enterprise property rights is state-owned enterprises, the policy environment is relatively relaxed, and the competition in the industry is relatively moderate.

In light of the findings, we propose the following policy recommendations:

First, to realize the ESG potential of digital transformation, it is imperative to accelerate the integration of advanced digital technologies—such as artificial intelligence, blockchain, and big data—into enterprise operations. Governments should introduce targeted incentive policies, cultivate a supportive institutional environment, and provide enabling infrastructure to foster digital innovation with ESG objectives in mind. Manufacturing enterprises, in turn, should actively adopt digital technologies to enhance clean production capacity, improve the transparency and accuracy of ESG information disclosure, and they should leverage digital platforms to strengthen public image and stakeholder trust. Moreover, digital tools can optimize resource allocation, reinforce resilience, and support evidence-based decision-making, thereby laying a solid foundation for long-term ESG improvement.

Second, digital transformation should be regarded not only as a technological upgrade but also as a catalyst for building resilient and adaptive ESG mechanisms. Firms are encouraged to align their digital evolution with strategic goals such as enhancing organizational resilience, fostering innovation capacity, and improving green total factor productivity. In particular, enterprises should embed ESG principles into corporate culture and governance frameworks, ensuring that digital upgrades translate into sustainable value creation. Concurrently, it is essential to strengthen collaboration between manufacturing firms, government agencies, and research institutions to drive market-oriented innovation, facilitate the development of high-end technologies, and enhance product competitiveness. State-owned enterprises, which are often key actors in national industrial strategies, should adopt a systematic digital transformation audit framework, conduct regular evaluations of transformation impacts, and focus on green productivity enhancement. This would promote not only resource efficiency but also high-quality ESG outcomes.

Third, given the heterogeneous effects of digital transformation across enterprise types and institutional contexts, it is essential to tailor ESG-related policies accordingly. Our findings reveal that property rights structure, policy environments, and industry competition levels significantly mediate the digital–ESG linkage. Relying solely on model ESG firms as policy references may risk policy misalignment due to imitation without contextual fit. Therefore, governments should avoid one-size-fits-all approaches and instead develop nuanced regulatory and incentive frameworks that reflect industrial characteristics, digital maturity, and regional disparities. For example, state-owned enterprises are generally better positioned to leverage digital transformation for ESG enhancement due to their access to policy resources and financing. In contrast, non-state-owned and small- to medium-sized manufacturers should adopt phased digital strategies, focus on critical ESG bottlenecks, and seek collaboration through digital ecosystems and supply chain networks. Moreover, ESG should be internalized as a growth-oriented strategic asset rather than a regulatory burden. Policymakers are further encouraged to introduce differentiated emission standards, offer tailored tax incentives, and lower digital adoption thresholds to facilitate the transition for SMEs. Strengthening intellectual property protection, embedding ESG considerations into public procurement and investment criteria, and fostering market-based ESG competition can collectively stimulate responsible innovation and create a virtuous cycle of economic efficiency and social value creation.

6.2. Limitations

This paper examines the impact of digital transformation on environmental, social, and governance factors from multiple perspectives; however, it presents certain limitations in the following areas. This paper examines the ESG performance of manufacturing companies, utilizing data derived from the scores of prominent research institutions. It excludes certain companies and does not account for unlisted firms or those with limited disclosure. Future research should incorporate a broader range of company types to enhance the dataset. Second, in selecting indicators, this study references various research findings to measure the level of digital transformation scientifically. However, the rapid emergence of digital technologies, particularly artificial intelligence, has resulted in frequent updates and iterations of intelligence theory. Consequently, this has led to ongoing changes in the measurement methods and assessment standards for digital transformation, necessitating further updates to the measurement approach. Finally, the research method employed was the commonly used empirical analysis. Future research can benefit from the integration of emerging artificial intelligence techniques to deepen analytical precision and expand the scope of inquiry.