The Mechanism of Enterprise Digital Transformation on Resilience from the Perspective of Financial Sustainability

Abstract

In the period following the global COVID-19 pandemic, financial sustainability has become critical for the survival of enterprises around the world. This paper investigates the mechanisms and implications of digital transformation on resilience from the perspective of financial sustainability, that is, financial resilience. Employing a sample of Chinese listed firms, the study documents that digital transformation was positively related to financial resilience in normal states due to improved internal corporate governance, increased analyst coverage, alleviated financial constraints, and reduced operation risk. The relationship was more significant for companies with digitally literate executives, non–state–owned companies, and high–tech companies. However, it was not significant during the period of the COVID-19 pandemic and for companies with consecutive losses and delisting warnings. These findings provide unique evidence to support the beneficial effects of digital transformation on corporate resilience and to identify strategies for strengthening financial sustainability.

1. Introduction

The concept of sustainability has gained considerable acceptance in the business world, with a growing number of companies pursuing the objective of enhancing the sustainable value of their enterprises. Notwithstanding the importance of corporate social responsibility (CSR) and environmental, social, and governance (ESG) considerations, financial sustainability, which is of paramount importance for the long–term success and resilience of any company, remains a pivotal aspect of sustainability.

The financial sustainability of a firm is defined as its capacity to create value for owners and ensure the long–term continuity of its operations through the strategic deployment of an optimal combination of investments and sources of financing [1]. Financial sustainability is not merely a matter of long–term financial stability; rather, it constitutes a crucial element of the overarching objective of sustainable development. It helps to avoid financial crises and provides a stable foundation for growth and development [2]. The concept of financial sustainability in relation to sustainability as its sub–concept. In light of the abovementioned, this paper takes the position that financial sustainability should be regarded as a component of the broader economic sustainability construct.

Among the conditions that are particularly important for ensuring financial sustainability is the high likelihood of business survival [3]. Thus, a firm’s resilience in the face of adversity is a key determinant of its long– and short–term economic sustainability. Resilience can be defined as an adaptive and dynamic capacity, representing an ongoing process of self–adjustment when faced with external shocks [4]. Despite the extensive research conducted on economic resilience [4,5], supply chain resilience [6,7], and enterprise resilience [8], there has been a paucity of discussion on the topic of financial resilience. This paper takes a financial sustainability perspective in examining the concept of financial resilience. Financial resilience can be defined as the capacity of an enterprise to avoid financial distress through a sustained high level of financial performance and a low level of operating volatility in the face of significant changes in the internal and external environment. This definition aligns with the concept of financial sustainability. In contrast with existing literature, which considers financial sustainability in relation to stock price decline [2], the ability of a company to generate consistent revenue and profits is of paramount importance for its continued existence [9]. It is therefore imperative to examine financial resilience from the standpoint of declining performance.

The advancement of digital technology and artificial intelligence has enhanced enterprises’ capacity to obtain resources [8] and mitigate risks of information asymmetry [6] and talent mismatch, thereby enabling them to achieve the optimal allocation of resources and a distinctive competitive advantage within a fiercely competitive landscape [10]. Furthermore, digital technology has significantly enhanced the decision–making process of corporate governance, facilitated the identification of development opportunities, enabled the adjustment of strategy during critical moments, and improved information transparency [11]. Consequently, enterprise digital transformation plays a significant role in improving enterprise financial resilience.

However, during the COVID-19 pandemic, digital transformation, although it can support the online office and production recovery in enterprises, requires huge initial and ongoing costs [12], which could increase the financial burden on firms, leading to uncertainty regarding their performance. In addition, from the standpoint of cost savings and reduced capital consumption [13,14,15,16], the substantial expense associated with digital transformation may not immediately yield benefits when enterprises are confronted with significant risks. The impact of digital transformation on the financial resilience of enterprises may vary in abnormal states.

This paper examines the mechanisms of digital transformation within the context of China. As indicated in the Digital China Development Report (2022), China’s digital economy reached 50.2 trillion yuan in 2022, representing a 41.5% share of the gross domestic product (GDP) and 8.1 TB of data production, accounting for 10.5% of the global total and ranking second in the world. As a country at the forefront of digital development, China can provide insights for other countries.

This paper makes the following contributions. First, although previous research has explored corporate digital transformation from various perspectives [17], fewer studies have examined the correlation between digital transformation and firms’ resilience. Our findings provide new empirical insights into digital transformation and resilience from the perspective of financial sustainability. Second, existing literature on financial sustainability tends to adopt a static perspective, focusing on revenue, ROE, and ROA [18,19]. In contrast, financial resilience measures a firm’s financial capability in the context of dynamic changes in revenue, ROE, and ROA in the face of a crisis. This represents a significant contribution to the field of financial sustainability research, establishing a crucial link between corporate resilience research and financial sustainability research. Furthermore, this paper puts forth a set of financial resilience indicators that are more fundamental measures of a company’s capacity to withstand financial risk than the stock price decline [20,21] or a combination of static financial indicators with financing ability, profitability, and liquidity ratios [22]. This contributes to the ongoing study of the measurement of resilience. Third, it is crucial to differentiate the influence of digital transformation in both normal and abnormal states in order to gain a more comprehensive understanding of the digital transformation strategy, as opposed to the studies that solely examine firm resilience in the context of adverse shocks [20,21].

2. Theoretical Background and Hypotheses

2.1. Digital Transformation and Financial Resilience

The theory of resource allocation posits that in a market economy, resources are automatically allocated by the market. Effective resource allocation can reduce transaction costs, enhance the efficiency of capital utilization [23], and facilitate the formation of distinctive competitive capabilities for companies. Nevertheless, both market monopoly and information asymmetry have the potential to impact the optimal allocation of resources. The digital transformation process can mitigate market information asymmetry, facilitate the uninhibited exchange of the free flow of digital, labor, and other necessary resources, and achieve the optimal allocation of resources.

The implementation of digital transformation can enhance business performance and financial resilience. It has the potential to fundamentally change the enterprise’s production model, rendering the traditional process more intelligent and information–based [24]. Consequently, production efficiency is enhanced, and the enterprise’s resilience is improved. However, the advancement of digital technology and artificial intelligence has strengthened enterprises’ ability to obtain resources [8]. Additionally, digital technology has significantly reduced communication costs between enterprises, thereby enhancing communication efficiency and mitigating the risks associated with information asymmetry and talent mismatch. Therefore, when confronted with a dilemma, enterprises can respond promptly and communicate effectively, thereby facilitating the resolution of the situation and the enhancement of their financial resilience. Considering the abovementioned analysis, this paper proposes the following hypothesis:

H1. 

Enterprise digital transformation has a significant positive impact on financial resilience.

The implementation of digital transformation during the COVID-19 pandemic has proven to be a double–edged sword. On the one hand, the use of digital technology ensures that enterprises have an online office and provides support for the resumption of production and work, which has led to an increasing number of traditional enterprises engaging in digital transformation. On the other hand, firstly, enterprises encounter numerous challenges during the process of digital transformation, with financial concerns [25] being the most significant. Indeed, enterprises have to make huge investments in both initial and ongoing costs, amounting to 5–15% of operating revenue per year, in order to implement digital transformation. During the COVID-19 pandemic, a high and prolonged initial investment in digital transformation could result in an increased financial burden for firms, potentially leading to uncertainty regarding their performance. Second, organizational learning theory posits that enterprises will response to changes such as digital transformation. Furthermore, the pressure to adapt is even more pronounced during the COVID-19 pandemic, where employees experience significant psychological distress, which in turn leads to a decline in productivity. Third, the process of digital transformation often necessitates the redesign and optimization of supply chains, which may increase uncertainty in the context of the global supply chain vulnerability during the COVID-19 pandemic.

Besides, when companies are in abnormal states, such as suffering consecutive losses and delisting warnings, the impact of digital transformation on financial resilience can also yield uncertain outcomes due to the contrasting forces of its pros and cons. With regard to cost savings and capital consumption reduction [13,14,15], the high cost of digital transformation may not yield immediate benefits when enterprises are confronted with severe shocks. Based on the abovementioned analysis, this paper proposes the following hypothesis:

H1a. 

In an abnormal state, enterprise digital transformation does not have a significant impact on financial resilience.

2.2. The Mechanism Variable of Corporate Governance

The implementation of improved corporate governance facilitates more effective digital transformation initiatives and enhances financial resilience for companies. In the era of big data, new business forms have emerged, leading to a significant impact on traditional governance. Nevertheless, digital technology provides new ideas for corporate governance in the new era. Economic organizations employ digital technology to accurately identify and analyze potential external governance entities and investors based on a massive amount of corporate governance big data [24]. The transition from the traditional “verticalized” governance model to a “flat” governance model establishes the foundation for breaking down governance data barriers and information asymmetry, thereby facilitating the reorganization of the corporate governance model. Corporate governance can be divided into two categories: internal corporate governance and external corporate governance.

In the context of internal corporate governance, enterprises rely on digital networks and technological advantages to promote corporate governance decision–making and organizational transformation, ultimately enhancing the quality of internal corporate governance.

A high level of corporate governance, in turn, enhances financial resilience. From a strategic perspective, this can assist the company in accurately identifying development opportunities and adjusting its strategies during critical moments. It can also aid in making crucial decisions regarding the company’s continued existence and expansion.

First, the implementation of a higher level of corporate governance can assist in the resolution of conflicts between corporate management, shareholders, and creditors by employing an effective internal control system. The internal control mechanism can mitigate the agency motives of management [26] and improve decision–making, incentives, and supervision [27]. This can enhance the efficiency and effectiveness of corporate decision–making, mitigate the impact of uncertain events that may lead to business distress, and ultimately improve financial resilience.

Second, a higher level of corporate governance corresponds to a higher level of business operations [28]. This allows enterprises to implement timely measures to facilitate transformation in the event of a general downturn in the business environment, thereby reducing the likelihood of future operational risks. At the same time, if the enterprise encounters difficulties, it is able to make prompt adjustments to facilitate the swift overcoming of such challenges [29].

Thirdly, the increasing prevalence of digitalization has resulted in a notable rise in the utilization of digital technology within the domain of corporate governance. This can markedly enhance the efficiency of various governance processes and the speed of decision–making [8]. Additionally, it enables quick identification and resolution of internal and external risks, thereby mitigating the adverse consequences of a crisis.

In the context of external corporate governance, it is expected that digital transformation will enhance external corporate governance, thereby improving financial resilience through strengthened supervision. Analysts, as experts in the analysis of highly sensitive information, play a crucial role as external monitors of listed companies. The digital transformation of enterprises can assist in reducing information asymmetry and improve information transparency, which in turn aids the monitoring of external stakeholders such as analysts [11]. Securities analyst tracking serves as a bridge between listed companies and external stakeholders, facilitating the mitigation of information asymmetry and the reduction of agency costs [30]. The accuracy of a firm’s future risk assessment and prediction of its financial resilience is enhanced with increased attention from analysts. This highlights the importance of objective analysis in risk management. In light of the abovementioned analysis, this paper proposes the following hypothesis:

H2. 

Corporate governance acts as a mediator between digital transformation and financial resilience.

2.3. The Mechanism Variable of Financial Constraints

While financial constraints diminish financial resilience, digital transformation may potentially alleviate financial constraints. From the perspective of the resource–based view, first, digital transformation is beneficial in that it reduces costs, thereby supporting sustainable competitive advantages and alleviating the financing constraints faced by these enterprises. Second, the digital transformation of enterprises is aligned with the development trend of the digital economy and national policies. Effective digital transformation serves to demonstrate to the capital market that the enterprise has a promising future [11]; therefore, enterprises that actively carry out digital transformation are more readily able to obtain bank loans, government grants, and support from the capital market. Third, the digitization of business processes can enable enterprises to continuously absorb new knowledge [31], enhance risk resistance, and improve financial resilience. Therefore, this paper proposes the following hypothesis:

H3. 

Financial constraints play a mediating role in the relationship between digital transformation and financial resilience.

2.4. The Mechanism Variable of Operation Risk

The term “resilience” is defined as the adaptive and dynamic capacity of an entity to withstand risks. In this context, the risk of business is measured by the cash flow volatility of the enterprises, which reflects the unpredictability of cash flows. Cash holdings serve as a proxy for an enterprise’s capacity to manage risks [32,33]. In light of the abovementioned analysis, this paper proposes the following hypothesis:

H4. 

Cash flow volatility serves as a mediator in the relationship between digital transformation and financial resilience.

3. Research Methods

3.1. Data and Research Sample

The initial sample comprised Chinese A–share listed firms from 2010 to 2022. Financial data were obtained from the CSMAR and WIND databases, while regional–level data were obtained from the National Bureau of Statistics of China. Based on the word root used by Yu and Shao [34], a precise search for word frequency was conducted on the core technologies of digital transformation on the annual reports to illustrate the impact of digital transformation. This paper employs a manual approach to data collection, focusing on word frequency searches related to digital transformation and keyword searches pertaining to digitalization in the background of executives. These searches were conducted in annual reports and executive biographies. The final sample comprised 19,062 firm–year observations, obtained after the exclusion of financial institutions, firms with missing data, special treatment companies (designated as “ST”), and companies that were given delisting warnings (designated as “*ST”). In addition, this paper presents the findings of an empirical study on ST or *ST companies, comprising 829 firm–year observations.

3.2. Variables

3.2.1. Dependent Variables

On the one hand, prior research has primarily indicated that capital market performance loss serves as a fundamental indicator for resilience studies [21,22,34]. Nevertheless, it is not appropriate to measure the resilience of listed companies based on the decline in stock prices. Given the influence of external factors such as climate change [20] on stock prices, it is challenging to demonstrate the resilience of a company when its stock price fluctuates. This is because resilience is not only about withstanding short–term declines but also about demonstrating the ability to recover and grow. This paper believes that a company’s ability to resist decline and recover corporate earnings is the fundamental factor that determines its survival and development in a crisis. On the other hand, the existing literature that studies financial sustainability adopts measurements from the static perspective of revenue, ROE, and ROA [18,19]. These static performance measures indicate the potential capacity to withstand the risks, but they cannot reflect the actual capacity of resilience. This paper measures the financial capability of a firm from the dynamic changes in revenue, ROE, and ROA in the face of a crisis and employs a methodology that utilizes declines in operating revenue (OR), the return on total assets (ROA), and the return on equity (ROE) as a means of assessing financial resilience. The following formula was adopted to calculate decline_OR, decline_ROA, decline_ROE:

Decline_Performance = (Performance_pre − Performance_after)/Performance_pre

(1)

where the “Performance” refers to the operating revenue (OR), the return on assets (ROA), and the return on equity (ROE).

In the normal state (2010–2019), this paper measures Performance_pre and Performance_after as the maximum quarterly performance in the prior year and measure as the minimum quarterly performance in the present year, respectively.

In the abnormal state, this paper tests the negative impact of the COVID-19 pandemic on firms’ performance. In China, the outbreak of COVID-19 was first reported in January 2020 and continued until the end of 2022, as officially declared by the government. Performance_pre and Performance_after are measured by the maximum quarterly performance of 2019 before the COVID-19 crisis and the minimum quarterly performance during the crisis (2020–2022), respectively. A large decline in performance indicates that the enterprise is experiencing a significant reduction in revenue and profit during the crisis, accompanied by a lack of resistance to risk. This reflects weak financial resilience.

3.2.2. Independent Variable

Existing literature measures digital transformation using the following three methods. First, it is measured by the ratio of digital technology–related intangible asset investment to total intangible assets at the end of the year disclosed in the notes to the financial reports [35]. Second, it is measured by constructing virtual variables to measure whether an enterprise has undergone digital transformation. Third, as exemplified by Yu and Shao [34], Python’s Jieba Chinese word–splitting function is used to conduct word frequency analysis on selected samples’ keywords related to digital development. Subsequently, the entropy value method is employed to determine the weights of each index, deriving the digital development index. In comparison to the initial two methods, the third method is the most persuasive. This paper makes reference to the methodology proposed by Yu and Shao [34] and employs the frequency of digitalization–related keywords in enterprise annual reports as a means of measuring enterprise digitalization. A precise search for word frequency was undertaken regarding the fundamental technologies underpinning digital transformation, including artificial intelligence, blockchain, cloud computing, big data, and digital technology. The total number of keyword frequencies plus 1 was taken as the logarithm to be a measure of digital transformation.

3.2.3. Mechanism Variables

Internal corporate governance (GCL) is evaluated through principal component analysis, which generates comprehensive indicators based on the following criteria: shareholding structure, incentives, and decision–making rights. Among the abovementioned indicators, the supervisory role of the board of directors is measured by the proportion of independent directors and the size of the board of directors. The supervisory role of the shareholding structure is measured by the proportion of institutional shareholding and equity balance degree. The incentive mechanism is measured by the compensation of the top management team (TMT) and the proportion of the executives’ shareholding. Finally, the capacity for decision–making is measured by the CEO–chairman of the board duality. Based on the seven abovementioned indicators, this paper uses principal component analysis to obtain the first principal component as a measure of the level of corporate governance of the comprehensive indicators.

Analyst: Analyst coverage is measured by the number of analysts or analyst groups that follow the company throughout the year.

WW: The WW index of firms’ external finance constraints was constructed by Whited and Wu [36] via the generalized method of moments (GMM) estimation of an investment Euler equation. It is a more comprehensive measure than the KZ index. It is consistent with the firm characteristics associated with external finance constraints. In this paper, the WW was employed by calculating it as follows:

λ_i,t = b₁TLTD_i,t + b₂DIVPOS_i,t + b₃SIZE_i,t + b₄SG_i,t + b₅ISG_i,t + b₆CF_i,t

(2)

where TLTD_i,t is the long–term debt over total assets, DIVPOS_i,t is a dummy variable equal to 1 if the company distributes dividends, SIZE_i,t is the natural logarithm of total assets, SG_i,t is the growth of the company, ISG_i,t is the industrial growth of the company, and CF_i,t is the cashflow over total assets.

According to Whited and Wu (2006) [36], the coefficients for each term are as follows:

WW = 0.021 TLTD_i,t − 0.062 DIVPOS_i,t − 0.044 SIZE_i,t − 0.035 SG_i,t + 0.102 ISG_i,t − 0.091 CF_i,t

(3)

CFV: Cash flow volatility is measured by the standard deviation of cash flows from year t − 2 to year.

3.2.4. Moderating Variables

Executives with a digital background: This paper employs a manual approach to select relevant keywords from the resumes of executives in sample companies based on digital keywords. If an executive in a company has a background or experience in big data, the internet, artificial intelligence, network engineering, computer science, etc., he or she is considered to have a digital background, and the company is marked as Digital–exe. Otherwise, the company is marked as Non–digital–exe.

The nature of ownership: If the company is a state–owned enterprise, the company is marked as SOE. Otherwise, it is marked as Non–SOE.

The nature of high–tech and non–high–tech companies: If the company is in high–tech industries, it is marked as High–tech. Otherwise, it is marked as Non–high–tech.

3.2.5. Control Variables

Following the existing literature [33,37], this paper mainly chose two kinds of control variables. First are the control variables, including the characteristics of the listed companies: Size of the firm (Size), accounts receivable ratio (REC), cashflow ratio (Cashflow), inventory ratio (INV), and age of listing (ListAge). Second are the control variables, including the characteristics of the companies’ corporate governance: the ownership structure (the proportion of shares held by the top five shareholders (Top 5)), the nature of ownership (SOE), agency cost of the second principal–agency relationship (the occupation of funds by the big shareholder (Occupy), and agency cost of the first principal–agency relationship (the management fee rate (Mfee)). This paper also controls the industry–fixed effects (Ind) and time–fixed effects (Year). The definitions of variables are shown in

Table 1. Definitions and measurement of variables.

Type of Variables	Variable	Definition
Dependent variable	DT	Natural logarithm of the total number of keyword frequencies, including those of five technologies, plus one.
Independent variables	decline_OR	decline_OR = (ORpre − ORafter)/ORpre, where OR is the operating revenue.
	decline_ROA	decline_ROA = (ROApre − ROAafter)/ROApre, where ROA is calculated as net profit divided by average total assets.
	decline_ROE	decline_ROE = (ROEpre − ROEafter)/ROEpre, where ROE is calculated as net profit divided by the average total owner’s equity.
Mechanism Variables	GCL	Internal corporate governance is measured by using principal component analysis to construct comprehensive indicators from the proportion of independent directors, the size of the board of directors, the proportion of institutional shareholding, equity balance degree, the compensation of TMT, the proportion of the executives’ shareholding, and CEO–chairman of the board duality.
	Analyst	The number of analysts or analyst groups that follow the company throughout the year.
	WW	It measures financial constraints by employing the model from Whited and Wu (2006) [36] that constructs an index of firms’ external finance constraints via the generalized method of moments (GMM) estimation of an investment Euler equation.
	CFV	Cash flow volatility is measured by the Standard deviation of cash flows from year t−2 to year.
Moderating variables	Digital–exe or Non–digital–exe	If an executive in a company has a background or experience in digital technology, the company is marked as Digital–exe. Otherwise, it is marked as Non–digital–exe.
	SOE or Non–SOE	If the company is a state–owned enterprise, the company is marked as SOE. Otherwise, it is marked as Non–SOE.
	High–tech or Non–high–tech	If the company is in high–tech industries, it is marked as High–tech. Otherwise, it is marked as Non–high–tech.
Control variables	SIZE	Natural logarithm of total assets.
	Cashflow	Cash flow from operating activities is divided by total assets.
	REC	Net Receivables divided by total assets.
	INV	Net inventory divided by total assets.
	ListAge	Natural logarithm of the number of years since a firm’s initial public offering plus 1.
	Top 5	The percentage of shares held by the largest five shareholders (decimal) divided by total shares.
	SOE	A dummy variable that equals 1 if the firm is a state–owned enterprise, and 0 otherwise.
	Occupy	Other receivables are divided by total assets.
	Mfee	Management fee divided by operating income.
	Ind	Industry dummy variables.
	Year	Year dummy variables.
Other variable using in robustness tests	Instru_hangzhou	Spherical distance between the listed companies’ location city and Hangzhou
	DTA	Natural logarithm of the total number of keyword “digital technology application” frequencies, plus one.
	GDP	Per capita GDP for each city in which the listed companies are located.
	Inst	The ratio of the added value of the tertiary industry to the added value of the secondary industry.
	R&D	The proportion of R&D expenditure to the GDP of the city.
	DTIA	Digital intangibles are defined as intangible assets related to the keywords of digital transformation and are measured by the ratio of intangibles invested in digitalization to total assets.

.

3.3. Empirical Model

This paper employed Model (4) to test the effect of digital transformation on financial resilience:

decline_Performance_i,t = α₀ + α₁DT_i,t + ΣControl_i,t + Ind_i,t + Year_t + ɛ_i,t

(4)

where decline_Performance_i,t refers to decline in revenue (decline_OR_i,t), decline in ROA (decline_ROA_i,t), and decline in ROE (decline_ROE_i,t). DT_i,t represents the digital transformation of the enterprise. Control_i,t denote control variables. Ind_i,t and Year_t control the industry–fixed effects and time–fixed effects, respectively. ɛ_i,t represents the error term. i and t denote enterprise i and year t, respectively.

To verify the hypothesis H2–H4, Equations (5) and (6) were set up to explore the mediating role of corporate governance, financial constraints, and cash flow volatility between digital transformation and financial resilience as follows:

Mechanism Variable_i,t = α₀ + α₁DT_i,t + ΣControl_i,t + Ind_i,t + Year_t + ɛ_i,t

(5)

decline_Performance_i,t = α₀ + α₁DT_i,t + + α₂Mechanism Variable_i,t + ΣControl_i,t + Ind_i,t + Year_t + ɛ_i,t

(6)

where Mechanism Variable_i,t refers to internal corporate governance (GCL), analyst coverage (Analyst), financing constraints index (WW), and cash flow volatility (CFV). α₂ denotes the coefficient for the influence of mechanism variables on financial resilience.

4. Empirical Results and Analysis

4.1. Descriptive Statistics

Table 2. Descriptive statistics.

Variable	Obs	Min	Mean	Median	Max	SD
DT	18,893	0.000	1.220	0.693	4.934	1.347
decline_OR	18,893	0.365	0.770	0.771	0.965	0.085
decline_ROA	18,893	0.228	1.052	0.827	6.745	0.843
decline_ROE	18,893	0.225	1.076	0.830	7.517	0.929
Size	18,893	19.520	22.360	22.202	26.4	1.304
Cashflow	18,893	−0.200	0.044	0.043	0.257	0.069
REC	18,893	0.000	0.112	0.088	0.507	0.101
INV	18,893	0.000	0.152	0.114	0.772	0.146
ListAge	18,893	0.000	2.394	2.485	3.332	0.668
Top 5	18,893	0.175	0.516	0.513	0.892	0.152
SOE	18,893	0.000	0.413	0.000	1.000	0.492
Mfee	18,893	0.008	0.095	0.074	0.766	0.083
Occupy	18,893	0.000	0.017	0.008	0.202	0.025

presents the main descriptive statistics results. The mean and median of digital transformation are 1.220 and 0.693, respectively. The mean and median of decline in operation revenue are 0.770 and 0.771, respectively. The mean value of the decline in the return on total assets is 1.052 with a median of 0.827; the decline in the return on equity is 1.076 with a median of 0.830.

4.2. Baseline Regression Results

Table 3. Effects of digital transformation on financial resilience.

	(1)	(2)	(3)	(4)	(5)	(6)
Variables	decline_ROA	decline_ROE	decline_OR	decline_ROA	decline_ROE	decline_OR
DT	−0.1331 ***	−0.1815 ***	−0.0013 **	−0.1258 ***	−0.1844 ***	−0.0014 **
	(−3.6950)	(−3.5459)	(−2.3117)	(−3.4690)	(−3.5915)	(−2.3274)
Size				−0.0759 **	−0.0741	−0.0003
				(−2.1152)	(−1.4567)	(−0.4681)
Cashflow				−6.4505 ***	−9.3870 ***	−0.0363 ***
				(−11.4996)	(−11.8166)	(−4.0437)
REC				0.9080 **	0.8955	0.0725 ***
				(2.0317)	(1.4149)	(10.1243)
INV				0.2433	0.2367	0.0200 ***
				(0.6961)	(0.4782)	(3.5644)
ListAge				0.2210 ***	0.2223 **	−0.0027 **
				(2.8698)	(2.0381)	(−2.1518)
Top 5				−1.1971 ***	−1.6985 ***	−0.0040
				(−4.4681)	(−4.4765)	(−0.9419)
SOE				0.1011	0.1435	0.0079 ***
				(1.1397)	(1.1425)	(5.5883)
Occupy				17.2334 ***	26.8169 ***	0.1263 ***
				(11.5444)	(12.6849)	(5.2799)
Mfee				6.4901 ***	8.8851 ***	0.1561 ***
				(12.3105)	(11.9005)	(18.4806)
Constant	1.8912 ***	2.2441 ***	0.7539 ***	2.5142 ***	2.6015 **	0.7336 ***
	(4.0192)	(3.3571)	(99.7002)	(2.8147)	(2.0566)	(51.2673)
Ind	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Observations	19,062	19,062	19,062	18,893	18,893	18,893
R–squared	0.029	0.030	0.139	0.063	0.063	0.167

**, *** indicate a significance level of 5% and 1%, respectively; robust standard errors are given in parentheses.

presents the results of Model (1) with the digital transformation (DT) as the main dependent variable and decline in performance (decline_OR_i,t, decline_ROA_i,t, and decline_ROE_i,t) as independent variables. Columns (1)–(3) display the regression outcomes without any additional control variables but with fixed industry and time effects, while columns (4)–(6) incorporate control variables. The results reveal that the digital transformation (DT) in columns (1), (2), (4), and (5) is significantly negatively related to the decline in ROA and the decline in ROE at the 1% level, which are significantly negatively related to the decline in revenue at the 5% level. This suggests that higher levels of digital transformation are correlated with smaller declines in performance, indicating the positive impact of digital transformation on a firm’s financial resilience.

To test the difference in the relationship between digital transformation and financial resilience in the normal state and abnormal state, this paper used the group regressions for the non–COVID-19 period and the COVID-19 period based on the outbreak of the COVID-19 pandemic in 2020. Columns (1)–(3) of

Table 4. Effects of digital transformation on financial resilience: normal period vs. COVID-19 crisis period.

	Normal Period 2010–2019			COVID-19 Crisis Period 2020–2022
	(1)	(2)	(3)	(4)	(5)	(6)
Variables	decline_ROA	decline_ROE	decline_OR	decline_ROA	decline_ROE	decline_OR
DT	−0.1209 ***	−0.1779 ***	−0.0015 **	−0.1945	−0.2751	0.0009
	(−3.3936)	(−3.5944)	(−2.3525)	(−1.1802)	(−1.1099)	(0.5639)
Size	−0.0840 **	−0.0987 **	0.0000	0.1849	0.4140 *	−0.0030 *
	(−2.3834)	(−2.0173)	(0.0495)	(1.1262)	(1.6762)	(−1.9061)
Cashflow	−5.6799 ***	−8.1739 ***	−0.0266 ***	−13.0786 ***	−19.6128 ***	−0.1116 ***
	(−10.4260)	(−10.8005)	(−2.7966)	(−4.5059)	(−4.4916)	(−4.0753)
REC	0.3065	0.1538	0.0729 ***	8.8958 ***	12.0644 ***	0.0568 ***
	(0.7037)	(0.2541)	(9.5928)	(4.0504)	(3.6514)	(2.7422)
INV	0.2516	0.1629	0.0209 ***	−0.2044	1.4102	0.0125
	(0.7408)	(0.3453)	(3.5306)	(−0.1123)	(0.5152)	(0.7308)
ListAge	0.3075 ***	0.3690 ***	−0.0028 **	−1.4677 ***	−2.4336 ***	0.0036
	(4.1699)	(3.6025)	(−2.2050)	(−2.7674)	(−3.0501)	(0.7290)
Top 5	−0.6771 ***	−0.9197 **	−0.0037	−4.5600 ***	−6.7034 ***	−0.0133
	(−2.5872)	(−2.5294)	(−0.8061)	(−3.5157)	(−3.4355)	(−1.0881)
SOE	0.1416	0.1823	0.0086 ***	−0.1381	−0.0523	0.0008
	(1.6271)	(1.5073)	(5.6778)	(−0.3429)	(−0.0864)	(0.2114)
Occupy	13.7592 ***	21.4902 ***	0.1101 ***	44.6402 ***	68.2459 ***	0.2470 ***
	(9.3912)	(10.5586)	(4.3033)	(6.5337)	(6.6397)	(3.8333)
Mfee	5.3108 ***	6.9997 ***	0.1500 ***	29.3630 ***	44.6885 ***	0.2114 ***
	(10.4446)	(9.9095)	(16.8987)	(9.3152)	(9.4238)	(7.1108)
Constant	2.7320 ***	3.2167 ***	0.7282 ***	0.0321	−3.6072	0.8567 ***
	(3.1190)	(2.6436)	(47.6196)	(0.0075)	(−0.5624)	(21.3063)
Ind	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Tes	Yes
Observations	17,072	17,072	17,072	1821	1821	1821
R–squared	0.046	0.047	0.123	0.179	0.178	0.247

*, **, *** indicate a significance level of 10%, 5%, and 1%, respectively; robust standard errors are given in parentheses.

show the regression results of the 2010–2019 samples, while columns (4)–(6) show the regression results of the 2020–2022 samples. This paper finds that the digital transformation (DT) coefficients were significantly negative at the 1% and 5% confidence levels, respectively, under normal circumstances, while they were insignificant under abnormal circumstances. It reveals that the impact of the COVID-19 pandemic will weaken the effect of digital transformation on financial resilience.

Furthermore, we adopted group regressions for “ST” and “*ST” companies to reflect the samples in the abnormal state.

Table 5. Effects of digital transformation on financial resilience: ST and *ST companies.

	(1)	(2)	(3)	(4)	(5)	(6)
Variables	decline_ROA	decline_ROE	decline_OR	decline_ROA	decline_ROE	decline_OR
DT	3.9696	−5.4203	−0.0150	1.8443	−0.0545	−0.0185
	(0.5568)	(−0.1837)	(−0.8007)	(0.2407)	(−0.0017)	(−0.9463)
Size				6.2984	−27.5231	0.0071
				(1.1419)	(−1.1582)	(0.4990)
Cashflow				−23.0215	−51.1796	−0.0332
				(−1.5853)	(−0.3931)	(−0.8978)
REC				77.0392	−251.5181	−0.0699
				(1.2315)	(−0.9120)	(−0.4381)
INV				−27.5159	−70.2549	0.0417
				(−0.5950)	(−0.3638)	(0.3519)
ListAge				−19.1498	4.3048	−0.0438
				(−1.0134)	(0.0550)	(−0.8650)
Top 5				−50.1142	39.4621	−0.0867
				(−1.0939)	(0.2009)	(−0.7350)
SOE				−8.4766	68.9093	−0.0200
				(−0.5712)	(1.0800)	(−0.5268)
Occupy				−85.6374	286.4415	0.1097
				(−1.0621)	(0.6459)	(0.5329)
Mfee				−0.0396	−9.9237	0.0001
				(−0.4894)	(−0.4625)	(0.5384)
Constant	−3.4530	−58.6771	0.7924 ***	−45.3760	424.9459	0.7986 **
	(−0.0582)	(−0.2527)	(5.1352)	(−0.3462)	(0.7665)	(2.3668)
Ind	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Observations	828	715	815	797	686	784
R–squared	0.153	0.102	0.071	0.165	0.108	0.078

**, *** indicate a significance level of 5% and 1%, respectively; robust standard errors are given in parentheses.

indicates that all the coefficients of digital transformation (DT) on financial resilience were insignificant, which demonstrates that the cost of digital transformation is a heavy burden compared with the benefit it brings for “ST” and “*ST” companies.

4.3. Mechanism Tests

4.3.1. Mechanism Test: Corporate Governance and Analyst Coverage

As shown in

Table 6. Mechanism test: mediating effect of internal corporate governance.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Variables	decline_ROA	decline_ROE	decline_OR	GCL	decline_ROA	decline_ROE	decline_OR
DT	−0.1258 ***	−0.1844 ***	−0.0014 **	0.0261 ***	−0.1276 ***	−0.1877 ***	−0.0013 **
	(−3.4690)	(−3.5915)	(−2.3274)	(4.7640)	(−3.5182)	(−3.6541)	(−2.2841)
GCL					0.0710	0.1269 *	−0.0009
					(1.4741)	(1.8600)	(−1.2068)
Size	−0.0759 **	−0.0741	−0.0003	−0.2047 ***	−0.0614 *	−0.0481	−0.0005
	(−2.1152)	(−1.4567)	(−0.4681)	(−37.6591)	(−1.6493)	(−0.9121)	(−0.7710)
Cashflow	−6.4505 ***	−9.3870 ***	−0.0363 ***	−0.8222 ***	−6.3921 ***	−9.2827 ***	−0.0371 ***
	(−11.4996)	(−11.8166)	(−4.0437)	(−9.6829)	(−11.3676)	(−11.6570)	(−4.1187)
REC	0.9080 **	0.8955	0.0725 ***	−0.2048 ***	0.9225 **	0.9215	0.0723 ***
	(2.0317)	(1.4149)	(10.1243)	(−3.0278)	(2.0638)	(1.4557)	(10.0953)
INV	0.2433	0.2367	0.0200 ***	0.1426 ***	0.2331	0.2186	0.0201 ***
	(0.6961)	(0.4782)	(3.5644)	(2.6961)	(0.6670)	(0.4416)	(3.5874)
ListAge	0.2210 ***	0.2223 **	−0.0027 **	−0.4180 ***	0.2507 ***	0.2754 **	−0.0030 **
	(2.8698)	(2.0381)	(−2.1518)	(−35.8566)	(3.1495)	(2.4426)	(−2.3871)
Top 5	−1.1971 ***	−1.6985 ***	−0.0040	−0.6749 ***	−1.1492 ***	−1.6129 ***	−0.0047
	(−4.4681)	(−4.4765)	(−0.9419)	(−16.6416)	(−4.2580)	(−4.2200)	(−1.0805)
SOE	0.1011	0.1435	0.0079 ***	−0.4672 ***	0.1343	0.2028	0.0075 ***
	(1.1397)	(1.1425)	(5.5883)	(−34.7910)	(1.4673)	(1.5650)	(5.1198)
Occupy	17.2334 ***	26.8169 ***	0.1263 ***	−0.1485	17.2439 ***	26.8357 ***	0.1261 ***
	(11.5444)	(12.6849)	(5.2799)	(−0.6571)	(11.5517)	(12.6945)	(5.2741)
Mfee	6.4901 ***	8.8851 ***	0.1561 ***	0.0308	6.4879 ***	8.8812 ***	0.1561 ***
	(12.3105)	(11.9005)	(18.4806)	(0.3858)	(12.3067)	(11.8960)	(18.4841)
Constant	2.5142 ***	2.6015 **	0.7336 ***	5.5435 ***	2.1204 **	1.8980	0.7388 ***
	(2.8147)	(2.0566)	(51.2673)	(40.9985)	(2.2745)	(1.4376)	(49.4644)
Ind	YES	YES	YES	YES	YES	YES	YES
Year	YES	YES	YES	YES	YES	YES	YES
Observations	18,893	18,893	18,893	18,893	18,893	18,893	18,893
R–squared	0.063	0.063	0.167	0.416	0.063	0.064	0.167
Sobel test					2.993 ***	3.303 ***	1.938 *

*, **, *** indicate a significance level of 10%, 5%, and 1%, respectively; robust standard errors are given in parentheses.

, columns (1)–(3) indicate the first step: the regression of digital transformation on the decline in performance, while column (4) presents the second step: the regression of digital transformation on internal corporate governance, and columns (5)–(7) reflect the third step: the regression of the digital transformation and internal corporate governance on the decline in performance. The digital transformation was significantly negatively related to the decline in ROA, decline in ROE, and decline in revenue at the 1% or 5% level, respectively. Digital transformation was positively related to internal corporate governance at the 1% level. Since the coefficients of internal corporate governance in columns (5) and (7) are not significant, this paper conducted the Sobel test and the results were all significant, indicating that the mediation effects are supported. Overall, these results show that digital transformation enhances financial resilience through internal corporate governance.

As shown in

Table 7. Mechanism test: mediating effect of analyst coverage.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Variables	decline_ROA	decline_ROE	decline_OR	Analyst	decline_ROA	decline_ROE	decline_OR
DT	−0.1069 ***	−0.1464 ***	−0.0015 **	0.4249 ***	−0.0300	−0.0307	−0.0012 *
	(−2.9779)	(−2.8830)	(−2.5739)	(5.4011)	(−0.9864)	(−0.7299)	(−1.8320)
Analyst					−0.0293 ***	−0.0378 ***	−0.0004 ***
					(−8.7839)	(−8.2085)	(−5.9411)
Size	−0.0632 *	−0.0674	0.0001	4.3285 ***	0.1439 ***	0.1969 ***	0.0022 ***
	(−1.7657)	(−1.3324)	(0.2063)	(52.0831)	(4.0936)	(4.0495)	(2.9698)
Cashflow	−5.7567 ***	−8.4725 ***	−0.0392 ***	33.5514 ***	−3.5821 ***	−5.1162 ***	−0.0470 ***
	(−10.2211)	(−10.6384)	(−4.3017)	(26.6209)	(−7.1788)	(−7.4154)	(−4.4863)
REC	0.8574 *	0.7076	0.0718 ***	6.4079 ***	0.2116	0.0128	0.0616 ***
	(1.9397)	(1.1320)	(10.0462)	(6.5035)	(0.5556)	(0.0242)	(7.7117)
INV	0.2960	0.3676	0.0157 ***	2.2539 ***	−0.2065	−0.3100	0.0112 *
	(0.8404)	(0.7381)	(2.7598)	(2.7909)	(−0.6626)	(−0.7192)	(1.7132)
ListAge	0.2171 ***	0.2214 **	−0.0029 **	−2.6727 ***	0.0669	0.0075	−0.0032 **
	(2.8421)	(2.0501)	(−2.3692)	(−16.0599)	(1.0312)	(0.0832)	(−2.3180)
Top 5	−1.2026 ***	−1.6965 ***	−0.0039	−2.7925 ***	−0.8708 ***	−1.3139 ***	−0.0028
	(−4.5313)	(−4.5205)	(−0.8975)	(−4.7654)	(−3.8478)	(−4.1989)	(−0.5853)
SOE	0.1305	0.2071 *	0.0085 ***	−1.6908 ***	0.1081	0.1735	0.0093 ***
	(1.4820)	(1.6631)	(5.9933)	(−8.4801)	(1.4011)	(1.6262)	(5.7322)
Occupy	15.6446 ***	24.7495 ***	0.1346 ***	−19.2459 ***	8.8800 ***	13.7865 ***	0.1585 ***
	(10.2662)	(11.4855)	(5.4607)	(−5.1994)	(6.2109)	(6.9739)	(5.2831)
Mfee	7.0724 ***	9.7533 ***	0.1735 ***	3.5794 ***	5.3669 ***	6.8973 ***	0.1498 ***
	(12.7018)	(12.3877)	(19.2663)	(2.6587)	(10.3285)	(9.6000)	(13.7398)
Constant	2.1884 **	2.3475 *	0.7236 ***	−78.1312 ***	−1.8619 **	−2.7747 **	0.6822 ***
	(2.4700)	(1.8738)	(50.4959)	(−37.9895)	(−2.2295)	(−2.4029)	(38.9275)
Ind	YES	YES	YES	YES	YES	YES	YES
Year	YES	YES	YES	YES	YES	YES	YES
Observations	18,131	18,131	18,131	13,517	13,517	13,517	13,517
R–squared	0.058	0.058	0.170	0.286	0.054	0.052	0.177

*, **, *** indicate a significance level of 10%, 5%, and 1%, respectively; robust standard errors are given in parentheses.

, columns (1)–(3) indicate the first step: the regression of digital transformation on the decline in performance, while column (4) presents the second step: the regression of digital transformation on analyst coverage, and columns (5)–(7) reflect the third step: the regression of the digital transformation and analyst coverage on the decline in performance. The digital transformation was significantly negatively related to the decline in ROA, decline in ROE, and decline in revenue at the 1% or 5% level, respectively. The digital transformation was positively related to analyst coverage at the 1% level. Analyst coverage was negatively related to the decline in ROA, decline in ROE, and decline in revenue at the 1% level. Overall, these results support that digital transformation enhances financial resilience through analyst coverage.

4.3.2. Mechanism Test: Financial Constraints

As shown in

Table 8. Mechanism test: mediating effect of financial constraints.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Variables	decline_ROA	decline_ROE	decline_OR	WW	decline_ROA	decline_ROE	decline_OR
DT	−0.1077 ***	−0.1469 ***	−0.0017 ***	−0.0148 ***	−0.1058 ***	−0.1409 ***	−0.0019 ***
	(−2.9665)	(−2.8519)	(−2.8228)	(−3.3750)	(−2.8281)	(−2.6518)	(−3.0688)
WW					0.1327 **	0.2081 **	0.0024 **
					(1.9689)	(2.1735)	(2.2011)
Size	−0.0634 *	−0.0708	0.0002		−0.0498	−0.0500	−0.0000
	(−1.7544)	(−1.3800)	(0.3056)		(−1.3238)	(−0.9350)	(−0.0641)
Cashflow	−6.1132 ***	−8.9444 ***	−0.0406 ***	−0.1229 *	−5.8827 ***	−8.5927 ***	−0.0402 ***
	(−10.7358)	(−11.0705)	(−4.4009)	(−1.7732)	(−10.0569)	(−10.3423)	(−4.1842)
REC	0.9005 **	0.7659	0.0706 ***	0.0715	0.0795	−0.2812	0.0635 ***
	(2.0104)	(1.2051)	(9.7326)	(1.3104)	(0.1716)	(−0.4275)	(8.3362)
INV	0.2436	0.3329	0.0143 **	−0.0685	0.1852	0.1785	0.0141 **
	(0.6856)	(0.6602)	(2.4793)	(−1.5956)	(0.5114)	(0.3470)	(2.3621)
ListAge	0.2059 ***	0.2168 *	−0.0025 *	−0.0328 ***	0.1817 **	0.1885 *	−0.0017
	(2.6164)	(1.9414)	(−1.9326)	(−3.5750)	(2.2962)	(1.6775)	(−1.3043)
Top 5	−1.2807 ***	−1.7843 ***	−0.0033	−0.2120 ***	−1.3120 ***	−1.8055 ***	0.0003
	(−4.7768)	(−4.6902)	(−0.7485)	(−6.7214)	(−4.7903)	(−4.6412)	(0.0595)
SOE	0.1530 *	0.2253 *	0.0082 ***	0.0177 *	0.2149 **	0.3046 **	0.0082 ***
	(1.7194)	(1.7843)	(5.7035)	(1.6560)	(2.3706)	(2.3648)	(5.5237)
Occupy	15.9096 ***	24.9855 ***	0.1390 ***	−0.3547*	16.0138 ***	25.5713 ***	0.1368 ***
	(10.2948)	(11.3944)	(5.5492)	(−1.8731)	(10.0045)	(11.2474)	(5.1978)
Mfee	6.8510 ***	9.7498 ***	0.1694 ***	0.3529 ***	5.7773 ***	8.5298 ***	0.1657 ***
	(12.1615)	(12.1977)	(18.5625)	(5.3991)	(9.9746)	(10.3683)	(17.3997)
Constant	2.1869 **	2.3038 *	0.7248 ***	−0.7911 ***	2.2919 **	2.4628 *	0.7218 ***
	(2.4377)	(1.8099)	(49.8609)	(−12.2832)	(2.4747)	(1.8723)	(47.3928)
Ind	YES	YES	YES	YES	YES	YES	YES
Year	YES	YES	YES	YES	YES	YES	YES
Observations	17,640	17,640	17,640	15,935	15,935	15,935	15,935
R–squared	0.172	0.059	0.059	0.014	0.057	0.058	0.171

*, **, *** indicate a significance level of 10%, 5%, and 1%, respectively; robust standard errors are given in parentheses.

, columns (1)–(3) indicate the first step: the regression of digital transformation on the decline in performance, while column (4) presents the second step: the regression of digital transformation on financial constraints, and columns (5)–(7) reflect the third step: the regression of the digital transformation and financial constraints on the decline in performance. The digital transformation was significantly negatively related to the decline in ROA, decline in ROE, and decline in revenue at the 1% or 5% level, respectively. The digital transformation was negatively related to financial constraints at the 1% level. Financial constraints were positively related to the decline in ROA, decline in ROE, and decline in revenue at the 5% level. Overall, these results support that digital transformation enhances financial resilience through releasing financial constraints.

4.3.3. Mechanism Test: Operation Risk

As shown in

Table 9. Mechanism test: mediating effect of operation risk.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Variables	decline_ROA	decline_ROE	decline_OR	CFV	decline_ROA	decline_ROE	decline_OR
DT	−0.1280 ***	−0.1877 ***	−0.0013 **	−0.0008 ***	−0.1258 ***	−0.1841 ***	−0.0013 **
	(−3.8061)	(−3.9357)	(−2.3131)	(−3.0933)	(−3.7302)	(−3.8488)	(−2.2857)
CFV					4.8628 ***	7.6161 ***	0.0333 **
					(4.8253)	(5.3288)	(2.0208)
Size	−0.0226	−0.0026	−0.0003	−0.0055 ***	−0.0043	0.0269	−0.0002
	(−0.6382)	(−0.0523)	(−0.5284)	(−20.6761)	(−0.1201)	(0.5278)	(−0.3146)
Cashflow	−6.3671 ***	−9.2975 ***	−0.0462 ***
	(−11.2786)	(−11.6138)	(−4.9893)
REC	0.0438	−0.3589	0.0814 ***	0.0041	0.8126 **	0.7617	0.0870 ***
	(0.1047)	(−0.6054)	(11.8622)	(1.3208)	(1.9658)	(1.2993)	(12.8500)
INV	0.2016	0.1439	0.0153 ***	0.0441 ***	0.6826 **	0.8235 *	0.0189 ***
	(0.5843)	(0.2939)	(2.6997)	(17.2938)	(1.9875)	(1.6906)	(3.3555)
Top 5	−1.1408 ***	−1.6027 ***	−0.0104 **	0.0160 ***	−1.4312 ***	−2.0349 ***	−0.0125 ***
	(−4.3379)	(−4.2973)	(−2.4180)	(8.1138)	(−5.4298)	(−5.4438)	(−2.8995)
SOE	0.1746 **	0.2420 **	0.0073 ***	−0.0042 ***	0.2531 ***	0.3589 ***	0.0079 ***
	(2.0134)	(1.9682)	(5.1468)	(−6.4710)	(2.9121)	(2.9113)	(5.5416)
ListAge	0.2080 ***	0.2229 **	−0.0029 **	0.0063 ***	0.1418 *	0.1229	−0.0033 ***
	(2.6872)	(2.0308)	(−2.2438)	(10.9382)	(1.8213)	(1.1135)	(−2.6043)
Occupy	15.1651 ***	23.7266 ***	0.1475 ***	0.0543 ***	16.8018 ***	26.0885 ***	0.1594 ***
	(9.9429)	(10.9697)	(5.8910)	(4.7782)	(11.0441)	(12.0917)	(6.4012)
Mfee	5.8622 ***	8.2892 ***	0.1545 ***	−0.0351 ***	6.8159 ***	9.7000 ***	0.1614 ***
	(10.7289)	(10.6980)	(17.2340)	(−8.6402)	(12.5128)	(12.5564)	(18.0964)
Constant	1.6605 **	1.4371	0.7379 ***	0.1588 ***	0.7520	0.0287	0.7316 ***
	(2.0654)	(1.2605)	(55.9260)	(26.3229)	(0.9149)	(0.0246)	(54.3631)
Ind	YES	YES	YES	YES	YES	YES	YES
Year	YES	YES	YES	YES	YES	YES	YES
Observations	17,633	17,633	17,633	17,633	17,633	17,633	17,633
R–squared	0.049	0.050	0.143	0.120	0.043	0.044	0.142

*, **, *** indicate a significance level of 10%, 5%, and 1%, respectively; robust standard errors are given in parentheses. To avoid multicollinearity, this paper did not control cash flow in the third step of the regression.

, columns (1)–(3) indicate the first step: the regression of digital transformation on the decline in performance, while column (4) presents the second step: the regression of digital transformation on operation risk, and columns (5)–(7) reflect the third step: the regression of the digital transformation and operation risk on the decline in performance. The digital transformation was significantly negatively related to the decline in ROA, decline in ROE, and decline in revenue at the 1% or 5% level, respectively. The digital transformation was negatively related to operation risk at the 1% level, and operation risk was positively related to the decline in ROA, decline in ROE, and decline in revenue at the 1% or 5% level, respectively. Due to digital transformation, information asymmetry is relieved and the predictability of information is enhanced, thereby strengthening financial resilience by reducing the risk of business.

5. Further Discussion

5.1. The Moderating Effect Analysis

5.1.1. Executives with a Digital Background

Since digital transformation not only has to be implemented but also secured and communicated, the digital transformation process calls for strong support from the top management team (TMT). Gerth and Peppard strengthen the importance of chief information officers’ (CIOs) and chief executive officers’ (CEOs) leadership in implementing digitization [38]. The CIO and the TMT should capture, integrate, and deliver information, and embed digital technologies into operational processes to achieve supply chain visibility. Furthermore, new information system governance policies will be established by the TMT to redesign the organizational structure. Both will improve supply chain resilience and organizational adaptability, which in turn will increase financial resilience. Accordingly, we predict that the executive with a digital background moderates the relationship between digital transformation and financial resilience.

Table 10. Mechanism test: moderating effect of the executives with a digital background.

	(1)	(2)	(3)	(4)	(5)	(6)
	Digital–Exe	Non–Digital–Exe	Digital–Exe	Non–Digital–Exe	Digital–Exe	Non–Digital–Exe
Variables	decline_ROA	decline_ROA	decline_ROE	decline_ROE	decline_OR	decline_OR
DT	−0.1359 ***	−0.0302	−0.1975 ***	−0.0291	−0.0012 **	−0.0002
	(−3.5977)	(−0.2153)	(−3.6796)	(−0.1532)	(−2.0536)	(−0.0732)
Constant	3.1744 ***	−2.0835	3.7137 ***	−5.1725	0.7308 ***	0.7977 ***
	(3.2930)	(−0.7283)	(2.7116)	(−1.3344)	(47.8549)	(15.7890)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Ind	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Observations	17,360	1533	17,360	1533	17,360	1533
R–squared	0.065	0.088	0.065	0.092	0.168	0.243

**, *** indicate a significance level of 5% and 1%, respectively; robust standard errors are given in parentheses.

reports the results of the subsample test regressions according to the digital background of the executives. Columns (1), (3), and (5) present the regression results of digitalization and the financial resilience for companies with digital background executives, while columns (2), (4), and (6) present the regression results for companies without digital background executives. From the regression results, the digital transformation (DT) coefficients of models (1), (3), and (5) were significantly negative at the 1% and 5% confidence levels, respectively, while the digital transformation (DT) coefficients of models (2), (4), and (6) were not significant. This indicates that with the rapid development of digital transformation, companies with executives with digital backgrounds have stronger financial resilience than companies without executives with digital backgrounds [39].

5.1.2. The Nature of Ownership

On the one hand, by assuming more social responsibility and political tasks, the entire management systems of state–owned enterprises (SOEs) tend to be stable, resulting in the fact that short–term reforms to realize digital transformation cannot be carried out. On the contrary, non–state–owned enterprises (Non–SOEs), which are mainly private enterprises and market–oriented, can perceive the updating development of new technologies. They are more courageous and likely to carry out rectification and use digital transformation to promote the improvement of financial resilience. On the other hand, state–owned enterprises rely on the government to support them in the face of crisis [40], while non–state–owned enterprises can rely mainly on themselves, and their demand for digital transformation is urgent. Accordingly, we anticipate that digital transformation will have a greater impact on the financial resilience of non–state–owned enterprises than that of state–owned enterprises.

Table 11. Mechanism test: moderating effect of the nature of ownership.

	(1)	(2)	(3)	(4)	(5)	(6)
	SOE	Non–SOE	SOE	Non–SOE	SOE	Non–SOE
Variables	decline_ROA	decline_ROA	decline_ROE	decline_ROE	decline_OR	decline_OR
DT	−0.0759	−0.1254 ***	0.0020	−0.1938 ***	0.0022 **	−0.0028 ***
	(−1.1334)	(−2.9306)	(0.8869)	(−3.1730)	(2.1931)	(−3.8611)
Constant	2.4573 *	4.3338 ***	0.9362 ***	4.8406 ***	0.7214 ***	0.7673 ***
	(1.7346)	(3.4800)	(19.8042)	(2.7223)	(34.3033)	(36.4045)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Ind	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Observations	7833	11,060	7829	11,060	7833	11,060
R–squared	0.052	0.093	0.237	0.092	0.199	0.167

*, **, *** indicate a significance level of 10%, 5%, and 1%, respectively; robust standard errors are given in parentheses.

reports the results of the subsample test regression by ownership, where the samples for columns (1), (3), and (5) are state–owned enterprises, and columns (2), (4), and (6) are non–state–owned enterprises. Digital transformation in columns (2), (4), and (6) was significantly negatively related to financial resilience at the 1% level. However, the digital transformation in column (5) was significantly positively related to financial resilience at the 5% level, while it was insignificant in columns (1) and (3). These results suggest that promoting digital transformation can enhance the financial resilience of non–state–owned enterprises, but not that of state–owned enterprises. Furthermore, the SUEST test showed that the coefficient of the grouping test was significant at the 1% level, indicating that digital transformation plays a stronger role in promoting the financial resilience of non–state–owned enterprises.

5.1.3. The Nature of High–Tech and Non–High–Tech Companies

According to the Industry Classification Guidelines for Listed Companies (2012 revised edition) issued by the China Securities Regulatory Commission and the “High and New Technology Fields Supported by the State” of “National Key Supported High–tech Fields in China, this paper defines Pharmaceutical Manufacturing (C27); Computer, Communication and Other Electronic Equipment Manufacturing (C39); Instrument Manufacturing (C40); Chemical Raw Materials and Chemical Products Manufacturing (C26); Chemical Fiber Manufacturing (C28); and Railway, Ship, Aerospace and Other Transport Equipment Manufacturing (C37) as high–tech industries. Otherwise, they are non–high–tech industries.

Table 12. Mechanism test: moderating effect of the nature of high–tech and non–high–tech firms.

	(1)	(2)	(3)	(4)	(5)	(6)
	High–Tech	Non–High–Tech	High–Tech	Non–High–Tech	High–Tech	Non–High–Tech
Variables	decline_ROA	decline_ROA	decline_ROE	decline_ROE	decline_OR	decline_OR
DT	−0.1307 **	−0.0655	−0.1837 **	−0.0847	−0.0002	0.0004
	(−2.0348)	(−1.5053)	(−2.0380)	(−1.3672)	(−0.3045)	(1.0108)
Constant	−1.3697	3.6531 ***	−1.9349	3.6378 **	0.7854 ***	0.8149 ***
	(−0.7920)	(3.0892)	(−0.7974)	(2.1599)	(49.7767)	(75.7748)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Ind	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Observations	4497	13,198	4497	13,198	4497	13,198
R–squared	0.050	0.061	0.047	0.061	0.299	0.253

**, *** indicate a significance level of 5% and 1%, respectively; robust standard errors are given in parentheses.

reports the results of the subsample test regression according to the nature of high–tech and non–high–tech companies. The digital transformation in columns (1) and (3) were all significantly negatively related to financial resilience at the 5% level. However, the digital transformation was not significant in columns (2), (4), and (6). These findings suggest that promoting digital transformation can enhance the financial resilience of high–tech companies, but not that of non–high–tech companies. Furthermore, the SUEST test result was 0.000***, showing that the coefficient of the grouping test was significant at the 1% level, indicating that digital transformation plays a stronger role in promoting financial resilience in high–tech companies.

High–tech companies, due to the nature of their work, are likely to be more digitized and more capable of using the digital technology of their firms to help them out of difficult situations. Non–high–tech companies, on the other hand, are not as digitized as high–tech firms, resulting in a lesser ability to leverage digitization. As a result, high–tech companies have a greater competitive advantage over non–high–tech companies, which manifests itself in a more significant effect of the degree of digitization on firm resilience.

5.2. Robustness Tests

This paper conducted a series of robustness checks to verify our main findings. First, we employed the two–stage least squares (2SLS) regression approach to address the endogeneity concern, as shown in

Table 13. Robustness test: IV 2SLS.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Variables	decline_ROA	decline_ROE	decline_OR	Instru_hangzhou	decline_ROA	decline_ROE	decline_OR
DT	−0.1069 ***	−0.1464 ***	−0.0015 **	0.4249 ***	−0.0300	−0.0307	−0.0012 *
	(−2.9779)	(−2.8830)	(−2.5739)	(5.4011)	(−0.9864)	(−0.7299)	(−1.8320)
Instru_hangzhou					−0.0293 ***	−0.0378 ***	−0.0004 ***
					(−8.7839)	(−8.2085)	(−5.9411)
Constant	2.1884 **	2.3475 *	0.7236 ***	−78.1312 ***	−1.8619 **	−2.7747 **	0.6822 ***
	(2.4700)	(1.8738)	(50.4959)	(−37.9895)	(−2.2295)	(−2.4029)	(38.9275)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Ind	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Observations	18,131	18,131	18,131	13,517	13,517	13,517	13,517
R–squared	0.058	0.058	0.170	0.286	0.054	0.052	0.177
Sobel test					−5.816 ***	−5.638 ***	−5.898 ***

*, **, *** indicate a significance level of 10%, 5%, and 1%, respectively; robust standard errors are given in parentheses.

. With reference to Ji [41], the “spherical distance between the listed companies’ location city and Hangzhou” (Instru_hangzhou) was chosen as the instrumental variable in this paper. The reason is that Alipay, which has been the most widely used mobile payment app globally since December 2021 (Juniper Research), originated in Hangzhou.

Second, this paper used time lag 1 for digital transformation and retested the model (1) as shown in columns (1)–(3) of

Table 14. Robustness test: Time lag 1 and PSM method.

	(1)	(2)	(3)	(4)	(5)	(6)
Variables	decline_ROA	decline_ROE	decline_OR	decline_ROA	decline_ROE	decline_OR
DT				−0.1032 **	−0.1513 ***	−0.0012 *
				(−2.4817)	(−2.6003)	(−1.7524)
DTt−1	−0.0742 *	−0.1072 *	−0.0014 **
	(−1.8785)	(−1.9111)	(−2.4206)
Constant	2.8094 ***	2.5813 *	0.7671 ***	0.9080	1.3106	0.7143 ***
	(2.8880)	(1.8681)	(54.0161)	(0.8603)	(0.8873)	(40.5638)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Ind	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Observations	16,465	16,465	16,465	12,874	12,874	12,874
R–squared	0.067	0.067	0.190	0.059	0.061	0.160

*, **, *** indicate a significance level of 10%, 5%, and 1%, respectively; robust standard errors are given in parentheses.

. Since digital transformation is a long–term process, its impact on the financial resilience of firms is not immediate.

Third, the propensity score matching (PSM) method was used to address the problem of sample selection bias, as shown in columns (4)–(6) in Table 14. The listed companies that have implemented digital transformation were seen as a treatment group, and others were seen as a control group. Then, the propensity score was calculated according to the firm size (Size), cashflow ratio (Cashflow), listing age (ListAge), and the nature of ownership (SOE) as confounders, and then 1:4 nearest neighbor matching was used to form the paired samples.

Fourth, the baseline regression was conducted again using another independent variable, DTA, as shown in columns (1)–(3) in

Table 15. Robustness test: changing dependent variables and adding control variables.

	(1)	(2)	(3)	(4)	(5)	(6)
Variables	decline_ROA	decline_ROE	decline_OR	decline_ROA	decline_ROE	decline_OR
DTA	−0.0760 *	−0.1038 *	−0.0027 ***
	(−1.9030)	(−1.8344)	(−4.1908)
DTIA				−0.1243 ***	−0.1830 ***	−0.0013 **
				(−3.4267)	(−3.5637)	(−2.2723)
Size	−0.0877 **	−0.0922 *	−0.0002	−0.0723 **	−0.0675	−0.0004
	(−2.4533)	(−1.8215)	(−0.2939)	(−2.0031)	(−1.3198)	(−0.7394)
Cashflow	−6.4434 ***	−9.3761 ***	−0.0364 ***	−6.4141 ***	−9.3673 ***	−0.0343 ***
	(−11.4845)	(−11.8001)	(−4.0474)	(−11.4231)	(−11.7793)	(−3.8103)
REC	0.8333 *	0.7815	0.0728 ***	0.9618 **	0.9712	0.0722 ***
	(1.8671)	(1.2364)	(10.1827)	(2.1478)	(1.5314)	(10.0688)
INV	0.2512	0.2496	0.0197 ***	0.2263	0.2063	0.0206 ***
	(0.7188)	(0.5042)	(3.5274)	(0.6471)	(0.4166)	(3.6729)
ListAge	0.2260 ***	0.2298 **	−0.0026 **	0.2073 ***	0.2058 *	−0.0028 **
	(2.9348)	(2.1067)	(−2.1392)	(2.6832)	(1.8809)	(−2.2526)
Top 5	−1.1857 ***	−1.6826 ***	−0.0037	−1.1764 ***	−1.6788 ***	−0.0036
	(−4.4238)	(−4.4328)	(−0.8648)	(−4.3863)	(−4.4197)	(−0.8451)
SOE	0.1097	0.1574	0.0077 ***	0.1079	0.1616	0.0073 ***
	(1.2358)	(1.2521)	(5.4333)	(1.2078)	(1.2768)	(5.1262)
Occupy	17.1885 ***	26.7453 ***	0.1272 ***	17.0342 ***	26.6202 ***	0.1221 ***
	(11.5117)	(12.6477)	(5.3209)	(11.3868)	(12.5647)	(5.0984)
Mfee	6.4433 ***	8.8140 ***	0.1562 ***	6.4689 ***	8.9212 ***	0.1511 ***
	(12.2247)	(11.8076)	(18.5085)	(12.1692)	(11.8498)	(17.7523)
GDP				0.0000	0.0000	−0.0000
				(0.8004)	(1.0219)	(−0.8060)
Inst				0.1005	0.0526	0.0063 ***
				(1.2144)	(0.4485)	(4.7483)
R&D				−14.3938 *	−15.4635	−0.3281 ***
				(−1.9255)	(−1.4606)	(−2.7412)
Constant	2.7293 ***	2.9346 **	0.7317 ***	2.4490 ***	2.4953 **	0.7350 ***
	(3.0631)	(2.3255)	(51.2840)	(2.7386)	(1.9702)	(51.3309)
Ind	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Observations	18,893	18,893	18,893	18,893	18,893	18,893
R–squared	0.062	0.063	0.168	0.063	0.064	0.168

*, **, *** indicate a significance level of 10%, 5%, and 1%, respectively; robust standard errors are given in parentheses.

. DTA is measured by the natural logarithm of the total number of keywords “digital technology application” frequencies, plus one.

Fifth, this paper employed digital–technology–related intangible assets (DTIA) instead of digital transformation. DTIA is measured by the ratio of intangible assets invested in digitalization to total assets in the current year, which are defined as the intangible assets related to the keywords of digital transformation. In addition, due to the varying levels of economic and scientific–technological development among cities, this paper added city characteristic variables such as per capita GDP (GDP), industrial structure (Inst), and R&D, as shown in columns (4)–(6) in Table 15.

The findings were found to be robust in all robustness tests.

6. Conclusions

Using a sample of 19,602 Chinese listed companies from 2010 to 2022, this paper empirically analyzed the impact of enterprise digital transformation on resilience from the perspective of financial sustainability in both the normal and abnormal states. In the normal state, firms with a higher degree of digital transformation exhibited stronger financial resilience. This finding remained robust when using IV 2SLS, time lag 1, and PSM, replacing the dependent variable, and adding control variables. However, the effect of firms’ digital transformation on financial resilience was not significant in abnormal states, such as the COVID-19 pandemic, consecutive losses, and delisting warnings. Mechanism tests indicated that digital transformation can enhance financial resilience by strengthening internal corporate governance and external governance, such as attracting analyst attention. Additionally, digital transformation can improve financial resilience by alleviating financing constraints and operation risk. Finally, digital transformation has a greater impact on financial resilience for companies with executives from digital–related backgrounds both for non–state–owned firms as well as for high–tech firms.

This study has several policy–related and practical implications. First, in the post–COVID-19 pandemic era, financial sustainability has become crucial for the survival of enterprises. Therefore, financial resilience is of great significance in achieving financial sustainability in times of crisis. Adopting the concept of financial resilience, which represents the capability of financial sustainability, this paper examined the role of digital transformation on financial resilience in normal and abnormal states. It is significant for exploring the strategic transformation of enterprises in abnormal states and discussing the impact of COVID-19 on the governance and finance of enterprises. This paper can expand the research on the emerging field of “enterprise resilience” from a financial sustainability perspective. Second, this paper proposed a set of financial resilience indicators based on revenue and earnings that can reflect the core competitiveness of enterprises, compared to the capital market stock indexes, which are widely used in the existing literature. Moreover, the related literature studies financial sustainability from the static perspective of revenue, ROE, and ROA [18,19], while this paper measured financial resilience by the dynamic decline of revenue, ROE, and ROA in the face of a crisis, which is an important addition to traditional financial sustainability research and establishes an important link between corporate resilience research and financial sustainability research. Third, this paper explored how digital transformation enhances financial resilience from perspectives of corporate governance, financing constraints, and operation risk, which proposes practical strategies to help firms strengthen their financial resilience. The research integrates digitalization, corporate governance, corporate finance, and operation risk and presents some interesting findings that support relevant theories.

Finally, the limitation of this paper is that it discusses only resilience and not recovery. In fact, resilience includes not only resistance but also recovery, which is the ability to remedy a bad situation quickly when one is in trouble. However, it is not the right time to test the recovery capability because many companies have not fully recovered from the COVID-19 pandemic to their original levels due to the increasing uncertainties from geopolitical complexities and the rise of anti–globalization. This paper argues that as the global economy recovers, scholars can initiate recovery–related research to improve the study of financial resilience in the future as firms realize performance recovery.