Does Digital Transformation Enhance the Sustainability of Enterprises: Evidence from China

Abstract

In the context of China’s economy being in transition, with the arrival of the digital economy era, it is of great significance to explore the issue of corporate sustainable development from the perspective of digital transformation. This study empirically examines the impact of digital transformation on corporate sustainable development and its mechanism of action with a sample of A-share listed companies in Shanghai and Shenzhen from 2004 to 2021. The findings show that corporate digital transformation can significantly improve their sustainable development and is heterogeneous. Specifically, enterprise digital transformation promotes the sustainable development level of underinvested enterprises more than that of overinvested enterprises and promotes the sustainable development of non-state-owned enterprises more than that of state-owned enterprises; enterprise digital transformation promotes the sustainable development of non-manufacturing enterprises more than that of manufacturing enterprises; and the higher the degree of marketization of an enterprise, the greater the impact of enterprise digital transformation on its sustainable development. Meanwhile, through a series of robustness test candidates such as endogeneity test, alternative explanatory variables, and explanatory variables, and adding macroeconomic characteristics, the research conclusions still hold. In addition, through the analysis of the mechanism of action, it is found that enterprise digital transformation can enhance the sustainable development of enterprises by reducing the short-sighted behavior of managers, reducing the cost of financing and leverage, and then enhancing the level of sustainable development of enterprises.

1. Introduction

The rapid evolution of “ABCD” technologies (AI, blockchain, cloud computing, big data) has propelled the global economy into a data-driven era. Despite pandemic-induced economic turmoil, the digital economy demonstrated resilience, contributing 45% to global GDP with 15.6% growth in 2021 (CAICT, 2022, https://cset.georgetown.edu/wp-content/uploads/t0442_AI_white_paper_2022_EN.pdf, accessed on 15 June 2025). Leading economies adopt distinct strategies: the U.S. leverages Silicon Valley to sustain a digital economy exceeding 60% of GDP, while the EU advances regulatory frameworks (e.g., Digital Markets Act) and Germany achieves 47% industrial digitalization under “Industry 4.0”. These cases underscore technology–industry integration as a competitive priority.

As the world’s second-largest economy, China has prioritized “deep integration of internet, AI, and big data with the real economy” as national strategy. In 2021, its digital economy reached 4.5 trillion RMB (39.8% of GDP), growing at 16.2% annually (CAICT, 2022, https://cset.georgetown.edu/wp-content/uploads/t0442_AI_white_paper_2022_EN.pdf, accessed on 15 June 2025), powered by 1.03 billion internet users and 4.1 million industrial internet platforms. This unique combination of massive market potential and diverse applications positions China as a global “innovation testing ground” for translating lab technologies into industrial practices.

Enterprises serve as critical nodes in bridging traditional and modern economies through digital transformation (Wu et al., 2021) [1]. Academic focus now centers on whether digital–business integration delivers measurable economic outcomes, particularly in investment efficiency—a key metric for macroeconomic growth and corporate sustainability.

Despite investment-driven growth (contributing 50.6% to GDP growth annually, 2001–2020), China faces persistent micro-level inefficiencies. Empirical studies reveal stark disparities: 39.3% of listed firms overinvest (exceeding optimal levels by 100.66%), while 60.7% underinvest (reaching only 46.31% of potential) (Zhang & Song, 2009) [2]. This “macro good, micro bad” dilemma (Xie, 2018) hampers resource allocation, exacerbating overcapacity, asset idling, and unsustainable growth [3]. As China transitions toward sustainable development, enhancing corporate investment efficiency emerges as imperative for systemic economic vitality.

Paradoxically, elevated investment volumes do not guarantee efficiency. Beneath the veneer of rapid economic progress lies a critical challenge: the suboptimal quality and efficiency of investments, which hinder the overall health, productivity, and sustainability of Chinese enterprises. Since the 1990s, China’s economy has been marked by a disparity where macro-level indicators, such as high GDP growth driven by investments, contrast starkly with micro-level inefficiencies. This dichotomy was aptly termed “macro good, micro bad” (Xie, 2018) [3].

Empirical evidence highlights persistent investment inefficiencies in Chinese firms. Zhang and Song’s (2009) analysis of 301 industrial firms revealed that 39.26% overinvested (exceeding optimal levels by 100.66% on average), while 60.74% underinvested (reaching only 46.31% of optimal thresholds) [2]. Such misallocations exacerbate overcapacity, idle assets, and systemic risks, undermining both firm sustainability and macroeconomic stability. As China transitions toward a new development paradigm, improving the efficiency of enterprise investment and the level of sustainable development has become an especially urgent issue for China’s economy.

Existing research indicates that internet development can optimize resource allocation (Bai et al., 2022) [4], while corporate investment efficiency is a critical manifestation of resource allocation efficiency and sustainable development at the micro level (Wang et al., 2022) [5]. Information asymmetry and principal–agent problems (Jensen, 1986; Myers & Majluf, 1984; Biddle, 2009; Chen, 2011) [6,7,8,9] have been identified as primary drivers of inefficient investment. Digital transformation, by enhancing information disclosure quality and internal governance, may serve as a key pathway to improving investment efficiency and advancing sustainable development. However, no studies have yet directly examined the causal mechanisms linking corporate digital transformation to sustainable development, leaving their relationship underexplored in direct research.

Amid global industrial restructuring and technological revolution, Chinese enterprises’ digital transformation confronts dual challenges: universal issues like data security and tech ethics, alongside distinctive complexities such as resource allocation in an ultra-large market and shifting growth drivers. Exploring the synergy between digital technologies and enterprise operations not only offers a “Chinese solution” for global transformation but also decodes China’s unique developmental logic, providing novel insights into digital economy-era growth paradigms. This dual significance underscores the international academic community’s growing focus on China’s digital economy research.

Therefore, this study selects data from listed companies between 2004 and 2021 to investigate the impact of digital transformation on corporate sustainable development levels. The marginal contributions of this research may include: (1) Systematically integrating digital transformation into the research framework of corporate sustainability drivers expands sustainability theory; (2) It reveals new pathways through which digitalization impacts sustainability by alleviating principal-agent conflicts (reducing managerial myopia) and mitigating information asymmetry (lowering financing costs, optimizing capital structures), deepening the application of agency theory and information asymmetry theory in the digital era; (3) It identifies and theoretically interprets heterogeneities in the impact of digital transformation on corporate sustainable development levels, including variations in property rights nature, industry characteristics, and marketization degrees. This not only enriches institutional theory and industrial organization theory but also provides practical guidance for enterprises to achieve digital governance and sustainable development. This translation maintains academic precision while adapting terminology (e.g., “principal-agent conflicts”, “information asymmetry”) and structural clarity for English readers. Key mechanisms and theoretical contributions are preserved through parallel structure and discipline-specific phrasing.

The following parts of the current investigation are arranged as: Section 2 presents the literature review and theoretical analysis; Section 3 details the research design, including data selection and model specification; Section 4 reports the empirical results, analyzing the impact of digital transformation on corporate sustainable development, its heterogeneity, and robustness checks; Section 5 examines the mechanisms through which digital transformation influences corporate sustainable development; and Section 6 concludes with key findings and policy recommendations.

2. Literature Review and Theoretical Analysis

2.1. Literature Review

Driven by scientific and technological innovation, mankind is stepping into the “digital” era, and the digital economy is constantly reshaping traditional industries and business models (Huang et al., 2025) [10], promoting the development of innovation and creativity. Many literatures have studied the impact of digital technology on economic development from a macro perspective. For example, the Internet can promote factor flow, thereby improving the efficiency of resource allocation (Bai et al., 2022) [4] and improving the total factor productivity of urban and manufacturing enterprises (Huang et al., 2019) [11]. The application of digital communication technology promotes innovation and improves labor productivity (Gaglio, 2022) [12]. At the same time, the government’s policy support for the digital economy is also conducive to promoting sustainable economic development (Teixeira & Tavares, 2022) [13]. Calabrese et al. (2024) studied the importance and urgency of industrial 4.0 (Digital Economy) policies at the EU level and at the level of individual EU Member States for the sustainable economic development of Member States [14].

Digital technology-related research has also begun to expand to the micro-enterprise level, especially in recent years. Enterprise digital transformation has received extensive attention from the academic community. Vial (2019) points out that digital transformation is essentially the process of deep integration of digital technology with the production and management of the enterprise and the process of bringing about profound transformational changes [15], which is not only the simple application of digital technology within the enterprise but also has a comprehensive impact on the enterprise structure and business model. It is not only the simple application of digital technology within the enterprise, but it also has a comprehensive impact on the enterprise structure, business model, employment model, production model, etc. (Yan et al., 2023) [16]. In the context of the digital economy, the enterprise organisational structure will gradually change from the traditional pyramid-like vertical and sectional features to networked and flat, which is not only conducive to promoting the effective transmission of data and information but also improves the enterprise’s ability to analyse digitally and respond to risks in a timely manner (Qi and Xiao, 2020) [17]. In terms of the specific impact of digitization on firms, studies have shown that digitization can optimize the human capital structure, improve innovation capacity, and, thus, increase firms’ total factor productivity (Huang et al., 2025) [18], as well as promote the division of labor in firms (Yan et al., 2022) [19], enhance stock liquidity (Wu et al., 2021) [1], and increase the demand for induced high-skilled labor, thus increasing labor income share (Xiao et al., 2022) [20], mitigating operational uncertainty, and reducing the amount of cash held in precautionary motives (Wu et al., 2023) [21]. In terms of measuring digital transformation, most of the literature refers to the methods of Zhao et al. (2020), Wu et al. (2021), or Yuan et al. (2021), etc., which use Python3.10 software to obtain the frequency of digitisation-related words in the annual reports of listed companies to measure the degree of digital transformation [1,22,23].

In recent years, the concept of corporate sustainability has continuously evolved and been enriched. The “Triple Bottom Line” theory proposed by Elkington (1997) remains a critical framework for understanding corporate sustainability [24], emphasizing that businesses must simultaneously address performance in three dimensions: economic, social, and environmental (Zhang et al., 2024) [25]. Scholars have adopted diverse methods and indicators to measure corporate sustainability levels. From the perspective of economic sustainability, commonly used indicators include corporate profitability, return on assets (ROA), and market share growth (Zaid et al., 2020; Dyllick & Muff, 2021) [26,27]. For instance, companies with high profitability are typically considered to have strong economic sustainability, as they can maintain stable operations and development over the long term. Regarding environmental sustainability, metrics encompass energy consumption intensity, greenhouse gas emissions, and waste disposal rates (Khaled et al., 2021; Sarkis et al., 2022) [28,29]. For example, reducing energy consumption intensity reflects a company’s ability to utilize energy resources more efficiently in production, thereby minimizing negative environmental impacts (Xu et al., 2024) [30]. Social sustainability is assessed through indicators such as employee satisfaction, labor productivity, and community engagement (Lim & Pope, 2022; Zhang & Yu, 2023) [31,32]. High employee satisfaction, for example, demonstrates effective human resource management practices, fostering long-term stability and development while highlighting a company’s commitment to employee rights (Crifo et al., 2019) [33]. This underscores the importance of social sustainability in corporate operations.

In recent years, there has been a gradual increase in research on the relationship between digital transformation and corporate sustainability. However, the conclusions of this research are still somewhat divergent. A number of studies have indicated that digital transformation has a beneficial impact on corporate sustainability. A study of Chinese listed enterprises by Li & Li (2024) found that digital transformation can significantly improve the productivity and resource utilisation efficiency of enterprise [34]. This, in turn, enhances the economic sustainability of enterprises whilst simultaneously reducing the environmental impacts of enterprises through the promotion of green technological innovation (Yan et al., 2020; Yan et al., 2023) and environmental sustainability (Yu et al., 2025) [16,35,36]. In terms of social sustainability, digital transformation has been shown to enhance communication and cooperation with stakeholders such as employees, suppliers, and customers and improve the social image and reputation of companies (Li, 2019) [37]. Nevertheless, a number of studies have identified the challenges and risks associated with digital transformation, including data security issues, the high cost of technology investment, and the mismatch between employees’ digital skills (Yang et al., 2025; Huang & Jiang, 2024) [38,39]. These factors have the potential to impede the achievement of corporate sustainability goals to a certain extent (Wamba et al., 2022) [40].

Based on existing research, digital transformation primarily influences enterprises by reducing information asymmetry and enhancing internal corporate governance. Meanwhile, corporate investment efficiency is mainly affected by information asymmetry and principal–agent problems. Mitigating these two issues can help improve investment efficiency and elevate the level of corporate sustainable development. This, to some extent, indicates a potential correlation between digital transformation and sustainable development in enterprises and also provides a foundation for the rationality of this research.

2.2. Theoretical Analysis

In essence, digital transformation impacts firms by reducing information friction, optimizing internal corporate governance, and, more specifically, reducing managerial short-sightedness, lowering the cost of corporate finance, reducing corporate leverage, and, thus, improving corporate sustainability. Specifically:

2.2.1. Enterprise Digital Transformation, Managers’ Short-Sighted Behavior and Sustainability Level

As the helmsman of the company’s strategy, managers play an important role in the sustainable development of the enterprise. According to the theory of Upper Echelons (Upper Echelons), the behavior of managers will determine the goals, behavior, and results of the enterprise (Hambrick and Mason, 1984) [41], and reasonable and effective capital expenditure is the basis for the survival and expansion of the enterprise, maintaining the endogenous growth dynamics, and maximizing enterprise value, and short-sighted behavior of managers will inevitably affect the enterprise investment and sustainable development to maximize corporate value, and managers’ short-sighted behavior will inevitably affect corporate investment and sustainable development. Under the influence of short-sighted behavior, enterprises will maximize short-term financial performance and stock prices at the expense of long-term interests and tend to choose short-term and high-yield projects (Holmstrom, 1999; Stein, 1989) when making investment decisions [42,43] while reducing long-term investments characterized by high investment costs, uncertain output, and intertemporal nature of returns, which will inevitably affect investment efficiency, resulting in overinvestment and underinvestment, which in turn is not conducive to the sustainable development of enterprises.

First, digital transformation can reduce firms’ perception of economic policy uncertainty and bring stable expectations to shareholders, making them more focused on the long-term development of the firm. Unlike the objective risk in investment, the perception of uncertainty is subjective (Bloom, 2014) [44], and this subjectivity can have a significant impact on investment decisions, as suggested by Keynes’ (1936) reference to “animal spirits” [45], and digital transformation can reduce this perception. This is because the causes of uncertainty perception can be attributed to information asymmetry and information processing ability (Fang et al., 2023) [46], while through digital transformation, on the one hand, enterprises can use ICT to enhance data integration and sharing between enterprises and suppliers and within enterprises to obtain more information. On the other hand, enterprises can effectively improve information processing capability through cloud platforms and artificial intelligence (AI) algorithms. Algorithms to effectively improve information processing capability, access, or predict various diversified demands, and they can more quickly adjust the strategic layout, control production costs, maintain supply chain security, and cope with various external shocks. Reduced uncertainty perception will bring stable expectations to shareholders and reduce short-term pressure on management evaluation, which is conducive to improving the efficiency of corporate investment.

Second, the digital transformation of enterprises can alleviate the principal–agent problem between shareholders and management. Under information asymmetry and incomplete contracts, shareholders as principals may not be able to select the optimal management team in advance, and it is difficult to effectively supervise and manage management after the fact, which makes management may take actions that are beneficial to themselves but not in the interests of shareholders and the enterprise, i.e., the problems of adverse selection and moral hazard coexist simultaneously (Jensen and Meckling, 1976) [47]. Digital transformation has significantly enhanced information transparency and enterprises’ ability to access information. Shareholders can leverage an active external candidate market, utilizing market mechanisms to select excellent management teams (Dou, 2017) [48], while also participating in corporate supervision and governance through platforms such as the internet and mobile devices at any time, thereby strengthening oversight of managers. Furthermore, by employing big data and algorithms, the scientific rigor and precision of enterprise decision-making can be greatly improved. This enables management to fully benefit from performance improvements within the company, reducing incentives to secretly pursue private interests. Therefore, alleviating principal-agent problems helps mitigate short-sighted behaviors among managers, ultimately enhancing corporate investment efficiency and the level of sustainable development.

2.2.2. Enterprise Digital Transformation, Financing Costs, and Sustainability Levels

The problem of “difficult and expensive financing” is the main problem faced by private enterprises in China (He et al., 2022) [49]. The problem of financing is also an important reason for the low investment efficiency of enterprises. The increase will also affect the investment strategy of enterprises, so that enterprises are biased towards investing in some high-yield but high-risk investments, which is not conducive to the sustainable development of enterprises (Falavigna & Ippoliti, 2022) [50].

The digital transformation of businesses can also reduce the cost of finance, thereby improving the sustainability of the business. Digital transformation can reduce the cost of financing for enterprises in three main ways. Firstly, the financing channels are broadened. With the help of digital technology and digital platforms, enterprises can obtain more financing channels. For example, through digital transformation, it is easier for enterprises to obtain green financial loan support, and the cost is lower (He et al., 2022) [51], which reduces the cost of corporate finance. Second, transaction costs are reduced. In the financing process, financial institutions need to process various types of information and assess financing risks, thereby incurring certain transaction costs. These costs are partially passed on to enterprises seeking financing, becoming the interest they must pay. Digital transformation improves the quality of enterprise information disclosure; financial institutions can conveniently obtain and handle information about the enterprise, and the corresponding reduction in transaction cost will also make the enterprise financing cost down. Third, the risk of default is reduced. In the case of information asymmetry, the enterprise may be due to investment failure caused by the funds that cannot be repaid, bringing the risk of default. In order to cope with this problem, financial institutions will increase the interest rate, which is the cost of financing, and, ultimately, the risk of default brought about to a certain extent by the default of the enterprise to bear, and the enterprise digital transformation to a certain extent will improve the financial institutions to the enterprise’s understanding of the information, reducing the enterprise financing costs. To a certain extent, this is conducive to the improvement of enterprise investment efficiency, which in turn promotes the improvement of the level of sustainable development.

2.2.3. Enterprise Digital Transformation, Leverage Ratio, and Sustainable Level

In circumstances where the enterprise leverage ratio is minimal, a suitable augmentation in the leverage ratio has the capacity to enhance the efficiency of enterprise investment. On the one hand, the enterprise has the option to augment leverage by means of borrowing and debt issuance to alleviate capital constraints and circumvent underinvestment. On the other hand, the enterprise can indirectly augment the proportion of management’s ownership of the shares by means of debt financing to mitigate the principal-agent problem (Jenson and Meckling, 1976) [47], which inhibits overinvestment. However, when a firm is highly leveraged, greater leverage can lead to diminished efficiency in investment. On the one hand, the existence of information asymmetry problems suggests that the higher the leverage of enterprises, the more likely they are to cause moral hazard problems. Research indicates that managers exhibit a propensity to allocate resources toward high-return, high-risk projects, primarily to transfer risk via overinvestment and maximize shareholder value over corporate interests (Lu et al., 2006) [52]. Stakeholders prioritize shareholder interests over corporate welfare, while creditors face challenges in effective oversight. Investment failures often lead to debtor enterprises being classified as non-performing assets by banks and creditors to evade claims, perpetuating a detrimental cycle of escalating leverage. On the other hand, elevated leverage increases corporate bankruptcy risk, default probability, and financing costs. (Unlike previously discussed market-driven financing costs, firm-specific behaviors can independently elevate financing costs). This phenomenon compromises firms’ financial risk-bearing capacity, adversely affecting investment efficiency and sustainable development. Digital transformation enhances enterprise productivity and profitability through improved production and management, thereby boosting self-financing capacity, reducing debt reliance, and lowering leverage ratios. Improved information accessibility and processing enhance creditor monitoring, effectively constraining firms’ leverage ratio increases. Optimizing leverage ratios within a reasonable range significantly improves enterprise investment efficiency and sustainable development.

Based on the above theoretical analyses, this research proposes the following research hypotheses:

Hypothesis 1.

The digital transformation of enterprises can improve the efficiency of investment.

Hypothesis 2.

Digital transformation will increase the sustainability level of firms by reducing the short-sighted behaviour of firm managers.

Hypothesis 3.

Digital transformation will increase the level of corporate sustainability by reducing the cost of corporate finance.

Hypothesis 4.

Digital transformation will increase the sustainability of firms by reducing leverage.

3. Research Design

3.1. Sample Selection and Data Sources

In this study, we select the panel data of A-share listed companies in Shanghai and Shenzhen from 2004 to 2021 and process the initial sample as follows: (1) exclude the data of financial industry and real estate companies (this is due to the strong financial attributes of China’s real estate enterprise industry, while the financial statement requirements of financial enterprises and general non-financial enterprises are quite different, according to the previous research literature, which will be excluded from the research samples of these two industries); (2) exclude the data of companies labelled as delisted companies; (3) exclude the data of companies with missing data and abnormalities; (4) exclude the data of companies labelled as “PT” and “ST” companies; (5) 1% and 99% deflated continuous variables; (5) exclude the data of enterprises labelled as “PT” and “ST”; and (6) Perform 1% and 99% winsorization on continuous variables As indicated by the aforementioned data processing methodology, following the elimination of 4158 data points, the initial sample size of 36,777 was reduced to 32,617 valid research samples. The annual financial reports of listed enterprises were obtained from Juchao Information Network, while the other enterprise data were sourced from China Stock Market & Accounting Research (CSMAR). The marketability index was retrieved from China Marketization Index Database.

3.2. Models

Based on the theoretical analysis above, in order to verify the impact of digital transformation on the total factor productivity of enterprises, this study refers to the research method of Chang et al. (2015) [53] and then constructs the following econometric model:

$$Invef{f}_{it}=\alpha +\beta Digita{l}_{it}+\gamma C_{it}+{\rho }_j+{\upsilon }_s+{\epsilon }_t+{\theta }_{it}$$

(1)

where subscripts i, j, s, and t are firm i, province j, industry s, and year t, respectively, and $Invef{f}_{it}$ is the level of firm sustainability. (This study employs investment efficiency to represent the sustainable development level of enterprises. This is because inefficient investment reflects the deviation of actual outcomes from the optimal state in capital allocation, precisely revealing the enterprise’s capability to balance short-term interests with long-term value. Consequently, this approach embodies the economic essence of sustainable development, which emphasizes efficient resource utilization and long-term strategic planning). $Digita{l}_{it}$ is the degree of digital transformation of the firm, C is a set of control variables, ${\rho }_j$ is the province fixed effect to control for regional characteristics that do not vary over time, ${\upsilon }_s$ is the industry fixed effect to control for industry characteristics that do not vary over time, ${\epsilon }_t$ is the year fixed effect to control for macroeconomic shocks that do not vary over time with the region and industry, and ${\theta }_{it}$ denotes the random error term.

3.3. Variable Definition

3.3.1. Level of Corporate Sustainability

In this study, we use the degree of inefficient investment to measure the sustainability level of enterprises (Inveff) according to the estimation method of Chen et al. (2011), and the smaller Inveff indicates the higher sustainability level of enterprises. Then, according to the research method of Chen et al. (2011), we use the book-to-market ratio (BM) as a Growth proxy variable (Penman, 1996) [9,54], and then use the same method to obtain the firm’s sustainability level (Inveff 2) for robustness testing.

3.3.2. Digital Transformation (Digital)

In view of the inadequacy of existing studies on digital transformation measurement and data availability, this study refers to the study of Pang & Liu (2022) [55], which used the ratio of digital investment (including software and hardware) to total assets to measure the digital transformation of enterprises (Digital).

3.3.3. Control Variable

Referring to previous studies (Chen et al., 2011) [9], the following control variables were added: (1) enterprise size (size), expressed as the logarithm of the enterprise’s total assets. (2) Cash ratio (cash), calculated as the ratio of cash and marketable securities to current liabilities. (3) Book-to-market ratio (BM), defined as (total equity minus debt) divided by (market price multiplied by total equity). (4) Shareholding ratio of top 10 shareholders (top10); the proportion of shares held by the top 10 shareholders was calculated by dividing their shareholdings by the total shares outstanding. (5) Number of board of directors (board), obtained by logarithmically adding 1 to the number of board members. (6) Ratio of sole director (Indep), defined as the ratio of independent directors to total directors. (7) Management cost ratio (Mfee), defined as administrative expenses divided by operating income. (8) Whether or not the two posts are combined (Dual), chairman and managing director takes the value of 1; otherwise it is 0. Descriptive statistics of the main variables are shown in

Table 1. Descriptive statistics of main variables.

Variable	Obs	Mean	Std. Dev.	Min	Max
Inveff	32,617	0.046	0.045	0.001	0.256
Lev	32,617	0.435	0.199	0.061	0.875
Size	32,617	22.078	1.254	19.833	26.018
Cash	32,617	0.771	1.245	0.022	8.247
BM	32,617	0.631	0.247	0.122	1.149
Top10	32,617	57.932	15.060	22.890	90.470
Indep	32,617	2.148	0.203	1.609	2.708
Mfee	32,617	0.372	0.053	0.300	0.571
Board	32,617	0.088	0.067	0.010	0.397
Dual	32,617	0.246	0.430	0.000	1.000
FirmAge	32,617	2.792	0.382	0.693	3.611
Growth	32,617	0.177	0.407	−0.737	4.330
Inveff	32,617	0.046	0.045	0.001	0.256
Lev	32,617	0.435	0.199	0.061	0.875
Size	32,617	22.078	1.254	19.833	26.018

. (9) Firm age (firmage), the natural logarithm of the firm’s establishment time plus 1. (10) Firm growth (growth), the growth rate of operating income.

4. Empirical Results and Analyses

4.1. Benchmark Regression Analysis

In order to test whether digital transformation affects the level of corporate sustainability, this study conducts a regression of based on model (1), and the empirical results are shown in

Table 2. Baseline regression results.

	(1)	(2)	(3)
	Inveff	Inveff	Inveff
Digital	−0.569 ***	−0.573 ***	−0.260 ***
	(0.054)	(0.055)	(0.060)
Size		−0.002 ***	0.000
		(0.000)	(0.000)
Cash		−0.001 ***	−0.001 ***
		(0.000)	(0.000)
BM		0.003 **	−0.005 ***
		(0.001)	(0.001)
Top10		0.000 ***	0.000 ***
		(0.000)	(0.000)
Board		0.005 ***	−0.002
		(0.002)	(0.002)
Indep		0.011 **	0.013 **
		(0.005)	(0.006)
Mfee		0.015 ***	0.025 ***
		(0.004)	(0.004)
Dual		0.002 **	0.003 ***
		(0.001)	(0.001)
Growth		−0.011 ***	−0.004 ***
		(0.001)	(0.001)
FirmAge		0.016 ***	0.015 ***
		(0.001)	(0.001)
Constant	0.047 ***	0.089 ***	0.039 ***
	(0.000)	(0.006)	(0.007)
Year FE	No	No	Yes
Province FE	No	No	Yes
Industry FE	No	No	Yes
Observations	32,617	32,617	32,617
R2	0.003	0.043	0.108

Note: **, *** represent the significance level at 5% and 10% respectively; Robust standard errors are in parentheses.

. As can be seen from column (1) of Table 2, when not adding control variables and not controlling for fixed effects, the coefficient of digital transformation is −0.569 and is significant at the 1% significant level, i.e., digital transformation improves the level of sustainable development of enterprises. After adding control variables and not controlling for fixed effects, the coefficient of digital transformation on the level of sustainable development of enterprises from column (2) is significant at the 1% level as −0.573, the research conclusion is still significant. After adding control variables, controlling for individual fixed effects and time fixed effects, the coefficient of digital transformation on the level of corporate sustainability from column (3) is −0.260 and is significantly negative at the 1% level of significance, which further suggests that the research conclusion is very robust, and this shows that digital transformation improves the level of corporate sustainability, and Hypothesis 1 is verified.

4.2. Robustness Check

4.2.1. Endogenous Analysis

In the baseline regression, this study includes fixed effects for year, province, and firm to some extent, mitigating endogeneity issues caused by omitted variables. To further address potential endogeneity concerns, this study employs an instrumental variable (IV) approach to reinforce the robustness of the findings. Specifically, the average digital transformation level of other firms in the same industry (excluding the focal firm) is used as the IV. The average digitalization of peer firms in the same industry reflects the sector’s overall digital awareness and adoption, which correlates with the firm’s digital transformation intensity (satisfying the relevance criterion). Meanwhile, individual firms’ investment efficiency is unlikely to reversely influence the industry-level digitalization trend, satisfying the exclusion restriction.

As shown in column (1) of

Table 3. Endogeneity analysis.

	(1)
	First Stage	Second Stage
	Digital	Inveff
Digital		−2.512 ***
		(1.312)
IV1	0.447 ***
	(0.087)
IV2
Control	Yes	Yes
Kleibergen-Paap rk LM	25.106 ***
Kleibergen-Paap rk Wald F	26.073 ***
Control	Yes	Yes
Province FE	Yes	Yes
Year FE	Yes	Yes
Industry FE	Yes	Yes
Observations	32,600	32,600

Note: *** represents the significance level at 10%, respectively, and the robust standard error is in brackets.

, the Kleibergen-Paap rk LM statistic rejects the hypothesis of tool variable underidentification, and the Kleibergen-Paap rk Wald F statistic rules out weak instrument issues, confirming the validity of the IV. In the second-stage regression, the coefficient of entrepreneurial firms’ digital transformation remains negative and significant, corroborating the reliability of the baseline estimates.

4.2.2. Replace the Explanatory Variables

As mentioned earlier, another measure of the level of corporate sustainability is obtained in the same way using BM as a proxy variable for growth (Inveff 2). The regression results are shown in column (1) of

Table 4. Robustness test.

	(1)	(2)	(3)	(4)
	Inveff 2	Inveff	Inveff	Inveff
Digital	−0.261 ***		−0.313 ***	−0.260 ***
	(0.061)		(0.076)	(0.069)
Digital 2		−0.128 ***
		(0.022)
Control	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Observations	32,617	32,583	26,572	27,999
R2	0.105	0.106	0.114	0.110

Note: *** represents the significance level at 10%, respectively; robust standard errors are in parentheses.

, and the coefficient of digital is −0.261 and is significant at the 1% significant level, which is basically consistent with the previous results, indicating that the conclusions of this study are well robust.

4.2.3. Substitution of Explanatory Variables

Referring to Wu et al. (2020) method [1], the logarithm of the word frequency of “digital” in the annual reports of enterprises was obtained through Python3.10 to measure the degree of digital transformation. The regression results are shown in column (2) of Table 4, and the coefficient of digital is −0.128 and is significant at 1% level of significance, indicating that the conclusions of this study are well robust.

4.2.4. Excluding Municipalities

Municipalities directly under the central government differ significantly from other cities in administrative status, economic development, and population size, potentially impacting the study’s conclusions. Therefore, the sample of firms located in four municipalities is excluded. The regression results are shown in column (3) of Table 4, and the coefficient of digital is −0.313 and is significant at the 1% level of significance, indicating that the conclusions of this research are well robust.

4.2.5. Excluding Data for 2020 and 2021

During 2020–2021, due to the impact of the epidemic, enterprise investment may be affected to some extent, and the accuracy of the conclusion may be affected. For the purpose of reserving samples as much as possible, the fixed effect is added to the benchmark regression to reduce possible bias, and it is eliminated in the robustness test. The regression results are shown in column (4) of Table 4. The coefficient of digital is −0.260, which is significant at the 1% significant level, indicating that the conclusion of this paper is very robust.

4.3. Heterogeneity Analysis

4.3.1. Heterogeneity of Inefficient Investments

That inefficient investment can arise from both underinvestment and overinvestment. Underinvestment occurs when the difference between actual and optimal investment expenditures is negative. Overinvestment occurs when the difference between actual and optimal investment expenditures is positive. This study argues that digital transformation can both mitigate underinvestment and discourage overinvestment. In order to test this conclusion, the sample was divided into two groups, underinvested and overinvested, according to the above criteria. The results are shown in

Table 5. Heterogeneity of inefficient investment and property rights.

	(1)	(2)	(3)	(4)
	Under Invest	Over Invest	NSOE	SOE
	Inveff	Inveff	Inveff	Inveff
Digital	−0.345 ***	−0.375 **	−0.316 ***	−0.165
	(0.033)	(0.154)	(0.073)	(0.109)
Control	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Observations	20,937	11,680	19,205	13,412
R2	0.189	0.152	0.121	0.130

Note: **, *** represent the significance level at 5% and 10% respectively; robust standard errors are in parentheses.

, with column (1) showing the regression results for the underinvestment sample. The coefficient of digital is −0.353 and significant at 1% level of significance, and column (2) shows the regression results for the sample of overinvestment; the coefficient of digital is −0.321 and significant at 10% level of significance, which validates the above conclusion; that is, digital transformation has a greater impact on the level of sustainability of underinvested firms than overinvested firms.

4.3.2. Property Rights Heterogeneity

SOEs often do not aim at profit maximization, and, being “quasi-official” in nature, have a natural endowment of political resources in terms of financing and risk tolerance. Non-state-owned enterprises, on the other hand, will pay more attention to the level of corporate sustainability due to the pressure of competition and the pursuit of profits, which leads to a difference in the level of attention paid to the level of corporate sustainability in digital transformation.

Therefore, according to the nature of property rights, the research divides the sample into two groups, State-Owned Enterprises (SOE) and Non-State-Owned Enterprises (NSOE), and conducts regression analyses separately, and the results are shown in Table 5. Column (3) shows the results of the regression of non-state firms; the coefficient of digital is −0.276 and is significant at 1% level of significance. Column (4) shows the results of the regression for non-state firms, and the coefficient of digital is negative but not significant, suggesting that digital transformation is more effective on the investment efficiency of non-state firms.

4.3.3. Industry Heterogeneity

Manufacturing and non-manufacturing enterprises have certain differences in production methods, supply chain characteristics, market competition, etc., so the effect of digital transformation on enterprise sustainable development may also be different. Based on this, the research divides the sample into two groups, manufacturing (NMF) and non-manufacturing (MF), according to the characteristics of the industry, and conducts regression analyses separately, and the results are shown in

Table 6. Industry heterogeneity and heterogeneity in degree of marketisation.

	(1)	(2)	(3)	(4)
	NMF	MF	LM	HM
	Inveff	Inveff	Inveff	Inveff
Digital	−0.125	−0.491 ***	−0.186 **	−0.339 ***
	(0.080)	(0.095)	(0.073)	(0.095)
Control	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Observations	10,416	22,201	16,560	16,057
R2	0.157	0.110	0.138	0.088

Note: **, *** represent the significance level at 5% and 10% respectively; robust standard errors are in parentheses.

. Column (1) shows the level of sustainability of non-manufacturing firms with a negative but insignificant coefficient, while column (2) shows the level of sustainability of manufacturing firms with a digital coefficient of −0.491 and significant at the 1% level of significance. It can be seen that digital transformation has a stronger impact on the level of sustainable development of manufacturing enterprises. The reason for this may be that, relative to non-manufacturing enterprises, manufacturing enterprises have a more complex industrial chain, higher technology intensity, and a greater space for the role of digital transformation, which in turn has a greater role in promoting the level of sustainable development of enterprises.

4.3.4. Heterogeneity in Degree of Marketisation

The degree of marketisation is closely related to a number of factors, including economic growth, political system, and social development, and it has a significant impact on microeconomic dynamics. Based on this, this study collects the marketisation indices published by the China Marketisation Index Database for the past years, divides the sample into the low marketisation group (LM) and the high marketisation group (HM) according to its median, and conducts regression analysis separately. The results are shown in Table 6; column (3) is the low marketisation group with a negative but insignificant digital coefficient, and column (4) is the high marketisation group with a digital coefficient of −0.339 and is significant at 1% level of significance. It is evident that digital transformation has a stronger effect on the level of sustainable development of enterprises. This is due to the fact that regions with high levels of marketisation have higher levels of factor development and allocation and sound institutions, which provide a good external environmental basis for the application and functioning of digital technologies.

5. Mechanism Test

Based on theoretical analysis, enterprise digital transformation influences sustainable development by mitigating managerial short-termism, reducing financing costs, and lowering leverage. To further validate this mechanism, OLS regression was employed for empirical testing.

5.1. Reducing Managerial Short-Sightedness

As shown earlier, digital transformation reduces economic uncertainty perception and mitigates shareholder–management agency issues, curbing managerial short-termism and, ultimately, enhancing enterprise sustainability. At the same time, it has been shown that language can reflect an individual’s perceptions, preferences, and personality, and that researchers can determine personal traits based on analysing the types of words and their frequency used by experimental subjects.

Therefore, this study refers to the research of Hu et al. (2021) [56], combined with the existing English “short-term horizon” thesaurus; the MD & A Chinese corpus is characterized. Developing a Chinese word set reflecting “managers’ short-sightedness” through machine learning, and based on the lexicographic method, the proportion of the total word frequency of the word is calculated and multiplied by 100 to get the short-sightedness indicator (myopia), where the larger the myopia value, the more myopic the manager’s behavioural decisions.

Then, further controlling for industry-fixed effects, province-fixed effects, and year-fixed effects, on top of controlling for the firm’s basic economic variables, OLS regressions are performed, and the empirical results are shown in columns (1) and (2) in

Table 7. Mechanism tests.

	(1)	(2)	(3)	(4)	(5)	(6)
	Myopia	Myopia	Cost	Cost	Lev	Lev
Digital	−0.643 ***		−0.032 *		−0.539 **
	(0.095)		(0.019)		(0.224)
Digital 2		−0.705 ***		−0.069 ***		−0.362 ***
		(0.039)		(0.007)		(0.079)
Control	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes	Yes	Yes
Observations	30,309	30,281	32,593	32,559	32,617	32,583
R2	0.139	0.145	0.266	0.268	0.540	0.540

Note: *, **, *** represent the significance level at 1%, 5%, and 10% respectively; robust standard errors are in parentheses.

.

As can be seen from column (1) in Table 7, the regression coefficient of myopia on digital was significantly negative at the 1% level. This shows that the digital transformation of companies significantly improves the sustainability of the organization. In order to further test the scientific validity and reasonableness of the findings of the study, this study further conducted an OLS regression of corporate managers’ myopic behaviour (Myopia) using digital 2, the regression coefficient of myopia on digital 2 remained significantly negative. Digital transformation improves a company’s sustainability credentials, which further validates the mechanism of action. This further validates the research hypothesis 2 of this study.

5.2. Reduced Financing Costs

Theoretical analysis indicates that digital transformation expands financing channels, lowers transaction costs, and reduces default risk, thereby decreasing financing costs and mitigating underinvestment and overinvestment issues. In order to test this mechanism of action, in addition to controlling for the firm’s basic economic variables, industry-, province-, and year-fixed effects were further included. An OLS regression of the cost of financing is then performed using the firm’s digital transformation variables. The empirical results are shown in columns (3) and (4) in Table 7. It should be noted here that the research refers to Zheng et al. (2013), this study defines corporate financing costs as the sum of interest expenses, handling fee expenses, and other financial expenses, divided by the total liabilities [57]. From columns (3) and (4), it can be seen, the coefficients of digital and digital 2 on cost are significantly negative at the 10 per cent level. This indicates that digital transformation lowers financing costs, validating the mechanism through which it enhances firm sustainability. This further validates the research hypothesis 3 of this study.

5.3. Reduced Leverage

Excessive leverage can lead to overinvestment or underinvestment. Digital transformation enhances production efficiency, increases profitability, reduces external financing needs, and strengthens creditor oversight, thereby maintaining leverage within a reasonable range and improving enterprise sustainability. To validate this channel of action, this study takes the enterprise debt ratio (Lev, total liabilities divided by total assets) as the proxy variable of the enterprise leverage ratio. OLS regression of firm leverage (Lev) is then performed using firm digital transformation (digital, digital 2). The empirical results are shown in column (5) and column (6) in Table 7. The regression coefficients of digital and digital 2 on Lev show that digital transformation significantly reduces enterprise leverage at the 1% level, supporting the theoretical mechanism that digital transformation enhances sustainable development by lowering leverage. This further validates the research hypothesis 4.

6. Conclusions and Implications

6.1. Conclusions

This study investigates the impact and mechanisms of digital transformation on enterprise sustainability, using A-share listed firms in Shanghai and Shenzhen from 2004 to 2021 as a sample, against the backdrop of digital technology-driven socio-economic changes and China’s challenges with inefficient and unsustainable development. The study demonstrates that digital transformation significantly enhances enterprise sustainability, a conclusion robust to various tests. Mechanism analysis reveals that digital transformation reduces managerial short-termism, lowers financing costs, and decreases leverage, thereby improving sustainability. Further analysis reveals that: (1) digital transformation alleviates both underinvestment and overinvestment; (2) its impact on sustainability is more pronounced in non-state-owned enterprises compared to state-owned ones; (3) the manufacturing sector benefits more significantly than non-manufacturing industries; (4) enterprises in high-marketization regions experience greater sustainability improvements than those in low-marketization areas, indicating that marketization enhances the positive effects of digital transformation.

6.2. Policy Implications

Based on the heterogeneous impacts and mechanisms of digital transformation on enterprises’ sustainable development identified in this study, the following policy recommendations are proposed:

Firstly, the implementation of diversified digital transformation support policies is recommended in order to optimise resource allocation. In accordance with the principle of “precise identification and categorized policy implementation”, the policy design process should prioritise addressing the imbalance in the economic implications of digital transformation. For instance, non-state-owned enterprises and manufacturing firms, which exhibit significant transformation advantages, should receive targeted policy incentives, such as tax reductions or subsidies.

The second point to consider is the reconstruction of governance frameworks oriented towards the market in order to remove the barriers to digital transformation that are of an institutional nature. In order to leverage the moderating role of marketisation in enhancing transformation efficacy, institutional innovations should focus on the elimination of obstacles to factor mobility and the improvement of governance rules. In the context of data element market construction, there is an urgent need for legislation that will establish a unified, open, and competitive market environment, thereby providing the soft infrastructure necessary for digital transformation. It is recommended that regulatory reforms adopt “principle-based supervision” for innovations in smart manufacturing and logistics whilst also establishing inter-departmental joint punishment systems to achieve closed-loop management of “data compliance-risk warning-joint disposal”.

Thirdly, the execution of gradual digital transformation strategies at the enterprise level is to be undertaken. In order to address the dual challenges posed by managerial cognitive biases and financing constraints, enterprises ought to establish a triple-integrated transformation system (“strategy–organization–technology”). From a strategic perspective, it is recommended that mandatory disclosure requirements for digital transformation reports be imposed on listed companies, with digital leadership integrated into executive performance assessments. The implementation of “digital twinning + real-time feedback” decision systems has the potential to mitigate the occurrence of managerial opportunism through the utilisation of virtual simulation experiments. From an organizational perspective, enterprises ought to optimise dynamic allocation mechanisms for digital innovation resources. In doing so, they should prioritise high-synergy domains based on organizational inertia and scene compatibility. In addition, it is imperative to develop an evaluation system that encompasses the dimensions of “digital transformation maturity–ESG performance”. This system should be integrated with data, carbon footprint accounting, and algorithmic ethics reviews, which are to be incorporated into corporate social responsibility disclosures. The objective of this integration is to nurture intrinsic incentives for the cultivation of sustainable digital ecosystems.

6.3. Limitations and Future Research

The present study is not without its shortcomings. Firstly, it does not fully incorporate important dimensions such as customer relationship management and supply chain synergy. This leads to an insufficiently comprehensive analysis of the mechanism of action. Secondly, it does not pay enough attention to the differences in the transformation of manufacturing segments (e.g., high-energy-consuming versus high-technology industries) and SMEs and non-listed enterprises.

Concomitantly, this constitutes a research direction for future studies. It will incorporate hitherto unexplored perspectives, such as industrial chain digitisation and consumer green preference. Moreover, cross-country comparative research will be conducted, and attention will be paid to the dynamic effects of emerging technologies, such as generative AI. The objective is to provide a more universal reference for the theory and practice of digital transformation and enterprise sustainable development.