Assessing the Impact of Financial Technology Innovations on the Sustainable Profitability of Listed Commercial Banks in China

Abstract

Commercial banks constitute a crucial segment of China’s financial system, and their efficient operation is directly linked to the development of other sectors within the national economy. The sustainable profitability of these banks is vital for maintaining the stability of China’s financial system. In the context of the current digital economy, it is of great theoretical and practical significance to conduct an in-depth analysis of the impact of financial technology (fintech) development on the sustainable profitability of commercial banks and its underlying mechanisms. Such research can promote the digital transformation of commercial banks, enhance risk supervision policies, and mitigate systemic financial risks. This study utilizes EViews software Version 13 to analyze annual data from 13 listed commercial banks in China over the period from 2011 to 2021. It examines the influence of fintech on the profitability of these banks, considering their unique characteristics and drawing insights from the existing literature on the mechanisms through which fintech affects bank profitability. Employing both a static panel fixed effects variable-intercept model and a dynamic panel generalized method of moments (GMM) model, the empirical findings indicate that fintech development significantly impacts the profitability of listed commercial banks.

1. Introduction

Financial technology (fintech), which represents the amalgamation of finance and technology, has introduced innovative solutions that enhance efficiency, accessibility, and customer experience within the financial sector [1]. As an information-intensive industry, the financial sector’s vast market provides a broad stage for the application of technology. Meanwhile, market demand has raised expectations for the digitization and intelligence of financial services, compelling the financial industry to increase its investment in technology. The investment potential of financial technology is widely recognized by investors and has become a popular field in the capital market. In addition, the outbreak of the COVID-19 pandemic has significantly impacted the production and life of countries worldwide. However, it has also presented new opportunities for digitalization and non-contact economic development [2]. Against this context, fintech is rapidly advancing and revolutionizing the financial services industry worldwide, including in China [3].

Fintech brings opportunities for the transformation and upgrading of commercial banks; the Internet is booming, so that banks can quickly access to the data, which has an important role in analyzing the business and customers for the bank, and brings a huge role for the bank to provide better services in order to achieve differentiated and customized services [4]. However, the development of fintech has also had a certain impact on the traditional business of commercial banks. Especially in credit, payment and settlement, and wealth management business, the rise of fintech has challenged the profitability of commercial banks to a certain extent [5].

Notably, in the current decade, China has started to lead the adoption and development of fintech. As of the beginning of 2020, China’s use of fintech has ranked first in the world, and the application rate in various fields has reached 69%, far exceeding the United States [6].

The commercial banks, the mainstay of China’s financial system, have been confronted by this pressing transformation and have become a frontier of fintech development [7]. Given the above backdrop, this study aims to provide a holistic exploration of the role of fintech innovation on the sustainable profitability of commercial banks, taking China as an example.

However, previous studies mainly discuss the impact of fintech on commercial banks theoretically, while the literature conducting the analysis from a quantitative perspective is limited. In addition, the limited studies of the impact of fintech on commercial bank profitability shows varying results [8,9]. Therefore, in order to contribute to the existing literature and clarify the relationship between fintech and commercial bank profitability under the context of rapidly expanding fintech, this study chooses the annual panel data of 13 representative listed commercial banks in China from 2011 to 2021, and empirically examines the impact of fintech innovation on the sustainable profitability of commercial banks, employing the static panel fixed effects variable-intercept model and the dynamic panel generalized method of moments (GMM). Unlike previous studies that only use fixed effects models [10], the application of the GMM model can alleviate the endogenous concerns occurring in the estimation based on the fixed effects model, thus ensuring the reliability of our empirical results [11].

The remaining of this paper is organized as follows: Section 2 provides the past literature and hypothesis development. Section 3 details the corresponding research methods and data. Section 4 offers empirical analysis and results. Section 5 discuss the results. Some conclusions and recommendations are provided in Section 6.

2. Literature Review and Hypothesis Development

In the current decade, fintech has gained increasing attention in the literature [12,13,14]. A considerable amount of studies investigate the various aspects of fintech, such as the connotation, evolution, and future of fintech [15,16,17,18], fintech and lending and payment innovations [19,20,21,22,23], fintech and sustainability [24], fintech development and risks [3,7,25], and technical dimensions in fintech [26]. Another group of articles focus on the effects of fintech on traditional financial institutions, especially in commercial banks. Specifically, Koffi [4] points out that fintech positively effects the traditional commercial banks by improving the service efficiency, lowering operation costs and credit thresholds, and increasing market competitiveness. Similarly, Li et al. [27] empirically confirm the positive impact of fintech start-ups on the share returns of retail banks, indicating the complementary effect between fintech and traditional banking. Wu, Luo, and Tao [11] find that fintech significantly promotes bank credit expansion. Chen et al. [28] conclude that fintech products positively affect the commercial bank’s non-financial performance, including the service quality and work efficiency, which is in line with Lee et al. [29], who discovered that the development of fintech can facilitate the efficiency of banks. In addition, from the perspective of risk taking, Li et al. [30] argue that fintech innovation reduces bank’s risk taking. Wu et al. [31] indicate that fintech innovation raises the risk of credit and liquidity, while lowering the insolvency risk.

Regarding the impact on the bank’s financial performance, previous studies show varying results, and there is still some hesitation regarding this subject in quantitative analysis. To be specific, Bömer and Maxin [32] report that the adoption of fintech may drive product upgrades and service innovation and further increase commercial bank profitability by employing the multiple-case study approach. Singh, Malik, and Jain [8] empirically investigate the data from Indian banks, concluding that improvement in the fintech innovation can promote the bank’s profitability. Nevertheless, some scholars have produced different results. They argue that the development of fintech erodes the profitability and asset quality of banks due to the loss of commercial bank competitiveness [9]. Additionally, Lv et al. [33] apply the error correction model (ECM) and Granger causality to data from the Industrial and Commercial Bank of China (ICBC) from 2011 to 2020 in order to examine how fintech development impacts the profitability of banks. They reveal that there is a “U”-shaped Granger causal relationship between fintech and the bank’s profitability, implying that fintech development initially reduces it, but increases it in later stages. Consistent with this, Song, Yu, and He [10] also find that fintech’s early development has a negative correlation with the profitability of commercial banks because the competitive effect outweighed the technological spillover effect. But, as fintech advances, technology spillovers become prominent, finally resulting in the enhancing of bank’s profitability.

Through the review of the literature, we found that there are positive and negative effects of fintech development on the profitability of commercial banks. With the rapid development of fintech, commercial banks have gradually achieved the establishment of data systems, which improve service quality and operational efficiency. However, the intermediary function of commercial banks is weakened due to the expansion of fintech firms. Therefore, the overall impact of fintech on the profitability of commercial banks remains unknown. Based on this, we posit the following two hypotheses:

H1a: 

Fintech has a positive impact on the profitability of commercial banks.

H1b: 

Fintech has a negative impact on the profitability of commercial banks.

3. Materials and Methods

3.1. Research Materials

3.1.1. Research Sample

Annual data were collected from 13 listed commercial banks in China. The sample comprises 5 state-owned banks and 8 joint-stock commercial banks. The specific institutions included in the analysis are the Bank of China, Agricultural Bank of China, Industrial and Commercial Bank of China, China Construction Bank, Bank of Communications, China Merchants Bank, Ping An Bank, Shanghai Pudong Development Bank, Industrial Bank of China, China Everbright Bank, China Minsheng Bank, China CITIC Bank, and Huaxia Bank. The assets held by these 13 banks contribute more than half of the total bank assets in China, suggesting a good representation of China’s banking sector.

3.1.2. Data Sources

The data are sourced from the Peking University Digital Inclusive Finance Index, compiled by Guo et al. [34], and the annual reports of various companies. The sample period is 2011–2021, including 143 observations.

3.1.3. Variable Definition

• Explained variables: ROE, ROA

The profitability of commercial banks refers to their ability to obtain profits through asset business, liability business, intermediary business, and narrow off-balance sheet business. The profitability indicator measures the ability of commercial banks to use funds to obtain profits, while controlling the cost and expense expenditures. Referring to Goddard et al. [35], return on equity (ROE) is selected as a proxy variable for bank profitability. The bank’s return on total assets (ROA) is also selected as a proxy variable to robustly test the findings of the study.

Return on equity is a core indicator for measuring the profitability of commercial banks, mainly used to measure the efficiency of commercial banks in using their own capital. The high or low return on total assets directly reflects the competitive strength and development ability of commercial banks, which stems from their ability to concentrate on the relationship between asset utilization efficiency and fund utilization effect. The use of total asset returns on investment indicators can analyze the stability and persistence of commercial bank profits, and can determine the specific scope and intensity of the impact of financial technology on listed commercial banks. By combining various indicators of return on total assets and return on net assets, the level of risk in a company’s operations can be explained.

• Core variable: FT

Guo, Wang, Wang, Kong, Zhang, and Cheng [34] construct the development level of digital inclusive finance at the provincial, municipal, and county levels in China from 33 indicator systems across 3 dimensions. This study uses the Peking University Digital Inclusive Finance Index at the prefecture level as a proxy variable for the development of financial technology, including the Overall Development Index (FinTech) and its components, namely coverage breadth, usage depth, and digitization. For local banks, prefecture-level city data are used, while, for national banks, national average data are utilized. This index comprehensively considers the new features of the Internet-based financial model, using massive data from financial technology enterprises, and constructs a digital inclusive finance index from multiple dimensions, such as coverage, depth of use, and digitization. This approach allows for a more comprehensive, multi-angled, and accurate reflection of the current state of regional financial technology development. Coverage breadth is mainly reflected by the number of electronic accounts, the depth of use is measured by the actual use of fintech services, and the degree of digitization considers the convenience and cost of users utilizing fintech services.

• Control variables

Referring to the existing research, select the bank’s cost-to-income ratio (CIR), equity-to-liability ratio (ETD), deposit-to-loan ratio (LDR), logarithmic total assets (LNA), non-performing loan ratio (NPL), and capital adequacy ratio (RAR) [36,37,38,39,40,41]. Specifically, the lower the cost-to-income ratio of a bank, the stronger its profitability. Therefore, the cost-to-income ratio is one of the important indicators for evaluating the profitability of commercial banks. A high debt-to-equity ratio indicates that the audited entity has a high level of debt capital and a low level of reliability in safeguarding that debt capital. Conversely, a low debt-to-equity ratio indicates that the audited entity has strong financial strength and a more reliable guarantee of debt capital. In addition, the natural logarithm of total assets can represent the total asset size of a commercial bank. When the total assets double, the natural logarithm of the total assets will increase by a fixed amount, providing a clear representation of the total asset growth. The non-performing loan ratio is used to evaluate the loan quality of banks, and is divided into three categories as follows: subprime, suspicious, and loss. A high non-performing loan ratio in a commercial bank will inevitably affect its profitability. The capital adequacy ratio reflects the degree to which a commercial bank can absorb losses with its own capital before the assets of depositors and creditors are compromised. This ratio is a key indicator of the bank’s ability to protect its stakeholders and ensure financial stability.

When conducting research and analysis, controlling variables is important [42]. These variables have undeniable relationships with each other, with core variables, and with the dependent variable, such as non-performing loan balances, leading to low capital adequacy ratios. Controlling variables is often crucial to various details in the analysis process and can be easily overlooked or mishandled. Therefore, when dealing with control variables, it is essential to maintain a rigorous attitude.

In summary, we identified the study variables.

Table 1. Definition of variables.

	Variable Name	Symbolic	Variable Definition
Explained variable	Return on Equity	ROE	Net Profit/Average Net Assets × 100%
	Return on Total Assets	ROA	Net Profit/Average Total Assets
Core variable	FinTech Index	FT	FinTech Index
Control variable	Equity-to-Debt Ratio	ETD	Total Debt/Shareholder Equity
	Cost-to-Income Ratio	CIR	Business and Management Expenses/Operating Income
	Loan-to-Deposit Ratio	LDR	Total Loan Amount/Total Deposit Amount
	Logarithm of Total Assets	LNA	The Natural Logarithm of Total Assets
	Risk-Adjusted Return	RAR	Capital/Risk Weighted Capital
	Non Performing Loan Ratio	NPL	(Secondary + Suspicious + Loss)/Total Loan Amount

summarizes the variables and their definitions.

3.2. Research Methods

3.2.1. Descriptive Statistics of Variables

Descriptive statistics is a statistical method employed to describe and summarize the fundamental characteristics of a dataset. It involves the calculation and presentation of statistical measures of central tendency, such as mean, median, and mode, as well as measures of dispersion, including variance, standard deviation, and range. Additionally, it encompasses the assessment of the data’s distribution shape through frequency distribution, percentiles, skewness, and kurtosis. Descriptive statistics enable a comprehensive understanding of the dataset’s basic characteristics and structure, thereby providing a foundation for further data analysis and statistical inference. The descriptive statistics in

Table 2. Descriptive statistics of variables.

	ROA	ROE	FT	ETD	CIR	LDR	LNA	RAR	NPL
Mean	1.029112	15.21804	284.7559	0.081361	0.546781	0.792067	29.57454	0.131383	0.014503
Median	1.021417	14.74079	286.37	0.074343	0.553824	0.761793	29.46653	0.1308	0.0146
Maximum	1.474801	25.31472	458.97	0.313018	0.798914	1.122249	31.19125	0.1802	0.074
Minimum	0.501373	6.180777	61.76	0.050231	0.291677	0.561094	27.84947	0.0878	0.0038
Std. Dev.	0.21836	4.544011	110.3788	0.035552	0.088587	0.12304	0.840221	0.019587	0.00774
Skewness	0.137742	0.303389	−0.320407	4.039348	−0.392243	0.737807	0.13314	0.407297	4.730397
Kurtosis	2.375518	2.164478	2.126157	21.72047	4.226105	2.984541	2.061503	2.730626	33.53947
Jarque–Bera	2.775803	6.353227	6.996538	2477.006	12.62423	12.97531	5.670439	4.386075	6090.403
Probability	0.249599	0.041727	0.03025	0	0.001814	0.001522	0.058706	0.111577	0
Sum	147.163	2176.18	40,720.1	11.63457	78.18961	113.2656	4229.159	18.7878	2.0739
Sum Sq. Dev.	6.770722	2932.021	1,730,053	0.179476	1.114356	2.149711	100.2479	0.054476	0.008507
Observations	143	143	143	143	143	143	143	143	143

are based on the study sample and data described above.

3.2.2. Model Setting and Method Selection

To clarify and investigate how fintech influences the profitability of commercial banks, this study conducts the fixed effects variable-intercept model and the dynamic panel generalized method of moments (GMM) model. In current practical application research, researchers mainly use the fixed effects variable-intercept model. The fixed effects variable-intercept model is a linear regression model primarily used for panel data analysis, and assumes that each individual has its own unique intercept, but that all individuals share the same regression slope. The fixed effects model controls for possible unobserved heterogeneity among individuals and eliminates bias due to these heterogeneities by including individual fixed effects in the regression equation. In addition, considering the possible dynamic characteristics of the explanatory variables, we then constructed a dynamic panel GMM model. This model is suitable for datasets containing multiple observations over time and uses the lagged values of the variables as instrumental variables in order to effectively address endogeneity.

4. Results

4.1. Static Panel Fixed Effects Variable-Intercept Model

4.1.1. Fixed Effects Variable-Intercept Model

Using Eviews 13.0 for empirical analysis, the fixed effects variable-intercept model was selected and the individual time-point fixed effects model was set as follows:

$$ROE/RO{A}_{it}={\beta }_0+{\beta }_{1}F{T}_{it}+\gamma Control{s}_{it}+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(1)

The results of the fixed effects variable intercept model in

Table 3. Fixed effects variable-intercept model.

Variable	Coefficient	Std. Error	t-Statistic	Prob.
C	5.517515	61.55978	0.089629	0.9287
FT	0.062074	0.029569	2.099272	0.038
CIR	−16.3785	2.240935	−7.308778	0
ETD	−1.367857	3.792664	−0.360659	0.719
LDR	−5.760939	1.837496	−3.135212	0.0022
LNA	0.250259	1.926603	0.129897	0.8969
NPL	−36.93244	15.78378	−2.339898	0.021
RAR	−4.945048	15.37978	−0.321529	0.7484
	Effects Specification
Cross-section fixed (dummy variables)
Period fixed (dummy variables)
R-squared	0.946854	Mean dependent var		15.77965
Adjusted R-squared	0.933215	S.D. dependent var		4.326471
S.E. of regression	1.118079	Akaike info criterion		3.245225
Sum squared resid	141.2614	Schwarz criterion		3.866801
Log likelihood	−202.0336	Hannan–Quinn criter.		3.497804
F-statistic	69.42182	Durbin–Watson stat		0.850538
Prob (F-statistic)	0

show that the estimated coefficient of FT is positive with a p-value of <0.05, indicating that the increase in the level of development of financial technology has brought about an increase in the profitability of commercial banks.

4.1.2. Hausman Test

A Hausman test was conducted to explain why we employ the fixed effects model rather than the random effects model. As shown in

Table 4. Hausman test.

Variable	Coefficient	Std. Error	t-Statistic	Prob.
C	61.64422	16.6082	3.711675	0.0003
FT	−0.01729	0.002703	−6.396362	0
CIR	−14.00207	2.302531	−6.081165	0
ETD	−5.924227	4.253078	−1.392927	0.1659
LDR	−8.756802	1.843773	−4.749393	0
LNA	−0.775457	0.580758	−1.335251	0.184
NPL	−57.20698	17.91695	−3.192898	0.0018
RAR	−16.01055	12.19526	−1.31285	0.1915
	Effects Specification
			S.D.	Rho
Cross-section random			1.301059	0.5752
Period random			0	0
Idiosyncratic random			1.118079	0.4248
	Weighted Statistics
R-squared	0.892891	Mean dependent var		3.957916
Adjusted R-squared	0.887338	S.D. dependent var		4.039416
S.E. of regression	1.355838	Sum squared resid		248.1702
F-statistic	160.7719	Durbin–Watson stat		0.830814
Prob (F-statistic)	0
	Unweighted Statistics
R-squared	0.684811	Mean dependent var		15.77965
Sum squared resid	837.7742	Durbin–Watson stat		0.246108

, the standard deviation of the period random effect is 0. This indicates that the period random effect does not fluctuate at different time points, making the sample unsuitable for a random effects model.

4.1.3. Fixed Effects Test

A fixed effects test was conducted, and the results in

Table 5. Fixed effects test.

Effects Test	Statistic	d.f.	Prob.
Cross-section F	21.459601	−12,113	0
Cross-section Chi-square	169.813454	12	0
Period F	3.712306	−10,113	0.0003
Period Chi-square	40.621657	10	0
Cross-Section/Period F	16.733251	−22,113	0
Cross-Section/Period Chi-square	207.171654	22	0

(p < 0.05) indicated that there were individual and point-in-time fixed effects, confirming that the model was set up correctly. The primary purpose of conducting a fixed effects test is to determine whether individual fixed effects and time fixed effects should be included in our panel data model. The result of the test is typically measured by the value of (p).

In the medium fixed effects test, the (p)-value was found to be less than 0.05, indicating a statistically significant result. This (p)-value provides sufficient evidence to reject the null hypothesis, suggesting that there are significant individual fixed effects and time fixed effects in the data. Consequently, it can be inferred that different individuals (such as various commercial banks or regions) and different time points (such as distinct years or quarters) have a significant and independent impact on the dependent variables (such as the profitability of commercial banks).

This finding further supports the appropriateness of our choice of the fixed effects model. By incorporating individual fixed effects and time fixed effects, it is possible to estimate the parameters of the model more accurately and capture the potential heterogeneity in the data. This heterogeneity may be attributed to specific factors associated with different individuals or time points, such as policy changes, economic cycles, technological progress, and other relevant variables.

In conclusion, by building the fixed effects model, we reveal that fintech has a significant positive impact on commercial banks’ profitability, validating Hypothesis 1a.

4.2. Dynamic Panel GMM Modeling

4.2.1. Dynamic Panel GMM Estimation

In analyzing the impact of the level of fintech (FT) development on the profitability of commercial banks, a dynamic panel model is constructed to account for the potential dynamic characteristics of the explanatory variables. This model not only considers the current period’s level of fintech development, but also incorporates the previous period’s value of the dependent variable (i.e., commercial bank profitability) to capture its possible lagged effect.

To address the potential endogeneity problem of the model, a dynamic panel generalized method of moments (GMM) estimation method is employed. This approach provides more accurate parameter estimates by introducing instrumental variables to control for potential bias. In contrast to the difference GMM estimation method proposed by Arellano and Bond [43], which produces a significant finite sample bias, Blundell and Bond [44] propose a system GMM estimator that combines the difference equation with the level equation. This method also incorporates a set of lagged difference variables as instruments corresponding to the level equation, thereby exhibiting better finite sample properties.

The estimation results in

Table 6. Dynamic panel GMM model.

Variable	Coefficient	Std. Error	t-Statistic	Prob.
WROE(−1)	−0.227659	0.411606	−0.553099	0.5813
FT	0.019055	0.005509	3.459129	0.0008
CIR	−54.17524	21.82047	−2.482772	0.0146
EDT	−4.607416	15.80605	−0.291497	0.7712
LDR	−34.54615	14.00687	−2.466372	0.0152
LNA	−12.27927	4.554011	−2.696364	0.0081
NPL	−186.6758	70.03215	−2.665573	0.0089
RAR	55.08682	18.89757	2.915021	0.0043
CR4	−0.464364	0.429547	−1.081054	0.2821

indicate that the coefficient of the level of fintech (FT) development is positive and significant at the 1% level. This finding suggests that the development of fintech has a substantial positive impact on the profitability of commercial banks. As the level of fintech continues to rise, the profitability of commercial banks also increases. This result substantiates the pivotal role of fintech in fostering the growth and development of commercial banks.

4.2.2. Sargan Over-Identification Constraint Test

To verify the validity of the instrumental variables used, a Sargan identification test is conducted. The null hypothesis of this test posits that the instrumental variable is not correlated with the error term. The test results are shown in

Table 7. Sargan over-identification constraint test.

Effects Specification
Cross-section fixed (first differences)
Mean dependent var	−1.099402	S.D. dependent var	1.194883
S.E. of regression	2.238466	Sum squared resid	541.1590
J-statistic	4.322867	Instrument rank	13
Prob (J-statistic)	0.364070

.

The test results indicate that the J-statistic is 4.32 and the p-value is 0.36, suggesting that we cannot reject the null hypothesis. Therefore, it can be considered that the instrumental variables used are valid and uncorrelated with the error term, thereby supporting our model specification and estimation results.

4.2.3. Arellano-Bond Autocorrelation Test

To check for serial correlation in the random error term of the first-order difference equation, the Arellano-Bond autocorrelation test was performed.

Table 8. Arellano-Bond autocorrelation test.

Test Order	m-Statistic	rho	SE (rho)	Prob.
AR(1)	−0.666685	−34.895574	52.341899	0.505
AR(2)	−0.603912	−35.498355	58.780705	0.5459

presents the results of the Arellano-Bond autocorrelation test, which is used to detect serial correlation in the random error term of the first-order difference equation. The key statistics reported in the table include the m-statistic, rho, standard error of rho (SE(rho)), and the probability value (Prob.). The results from both AR(1) and AR(2) tests indicate that there is no significant serial correlation in the random error term at either the first or second order, as evidenced by the high p-values (0.505 and 0.5459, respectively). Therefore, we can conclude that the model does not suffer from serial correlation issues, supporting its validity. In our analysis, the Arellano-Bond autocorrelation test indicated a p-value of 0.5459 for AR(2), which exceeds common significance thresholds, such as 0.1 and 0.05. Consequently, the null hypothesis that there is no second-order serial correlation in the random error term of the first-order difference equation cannot be rejected, indicating that we are using the correct model set up and valid instrumental variables.

To sum up, by building a dynamic panel model and using GMM estimation, we conclude that the development of financial technology in the previous period has a positive impact on the profitability of commercial banks in the current period, further validating Hypothesis 1a. These findings provide evidence for a deeper understanding of the relationship between financial technology and the profitability of commercial banks.

5. Discussion

Using EViews software, this study analyzes the annual data of 13 listed commercial banks in China. It empirically examines the static panel fixed effects variable-intercept model and the dynamic panel GMM model, clearly demonstrating that the development of fintech has a significant positive impact on the profitability of listed commercial banks. This finding is consistent with the results of Bömer and Maxin [32] and Singh, Malik, and Jain [8], which also confirm that fintech development significantly improve banks’ return on assets.

With the rise and development of fintech, its impact on commercial banks has been twofold. On one hand, it has weakened the intermediary function of commercial banks, causing a certain impact on their traditional business models and leading to a decline in profitability [16,45]. On the other hand, the application of FT has brought about technological spillover effects, enabling commercial banks to improve their profitability by reducing operating costs, enhancing operational efficiency, and broadening customer channels [18]. Given these opposing factors, the overall profitability of the 13 listed commercial banks in China exhibits a more positive trend.

Theoretically, this study strengthens the understanding of the role of fintech in the modern financial system, corroborates the association between technological innovation and financial service efficiency, and motivates efforts to innovate and improve productivity. On the practical side, this study suggests that banking managers should break through traditional operational models, actively learn new business paradigms, and explore the potential applications of fintech, such as blockchain and artificial intelligence, to optimize customer service and management, thereby enhancing market competitiveness and profitability.

Additionally, this study provides valuable information for investors in the banking industry, indicating that investing in fintech innovation is an effective way to enhance bank performance. Banks adept at building a robust FT service system have greater growth potential. For policymakers, this study underscores the importance of formulating policies that support the development of FT, which can contribute to the healthy and sustainable development of the entire banking industry. In the long run, the positive utility of fintech will lead to the rise and development of FT enterprises and bring about significant changes in financial services.

Although this study provides valuable insights, there are some limitations. For example, the data in this study are based on large commercial banks, while small and medium-sized banks have difficulties in transitioning from traditional to FT services due to capital constraints, immature technology, and regulatory compliance. Therefore, commercial banks with different asset sizes may have different findings. Future research could be extended to unlisted and small banks, and could be compared with fintech banks to fully assess the impact of fintech on different types of banks. This study uses the FinTech Index as the core variable, which only shows the development in the overall picture of FT. In the future, the impact of different dimensions of the FT development status on banks can be examined in more detail to provide more information for bank managers, investors, and policymakers. In addition, future research could explore the effects of FT in different economic environments and application scenarios, as well as the specific impact of different FT tools (e.g., digital currencies, online payments, etc.) on bank performance.

6. Conclusions

This study chooses the annual panel data of 13 representative listed commercial banks in China from 2011 to 2021, empirically examining the impact of fintech innovation on the sustainable profitability of commercial banks by employing the static panel fixed effects variable-intercept model and dynamic panel GMM model. Previous studies mainly discuss the impact of fintech on commercial banks theoretically, while limited research conducted the analysis from a quantitative perspective. Against this context, this study contributes to the existing literature, clarifying the relationship between fintech and commercial bank profitability under the context of rapid expanding fintech.

The findings indicate that FT innovations play a significant role in enhancing the operational efficiency and profitability of these banks. Key areas such as digital payment systems, online banking platforms, and blockchain technologies have been identified as critical drivers of this positive effect. The empirical evidence suggests that banks adopting FT innovations are better positioned to reduce operational costs, improve customer service, and offer more diversified financial products. These advantages contribute to a more robust and sustainable profit model, which is crucial in an increasingly competitive financial landscape. Moreover, the Arellano-Bond autocorrelation test results validate the appropriateness of our model and instrumental variables, further strengthening the reliability of our conclusions. Specifically, the absence of second-order serial correlation confirms that our dynamic panel data model effectively captures the essential features of the data without introducing bias. In summary, the integration of FT innovations appears to be a vital strategy for listed commercial banks in China aiming to achieve sustainable profitability. Future research should continue to explore this relationship, considering the potential risks and regulatory challenges associated with rapid technological advancements in the financial sector.