Digital Finance, ESG Performance, and Financial Performance in Chinese Firm Levels: The Pathway to Sustainability

Abstract

Given that digital finance is critical for achieving sustainability, this study seeks to probe the mechanisms for using digital finance to solve the triple-bottom-line dilemma of sustainability. This paper examines the inner influence mechanisms of digital finance on ESG (environmental, social, and governance) and financial performance. The results show digital finance is conducive to ESG performance while indirectly enhancing firms’ short- and long-term financial performance via ESG. Further, digital finance usage depth is the primary enabler for ESG and short-term financial performance. The mechanism analysis reveals that the positive relationship between digital finance and ESG will be enhanced by upgrading green innovation and boosting digital transformation. Moreover, heterogeneous analysis states that digital finance usage depth has a more pronounced positive role on ESG and financial performance in large firms and SOEs (state-owned enterprises) while coverage breadth positively affects ESG and is more pronounced in small firms and non-SOEs. This paper expands knowledge about digital finance via sustainability practice pathways.

1. Introduction

Digital finance is a novel paradigm of finance service used for achieving high coverage, low cost, and sustainability and has been widely used around the world [1]. The rapid development of the global economy is accompanied by a high level of pollutant emissions and resource consumption [2]. As environmental pollution and resource scarcity have become a serious challenge for countries around the world, especially China, sustainable growth has grown into an inevitable tendency [3]. The essential role of digital finance in achieving sustainable growth has been emphasized by the G20 Digital Economy Development and Cooperation Initiative. Meanwhile, the 14th Five-Year Plan emphasizes that digital finance is a key driver of Chinese sustainability. As a practice pathway to achieving sustainability, it is necessary to examine whether digital finance can produce environmentally, socially, and economically sustainable bonuses for Chinese firms. If so, what are the underlying mechanisms?

In the era of sustainability, the potential pros and cons of digital finance have become vital research topics. Most scholars have acknowledged that digital finance can offer an all-win sustainable solution for the economy, society, and the environment [4]. Economists have highlighted that digital finance can break the barriers of information asymmetry with its resource and information effects [5]. In addition, the inclusive and long-tail effects of digital finance can add vitality to the market economy [6]. However, research has also stated that digital finance is an uncertainty. Specifically, digital finance’s development has non-negligible risks like illegal fundraising, financial fraud, and cybercrime [7]. Meanwhile, the application of digital technology will accelerate the proliferation of risks, lead to resource mismatches, and crowd out sustainable development factor resources [8]. As a whole, studies have neglected the varying sub-dimensional heterogeneity of digital finance and the underlying mechanisms between digital finance and sustainability, causing inconsistent evidence about digital finance. Accordingly, this work expands existing research about the role of digital finance by examining the influence mechanisms of heterogeneous digital finance on sustainability.

Recently, ESG (environmental, social, and governance), which reflects firms’ environmental, social, and governance performance, has become the main indicator of firm sustainability. Firms can provide ESG information for measuring the contribution of firms to sustainable value creation [9]. However, the WCED (World Commission on Environment and Development) emphasizes that sustainability considers the coordination of development among society, the economy, and the environment. It indicates that firms need to follow the “triple bottom line” principle, that is, balance environmental, social, and economic activities. Studies have also highlighted that firms’ environmental and social activities are always expensive, creating administrative burdens and exacerbating agency conflicts. It is worth noting that financial performance considers the firm’s economic status, which is both a survival precondition and an operating goal. As a result, environmental, social, and financial performances fall into a game dilemma. Against this background, we construct a theoretical framework that integrates digital finance, ESG, and financial performance to investigate the correlation among digital finance, ESG, and financial performance in Chinese firms. Given that Chinese digital finance is globally leading and has a relatively mature measurement index, both the industry and scholars have increasingly focused on the development of digital finance and its mechanisms. In summary, this paper answers the following questions and reports the relevant management implications. Does heterogeneous digital finance benefit both firms’ ESG and financial performance? What inner mechanisms are involved? Are there size and ownership heterogeneous in digital finance that affect ESG and financial performance?

In detail, this paper has the following contributions. First of all, we extend our cognition of the impact of digital finance by incorporating digital finance, ESG, and financial performance into a comprehensive theoretical framework. Given that financial performance is an existential prerequisite and operating goal for firms [10], the nonlinear mechanism of ESG in the connection between digital finance and financial performance is examined to enrich the prevalent study. We also expand the existing research by exploring the sub-dimensional heterogeneity of digital finance and considering both short- and long-term financial performance. Secondly, we enrich the investigation of the internal mechanisms through which digital finance moves ESG. This paper indicates the vital utility of green innovation and firm digital transformation in transmitting the avail of digital finance. We also incorporate ownership and firm size into heterogeneity analysis, amplifying the research depth by considering firm resource endowments. Thirdly, sustainability is of great significance to emerging economies, and ESG and financial performance systematically cover the “triple bottom line” principle of sustainable development. This study clarifies the transmission pathway of digital finance in enhancing ESG and financial performance in emerging economies, i.e., China. It provides a theoretical foundation and empirical insights for the implementation of digital economy strategies and policies promoting environmental, social, and economic sustainable development.

2. Literature Review

Digital finance uses emerging technologies, i.e., cloud computing and artificial intelligence, to blaze new trails in financial products and service processes, realize credit, payments, investments, insurance, etc., and financial activities, which are a new financial development mode contributing to fostering sustainability [4,11]. As a product of the information age, the digital economy can overcome traditional finance’s reliance on physical branches and provide greater geographic coverage and cost advantages; it has garnered attention from global industry and scholars. The popularity of digital finance has also been increasing annually [6]. Ozili defined digital finance as all products, services, technologies, and/or infrastructures that allow individuals and companies to access payment, savings, and credit services via the internet without needing to visit bank branches or deal directly with financial service providers [12]. Guo et al. have noted that, broadly, digital finance refers to financial activities conducted by traditional financial institutions and internet companies using digital technologies, while narrowly, it generally refers to new financial models developed by internet companies [13]. Digital finance measurement, as the basis on which firms formulate business strategies, has become topical for scholars. There are two main methods for measuring digital finance. One is text mining based on search engine news databases, which constructs financial indices by calculating the frequency of keywords related to digital finance [14,15]. However, this index can only be constructed at the national level, and the selection of keywords is subjective, lacking a standardized and comparable benchmark [16]. The other method is the Digital Financial Inclusion Index compiled by the Digital Finance Center at Peking University [1,5,17,18,19]. Its data are provided by Ant Group, the leading digital finance company globally, making it highly representative and reliable, and it has become a mainstream indicator for measuring digital finance [13]. Although there is much literature on the research of digital finance, the main focus of scholars is the general impact of digital finance on the economy and environment. Existing research presents varying outcomes regarding the interplay among digital finance, ESG, and financial performance.

Concerning digital finance and ESG performance, there are two views. The first view is based on RBV theory, which sustains that digital finance positively affects ESG, positing that digital finance can bring unique resources and capabilities, thereby reducing information asymmetry, relieving financial constraints, and guiding firms to participate in more ESG activities [1]. Moreover, the inclusive features of digital finance can facilitate firms to increase investment in green technologies [20]. This fosters the production and use of clean technologies [21], contributing to mitigating environmental degradation [22,23], improving firm goodwill [24], and thereby being beneficial to firm ESG performance. The second view asserts that the long-tail effect of digital finance makes its ESG outcomes uncertain. Specifically, digital finance provides financial support to more firms, especially industrial firms and small–medium-sized firms. This facilitates their access to funding for production and services. However, the increased production inevitably accompanies an upswing in energy consumption, whose ESG benefits are questionable [25]. It has also been pointed out that digital finance stimulates market consumption, leading to a surge in the usage of consumables like air conditioners and automobiles, thereby escalating environmental pressure [26]. Meanwhile, the rapid growth of digital finance brings about risks such as usury and illegal fundraising, which are not conducive to ESG [7]. Altogether, these two views hold that digital finance invokes both challenges and opportunities for firm ESG.

There are three main perspectives on digital finance and financial performance. Most studies advocate that digital finance will bring new economic growth points [27]. Specifically, with the features of wide coverage, strong penetration, and high innovation, digital finance can boost firms’ financial performance by realizing scope economy and scale economy [28], which is supported by the RBV and economic growth theories. Conversely, some research indicates that digital finance may lead to increased financial investment and accelerate risk diffusion due to its inclusive and digitized nature, which will result in resource misallocation and firms’ financial burden [29]. The digital divide also cannot be ignored. Recent studies claim that the utility of digital finance for financial performance is uncertain and controversial, mainly through varying nonlinear mechanisms and heterogeneity among digital finance types [30,31]. Liu et al. stated that the nonlinear mechanisms of digital finance on financial performance are mainly related to firm resource endowment and external environmental factors [32]. The disputable link between digital finance and financial performance is that firms have various resource endowments and external surroundings [33,34]. Additionally, research has also detected that the heterogeneity of digital finance’s measurement indicators will also cause differences in the impact on financial performance [35]; thus, the topic of digital finance and financial performance warrants further discussion.

Recently, the literature has realized the pivotal role that digital finance plays in sustainable development [22,36]. Studies have presented an increased focus on the influence of digital finance on sustainable development, which integrates environmental, social, and economic considerations. Cao et al. argued that digital finance will boost green technology innovation by broadening access to finance, reducing information asymmetry, and spreading risks, thereby improving energy–environmental performance and contributing to sustainability [6]. Research has further proposed that digital finance can drive financial resources toward low energy consumption and emissions via technological innovation and transformation projects, which contribute to achieving green growth [37]. Nevertheless, existing research primarily adopts single factor indexes to evaluate sustainable development, for instance, ESG [1], carbon emissions [22], or building an integrated index based on green total factor productivity [15,38] and sustainable development index systems [10,39]. Few studies have examined whether digital finance can simultaneously produce social, environmental, and economic dividends, and consider the interactions between the environment, society, and economy. Considering that existing studies have not yet considered the intrinsic mechanisms and heterogeneous conditions of the role of digital finance, we adopt ESG and financial performance as indicators for an alternative dependent variable and investigate the inner mechanisms of digital finance in enhancing the all-win situation of ESG and financial performance across types and nonlinear perspectives.

3. Research Hypotheses

3.1. Digital Finance and ESG Performance

As energy and environmental pressures intensify, scholars are increasingly focusing on the influence of digital finance on ESG. Digital finance provides a variety of mechanisms to achieve ESG. Building on the RBV, we argue that digital finance gives firms the resource and capability advantages that benefit ESG. Firstly, digital finance can effectively reduce information asymmetry via the use of digital technology [40,41]. Specifically, digitization provides stakeholders with more timely and transparent information to protect their rights and interests and enables firms to gain green guarantees and good trust, thereby boosting CSR (corporate social responsibility) and facilitating corporate governance. In the same vein, digital finance helps to reduce agency costs and promote goodwill, which in turn increases ESG [24].

Secondly, digital finance broadens funding sources and eases financing constraints [22,42], which can bankroll firms to implement green innovation [43]. The ultimate goal is to guide production toward environmental goals [2]. Therefore, digital finance will generate sound ESG, and the first hypothesis is as follows:

H1. 

Digital finance positively affects ESG performance.

3.2. Digital Finance and Financial Performance

Digital finance, as the core of modern economic development, has a vital impact on financial performance, and existing research has provided mixed results. Despite the controversial background on this topic, we advocate that digital finance improves financial performance via resource information advantage and technology advantage mechanisms. First, digital finance possesses traits such as digitization, low cost, and inclusivity, which can bring it an information and resource effect that reduces economic risks [5,44]. The adoption and popularization of digital finance have helped firms develop scale economy and scope economy advantages, thereby increasing their financial benefits [39].

Second, the literature has indicated that digital finance depends on information technology, i.e., big data, blockchain, etc. [45]. As the product of digital technology innovation, digital finance utilizes the long-tail effect of technology platforms to significantly reduce marginal costs, which can also stimulate market vitality by easing financing constraints and expanding service scope [39]. Meanwhile, the research advocates that digital finance can drive financial service products’ upgrading and facilitate industrial transformation, which can maintain competitiveness and become a new engine of economic growth [46]. Taking into account the analysis, we formulate the second hypothesis as follows:

H2. 

Digital finance positively affects financial performance.

3.3. The Mediating Role of ESG

ESG, which is derived from ethical and responsible investing, represents an important guide for the capital market in assessing the investment worthiness of a firm based on the performance of environmental, social, and corporate governance [47]. Focusing on ESG helps firms shift from a sole focus on profit maximization to a diversified set of objectives in environmental protection and social responsibility and pursue sustainability, which can be realized by controlling product and service quality, enhancing stakeholder confidence, and reducing financing costs [48]. Nevertheless, given that profitability is a business survival prerequisite and business objective [10], determining whether ESG contributes to financial performance is a necessary and hot topic [49]. Existing research points out that fulfilling ESG performance can assist firms in improving goodwill and forming resource advantages, which boost social value and firm value [50]. Nevertheless, empirical investigations examining the impact of ESG on financial performance remain controversial. ESG, as a sustainable development practice, requires long-term investment, increasing shareholder costs and limiting investment opportunities, thus adversely affecting profitability [51,52,53].

We hypothesize that digital finance is critical for enhancing firms’ ESG performance and improving financial performance, as discussed in H1 and H2. When ESG performance is supported by effective digital finance, this socially friendly digital finance provides firms with a stronger competitive advantage. Meanwhile, the comprehensive perspective in the study of ESG and financial performance is further endorsed by stakeholder theory. This theory asserts that the success of digital finance hinges on meeting the legitimate needs of all stakeholders, encompassing the environment, society, and economy, to form sustainable mechanisms for achieving financial performance [47]. Generally speaking, digital finance simplifies the firm financing and sales process while providing more transparent and safer financing and sales channels, which can effectively reduce firm risks and break the geographical boundaries of sales and use of products and services. Adequate and reliable capital investment helps firms focus on the marketization of products and services, taking into account the needs of internal and external stakeholders.

Overall, the study states the potential for an indirect impact of digital finance on financial performance by ESG performance. We hold that digital finance has the capacity to enhance firms’ ESG performance, which can reduce agency costs and advance reputation [24]. Total factor productivity can be improved by optimizing the production and service processes, which will transform into the firms’ survival advantages and benefits. Therefore, we propose that ESG entails how digital finance can boost firms’ financial performance and formulate the third hypothesis as follows:

H3. 

ESG performance mediates the relationship between digital finance and financial performance.

3.4. The Mechanism Effect

As discussed in H1 above, green innovation and digital technologies may be crucial factors in enhancing firms’ ESG performance. On one hand, digital finance significantly alleviates the financing constraints and broadens funding sources. This enables firms to undertake green innovation activities, such as saving energy, reducing emissions via eco-friendly technology transformations, and increasing productivity [2]. Therefore, digital finance supports firms in implementing green innovations, which can improve ESG performance by reducing ecological pressure on the firms.

On the other hand, another potential mechanism through which digital finance can promote ESG is by facilitating the firm’s digital transformation. Digital finance relies on the implementation of digital technology, and firms’ digital transformation has become essential for the effective development and application of digital finance in business activities. Digital transformation not only enhances employee satisfaction and work efficiency by streamlining office processes but also improves stakeholder satisfaction via more timely and transparent information exchange. This, in turn, helps firms build goodwill and trust, thereby enhancing ESG performance. Accordingly, H4 is proposed.

H4. 

The positive impact of digital finance can be achieved by promoting green innovation and supporting digital transformation.

3.5. The Heterogeneous Effect

Research emphasizes that the utility impacts of firm development vary with firm organizational characteristics such as firm size and ownership [46]. Specifically, compared to small firms, large firms have stronger financial resources, technological foundations, and workforce scales, which makes them more likely to achieve economies of scale and benefit from cumulative advantages. Digital finance can alleviate financing constraints for small firms by providing new funding channels and financial services.

Ownership affects a firm’s resource endowment, strategic positioning, and governance structure [3]. SOEs, due to their political connections, receive more policy support and resource allocation, while also bearing greater social responsibilities. In contrast, non-SOEs face intense market competition, and digital finance can effectively address financial barriers such as investment budget constraints, thereby benefiting their survival and development. Consequently, H5 is postulated.

H5. 

Digital finance has heterogeneous impacts on ESG and financial performance across firms’ size and ownership types.

4. Methodology and Variables

4.1. Variables

4.1.1. Explained Variables

A firm’s ESG performance (ESG) is measured using the CESG (Environmental, Social and Governance Database of Listed Company) Database from the CNRDs (Chinese Research Data Services) platform. CESG referred to the design ideas of the internationally important ESG rating databases: MSCI (Morgan Stanley Capital International) ESG database and Bloomberg’s ESG Rating System [9] and combined them with the Chinese-specific context. Refer to Appendix A for detailed information. The ESG performance index classifies six dimensions: Charity, volunteer activities and social controversy, Corporate governance, Diversity, Employee relations, Environment, and Products. Each aspect encompasses strengths and concerns data. Following [54], we construct an ESG evaluation that pluses strengths and decreases concerns for each firm year. By standardizing annual strengths (concerns) for each firm and category, each category index ranges from −1 to 1. Thus, the firm ESG index, as the explained variable, ranges from −6 to 6, while the mean value is a mere 2.3076 (the mean values of sub-category indicators of ESG are as follows: charity, volunteer activities, and social controversy are 0.3346; corporate governance is 0.3486; diversity is 0.2668; employee relations is 0.5876; environment is 0.3348; and products is 0.4252), implying that Chinese firms are still relatively behind in implementing environmental and social governance.

Firms’ financial performance is examined through short-term and long-term financial performance in our study. Firms’ short-term financial performance is evaluated via the widely employed return on assets (ROA), a common measure in research on Chinese listed firms [46] and reflects the short-term profitability of a business using its resources and assets to create value. ROA, defined by net profit to total assets, is a critical index that represents the firm’s organizational ability due to its stability and reliability [10,53]. Firms’ long-term financial performance is measured by economic value added (EVA), which reflects the economic profits generated from the investment and can help investors identify investment opportunities. Referring to [55], we define EVA as the logarithm of the net operating profit after tax less total capital multiplied by the weighted average cost of capital, which is an important financial index that reflects the value created by the firms for shareholders.

4.1.2. Explanatory Variables

Digital finance (Df) is measured using the digital finance inclusion index compiled by Peking University’s Digital Finance Center. Given that it employs Ant Financial transaction account big data, the digital finance inclusion index is widely used in China, a common measure in research about digital finance on Chinese firms [1,11,17,18,19,46,56] and regions [5,6,44]. Ant Financial is considered the top 1 global digital finance firm with its transaction data having advantages of high coverage, strong pertinence, low error, and full dimension. Using its data to construct a digital finance index is reliable. Following [1,56], we adopt the logarithm of the prefecture-level city digital finance inclusion index where the firm is located to measure firm digital finance. (Digital finance leverages technologies such as big data, blockchain, and artificial intelligence to drive significant transformations in insurance, payments, currency, and other financial sectors. The development and application of firms’ digital finance are closely linked to the construction of regional technology platforms, which can be assessed via regional transaction data. Existing research primarily utilizes the digital finance index of the city or province in which firms are located to gauge their digital finance level. However, we have to acknowledge that the digital finance index on a firm-by-firm basis remains a key focus for ongoing attention and research) Peking University Institute of Digital Finance constructed a sub-evaluation index that includes the coverage breadth index (DfI), the usage depth index (DfII), and the digitization degree index (DfIII) [6,13]. The mean Df value is 5.2899, increasing with each passing year from 4.2451 in 2011 to 5.7708 in 2021, meaning that Chinese digital finance remains in a rapid development stage.

Control variables. To capture organizational factors, we employ firm size (Size), labor intensity (Labor), ownership concentration (Con), and state-owned enterprises (SOE) as the control variables. Refer to

Table 1. Variables definition and descriptive statistics.

Variables	Definitions	Mean	SD	Min	Max
ESG	ESG performance index from the CNRDs platform	2.308	0.959	−1.580	5.604
ROA	Net profit to total asset	0.044	0.050	−0.467	0.482
EVA	Logarithm of economic value added	15.394	0.232	0.000	15.834
Df	Logarithm of municipal-level digital finance inclusion index	5.282	0.455	3.193	5.885
DfI	Logarithm of the coverage breadth index	5.290	0.446	1.856	5.918
DfII	Logarithm of the depth of usage index	5.253	0.465	2.639	5.870
DfIII	Logarithm of the digitization degree index	5.253	0.650	1.221	6.365
Size	Logarithm of a firm’s total assets	14.225	1.446	10.823	19.083
Labor	Number of employees to revenue	0.010	0.011	0.000	0.188
SOE	A dummy variable, setting a value of 1 for state-owned and 0 for others	0.672	0.470	0.000	1.000
Con	Shareholding ratio of the largest shareholder	0.376	0.155	0.034	0.865

for the definitions.

4.2. Data Sources

This paper surveys listed firms on the A stock markets of the Shanghai and Shenzhen Stock Exchanges. The ESG and innovation data are obtained from the CNRDs platform; firm digitization degree data are extracted from the CnOpenData database, and general information on financial performance and control variables is obtained from the CSMAR database. The digital finance data are provided by the Peking University Institute of Digital Finance. After excluding samples (1) of ST and PT, (2) with too many missing values, and (3) of the financial industry, the final research sample contains 425 listed firms from 2011 to 2021 or 4675 firms years. Table 1 shows the variable descriptive statistics. We can see that the mean, standard deviation, minimum and maximum values of ESG are 2.308, 0.959, −1.580, and 5.604, respectively. The results show that there is variability in ESG across firms. The mean and the standard deviation values of ROA are 0.044 and 0.050, indicating that there exist significant differences in profitability among listed firms. The mean value of EVA is 15.394, while the minimum value is 0 and the maximum value is 15.834, meaning that the economic added value of different enterprises differs greatly, and the economic added value of the sample listed enterprises is relatively high. The mean, standard deviation, minimum, and maximum values of DF are 5.282, 0.455, 3.193, and 5.885, respectively, indicating that Chinese cities attach importance to the development of digital finance, and the sample level is in the middle to upper level.

4.3. Econometric Specification

Following [57], we consider the dataset’s characteristics and endogeneity interference due to omitted variable bias. Thus, this paper employs the two-way fixed-effect panel data model to explore the impact of digital finance on ESG and financial performance:

$$y_{it}=\alpha +x_{it}\beta +z_{it}\sigma +{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(1)

where y_it means the explained variables (ESG, ROA, and EVA, respectively); x_it is the explanatory variable (Df), and sub-indicators, DfI, DfII, and DfIII, are further adopted to investigate the heterogeneous role mechanisms of digital finance; z_it is the control variables. ${\gamma }_t$ and ${\mu }_i$ denote time and individual fixed effect, respectively; and ${\epsilon }_{it}$ is the random error. $\alpha$, $\beta$, and $\sigma$ are parameters to be estimated. Subscripts i and t proxy the firm and year.

This paper aims to construct a theoretical framework linking digital finance, ESG, and financial performance. Zhang et al. stated that financial performance serves as both a prerequisite and ultimate goal for the survival of a firm [10], while Truant et al. stated that ESG considers the firm’s environment, social, and governance performance, which reflect firm sustainability [58]. Hence, it is crucial to investigate the role of ESG in the digital finance–financial performance relationship. It can help achieve a firm’s economic growth in a socially, environmentally, and governance-friendly way. Specifically, this paper uses the stepwise regression approach to assess the mediating effect of ESG. The models used are

$$Y_{it}=\alpha +x_{it}\beta +z_{it}\sigma +{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(2)

$$ES{G}_{it}=\alpha +x_{it}\beta +z_{it}\sigma +{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(3)

$$Y_{it}=\alpha +x_{it}\beta +ES{G}_{it}\chi +z_{it}\sigma +{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(4)

where Y_it means the financial performance variables (ROA and EVA, respectively).

5. Results and Discussion

5.1. Benchmark Regression Results

Table 2. Benchmark estimation results.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
	ESG		ROA		EVA
Df	0.337 *** (0.032)		−0.010 *** (0.002)		−0.005 *** (0.002)
DfI		0.314 *** (0.080)		−0.010 ** (0.004)		−0.003 (0.004)
DfII		0.428 *** (0.081)		0.011 ** (0.005)		0.006 (0.004)
DfIII		−0.260 *** (0.041)		−0.007 *** (0.002)		−0.005 ** (0.002)
Size	0.236 *** (0.034)	0.223 *** (0.034)	−0.010 *** (0.002)	−0.011 *** (0.002)	0.002 (0.002)	0.001 (0.002)
Labor	−2.417 (2.115)	−1.571 (2.104)	−1.126 *** (0.117)	−1.107 *** (0.117)	−0.542 *** (0.111)	−0.528 *** (0.111)
SOE	−0.017 (0.096)	−0.027 (0.096)	0.003 (0.005)	0.003 (0.005)	0.002 (0.005)	0.002 (0.005)
Con	−0.189 (0.199)	−0.148 (0.198)	−0.015 (0.011)	−0.013 (0.011)	−0.003 (0.011)	−0.002 (0.011)
Constant	−2.795 *** (0.415)	−3.387 *** (0.425)	0.251 *** (0.023)	0.245 *** (0.024)	15.404 *** (0.022)	15.397 *** (0.022)
Fixed Effect	Yes	Yes	Yes	Yes	Yes	Yes

Standard error statistics in parenthesis. *** and ** indicate the significance levels of 1% and 5%, respectively.

displays the regression results of Equation (1). The results of Model 1, Model 3, and Model 5 show that digital finance has a significant and positive impact on ESG, while it exhibits a significant and negative impact on ROA and EVA. Thus, H1 is supported, but H2 is not, meaning that digital finance’s development will drive higher Chinese ESG performance but is not conducive to their short- and long-term financial performance. This is primarily attributed to the fact that digital finance is accompanied by high investment costs, increased market uncertainty, and increased market competition, which affect firm earnings [59]. Additionally, the characteristics of digital finance, such as social inclusion, environmental sustainability, and governance transparency, make it have a positive impact on firm ESG.

Digital finance is composed of the sub-evaluation indices of the coverage breadth (DfI), the usage depth (DfII), and the digitization degree (DfIII). We then further investigate the heterogeneous effect of the three aspects. In terms of the impact of digital finance on ESG, Model 2 results indicate that DfI and DfII have a significantly positive impact on ESG. In contrast, DfIII does not. The findings indicate that enhancing coverage breadth and usage depth is beneficial to ESG. Digital finance has the long-tail effect, which makes it supply more bankrolls for firm ESG, with an energetic bearing on ESG. However, augmenting the digital financial digitization is detrimental to ESG, in contrast to [60]. This may be due to the speedy progression of Chinese digital technology, which is enough to support the growth of digital financial services. Concerning the impact of digital finance on financial performance. We can conclude from Model 4 that DfI and DfIII have a significant and negative impact on ROA, whereas DfII has a significant and positive impact on ROA. According to Model 6, we can obtain that DfIII has a significant and negative impact on EVA, whereas DfI and DfII have no significant impact on EVA. Overall, although the digital finance coverage breadth will have a direct negative impact on the short-term financial performance, it does not have a significant direct impact on the long-term financial performance, so the impact of digital finance on the firm’s financial performance still needs to be further discussed. These results expand the insights of [46], pointing out that the rise in digital financial usage depth can effectively expand firms’ financial resources and bring good immediate financial benefits.

Additionally,

Table 3. Mediating effects tests results of ESG on digital finance and short- term financial performance (ROA).

Panel A: Stepwise Regression Estimation	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9	Model 10	Model 11	Model 12
	ROA				ESG				ROA
ESG									0.003 *** (0.001)	0.003 *** (0.001)	0.003 *** (0.001)	0.003 *** (0.001)
Df	−0.010 *** (0.002)				0.390 *** (0.027)				−0.011 *** (0.002)
DfI		−0.010 *** (0.002)				0.420 *** (0.028)				−0.012 *** (0.002)
DfII			0.010 *** (0.001)				0.386 *** (0.030)				0.009 *** (0.001)
DfIII				−0.008 *** (0.001)				−0.186 *** (0.017)				−0.007 *** (0.001)
Control Vars.	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Panel B: mediating effects analysis
Mediating variable: ESG	Total effects				Indirect effects				Direct effects
	−0.010	−0.011	0.010	−0.008	0.001	0.001	0.001	−0.001	−0.011	−0.012	0.009	−0.007
	Proportion of Total effect mediated (%)				Indirect to direct effect (%)				Total to direct effect (%)
	10.000	9.091	10.000	12.500	9.091	8.333	11.111	14.286	90.909	91.667	111.111	114.286

Standard error statistics in parenthesis. *** indicate the significance levels of 1%. Referring to [62], the indirect effects of ESG occupy a relatively high proportion of the total effect mediated (10.000%, 9.091%, 10.000%, and 12.500%, respectively); therefore, ESG is an eligible mediate variable in the relationship between digital finance and ROA.

and

Table 4. Mediating effects tests results of ESG on digital finance and long- term financial performance (EVA).

Panel A: Stepwise Regression Estimation	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9	Model 10	Model 11	Model 12
	EVA				ESG				EVA
ESG									0.003 *** (0.001)	0.003 *** (0.001)	0.003 *** (0.001)	0.003 *** (0.001)
Df	−0.005 *** (0.001)				0.390 *** (0.027)				−0.006 *** (0.001)
DfI		−0.003 (0.004)				0.420 *** (0.028)				−0.004 (0.003)
DfII			−0.006 (0.004)				0.386 *** (0.030)				−0.007 (0.005)
DfIII				−0.004 ** (0.002)				−0.186 *** (0.032)				−0.005 *** (0.002)
Control Vars.	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Panel B: mediating effects analysis
Mediating variable: ESG	Total effects				Indirect effects				Direct effects
	−0.005	−0.003	−0.006	−0.004	0.001	0.001	0.001	−0.001	−0.006	−0.004	−0.007	−0.005
	Proportion of Total effect mediated				Indirect to direct effect				Total to direct effect
	20.000	33.333	16.667	25.000	16.667	25.000	14.286	20.000	83.333	75.000	85.714	80.000

Standard error statistics in parenthesis. *** and ** indicate the significance levels of 1% and 5%, respectively. Referring to [62], the indirect effects of ESG occupy a relatively high proportion of the total effect mediated (20.000%, 33.333%, 16.667%, and 25.000%, respectively), therefore, ESG is an eligible mediate variable in the relationship between digital finance and EVA.

display the regression results of Equations (2)–(4), the results illustrate that ESG serves as a mediator in the relationship between Df and short-term financial performance (ROA) and long-term financial performance (EVA). On one hand, the coefficients of Df on ROA and EVA are notably significant and negative (see Model 1 in Table 3 and Table 4, respectively), while the coefficients of Df on ESG are significantly positive (see Model 5 in Table 3 and Table 4, respectively). On the other hand, the coefficients of ESG on ROA and EVA are significantly positive (see Model 9 in Table 3 and Table 4, respectively), aligning with the findings of [61]. These results support H3, showing that digital finance negatively affects both short- and long-term financial performance directly, while it can also indirectly promote financial performance via ESG; thus, we can conclude that H2 is only partially supported. In essence, ESG offsets the inhibitory impact of digital finance on financial performance via a partial intermediary effect. We can see that committing to firm ESG performance proves to be an effective strategy for digital financial service firms to improve financial benefits. A better ESG can produce a lower level of debt financing cost and a higher level of financial flexibility to provide support for a firm’s financial performance [61].

Considering the heterogeneity of DF sub-dimensions, we further explore the heterogeneous impact mechanism of Df on short-term and long-term financial performance via ESG. With respect to short-term financial performance–ROA, the coefficients of DfII on ESG and ROA are observed to be 0.386 and 0.009, respectively, and significant. This suggests that ESG can enhance the positive association between digital finance usage depth and ROA. The coefficients of DfI on ESG and ROA are 0.420 and −0.012, respectively, and both are strongly significant. This indicates that ESG offsets the inhibitory impact of the digital finance coverage breadth on firms’ short-term financial performance via a partial intermediary effect. Furthermore, the coefficients of DfIII on ESG and ROA are −0.186 and −0.007, respectively, and both are strongly significant. This implies that ESG can exacerbate the negative impact of the digital finance digitization degree on ROA. With respect to long-term financial performance–EVA, DfI and DfII have no direct impact on EVA but can indirectly and positively affect EVA via ESG. This indicates that focusing on ESG performance will contribute to the impact of digital finance coverage breadth and usage depth and long-term financial performance. DfIII has not only direct but also indirect negative impacts on EVA via ESG, meaning that ESG can exacerbate the negative relationship between digital finance digitization and EVA. Those results confirm that ESG mediates the relationship between digital finance and financial performance. These are mainly due to the fact that, first, digital finance’s development needs the support of digital technology tools, which demands a substantial investment cost and may seize the firms’ financial resources both in the short and long term. Second, Chinese digital finance is already a world leader, and its digital infrastructure is relatively mature [6]. By contrast, focusing on improving the digital financial coverage breadth, especially the usage depth, can bring benefits to environmental, social, governance, and economic development.

It is worth noting that ESG has a significant positive impact on both short- and long-term financial performance (see Models 9–12 in Table 3 and Table 4, respectively), meaning that ESG activities bring immediate and delayed financial benefits and also reflect that the firm’s commitment to environmental and social sustainability is not incompatible with the improvement of financial performance. In contrast to [53], our result regarding the impact of ESG is consistent with that obtained by [63]. Combining Df can drive ESG performance, and it can be said that digital finance is conducive to a firm’s social, economic, and environmental dimensions of sustainability. On one hand, the disclosure of ESG reduces the information asymmetry between firms and stakeholders and enables firms to have good trust and a green guarantee, which helps to improve firms’ market competitiveness and enhance financial performance [61]. An environmentally responsible investor would especially only invest in firms whose policies are environmentally and socially friendly [64]. On the other hand, ESG requires a large investment, and its materialization usually takes a long time, so long-term financial performance evaluation is necessary. Green firms enjoyed better business opportunities, higher employee satisfaction, cost savings, risk reduction, reputation enhancement, and financing advantages, which would reap higher long-term financial performance than other firms in the long run [65].

5.2. Robustness Tests

Although controlling for the two-way fixed effects and introducing control variables in the benchmark regression partially mitigates the endogeneity issue, the presence of reverse causality, measurement errors, and omitted variables may still influence the robustness of the regression results. Considering that Instrumental variable (IV) estimation is an effective method to mitigate the impacts of endogenous problems [60]. Following [32] suggestions on endogeneity, we use provincial mobile phone penetration as an instrumental variable to represent internet penetration, implementing a 2SLS regression approach. We can see that the IVs are effective in

Table 5. Endogeneity test results.

Variables	Phase II	Phase I	Phase II	Phase I	Phase II	Phase I
	ESP		ROA		EVA
Df	1.127 *** (0.102)		−1.251 *** (0.264)		−0.019 *** (0.005)
Phone		0.366 *** (0.016)		−0.501 *** (0.105)		0.377 *** (0.016)
Control Vars	Yes	Yes	Yes	Yes	Yes	Yes
Fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Kleibergen-Paap rk LM statistic	384.511		21.368		395.889
Kleibergen-Paap rk LM wald statistic	524.156		22.882		545.852

Standard error statistics in parenthesis. *** indicate the significance levels of 1%.

. Meanwhile, the impact of Df on ESG, ROA, and EVA are both basically in line with the benchmark results. Overall, we can infer that the findings are still robust after taking into account the endogeneity.

Robustness tests are further conducted in three ways to verify the above results (See

Table 6. Robust estimation results for ESG.

Variables	Quantile Regression			Exclude Odd Values of Variables	Lagging ESG by First-Order
	0.1	0.5	0.9
Df	0.465 *** (0.085)	0.322 *** (0.078)	0.410 *** (0.149)	0.339 *** (0.032)	0.457 *** (0.033)
DfI	0.765 *** (0.233)	0.307 *** (0.101)	0.388 * (0.227)	0.433 *** (0.094)	0.255 * (0.132)
DfII	0.174 (0.232)	0.418 *** (0.127)	0.286 (0.285)	0.464 *** (0.087)	0.755 *** (0.095)
DfIII	−0.270 * (0.158)	−0.247 *** (0.065)	−0.155 (0.158)	−0.336 *** (0.044)	−0.279 *** (0.041)
Control Vars.	Yes	Yes	Yes	Yes	Yes
Fixed effect	Yes	Yes	Yes	Yes	Yes

Standard error statistics in parenthesis. *** and * indicate the significance levels of 1% and 10%, respectively.

,

Table 7. Robust estimation results for ROA.

Variables	Quantile Regression			Exclude Odd Values of Variables	Lagging ROA by First-Order
	0.1	0.5	0.9
Df	−0.004 ** (0.002)	−0.007 ** (0.003)	−0.010 ** (0.005)	−0.009 *** (0.001)	−0.015 *** (0.002)
DfI	−0.006 (0.006)	−0.011 (0.009)	−0.031 *** (0.010)	−0.011 *** (0.004)	−0.009 ** (0.005)
DfII	0.015 *** (0.005)	0.018 *** (0.007)	0.027 *** (0.008)	0.011 *** (0.004)	0.011 ** (0.005)
DfIII	−0.009 *** (0.003)	−0.012 *** (0.003)	−0.009 (0.011)	−0.007 *** (0.002)	−0.003 (0.002)
Control Vars.	Yes	Yes	Yes	Yes	Yes
Fixed effect	Yes	Yes	Yes	Yes	Yes

Standard error statistics in parenthesis. *** and ** indicate the significance levels of 1% and 5%, respectively.

and

Table 8. Robust estimation results for EVA.

Variables	Quantile Regression			Exclude Odd Values of Variables	Lagging EVA by First-Order
	0.1	0.5	0.9
Df	0.001 (0.034)	−0.001 *** (0.000)	−0.004 * (0.002)	−0.005 *** (0.001)	−0.002 ** (0.001)
DfI	−0.001 (0.001)	−0.001 (0.001)	−0.010 (0.086)	−0.003 (0.003)	−0.005 (0.004)
DfII	0.004 (0.003)	0.002 * (0.001)	0.004 (0.106)	0.004 (0.003)	0.003 (0.004)
DfIII	−0.002 ** (0.001)	−0.001 ** (0.000)	−0.001 ** (0.000)	−0.004 ** (0.002)	−0.003 * (0.002)
Control Vars.	Yes	Yes	Yes	Yes	Yes
Fixed effect	Yes	Yes	Yes	Yes	Yes

Standard error statistics in parenthesis. ***, **, and * indicate the significance levels of 1%, 5%, and 10%, respectively.

). Firstly, the regression model was changed. Following [57], we employed the quantile regression model to fit the sample data. Secondly, we excluded odd values of explanatory variables. To diminish the outliers’ intervention with the results, a 1% double-tailed retraction was employed for the digital finance, ESG, and financial performance variables. Thirdly, we substituted the original digital finance variable with a first-order lagged digital financial level. Table 5 and Table 6 reveal that the coefficient sign of digital finance aligns with the baseline results, further verifying the study’s robustness.

6. Further Analyses

6.1. Mechanism Analysis

The above research integrates digital finance, ESG, and financial performance into a theoretical framework, providing empirical evidence that digital finance promotes ESG and that ESG can effectively mitigate the negative impacts of digital finance on financial performance. Further investigation is conducted to explore the mechanism by which digital finance matters for ESG via green innovation and firm digital transformation, as shown in

Table 9. The impact mechanism of digital on ESG.

Variables	Total Effects	Green Innovation		Digital Transformation
	ESG	GInno	ESG	Dig	ESG
Df	0.390 *** (0.027)	0.110 *** (0.023)	0.382 *** (0.027)	0.822 *** (0.028)	0.290 ** (0.029)
SA
GInno			0.085 *** (0.018)
Dig					0.110 *** (0.014)
Control Vars.	Yes	Yes	Yes	Yes	Yes
Fixed effect	Yes	Yes	Yes	Yes	Yes
Soble		2.256 ** (0.024)		10.070 *** (0.000)

Standard error statistics in parenthesis. *** and ** indicate the significance levels of 1% and 5%, respectively.

. Following [57], green innovation is measured by the logarithm of “1 + the patent applications of green inventions and green utility models”. The firm’s digital transformation is assessed using text analysis methodology in our study. Referring to [15,66], based on annual report texts, keywords and their frequencies related to digital transformation are extracted in the following respects: artificial intelligence technology, big data technology, cloud computing technology, blockchain technology, and digital technology application. Then, the above respects’ word frequencies are aggregated and processed. The results in Table 9 support H4.

Concerning the role of green innovation, Column (3) reveals a significant and positive impact of Df on GInno, meaning that digital finance fosters green innovation, which, combined with the findings presented in columns (2) and (4). It indicates that digital finance promotes ESG by fostering green innovation. We can see that digital finance has a green innovation effect, providing firms with more R&D funds, saving energy, reducing emissions by improving the efficiency of production and service, and making technologies ecologically friendly. While easing the ecological pressure on firms, it also boosts their ESG performance.

Concerning the role of firm digital transformation, Column (5) displays a significant and positive impact of Df on Dig, meaning that digital finance contributes to firm digital transformation. Combining the findings presented in columns (2) and (6), it reveals that digital finance promotes ESG by boosting firms’ digital transformation. Digital finance depends on digital information technology for development, and digital financial development forces firms to boost digitization [15]. The firm’s digital transformation advances the digital office through the paperless office to standardize process management while reducing energy consumption, thereby improving the level of corporate governance.

6.2. Heterogeneity Analysis

6.2.1. Firm Size

Digital finance is characterized by a long tail. Research points to the importance of its development for small firms [6]. Small firms account for a large market share, and they are considered to be the main drivers of economic and social development. In contrast, large firms have resource advantages and are more likely to form economies of scope and scale [3]. Following [57], we categorize the sample into small and large firms, determined by the median value of the firm size.

Table 10. Regression results under different firm sizes.

Variables	Large Scale Firms Group			Small Scale Firms Group
	ESG	ROA	EVA	ESG	ROA	EVA
Df	0.470 *** (0.051)	−0.010 *** (0.002)	−0.011 *** (0.004)	0.230 *** (0.040)	−0.008 *** (0.002)	−0.002 *** (0.000)
DfI	0.283 * (0.145)	−0.011 ** (0.005)	−0.010 (0.009)	0.637 *** (0.119)	−0.010 (0.007)	−0.002 * (0.001)
DfII	0.753 *** (0.132)	0.012 ** (0.005)	0.014 (0.009)	0.120 (0.111)	0.010 (0.007)	0.001 (0.001)
DfIII	−0.371 *** (0.067)	−0.008 *** (0.002)	−0.011 ** (0.005)	−0.305 *** (0.056)	−0.006 * (0.003)	−0.001 (0.001)
Control Vars.	Yes	Yes	Yes	Yes	Yes	Yes
Fixed effect	Yes	Yes	Yes	Yes	Yes	Yes

Standard error statistics in parenthesis. ***, **, and * indicate the significance levels of 1%, 5%, and 10%, respectively.

then investigates the heterogeneous impact of digital finance on ESG, ROA, and EVA across different firm scales.

The findings indicate that digital finance exerts a greater positive effect on ESG in large firms while yielding a greater negative impact on ROA and EVA in large firms, thereby supporting H5. As for the sub-dimensions of Df, we can see that the digital finance usage depth is more pronounced in boosting large firms’ ESG and ROA. The digital finance coverage breadth is more effective in enhancing small firms’ ESG. For one thing, digital finance provides large firms new financing opportunities and supports the development of new businesses. However, large firms have greater capital needs and higher risk uncertainty [46]. For another, the inclusiveness of digital finance could offer finance services to the “long tail”, thereby exerting a digital financial inclusive effect. In general, digital finance has a heterogeneous impact on firms of varying sizes, with large firms needing to work on improving the digital finance usage depth and small firms needing to work on popularizing the digital finance coverage breadth.

6.2.2. Ownership

Ownership type, as an important institutional factor, may result in differences in strategic positioning, governance systems, and resource allocation among firms, leading to heterogeneity in digital finance effects. Concerning political connections, state-owned enterprises (SOEs) have abundant resource allocation superiority and preferential government treatment, which will reduce the marginal effect of digital finance. In addition, SOEs are in the hands of the government and usually focus more on political and social objectives [3]. In contrast, non-SOEs bring strong dynamism to the market and stimulate China’s socio-economic development. On this basis, we delve into the heterogeneous impact of digital finance on ESG, ROA, and ESG with different ownership types in

Table 11. Regression results under different ownership.

Variables	SOEs			Non-SOEs
	ESG	ROA	EVA	ESG	ROA	EVA
Df	0.370 *** (0.039)	−0.011 *** (0.002)	−0.006 *** (0.002)	0.265 *** (0.058)	−0.008 *** (0.003)	0.001 (0.002)
DfI	0.269 ** (0.114)	−0.015 *** (0.005)	−0.005 (0.006)	0.803 *** (0.163)	−0.004 (0.009)	0.002 (0.005)
DfII	0.578 *** (0.106)	0.016 *** (0.004)	0.007 (0.006)	0.146 (0.149)	0.001 (0.004)	0.003 (0.005)
DfIII	−0.303 *** (0.054)	−0.009 *** (0.002)	−0.005 * (0.003)	−0.399 *** (0.075)	−0.003 (0.004)	−0.003 (0.002)
Control Vars.	Yes	Yes	Yes	Yes	Yes	Yes
Fixed effect	Yes	Yes	Yes	Yes	Yes	Yes

Standard error statistics in parenthesis. ***, **, and * indicate the significance levels of 1%, 5%, and 10%, respectively.

.

The outcomes reveal that digital finance exerts a greater positive impact on ESG and a negative impact on ROA in SOEs while yielding no significant impact on EVA in non-SOEs, thereby supporting H5. As for the sub-dimensions of Df, we can obtain that the digital finance coverage breadth greatly promotes ESG in non-SOEs. Non-SOEs are more flexible and innovative and have stronger competitive incentives. This enables them to launch a broader range of digital financial products and services more quickly to meet market demand. Concerning the digital finance usage depth, the coefficient of DfII in SOEs is significant and exceeds that in non-SOEs, suggesting that the digital finance usage depth greatly promotes ESG and ROA in SOEs. In contrast, the coefficient of digital finance usage depth by non-SOEs is not significant. The results are inconsistent with [46]. A possible reason for this is that our study further focuses on the digital financial sub-dimensions and expands on existing research. This declares that the digital finance usage depth can better enhance ESG and ROA in SOEs, suggesting that SOEs always take on more social responsibilities to focus more on the environment and social friendliness. In other words, SOEs can facilitate the digital finance usage depth to act as a positive function in social, environmental, and short-term economic development, while non-SOEs can facilitate the digital finance coverage breadth to act as a positive function in improving the level of social and environmental governance.

7. Conclusions and Policy Implications

This paper investigates the internal influence mechanisms of digital finance, ESG, and financial performance across types, nonlinear, and heterogeneity perspectives. The empirical conclusions can be drawn as follows: First, the coefficients of digital finance on ESG, ROA, and EVA are 0.337, −0.010, and −0.005, respectively, and significant, stating that ESG is beneficial to ESG but not conducive to both short- and long-term financial performance. The coefficients of ESG on ROA and EVA are 0.003 and significant, meaning that digital finance can indirectly enhance firms’ short- and long-term financial performance via ESG. Second, there exists a heterogeneous impact of digital finance from different sub-dimensions. The digital finance usage depth is identified as the primary driver, which directly, significantly, and positively affects both ESG and short-term financial performance while digital finance coverage breadth is only directly positive for ESG. The coverage breadth and usage depth exhibit a significant, indirect, and positive impact on both short- and long-term financial performance via ESG. Third, mechanism analysis indicates that the positive impact of digital finance on ESG is enhanced by upgrading green innovation and boosting digital transformation. Finally, the analysis of heterogeneity indicates that digital finance affects ESG and ROA heterogeneously under varying firm sizes and ownership. Digital finance usage depth exhibits a more significant positive impact on both the ESG and ROA of large firms and SOEs, while coverage breadth has a more pronounced positive impact on the ESG of small firms and non-SOEs.

The policy implications can be drawn as follows. In general, policymakers should actively promote the development of digital finance. Digital finance is seen as a facilitator of firm sustainability, making contributions to ESG and financial performance. Meanwhile, policymakers need to be aware of digital financial development and use. Managers need to prioritize the utility of digital finance and make it conform to the firm’s resource allocation situation and development concept.

Based on the empirical results, there are also specific implications for boosting the utility of firms’ digital finance. Firstly, given that the coefficients of ESG on ROA and EVA are positive (0.003) and significant, meaning that ESG is conducive to both short- and long-term financial performance, policymakers should focus on the environmental and social preferences of digital finance to meet stakeholder needs. In the procedure for guiding digital financial development, the government should focus more on the allocation of financial resources associated with green innovation and firm digital transformation and make them approved and disbursed by stakeholders and the market to produce sustainable benefits. The digital finance effects should not stop at generating good goodwill for firms. In addition, the government should formulate diverse digital financial inspire policies, enlarge the investment guarantee for digital finance usage depth, and play its main driving role.

Secondly, considering that the coefficients of digital finance use depth on ESG, ROA, and EVA in large firms (0.753, 0.012, and 0.014) are greater than those in small firms (0.120, 0.010, and 0.001). Meanwhile, the coefficient of digital finance coverage breadth on ESG in small firms (0.637) is greater than that in large firms (0.283). The results reveal that the digital finance use depth is conducive to both ESG and financial performance in large firms, while the digital finance coverage breadth is conducive to ESG in small firms, policy, and financial support for the utility depth of large firms and the coverage breadth of small firms should be strengthened. Firms of different sizes are encouraged to develop distinctive digital financial business models based on their resource endowments to encourage small firms to exert their long-tail effect and large firms to exert their scale effect.

Thirdly, considering that the positive connection between digital finance usage depth vis à vis ESG and ROA is more evidenced in SOEs (0.578 and 0.016) than in non-SOEs (0.146 and 0.001) and that the positive connection between digital finance coverage breadth and ESG is more evidenced in non-SOEs (0.803) than that in SOEs (0.269), the digital financial incentive policies for the usage depth in SOEs and the coverage breadth in non-SOEs should be strengthened and adapted. Managers should encourage SOEs to develop digital finance use depth by adjusting tax credits, smooth financing channels, etc.., to simultaneously bring environmental, social, and economic benefits. Furthermore, non-SOEs are encouraged to expand their coverage breadth, launch a wider range of digital financial products that serve to meet stakeholders’ needs and stimulate market vitality.

Despite the fact that this paper has critical implications for promoting digital finance to achieve sustainability, there are still some limitations that need to be further explored. First, given the availability of sample data, we concentrate on Chinese firms to scrutinize the influence mechanisms of digital finance in enhancing ESG and financial performance. Future studies could expand the study scope to include data from more economies, providing more validation to obtain a more generalized view. Future studies can also conduct comparative analyses across multiple economies to gain a deeper understanding of the role of digital finance in different economies. Second, following the existing literature, we employ ROA and EVA to measure the short- and long-term financial performance, which is as employed as criticized. An effective measurement of firms’ financial performance requires further optimization and continuous improvement in the future study.