Does Fintech Development Reduce Corporate Earnings Management? Evidence from China

Abstract

This paper takes A-share companies listed on the Shanghai and Shenzhen stock exchanges from 2011 to 2020 as the research object and empirically tests the impact of fintech development on corporate earnings management and its mechanism. It is found that fintech development significantly reduces corporate earnings management. This conclusion still holds after a series of robustness tests. The mechanism test shows that fintech development reduces corporate earnings management by alleviating information asymmetry and easing corporate financing constraints. A heterogeneity test shows that fintech development has a greater effect on reducing earnings management in non-state-owned enterprises, small-scale enterprises and enterprises with low profitability, as well as enterprises in non-eastern China and regions with low marketization levels. This study clarifies the impact of fintech development on the field of micro corporate governance and provides rewarding policy implications for reducing corporate earnings management behavior, improving the level of corporate governance and facilitating the high-quality development of the capital market.

1. Introduction

The quality of enterprise accounting information, especially earnings information, is an important part of the high-quality development of the capital market. In recent years, earnings management, performance manipulation and even financial fraud are often seen among Chinese listed companies, which have produced a seriously negative impact on the legitimate rights and interests of investors and the healthy development of China’s capital market. Corporate earnings management will not only cause serious negative externalities to stakeholders including investors, creditors, regulators and enterprise employees, but also reduce investments with long-term benefits, such as investments in enterprise research and development, employee training, advertising, maintenance, etc. [1,2], which results in a weakening of enterprises’ market competitiveness and sacrificing of their long-term development. Existing studies on earnings management mainly focus on the implementation motivation [3,4,5,6,7,8] and governance mechanism [9]. However, in practice, the problem of corporate earnings management has not been solved well. The main reason is that “contract friction” (imperfect system) and “communication friction” (asymmetric information environment) exist in the process of enterprise operation [10]. Therefore, how to effectively supervise and constrain corporate earnings management is a problem with both theoretical value and practical significance, which deserves extensive attention from academia and industry.

In recent years, driven by emerging cutting-edge technologies such as big data, cloud computing, blockchain and artificial intelligence, fintech is developing rapidly in China and is reshaping the existing financial and economic ecology. The existing research shows that fintech development can strengthen banks’ risk control ability [11,12,13,14], improve commercial banks’ efficiency [15,16] and promote competition between banks [17] through “enabling”, thus forcing commercial banks to make digital transformations and change the traditional financial service mode. At the same time, fintech development can effectively ease corporate financing constraints [18] and promote entrepreneurship [19], innovation [20,21] and total factor productivity [22,23], thereby enhancing the ability of the financial sector to serve the real economy and facilitating the high-quality development of the economy.

However, few studies pay attention to the impact of fintech development on corporate governance, especially on earnings management. We focus on the impact of fintech development on the earnings management of entity enterprises (instead of financial sectors). The essence and core of many financial innovations based on fintech are to solve the problem of information asymmetry by means of emerging technologies. The development of fintech has changed the way of information dissemination and transmission efficiency. By reducing information asymmetry and alleviating “communication friction”, fintech provides possibilities for restraining corporate earnings management. For example, data mining, machine learning and fintech supervision have shown strong vitality in identifying enterprise financial information. Therefore, it is necessary to clarify the relationship between fintech and earnings management theoretically and empirically.

Based on the above analysis, this paper selects A-share companies listed on the Shanghai and Shenzhen stock exchanges from 2011 to 2020 as the research object to investigate the impact of fintech development on corporate earnings management and its mechanism. The research finds that fintech development significantly reduces corporate earnings management, and its main mechanism is to alleviate information asymmetry and ease corporate financing constraints; the effect of fintech development on reducing corporate earnings management is stronger in non-state-owned enterprises, small-scale enterprises and enterprises with low profitability, as well as enterprises in non-eastern China and regions with low marketization level.

This paper contributes to several aspects of the literature. First, this study is related to the economic effect of fintech development. Many existing studies focus on the impact of fintech development on traditional financial institutions [11,12,13,14,15,16,17]. Emerging studies find that fintech has positive effects on corporate financing, innovation and efficiency [18,19,20,21,22,23]. Few papers study the impact of fintech development on corporate governance. This study focuses on the impact of fintech development on the earnings management of entity enterprises (instead of financial sectors). Our findings supplement the strategic reactions of corporate information disclosure to local fintech development, which provides an overall understanding of how fintech curbs the earnings management behaviors of the management.

Second, we enrich the existing literature on financial development by revealing how emerging technologies in financial services sectors improve the quality of earnings information. Previous studies discuss the function of traditional financial intermediation from the aspects of the number of local bank branches [24] and private credit accessibility [25]. We highlight that regional fintech development can reduce earnings management and improve the quality of earnings information. Compared with the research of Gao et al. [26], we emphasize the impact of fintech development on reducing accrued earnings management (instead of real earnings management) from the two aspects of improving the information environment and reducing implementation motivation, because accrued earnings management is more sensitive to fintech development in theory.

Third, we provide new ideas for the governance of corporate earnings management. A large number of studies argue that various internal and external governance systems of enterprises, including the internal control system, supervisory board system, audit committee system, independent director system, external audit system and external governance environment [9], can constrain earnings management to a certain extent. However, we focus on the indispensable role of fintech in alleviating information asymmetry and easing financing constraints. We verify that fintech development improves the quality of earnings information by providing empirical evidence on the impact of fintech development on accrued earnings management.

2. Theoretical Analysis and Research Hypothesis

2.1. Fintech and Earnings Management

Corporate earnings management refers to the behavior that the management, on the basis of following accounting standards, uses accounting methods or operating activities to adjust or control accounting income information on an enterprise’s financial report, so as to mislead the stakeholders about the enterprise’s operating conditions and maximize the interests of the management or controlling shareholders. The occurrence of corporate earnings management mainly depends on the cost and benefit of doing so. When the benefit of earnings management exceeds the cost, enterprises will have incentives to implement earnings management. Benefits of earnings management include increasing the private interests of the management or major shareholders, raising the share price and obtaining more favorable financing conditions. However, from the perspective of a creditor, enterprise accounting information, especially earnings information, is an important channel for them to understand and judge the business status of enterprises and prevent credit risks. If an enterprise is labeled as implementing high-intensity earnings management, it will face stricter financing conditions and higher financing costs [27]. Therefore, the enterprise needs to balance the cost and benefit when deciding its earnings management level. Existing studies find that fintech development has positive effects on corporate governance [12,20,28,29,30]. On the one hand, as an effective external corporate governance mechanism, fintech can help identify corporate earnings management actions, give full play to its role of external supervision, and then increase the risk and cost of implementing earnings management. On the other hand, fintech development helps improve enterprise operating performance, thereby reducing the motivation and benefits of the enterprise to implement earnings management based on performance pressure and financing needs. To sum up, fintech development leads to an increase in the cost of implementing earnings management and a decrease in the benefit of doing so, which inhibits corporate earnings management to a certain extent. Therefore, this paper proposes the research hypothesis H1.

Hypothesis 1 (H1).

Fintech development reduces corporate earnings management.

2.2. The Influence Mechanism of Fintech on Corporate Earnings Management

Considering a firm’s conditions and motivations to implement earnings management, fintech development may reduce earnings management by alleviating information asymmetry and easing corporate financing constraints.

2.2.1. Alleviating Information Asymmetry

According to the principal–agent theory, information asymmetry is the source of earnings management by the enterprise management [31]. Alleviating information asymmetry (solving communication friction) can be regarded as a way of ex ante supervision because it raises the cost of implementing earnings management, which helps reduce earnings management. Fintech development can alleviate the information asymmetry in the process of information release, information transmission and information reception in the principal–agent relationship, which helps to improve the quality of enterprise information disclosure and curb the behavior of corporate earnings management [32]. (1) From the perspective of information release, fintech development can reduce the private ownership of corporate earnings information and attenuate the ability and motivation of managers to hide private information through mechanisms of constraints and incentives. By virtue of its powerful data search and analysis capabilities, fintech can obtain all kinds of soft and hard enterprise information, leading to a decline in the ability of the management to hide private earnings information, thus restraining the corporate earnings management. At the same time, fintech development has promoted enterprise innovation and improved total factor productivity, thereby intensifying the competition between enterprises. The management will voluntarily improve the quality of information disclosure in order to pass on the competitive advantages of company value and product quality [33], and their motivation to hide private information decreases. In this way, the effect of incentives reduces corporate earnings management. (2) From the perspective of information transmission, fintech development can not only help external stakeholders collect more enterprise soft information, but also “harden” soft information, reduce manual collection and decision-making processes and improve the transmissibility of information [34]. At the same time, with the help of the Internet and social media platforms, fintech has accelerated the spread of information flow, lowered the loss and lag of earnings information in the process of transmission and disclosure and effectively alleviated the formation of insider information in the traditional transmission channel. In addition, external stakeholders are enabled to effectively supervise and restrain the earnings management behavior of enterprises because now they can efficiently communicate with each other and receive feedback information through online comments. (3) From the perspective of information reception, fintech helps external enterprise stakeholders reduce the cost of searching for earnings information and improve their ability of analyzing information. Fintech relies on big data to build massive database resources, which can obtain not only hard information such as enterprise finance, transaction and operation information, but also the unquantifiable soft information of enterprises. Meanwhile, it expands the scope of information sharing, greatly enriches the sources and types of earnings information and reduces the cost of information acquisition. In addition, fintech can effectively make up for the lack of information analysis ability of enterprise external stakeholders through artificial intelligence algorithms such as machine learning. In summary, this paper proposes the research hypothesis H2.

Hypothesis 2 (H2).

Fintech development reduces corporate earnings management by alleviating information asymmetry.

2.2.2. Easing Financing Constraints

From the perspective of behavioral motivation, financing motivation is one of the important motivations for Chinese enterprises to implement earnings management. In order to meet the profit requirements for external financing, enterprises are motivated to adjust their profits through earnings management, so as to obtain more external funds at a lower cost [35]. The easing of corporate financing constraints will also lead to a reduction in the possibility of earnings management by the management out of financing motivation. As a newcomer in the field of finance, fintech is the product of deep integration of finance and technology. Moreover, fintech development has improved the traditional service mode of financial institutions, which can now accurately identify enterprises’ financing needs and then ease their financing constraints. Specifically, fintech development eases financing constraints from the following three aspects: First of all, financial institutions use fintech to integrate and process all kinds of data in batches, which helps reduce costs in customer acquisition, information processing, risk control and customer maintenance. The application of fintech has greatly reduced the marginal cost of financial institutions to provide financial services, especially for a large number of small and micro enterprises, which need low-cost financial services. Secondly, “Big Tech Lending”, with the big science and technology ecosystem and big data risk control model as its core, has the outstanding advantage of not relying on mortgage assets. It breaks through the traditional credit risk management framework and greatly improves the availability of financing. For those small- and medium-sized enterprises that lack complete financial data and sufficient mortgage assets, “Big Tech Lending” is particularly helpful in providing them with sustainable credit support [36]. Finally, relying on information technology, fintech can deeply excavate enterprise information, guide the flow of credit resources to high-quality enterprises, especially small and medium-sized private enterprises, and optimize the allocation of credit resources. The “enabling” of fintech has significantly improved the credit identification ability of traditional financial institutions, which can promote the continuous improvement of the financial system and reduce the degree of financial resource mismatch, thus playing a positive role in reducing corporate financing constraints [37]. To sum up, fintech development can improve the traditional financial service mode, reduce the financing cost of enterprises, raise the availability of financing and the efficiency of financial resource allocation, effectively ease corporate financing constraints and ultimately reduce the motivation of the management to implement earnings management. Therefore, this paper proposes the research hypothesis H3.

Hypothesis 3 (H3).

Fintech development reduces corporate earnings management by easing corporate financing constraints.

3. Study Design

3.1. Data Source and Sample Selection

Since 2011, a series of Internet financial products represented by “Yu’e Bao” have provided a wide range of financial services for the public. At the same time, driven by emerging cutting-edge technologies such as big data, cloud computing, blockchain and artificial intelligence, China’s fintech has also developed rapidly. Moreover, the published data of the Digital Financial Inclusion Index by the Institute of Digital Finance at Peking University starts from 2011. Therefore, our research sample consists of A-share companies listed on the Shanghai and Shenzhen stock exchanges from 2011 to 2020. The samples are screened according to the following principles: (1) exclude financial companies; (2) exclude ST, ST * and PT companies; (3) exclude samples with missing values. All continuous variables are winsorized at 1% in both tails to avoid the influence of extreme values. We obtain a final sample of 24,774 firm-year observations. The data of firm-level variables are obtained from the China Stock Market & Accounting Research (CSMAR) Database. The fintech data (the Digital Financial Inclusion Index) are from the Institute of Digital Finance at Peking University and Ant Financial Services Group [38].

3.2. Empirical Model Design

3.2.1. Baseline Model

To analyze the impact of fintech development on corporate earnings management, based on the previous theoretical analysis and research hypothesis H1, this study sets the following panel econometric model:

$$D{A}_{i,t}={\alpha }_0+{\alpha }_{1}FINTEC{H}_{m,t}+{\alpha }_{2}CV{S}_{i,t}+{\displaystyle \sum IND}+{\displaystyle \sum YEAR}+{\epsilon }_{i,t}$$

(1)

In Equation (1), the dependent variable DA_i,t represents the degree of earnings management of enterprise i in year t. The independent variable FINTECH_m,t indicates the level of fintech development in region m in year t. If the regression coefficient α₁ is significantly negative, it indicates that fintech development significantly reduces corporate earnings management. CVS_i,t represents a series of control variables that affect earnings management. ΣIND and ΣYEAR represent industry fixed effect and year fixed effect, respectively, which are used to control the influence of industry-level factors that do not change over time and year-level factors that do not change with individuals. ε_i,t represents a random error term.

3.2.2. Mediation Effect Model

To further verify the mechanism of the impact of fintech development on corporate earnings management, this paper follows the research of Wen et al. [39] and uses the mediation effect model to verify research hypothesis H2 and research hypothesis H3. The model settings are as follows:

$$ME{D}_{i,t}={\beta }_0+{\beta }_{1}FINTEC{H}_{m,t}+{\beta }_{2}CV{S}_{i,t}+{\displaystyle \sum IND}+{\displaystyle \sum YEAR}+{\epsilon }_{i,t}$$

(2)

$$D{A}_{i,t}={\gamma }_0+{\gamma }_{1}FINTEC{H}_{m,t}+{\gamma }_{2}ME{D}_{i,t}+{\gamma }_{3}CV{S}_{i,t}+{\displaystyle \sum IND}+{\displaystyle \sum YEAR}+{\epsilon }_{i,t}$$

(3)

In Equation (2), MED_i,t represents mediating variables. Specifically, in this study, information asymmetry (ASY) and corporate financing constraints (FCs) are the two mediating variables. The specification of other variables is consistent with that in Equation (1). According to the mediation effect test method, when α₁ is significant: if β₁ and γ₂ are significant but γ₁ is not significant, then there is a complete mediation effect; if β₁, γ₂ and γ₁ are all significant, then there is a partial mediation effect; if only one of β₁ and γ₂ is significant, then a Sobel test is needed to test whether a mediation effect exists.

3.3. Definition and Measurement of Variables

3.3.1. Dependent Variable

The dependent variable of this paper is corporate earnings management (DA).

Table 1. Main variable names and definitions.

Type	Symbol	Name	Definition
Dependent variable	DA	Corporate earnings management	Discretionary accrual, see full article for details
Independent variable	FINTECH	Fintech	Digital Financial Inclusion Index by Peking University/100
Mediating variables	ASY	Information asymmetry	Refer to the research of Bharath et al. [27]
	FCs	Corporate financing constraints	Absolute value of SA Index
Controlled variables	SIZE	Company size	Natural logarithm of enterprise’s total asset
	LEV	Asset-liability ratio	Total liabilities/total asset
	ROA	Profitability	Net profit/[(total assets at the beginning of the year + total assets at the end of the year)/2]
	GROW	Growth rate	(Current operating income—previous operating income)/previous operating income
	EQUITY	Equity concentration	Shareholding ratio of the largest shareholder
	AUDIT	Audit quality	If the auditor’s report is from the “Big Four” (PWC, DTT, EY and KPMG), then the value is 1, otherwise the value is 0
	MWAGE	Management remuneration	Natural logarithm of the total remuneration of the top three directors, supervisors and senior executives
	DUAL	Duality of chairman and general manager	If the chairman of the board of directors and the general manager is the same person, then the value is 1, otherwise the value is 0
	MSHARE	Management shareholding ratio	Ratio of the number of shares held by directors, supervisors and senior executives to the total number of shares
	INDEP	Independent director ratio	Ratio of the number of independent directors to the total number of directors

shows the variable names and definitions of this study. Following the revised Jones Model by Dechow et al. [7], discretionary accrual is used in this paper as the proxy for corporate earnings management. The model to calculate discretionary accrual is as follows:

$$\frac{T{A}_{i,t}}{A_{i,t-1}}={\beta }_1\frac{1}{A_{i,t-1}}+{\beta }_2\frac{\Delta RE{V}_{i,t}-\Delta RE{C}_{i,t}}{A_{i,t-1}}+{\beta }_3\frac{PP{E}_{i,t}}{A_{i,t-1}}+{\epsilon }_{i,t}$$

(4)

In Equation (4), TA_i,t refers to the difference between the enterprise net profit and the net cash flow from operating activities; A_i,t−1 represents the total assets at the end of period t − 1; ΔREV_i,t represents the increment of the enterprise operating income; ΔREC_i,t indicates the increment of enterprise accounts receivable; PPE_i,t represents enterprise fixed assets.

For each industry and accounting year in the research sample (excluding sample groups with less than 10 observations after industry classification and with missing relevant data), we estimate using Equation (4), respectively, and define the residual error obtained from the regression as the corporate discretionary accrual. Because this paper puts emphasis on the extent of corporate earnings management, the absolute value of discretionary accrual is used as the proxy of corporate earnings management. The larger the value, the more serious the corporate earnings management. In the part of robustness test, we use real earnings management (REM) as the proxy of earnings management to further verify the reliability of our conclusions.

3.3.2. Independent Variable

The independent variable of this paper is fintech (FINTECH). Following the research by Ding et al. [20] and Chen et al. [23], this paper selects the provincial level Digital Financial Inclusion Index of China released by the Institute of Digital Finance, Peking University as the proxy of the level of fintech development. The Digital Financial Inclusion Index includes three dimensions: coverage, depth of use and degree of digitalization, which, respectively, reflect the coverage of digital accounts, the real use of digital financial products and services, and the customer experience of the facilitation, materialization and credit of financial services. Compared with other single-dimensional fintech proxies, such as the number of fintech companies in each province [22] and the number of fintech patents per capita [26], the Digital Financial Inclusion Index depicts China’s fintech development level from a multidimensional and comprehensive perspective [38]. At the same time, the index is constructed based on hundreds of millions of bottom-level data of Ant Financial Trading Account, which can accurately reflect the level of regional fintech development. Specifically, the Digital Financial Inclusion Index is one single index, which is consolidated from 33 indicators of digital financial inclusion from the above three dimensions. The calculation methodology includes non-dimensionalization, analytical hierarchy process and index synthesis. The published data of the index start from 2011. The index is a representative measure of fintech, which is widely used in a large number of studies [20,23,32]. In 2011, the mean of the fintech (FINTECH) was 0.40, then it grew to 2.20 in 2015 and further rose to 3.41 in 2020. In the robustness test part, this paper uses the number of fintech companies to replace the Digital Financial Inclusion Index to measure the level of fintech development and further tests the reliability of this conclusions.

3.3.3. Mediating Variables

The mediating variables selected in this paper are information asymmetry (ASY) and corporate financing constraints (FCs). The specific measures are as follows: (1) The measurement of information asymmetry (ASY). Drawing on the variable design of Bharath et al. [40], this paper uses ASY to measure the degree of information asymmetry in the principal–agent relationship. The larger the value of ASY, the higher the degree of information asymmetry. (2) The measurement of corporate financing constraints (FCs). Using the research of Hadlock and Pierce [41] for reference, this paper uses the SA index (SA = −0.737 × size + 0.043 × size² − 0.04 × age) constructed with two strongly exogenous variables, enterprise size and enterprise age, as the proxy of corporate financing constraints. The SA index has the characteristics of overcoming endogeneity, easy calculation and strong robustness. The SA index is negative, so it is treated as an absolute value. The greater the absolute value, the higher the financing constraints faced by the enterprise.

3.3.4. Controlled Variables

Following the relevant literature, this study has added a series of variables that can influence corporate earnings management [42,43]. The controlled variables include company size (SIZE), asset liability ratio (LEV), profitability (ROA), growth rate (GROW), equity concentration (EQUITY), audit quality (AUDIT), management remuneration (MWAGE), duality of chairman and general manager (DUAL), management shareholding ratio (MSHARE) and independent director ratio (INDEP).

3.4. Summary Statistics and Correlations

Table 2. Summary statistics.

Variable	N	Mean	SD	Min	Median	Max
DA	24,774	0.068	0.071	0.001	0.047	0.416
REM	24,774	0.138	0.140	0.002	0.096	0.757
FINTECH	24,774	2.584	0.986	0.324	2.681	4.319
ASY	24,774	−0.034	0.612	−1.957	0.019	1.756
FCs	24,774	3.796	0.244	3.126	3.797	4.393
SIZE	24,774	22.249	1.290	19.969	22.064	26.230
LEV	24,774	0.428	0.205	0.055	0.421	0.887
ROA	24,774	0.040	0.061	−0.229	0.038	0.209
GROW	24,774	0.160	0.392	−0.561	0.100	2.439
EQUITY	24,774	0.345	0.149	0.086	0.324	0.748
AUDIT	24,774	0.060	0.237	0.000	0.000	1.000
MWAGE	24,774	14.498	0.689	12.899	14.463	16.470
DUAL	24,774	0.267	0.442	0.000	0.000	1.000
MSHARE	24,774	0.128	0.192	0.000	0.004	0.674
INDEP	24,774	0.376	0.054	0.333	0.364	0.571

reports the summary statistics of variables used in the model. The mean value of corporate earnings management (DA) is 0.068, the standard deviation is 0.071, the minimum value is 0.001 and the maximum value is 0.416. The mean and standard deviation of DA are quite close. This reflects that there is significant difference in the accrual earnings management among Chinese listed companies, especially among parts of listed companies. The mean value of real earnings management (REM) is 0.138 and the standard deviation is 0.140, indicating that there is also significant difference in real earnings management (REM) among Chinese listed companies. Therefore, it is of great practical significance to explore the influencing factors of the earnings management of Chinese listed companies. In terms of fintech, the average value of fintech (FINTECH) is 2.584, the standard deviation is 0.986, the minimum value is 0.324 and the maximum value is 4.319. To be specific, the average level of fintech is approximately 60% of the maximum value. This indicates that the level of fintech development varies greatly among different provinces in China, while the average level is relatively high. Summary statistics of other variables are similar to previous studies.

Table 3. Correlation matrix and VIF values.

Variable	DA	FINTECH	SIZE	LEV	ROA	GROW	EQUITY	AUDIT	MWAGE	DUAL	MSHARE	INDEP
DA	1
FINTECH	−0.0723 ***	1
SIZE	−0.0646 ***	0.1029 ***	1
LEV	0.0708 ***	−0.0579 ***	0.5196 ***	1
ROA	−0.1306 ***	−0.0208 ***	−0.0053	−0.3592 ***	1
GROW	0.0988 ***	−0.0486 ***	0.0322 ***	0.0210 ***	0.2495 ***	1
EQUITY	0.0316 ***	−0.0813 ***	0.2083 ***	0.0660 ***	0.1225 ***	−0.0018	1
AUDIT	−0.0418 ***	0.0245 ***	0.3511 ***	0.1116 ***	0.0412 ***	−0.0171 ***	0.1477 ***	1
MWAGE	−0.0502 ***	0.3524 ***	0.4468 ***	0.1077 ***	0.1921 ***	0.0309 ***	−0.0158 **	0.2238 ***	1
DUAL	0.0113 *	0.0896 ***	−0.1846 ***	−0.1342 ***	0.0451 ***	0.0304 ***	−0.0522 ***	−0.0704 ***	−0.0321 ***	1
MSHARE	−0.0068	0.0915 ***	−0.3493 ***	−0.3175 ***	0.1645 ***	0.0790 ***	−0.1097 ***	−0.1245 ***	−0.0888 ***	0.2495 ***	1
INDEP	−0.0147 *	0.0619 ***	0.0052	−0.0074	−0.0191 ***	−0.0022	0.0387 ***	0.0338 ***	−0.0134 **	0.1143 ***	0.0684 ***	1
VIF		1.22	2.11	1.75	1.42	1.10	1.11	1.17	1.55	1.10	1.26	1.03

Notes: ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

reports the correlation coefficient values between the main variables and VIF values. Firstly, the correlation coefficient between variables DA and FINTECH is significantly negative, indicating that there is a negative correlation between fintech development variables and the corporate earnings management variable. Secondly, the correlation coefficients between the independent variable and controlled variables are not high and the VIF values of the variables are all less than 10, indicating that there is no problem of multicollinearity between variables.

4. Empirical Results and Analysis

4.1. Baseline Regression

To verify the research hypothesis H1, this paper estimates Equation (1) using the stepwise regression method. At the same time, considering the strong correlation between enterprises in the same industry, the statistical significance is corrected by using robust standard errors clustered at the industry level in the regression.

Table 4. Baseline regression results.

Variables	(1)	(2)	(3)
	DA	DA	DA
FINTECH	−0.00263 ***	−0.00197 ***	−0.00185 ***
	(0.00040)	(0.00048)	(0.00043)
SIZE		−0.00641 ***	−0.00623 ***
		(0.00161)	(0.00154)
LEV		0.01912 ***	0.01868 ***
		(0.00355)	(0.00349)
ROA		−0.17208 ***	−0.17311 ***
		(0.04327)	(0.04173)
GROW		0.02166 ***	0.02188 ***
		(0.00272)	(0.00254)
EQUITY		0.00077	0.00134
		(0.00431)	(0.00446)
AUDIT		−0.00203	−0.00193
		(0.00280)	(0.00276)
MWAGE		−0.00276 ***	−0.00246 ***
		(0.00081)	(0.00084)
DUAL		0.00064	0.00050
		(0.00089)	(0.00091)
MSHARE		−0.00480 **	−0.00557 **
		(0.00222)	(0.00254)
INDEP		−0.01629 *	−0.01433 *
		(0.00813)	(0.00767)
Constant	0.07493 ***	0.16102 ***	0.16246 ***
	(0.01044)	(0.02317)	(0.02423)
Industry FE	YES	YES	YES
Year FE	YES	YES	YES
Industry × Year FE	NO	NO	YES
Observations	24,774	24,774	24,774
Adj.R2	0.2751	0.2835	0.5023

Notes: Robust standard errors clustered at the industry level are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively. According to most previous studies, p-values to determine statistical significance are from a two-tailed test. Usually, Stata regression results only show two-tailed p-values. We can calculate the one-tailed p-values directly from the regression output by using the following formula. Column (1): H0: α₁ = 0, p-value =0.00004, H0: α₁ ≤ 0, p-value = 1 − (0.00004/2) = 0.99998; Column (2): H0: α₁ = 0, p-value =0.001, H0: α₁ ≤ 0, p-value =1 − (0.001/2) = 0.9995; Column (3): H0: α₁ = 0, p-value =0.001, H0: α₁ ≤ 0, p-value = 1 − (0.001/2) = 0.9995. p-values from a one-tailed test still yield the same result.

reports the estimation results of the impact of fintech development on corporate earnings management. In Column (1), without controlled variables, every 1 unit increase in a province’s fintech development is associated with a 0.263% decrease in local firms’ accrued earnings management (p < 0.01). The proportion of the firm’s accrued earnings management in total assets has increased 0.263%, while the average ROA of the firm is only 4%. By comparison, we can see that regional fintech development has an important impact on corporate earnings management in the corresponding province. In addition, for an average-sized firm in our sample, its total asset is approximately 4.6 billion (exp (22.249)) RMB. Thus, a 0.263% decrease in accrued earnings management, which is measured in proportion to total asset, means an approximately 12 million (4.6 billion × 0.263%) RMB earnings adjustment. In Column (2), we add firm-level controlled variables. Although the coefficient has changed after being absorbed by the controlled variables, every 1 unit increase in a province’s fintech development is still associated with a 0.197% decrease in local firms’ accrued earnings management. According to the above analysis, it means an about 9 million (4.6 billion × 0.193%) RMB earnings adjustment. In Column (3), we control Industry × Year fixed effects to rule out the impacts from unobservable industry-level factors that change over time. After controlling for Industry × Year fixed effects, the effect of a 1 unit increase in a province’s fintech development is associated with a 0.185% decrease in local firms’ accrued earnings management (p < 0.01), which means an about 8.5 million (4.6 billion × 0.185%) RMB earnings adjustment. To sum up, we believe that fintech development increases the cost and decreases the income of implementing earnings management, which in turn plays a positive role in reducing earnings management. That is, fintech development helps to reduce earnings management. It fails to reject H1. The above results infer the effectiveness of fintech development as an external corporate governance mechanism. To further take advantage of the positive effect of fintech on strengthening corporate governance and improving the quality of corporate information disclosure, we should promote a deeper integration of finance and technology and wider application of fintech. Regarding the estimates of the coefficients of the controlled variables, all of them are consistent with the literature.

4.2. Robustness Test

To further verify the reliability of the above empirical results, this paper uses methods such as replacing the measure of independent variable and dependent variable, excluding special samples, using one-period lagged fintech development and using instrument variable estimation to test the robustness.

Table 5. Robustness test results.

Variables	(1)	(2)	(3)	(4)	(5)
	DA	REM 1	DA	DA	DA
FINTECH	−0.00023 ***	−0.00436 ***	−0.00317 ***	−0.00225 ***	−0.00247 ***
	(0.00008)	(0.00121)	(0.00059)	(0.00064)	(0.00061)
Control	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES
Observations	16944	24,774	19848	24,774	24,774
Adj.R2	0.2898	0.2683	0.2975	0.2735	0.2923

Notes: Robust standard errors clustered at the industry level are reported in parentheses. *** indicates statistical significance at the 1% level. ¹ Real earnings management (REM) mainly includes production manipulation (PROD), sales manipulation (CFO) and discretionary expenditure manipulation (DISEXP). This paper follows the research of Roychowdhury [45], Cohen and Zarowin [46] to measure them, respectively, and uses the formula (REM = PROD-CFO-DISEXP) to measure the degree of corporate real earnings management. The greater the absolute value of this indicator, the higher the corporate real earnings management. The data are from the China Stock Market & Accounting Research (CSMAR) Database. Due to space limitation, the regression results of control variables are not listed in Table 5, but are available on request. (The same below).

reports the robustness test results.

4.2.1. Replace Measure of Fintech Development

To eliminate the influence of the selection error of the independent variable on the research results, following the research of Song et al. [22], this paper uses the number of fintech companies in each province to measure the level of regional fintech development and re-estimates Equation (1). In Column (1) of Table 5, every 1% increase in the number of provincial fintech companies is associated with a 0.023% decrease in local firms’ accrued earnings management (p < 0.01). The previous conclusion is still valid.

4.2.2. Replace Measure of Corporate Earnings Management

To eliminate the influence of the selection error of the dependent variable on the research results, this paper uses real earnings management (REM) to replace discretionary accruals (DA) to measure the degree of corporate earnings management and re-estimates Equation (1). In Column (2) of Table 5, every 1 unit increase in a province’s fintech development is associated with a 0.436% decrease in local firms’ real earnings management (p < 0.01). The previous conclusion is still valid.

4.2.3. Exclude Special Samples

The municipality directly under the Central Government has its particularity in China’s administrative division. At the same time, the fintech of the municipality has developed to a decent level, while their overall corporate earnings management may be low. As a result, the reverse causality problem can be serious. Therefore, this paper removes the research samples of Beijing, Shanghai, Tianjin and Chongqing and re-estimates Equation (1). In Column (3) of Table 5, every 1 unit increase in a province’s fintech development is associated with a 0.317% decrease in local firms’ accrued earnings management (p < 0.01). The previous conclusion is still valid.

4.2.4. Use One-Period Lagged Fintech Development

Fintech development and corporate earnings management will have mutual effects on each other in the current period, but the corporate earnings management in the current period is unlikely to affect the fintech development in the previous period. Therefore, to alleviate this reverse causality, this paper uses one-period lagged fintech (L.FINTECH) as the independent variable. In addition, using one-period lagged fintech also takes into account the possibility of a lag in the effect of fintech development on corporate earnings management. In Column (4) of Table 5, every 1 unit increase in a province’s fintech development is associated with a 0.225% decrease in local firms’ accrued earnings management (p < 0.01). The previous conclusion is still valid.

4.2.5. Use Instrument Variable Estimation

To address the potential endogeneity of the independent variable (FINTECH) as much as possible, this paper follows the research method of Chong et al. [44] and Song et al. [22] to construct instrument variables. To be specific, the average value of the Digital Financial Inclusion Index of the three provinces with the closest real GDP per capita to a certain province (IV) is used as the instrument variable of the province’s fintech development. On the one hand, the provinces with close per capita real GDP have similar characteristics, such as economic and financial development levels and industrial layouts; therefore, their fintech development levels are strongly correlated. On the other hand, there is a significant geographical segmentation between provinces in China. The level of fintech development in three provinces with similar per capita GDP is unlikely to directly affect the corporate earnings management in another province. Therefore, the instrument variables reconstructed in this paper conform to the assumptions of relevance and exogeneity to a certain extent. In this paper, the two-stage least-square method (2SLS) is used for estimation. First, the DWH endogeneity test is conducted for FINTECH. The DWH statistic value is 11.76521 and the p-value is 0.00000, indicating that FINTECH in Equation (1) has an endogenous problem. In the first stage of regression, the coefficient of the instrument variable (IV) is 1.01248 (p < 0.01), indicating that the instrument variable (IV) is positively correlated with the independent variable (FINTECH). It verifies that the selection of the instrument variable meets the correlation assumption. Column (5) of Table 5 reports the results of the second-stage regression. Every 1 unit increase in a province’s fintech development is associated with a 0.247% decrease in local firms’ accrued earnings management (p < 0.01), indicating that the conclusion that fintech development significantly reduces corporate earnings management is still valid after considering endogenous issues.

5. Further Analysis

5.1. Influence Mechanism Test

The study above shows that fintech development significantly reduces corporate earnings management. However, it must be noted that it only describes the relationship between fintech and earnings management as a whole, and does not test the mechanism inside the black box. To further reveal the internal mechanism of the impact of fintech development on corporate earnings management, this paper follows the theoretical analysis and uses the mediation effect model to test the influence channels of fintech development from two aspects: information asymmetry and financing constraints.

5.1.1. Information Asymmetry Channel

To verify the research hypothesis H2 and test whether fintech development reduces corporate earnings management by alleviating information asymmetry, this paper estimates Equations (1)–(3) by treating information asymmetry (ASY) as the mediating variable.

Table 6. Information asymmetry mechanism test results.

Variables	(1)	(2)	(3)
	DA	ASY	DA
FINTECH	−0.00197 ***	−0.01461 ***	−0.00172 ***
	(0.00048)	(0.00485)	(0.00051)
ASY			0.01680 ***
			(0.00401)
Control	YES	YES	YES
Industry FE	YES	YES	YES
Year FE	YES	YES	YES
Observations	24,774	24,774	24,774
Adj.R2	0.2835	0.2432	0.2998

Notes: Robust standard errors clustered at the industry level are reported in parentheses. *** indicates statistical significance at the 1% level.

reports the information asymmetry mechanism test results. In Column (2) of Table 6, the coefficient of fintech (FINTECH) on information asymmetry (ASY) is significantly negative, indicating that fintech development significantly alleviates information asymmetry. In Column (3) of Table 6, the coefficient of information asymmetry (ASY) on corporate earnings management (DA) is significantly positive, indicating that the alleviation of information asymmetry significantly reduces corporate earnings management. According to the mediation effect test rules, the information asymmetry channel passes the test and is a partial mediating factor. To sum up, fintech development can reduce corporate earnings management by alleviating the information asymmetry. It fails to reject H2.

5.1.2. Financing Constraints Channel

To verify the research hypothesis H3 and test whether fintech development reduces corporate earnings management by easing corporate financing constraints, this paper estimates Equations (1)–(3) by treating corporate financing constraints (FCs) as the mediating variable.

Table 7. Financing constraints mechanism test results.

Variables	(1)	(2)	(3)
	DA	FCs	DA
FINTECH	−0.00197 ***	−0.04684 ***	−0.00169 ***
	(0.00048)	(0.01180)	(0.00036)
FCs			0.00594 ***
			(0.00147)
Control	YES	YES	YES
Industry FE	YES	YES	YES
Year FE	YES	YES	YES
Observations	24,774	24,774	24,774
Adj.R2	0.2835	0.2378	0.2969

Notes: Robust standard errors clustered at the industry level are reported in parentheses. *** indicates statistical significance at the 1% level.

reports the financing constraints mechanism test results. In Column (2) of Table 7, the coefficient of fintech (FINTECH) on corporate financing constraints (FCs) is significantly negative, indicating that fintech development significantly eases corporate financing constraints. In Column (3) of Table 7, the coefficient of corporate financing constraints (FCs) on corporate earnings management (DA) is significantly positive, indicating that the easing of corporate financing constraints significantly reduces corporate earnings management. According to the mediation effect test rules, the financing constraints channel passes the test and is a partial mediating factor. To sum up, fintech development can reduce corporate earnings management by easing corporate financing constraints. It fails to reject H3.

5.2. Heterogeneity Test

The study above empirically tests the positive role of fintech development in reducing corporate earnings management and deeply analyzes the mechanisms of the impact of fintech development on corporate earnings management from the perspective of information asymmetry and financing constraints. However, it is worth noting that previous studies have shown that enterprises’ internal and external environments have an important impact on earnings management. So, will the effect of fintech development on reducing earnings management exhibit heterogeneity? Therefore, this paper further analyzes the heterogeneous impact of fintech development on earnings management from two aspects: enterprise characteristics (nature of property right, enterprise size, profitability) and external environment (location, marketization level).

5.2.1. Heterogeneity Test Based on Enterprise Characteristics

The heterogeneity test of property right, enterprise size and profitability introduces the interaction term of fintech and property right (FINTECH $\times$ SOE), the interaction term of fintech and enterprise size (FINTECH $\times$ SIZE) and the interaction term of fintech and profitability (FINTECH $\times$ ROA), respectively, into Equation (1) for estimation.

Table 8. Heterogeneity test results based on enterprise characteristics.

Variables	(1)	(2)	(3)
	Property Right Difference	Size Difference	Profitability Difference
	DA	DA	DA
FINTECH	−0.00295 ***	−0.05766 ***	−0.00351 ***
	(0.00058)	(0.01748)	(0.00086)
FINTECH $\times$ SOE 1	0.00224 **
	(0.00095)
FINTECH $\times$ SIZE		0.00223 **
		(0.00089)
FINTECH $\times$ ROA			0.01643 **
			(0.00821)
Control	YES	YES	YES
Industry FE	YES	YES	YES
Year FE	YES	YES	YES
Observations	24,774	24,774	24,774
Adj.R2	0.2893	0.2955	0.2916

Notes: Robust standard errors clustered at the industry level are reported in parentheses. *** and ** indicate statistical significance at the 1% and 5% levels, respectively. ¹ SOE is a dummy variable. When an enterprise belongs to state-owned enterprises (SOEs), SOE = 1; when an enterprise belongs to non-state-owned enterprises (non-SOEs), SOE = 0.

reports heterogeneity test results based on enterprise characteristics.

In Column (1) of Table 8, the coefficient of the interaction term (FINTECH $\times$ SOE) is positive (p < 0.05), indicating that, compared with SOEs, the effect of fintech development on reducing earnings management is stronger in non-SOEs. In Column (2) of Table 8, the coefficient of the interaction term (FINTECH $\times$ SIZE) is positive (p < 0.05), indicating that, compared with large-scale enterprises, the effect of fintech development on reducing earnings management is stronger in small-scale enterprises. The main reason is that, compared with SOEs and large-scale enterprises, non-SOEs and small-scale enterprises face more serious “ownership discrimination” and “scale discrimination” in the financing process. Thereby, their financing constraints are more serious. Fintech development eases the financing constraints of non-SOEs enterprises and small-scale enterprises to a greater level, which will lead to a sharper decline in the motivation of earnings management. To sum up, the effect of fintech development on reducing earnings management is stronger in non-SOEs and small-scale enterprises.

In Column (3) of Table 8, the coefficient of the interaction term (FINTECH $\times$ ROA) is positive (p < 0.05), indicating that the lower the profitability of an enterprise, the greater the effect of fintech development on reducing earnings management. This is because enterprises with lower profitability are faced with greater pressure of performance assessment, supervision, capital market performance and so on. The management has greater motivation to implement earnings management, resulting in a higher level of earnings management. However, fintech development greatly alleviates the information asymmetry, improves the transparency of enterprise information and plays an effective role in external supervision, which reduces the space for enterprises to implement earnings management and increases its cost and the risk. To sum up, the effect of fintech development on reducing earnings management is stronger in enterprises with low profitability.

5.2.2. Heterogeneity Test Based on External Environment

The regional heterogeneity test and marketization heterogeneity test are to introduce the interaction terms of fintech and region (FINTECH $\times$ EAST) and fintech and marketization (FINTECH $\times$ MARKET) into Equation (1) for estimation.

Table 9. Heterogeneity test results based on external environment.

Variables	(1)	(2)
	Region Difference	Marketization 1 Difference
	DA	DA
FINTECH	−0.00282 ***	−0.00428 **
	(0.00073)	(0.00169)
FINTECH $\times$ EAST 2	0.00159 ***
	(0.00053)
FINTECH $\times$ MARKET		0.00186 **
		(0.00076)
Control	YES	YES
Industry FE	YES	YES
Year FE	YES	YES
Observations	24,774	24,774
Adj.R2	0.2895	0.2679

Notes: Robust standard errors clustered at the industry level are reported in parentheses. *** and ** indicate statistical significance at the 1% and 5% levels, respectively. ¹ Based on previous studies, this paper uses Fan Gang index to measure regional marketization level. The index comprehensively reflects the marketization level of various regions in China from five aspects: the relationship between the government and the market, the development of the non-state-owned economy, the development of the product market, the development of the factor market and the development of intermediary organizations and the law. At the same time, the index is often used by the academic community to evaluate institutional quality in various regions. The data come from the report of Wang et al. [47]. ² EAST is a dummy variable. When an enterprise is located in eastern China, EAST = 1; when an enterprise is located in non-eastern China, EAST = 0. Eastern China includes Beijing, Shanghai, Tianjin, Jiangsu, Zhejiang, Guangdong, Fujian, Hainan, Shandong, Liaoning and Hebei; non-Eastern China includes the other provinces. Eastern China firms are identified according to their registration place.

reports heterogeneity test results based on external environment.

In Column (1) of Table 9, the coefficient of the interaction term (FINTECH $\times$ EAST) is positive (p < 0.01), indicating that, compared with eastern China, the effect of fintech development on reducing earnings management is stronger in non-eastern China. In Column (2) of Table 9, the coefficient of the interaction term (FINTECH $\times$ MARKET) is positive (p < 0.05), indicating that the lower the level of marketization, the greater the effect of fintech development on reducing earnings management. This is because non-eastern China and regions with low levels of marketization have relatively poor external environments such as government intervention, marketization and rule of law, and the overall level of corporate governance is also relatively poor. Fintech development plays a “timely role” in reducing corporate earnings management in this region by alleviating information asymmetry and easing financing constraints, while it only plays an “icing on the cake” role in reducing corporate earnings management in eastern China and regions with higher marketization levels. To sum up, the effect of fintech development on reducing earnings management is stronger in enterprises in non-eastern China and regions with low levels of marketization.

6. Conclusions

This paper selects A-share companies listed on the Shanghai and Shenzhen stock exchanges from 2011 to 2020 as the research object. Based on the systematic theoretical analysis of the relationship between fintech development and earnings management, this paper empirically tests the impact of fintech development on earnings management and its mechanism. The main conclusions are as follows: (1) Fintech development significantly reduces corporate earnings management. This conclusion is still valid after a series of robustness tests, including replacing the measure of fintech development and corporate earnings management, excluding special samples, using one-period lagged fintech development and using instrument variable estimation. (2) Fintech development reduces corporate earnings management by alleviating information asymmetry and easing corporate financing constraints. It shows that fintech development improves the information flow, alleviates the information asymmetry in the principal–agent relationship, improves the information transparency and effectively plays the role of external supervision. At the same time, fintech development reduces the motivation of the management to implement earnings management by easing financing constraints, thereby reducing corporate earnings management. (3) Enterprise characteristics and external environment will affect the performance of fintech development in governing corporate earnings management. That is, fintech development has a greater effect on reducing earnings management in non-state-owned enterprises, small-scale enterprises and enterprises with low profitability, as well as enterprises in non-eastern China and regions with low marketization levels.

The findings of this study provide the following rewarding implications for improving the effect of fintech on reducing earnings management, strengthening corporate governance and facilitating the high-quality development of China’s capital market. First, for the government, it should promote the overall and deep integration of finance and technology. The government should improve the information infrastructure and increase the investment in research and development of the bottom technology of fintech, so as to comprehensively promote the application of fintech in a wider range. At the same time, the government should actively guide the effective flow of information among various departments and break the “isolated data island” dilemma. Second, for enterprises, they should be fully aware of the positive role of fintech development. Listed companies should actively adapt to and integrate into the development trend of the “digital economy” era, strive to improve the corporate governance system and curb the earnings management behaviors. Third, for regulatory authorities, they should make full use of the powerful information search and analysis capabilities of fintech to effectively supervise corporate earnings management and strive to create a healthy environment with better data security and privacy protection. Fourth, for investors and creditors, they should improve their abilities of using fintech. On the one hand, they should deeply understand the connotation of fintech and make full use of advanced information technology to improve their ability to search for and identify information. On the other hand, they should take advantage of the Internet and social media platforms, which facilitate information exchange and transmission, to fulfill their responsibilities of external governance.

This study has certain limitations. We only consider the impact of fintech development on corporate earnings management. However, earnings management is only one aspect of the quality of information disclosure. There are many other aspects, such as social responsibility, earnings forecast and environmental information disclosure, which are not covered in previous studies and this study. Therefore, we can follow up with future research on information disclosure from other perspectives.