A Digital Analysis Approach to the Performance Enhancement of Chinese Listed Commercial Banks by Interest Rate Marketization and Diversification

Interest rate marketization refers to the abandonment of the central bank to implement strict control under the interest rate level of the money market operating financing, the market demand and supply of funds to form the interest rate decision and transmission mechanism [27]. From the level of interest rate decisions, the central bank gradually liberalizes the restrictions on deposit interest rates. The bank-specific deposit interest rates have the right to decide independently. From the point of view of the bank itself, the deposit business is its main source of liabilities, and thus the high and low interest rates also directly affect the high and low operating profits of the bank. Therefore, from the perspective of maximizing its own interests, the bank should give full play to its own advantages and its market position, as well as its own assets and liabilities term structure, risk structure, etc., analyze and measure its capital needs, real-time adjustment of the level of deposit interest rates. Observing the world interest rate marketization development law, there are four worthy of attention, one is the main body of financial transactions can be completely free to determine the market interest rate, the resulting interest rate must be helpful to the stability of the capital market. The second reason is that the market is the dominant factor in deciding interest rates, the transmission mechanism, risk, and term structure. Third, the interbank lending rate can be used as a basic indicator for determining interest rates. Fourth, the central bank does not directly control the level of interest rates on financial assets.

Diversification involves reducing operating costs by using existing products and services while expanding into businesses that are related to traditional businesses. Diversification also includes the enterprise entering a market unrelated to its original business. For a special enterprise such as commercial banks, the concept of diversification refers to the expansion of new businesses to meet the different needs of customers in order to make up for the shortcomings of the traditional deposit and loan business [28]. Through innovation and expansion of business scope to expand new revenue channels, and then improve their competitiveness in the industry internal and external competitive pressure. The main businesses include settlement and clearing, agency business, and other diversified businesses that increase non-interest income.

Commercial banks’ innovative business diversification mode mainly relies on the two modes of increasing income channels and diversifying services. In order to meet the increasing personalized needs of customers, commercial banks should develop new products, develop new businesses, and add new elements to traditional businesses to enrich the business structure and achieve performance improvement. The research perspective of this paper focuses on the increase in non-interest income that banks can achieve through diversification, which can provide banks with a variety of income streams. Overall, the performance of diversified operations is usually manifested in the increase of non-interest income.

The performance of an enterprise refers to the enterprise’s favorable return to the enterprise through a time period of production, distribution, service and other operations, commercial banks are the main business of operating deposits and loans, in order to realize the return and make profits for the purpose of special enterprises. At present, related scholars have four main methods to measure the performance of commercial banks, which are the balanced scorecard method, single financial analysis method, factor analysis method, and efficiency frontier analysis method.

For commercial banks need to use scientific evaluation methods and analytical techniques, through certain standards, procedures and indicator systems to make qualitative analysis and quantitative analysis of the operating behavior and operating results of commercial banks during the survival period, and to make objective, fair and reasonable performance evaluation, so as to provide a strong reference for commercial banks to improve market competitiveness [29]. The superiority of the data envelopment analysis method in evaluating the performance of commercial banks lies in. First, there is no need to determine the weight of each indicator, and the efficiency of the effective unit can be determined through the actual data, which can avoid the interference of subjective factors. Secondly, both monetary and physical dimensions can be measured as input-output indicators when evaluating performance with the DEA method, so some important non-financial indicators can be added to measure the performance of commercial banks. Thirdly, the use of data envelopment analysis can analyze the efficiency of the use of each input, but also to derive the relative efficiency value of the sample unit and the best sample of inputs, the resulting efficiency results can be divided into technical efficiency scale efficiency and pure technical efficiency, and this is analyzed in detail.

At this stage, China’s interest rate marketization has made a decisive leap forward, and scholars have carried out in-depth research on the impact brought by commercial banks, in which performance is the root of commercial banks’ vitality, and a systematic analysis of its impact is of great value. Scholars from multiple perspectives on its research, generally agreed that the marketization of interest rates in the short term to a certain extent will reduce the performance of commercial banks. The marketization of interest rates makes commercial banks more competitive and increases their business risks, which can significantly reduce their business performance.

By analyzing the mechanism of interest rate marketization on the performance of commercial banks and the existing research results, this paper argues that interest rate marketization is the People’s Bank of China to abolish the control of interest rates, the financial institutions according to the market demand and supply of funds to independently determine the level of interest rates. With the advancement of reform, the intensification of market competition leads to the continuous narrowing of deposit and loan spreads, thus affecting the performance of commercial banks. Through the above analysis, this paper proposes the following hypotheses:

H1: Interest rate marketization causes price competition and has a significant negative impact on commercial bank performance.

Although the development of marketization of interest rates will have a negative impact on the performance of commercial banks, in the short term, the increase in competitive pressures and risks in the banking sector will be more explicitly surplus. This is mainly because the interest rate risk of commercial banks will gradually increase with the general marketization of interest rates, because the marketization of interest rates will exacerbate interest rate volatility, and interest rate volatility and the increase in the number of fluctuations in the interest rate will make commercial banks to interest rate maturity as the basis of the success of the construction of the relevant financial products to a certain degree of difficulty.

In addition, the continuous reform of interest rate marketization will also force commercial banks to reform and innovate the financial model of intermediate business, especially in the renewal of the banking industry’s derivative business, which, to a certain extent, promotes the optimization of the diversified business model of commercial banks. In the process of the gradual promotion of interest rate marketization, commercial banks not only have the right to price their own products, but also for the emergence of new financial products to lay the road, and further promote the development of diversified business of commercial banks. Based on this, this paper proposes the following hypothesis:

H2: Interest rate marketization pushes commercial banks to carry out financial innovation and realize diversified operation.

The in-depth reform of interest rate marketization has created a free financial environment, facilitated the adjustment of the asset and business structure of commercial banks, and accelerated the process of diversification of the banking industry to a certain extent. Interest rate marketization has broken the pattern of commercial banks relying on deposit and loan business to obtain income. Based on this, commercial banks have to innovate financial products to increase their sources of income through investment wealth management, interbank market and other businesses, and these innovative financial products are regarded as alternative derivatives to make up for the loss of deposit and loan business. Interest rate marketization has narrowed the net interest margin, leading to an increased dependence of commercial banks on financial derivatives. Although financial derivatives have improved performance, they can also pose a number of problems, such as an increase in risk due to rigid payment. On the other hand, it has increased the competition in the business of investment and wealth management and the interbank market. With the continuous enrichment of investment and wealth management, interbank market and other business types, the marketization of interest rates will inevitably intensify the competition in the financial market, affecting the performance of commercial banks. Through the above analysis, this paper proposes the following hypotheses:

H3: Diversification plays a moderating role in the relationship between interest rate marketization and commercial bank performance, and diversification can weaken the negative impact of interest rate marketization on commercial bank performance.

In this paper, the panel data of 36 commercial banks (18 large and medium-sized banks and 8 city commercial banks) listed in Shanghai and Shenzhen A-shares for the years 2012~2024 are selected as the research samples. Considering the data availability and completeness, the panel data of the above 36 commercial banks in the year 2012~2024 are selected as the research object. The empirical data of this paper comes from the Cathay Pacific database, WIND database and the annual reports of major banks, and the interest rate marketization aspect and the business data aspect are constructed based on the keyword word frequency crawled from Baidu News of Baidu search engine.

Regression analysis is a statistical method that deals with the correlation of variables and can represent the mathematical relationship between independent and dependent variables [30]. By using the regression analysis method, we are able to extract the core information from the collected statistical data and use its main characteristics as independent variables, and then predict the results of dependent variables based on the range of these variables, thus obtaining the mathematical relationship between the two. Linear regression models are widely used due to their concise form and ease of construction. At the same time, the model also contains the basic concepts and ways of thinking of machine learning, in which the regression coefficients indicate the value of reflecting each attribute, so it has a good degree of comprehensibility.

In this paper, the multiple linear regression model is used as the analysis model, and its general form is: 1 y=β0+β1x1+β2x2+…+βpxp+ε

where β₀,β₁,…β_p is the regression coefficient, x₁,x₂,…,x_p is a p general variable to obtain precise measurements through data collection, often referred to as the independent variable, while y is the predicted value, referred to as the dependent variable, and ε is the random error term.

Faced with a specific problem, if n set of valid data x₁₁,x_i2,…,x_ip , y_i(i = 1,2,…,n) is obtained and the model is expressed in terms of the p observations of y and the x₁,x₂,…,x_p values associated with them, the model can be expressed as follows: 2 { y1=β0+β1x11+…+βpx1p+ε1y2=β0+β1x21+…+βpx2p+ε2…yn=β0+β1xn1+…+βpxnp+εn

The following properties are usually assumed: 3 E(εj)=0 4 Var(εj)=σ2(Constant) 5 Cov(εj,εk)=0,j≠k

The above equation (2) can be written as: 6 Y=Xβ+ε

Among them: 7 X=[ 1x11x12…x1p1x21x22…x2p…….1xn1xn2…xnp ] 8 Y=(y1,y2,…,yn)T,β=(β0,β1,…,βp)T,ε=(ε1,ε2,…,εn)T

Matrix X is a n×(p+1) nd order matrix. X is the design matrix.

8 Stepwise-multiple regression is an analytical method introduced to address the problem of multiple variables and the possibility of multicollinearity between them. First, each independent variable is introduced and an F-test is performed on it, then a t-test is performed for each independent variable that has been included, and if the previous independent variable is no longer significant due to the addition of the new independent variable, it should be removed to ensure that only significant variables are present in the regression model. This cycle is repeated until no more variables can be added or removed, so that the final selection of the most optimized explanatory variables can be determined.

For the performance measurement of listed commercial banks, it is mainly quantified by Tobin’s Q, economic value added or financial related indicators. In order to obtain research data more conveniently and calculate the simplicity, this paper mainly adopts the return on total assets (ROA) to characterize the performance of listed commercial banks, and takes it as the explanatory variable.

For the characterization of interest rate marketization, this paper combines the description of net interest margin changes in existing related studies, and selects net interest margin (NET) of listed commercial banks as an alternative to interest rate marketization, and takes it as an explanatory variable. The smaller the net interest spread of listed commercial banks, the higher the degree of interest rate marketization.

For the diversification measurement indexes of listed commercial banks, they mainly include entropy index, Herfindahl index and specialization ratio method. Among them, the entropy index fully takes into account the contribution of different business share of the enterprise, which can be expressed as: 9 EI=∑ ri∗(ln1ri)

where r_i denotes the share of the firm’s i nd business. The use of entropy index to measure the degree of diversification of enterprises can solve the problem of multicollinearity in the regression process. Therefore, it is more appropriate to choose the entropy index (EI) to measure the degree of diversification of listed commercial banks and use it as a moderating variable.

In addition, this paper also chooses economic growth rate (GDPR), asset size (LNA), cost-to-income ratio (CIR), capital adequacy ratio (CA), deposit-to-lending ratio (LDR), and non-performing loan ratio (NPL) as the control variables to ensure the validity of the factors that interest rate marketization and diversification affect the performance of listed commercial banks.

Since the subject of this paper is the relationship between interest rate marketization and commercial bank performance, commercial bank performance (ROA) is used as the explanatory variable, and net interest margin (NET) of listed commercial banks is used to measure interest rate marketization. The model designed to test hypothesis H1 is as follows: 10 ROAi,t=α0+α1NETi,t+α2∑ Controli,t+εi,t

where α₀ is the constant term, α₁ ~ α₂ is the regression coefficient of each variable, Control_i,t is each control variable, and ε_i,t is the random error term.

H2 mainly includes interest rate marketization and diversification, for which this paper establishes the following model: 11 EIi,t=β0+β1NETi,t+β2∑ Controli,t+εi,t

Where, β₀ is the constant term, β₁ ~ β₂ is the regression coefficient of each variable, Control_i,t is each control variable, and , ε_i,t is the random error term.

In order to verify the moderating effect of diversification between interest rate marketization and the performance of listed commercial banks, i.e. to verify H3, this paper constructs the moderating effect model as follows: 12 ROAi,t=γ0+γ1NETi,t+γ2EIi,t+γ3NETi,t∗EIi,t+γ4∑ Controli,t+εi,t

where γ₀ is the constant term, γ₁ ~ γ₄ is the regression coefficient of each variable, NET_i,t*EI_i,t is the interaction term between interest rate marketization and diversification of listed commercial banks, and the rest of the variables are defined as before.

Based on the financial panel data of 36 listed commercial banks from 2012 to 2024, they were entered into Stata software for standardization and descriptive statistical analysis was carried out for each variable. Table 1 shows the results of descriptive statistical analysis of each variable. Based on the results of descriptive statistics, it can be seen that the mean value of the performance of listed commercial banks (ROA) is 0.165, and the standard deviation is 0.046, indicating that the distribution of the performance level of each listed commercial bank in the sample is relatively moderate in terms of volatility. The extreme variance (the difference between the maximum and minimum values) of commercial bank performance is 0.229, which indicates that there is a high degree of variability in the performance level among the listed commercial banks in the sample, and therefore the conclusions drawn from this study will be somewhat generalizable. The mean values of interest rate marketization (NET) and diversification (EI) are 0.038 and 0.026, respectively, and the standard deviations are 0.043 and 0.017, respectively. These data indicate that the level of interest rate marketization and diversification of the sample listed commercial banks fluctuates widely within the observed interval, showing that there are significant differences in the level of interest rate marketization and diversification among different listed commercial banks.

Before carrying out regression analysis, it is necessary to test the correlation and multicollinearity of each variable in the model to ensure that the model will not appear pseudo-regression, and better ensure the stability of the regression results. In order to avoid the phenomenon of false regression caused by the existence of multicollinearity in the research model, this paper needs to further calculate the variance inflation factor (VIF) of each variable, i.e., through the regression method of the analysis of covariance to further explore whether there is a problem of multicollinearity in the model. Table 2 shows the results of variable correlation test and multicollinearity test, in which *,**,*** indicate that they are significant at the 10%, 5%, and 1% levels, respectively.

As can be seen from the table, the explanatory variable ROA shows a significant negative correlation with the core explanatory variable NET in the model, and the level of diversification significantly reduces the impact of interest rate marketization on listed commercial banks. In addition, from the correlation analysis of the performance level of listed commercial banks with each other variable, it can be seen that the performance of listed commercial banks has a significant correlation between the interest rate marketization and all other variables in the development process, i.e., the indicators of several other variables are affected by the interest rate marketization and interact with the performance of listed commercial banks. In addition, the VIF value of each research-related variable is less than 5, which proves that the model designed in this paper does not have serious multicollinearity problems, does not cause spurious regression, and can be used for benchmark regression analysis.

This paper takes the financial panel data of 36 listed commercial banks from 2012 to 2024 as the research sample, and in order to further avoid problems such as spurious regression, the smoothness test of the performance panel data of listed commercial banks. This paper mainly uses the current academic mainstream LLC and IPS panel unit root test method for analysis. Table 3 shows the results of the smoothness test of the variables.

The return on total assets (ROA), interest rate marketization (NET), diversification (EI), economic growth rate (GDPR), cost-to-income ratio (CIR), and capital adequacy ratio (CA) in the LLC test can all reject the original hypothesis at the 1% significant level, while the asset size (LNA), deposit-to-loan ratio (LDR) and non-performing loan ratio (NPL) reject the original Hypothesis. In the IPS test, except for interest rate marketization (NET) and diversification (EI), which reject the original hypothesis at a 5% significant level, all other variables can reject the original hypothesis at a 1% significant level. The above unit root test results indicate that the panel data in the sample are smooth and can be analyzed by subsequent regression empirical analysis. In addition, the output statistics of model 1~3 in the F-test are 45.271, 41.639, and 40.852, respectively, and their P-values are less than 0.01, which can reject the original hypothesis at 1% significant level, and the panel data should be constructed as a fixed effect model. While in Hausman test, the statistics obtained from model 1 and model 2 are 24.026, 22.148, whose P-values are less than 0.01, which can reject the original hypothesis at 1% significant level, and the panel data should be constructed as a fixed effect model. And the statistic of model 3 is 18.974, the P-value is less than 0.05, which can reject the original hypothesis at 5% significant level, and the panel data should be constructed with random effect model.

Based on the Hausman test, it can be seen that this paper uses the fixed effects model for the full sample group (1), the large and medium-sized bank group (2), and the urban commercial banks (3) when studying the impact of interest rate marketization on the performance of listed commercial banks. The benchmark regression of interest rate marketization and commercial bank performance is shown in Table 4. In the table *,**,*** denotes significant at the 10%, 5%, and 1% levels, and t-values are in parentheses.

In the regression results of the full sample, large and medium-sized banks, and urban commercial banks, the impact of net interest margin on ROA is negatively significant at the 1% level, i.e., there is a significant negative impact of interest rate marketization on the performance of listed commercial banks, and H1 is verified. With the deep development of interest rate marketization, the net interest margin shrinks, competition in the banking industry becomes more intense, and the performance of commercial banks decreases under the existing business model. Meanwhile, the regression coefficient of urban commercial banks is larger (-0.236), which indicates that urban commercial banks are more affected and their ability to fight against the impact of interest rate marketization is weaker compared to large and medium-sized banks. The reason for this difference may lie in the fact that large and medium-sized banks are relatively less affected by interest rate marketization because of their greater strength, relatively sound internal governance and broader business. Urban commercial banks, with their narrower scope of operation and greater reliance on the spreads generated by their traditional credit business, have been more affected by the marketization of interest rates.

Among the control variables, asset size (LNA) is significantly negative at the 1% level in the regressions for the full sample and the sample of large and medium-sized banks, and at the 5% level in the regression for the sample of urban commercial banks. This does not seem to be common sense, and the reason for this may be due to the fact that the current asset size of commercial banks has already reached a certain level, and any further expansion of the asset size on this basis will exceed the existing level of operation and management, and generate more non-performing assets, thus affecting the bank’s performance. The deposit and loan ratio (LDR) shows a significant negative correlation in the regression of the full sample and the sample of large and medium-sized banks. Generally speaking, the higher the deposit-to-loan ratio, the higher the profitability of commercial banks, but under the marketization of interest rates, relying solely on the expansion of traditional credit business is no longer conducive to the improvement of the performance of large and medium-sized commercial banks, and it is urgent to improve the level of diversification and expand non-credit business. The non-performing loan ratio (NPL) shows a significant negative correlation in the regression of all three samples, indicating that the risk level of commercial banks’ lending business has a greater impact on their performance, and that they need to further improve their risk-resistant ability.

The hypothesis of this paper on the analysis of diversification and the performance of listed commercial banks is that the reform of interest rate marketization promotes the implementation of the diversification strategy of commercial banks, and the degree of diversification of commercial banks will increase accordingly with the increase in the degree of interest rate marketization. In this regard, the fixed effect model is used for regression analysis, and the regression results for the full sample group (1), the large and medium-sized bank group (2), and the urban commercial banks (3) are shown in Table 5.

In all three groups of regression results, the effect of net profit margin on the level of diversification shows a negative and significant effect at the 1% level. This is a good indication that commercial banks will take active measures to enhance their diversification level when the net interest margin decreases. In the full sample group, when the degree of interest rate marketization decreases, the level of diversification of listed commercial banks will increase, and its regression coefficient is 0.115, which shows a negative significant effect at the 1% level. That is, for every 1 percentage point increase in the degree of interest rate marketization, the diversification level of listed commercial banks will increase by 0.115 percentage points, according to which H2 can be verified.The regression coefficient value of capital adequacy indicator (CA) under all three groups has a positive and significant effect on the level of diversification at the 1% level. It indicates that when there is relatively sufficient capital, listed commercial banks can provide strong financial support for the development of various businesses, which in turn will enable the banks to achieve a higher level of performance. According to the model fit, the adjusted R² value is 0.615, which indicates that the model fits 61.5% well. Therefore, when the level of interest rate marketization gradually increases, listed commercial banks, in order to stand out from the fierce competition, will promote performance growth through business innovation, increasing product variety, expanding the scope of business and other diversified operation methods. In addition, the regression coefficients of large and medium-sized listed commercial banks are larger than those of urban commercial banks, which is mainly due to the fact that large and medium-sized banks have better basic conditions and are more capable of implementing diversified operations in the face of the impact of the interest rate marketization, thus laying a foundation for promoting the performance of commercial banks.

Based on the previous analysis, it is known that commercial bank performance is negatively affected by interest rate marketization, and Model 1 tests the impact of interest rate marketization on commercial bank performance. In order to further test whether the relationship between interest rate marketization and the level of diversification affects commercial bank performance, this paper adds an interaction term to estimate model 3 for the full sample group, the large and medium-sized bank group (2), and the urban commercial bank group (3) samples respectively. In order to test the moderating effect of diversification on the relationship between interest rate marketization and commercial bank performance. Table 6 displays the regression results after adding the interaction term.

In the full sample, after adding the interaction term between net interest margin and diversification, there is no significant change in the effect and significance of the variables on the performance of commercial banks compared to the test results in Table 4. After the introduction of the interaction term, the net interest margin and the interaction term are both significant at the 1% level, and the coefficient of the interaction term has the opposite sign of the coefficient of diversification (EI), which indicates that the interaction of net interest margin (NET) and diversification (EI) reduces the negative impact of the interest rate marketization itself on the performance of the commercial banks, and that there is a moderating effect of diversification on the interest rate marketization and the performance of the commercial banks, and that H3 holds. For joint-stock commercial banks, after adding the interaction term between the diversification variable and net profit margin, the ROA indicators are all significantly affected by the interaction term, and the coefficients of the interaction term are significantly negative at the 1% level, contrary to the sign of the coefficients of diversification in Table 5. It indicates that the interaction between net profit margin and diversification slows down the negative effects of interest rate marketization on the performance of large and medium-sized commercial banks. On the other hand, the effects of the variables on the performance of commercial banks are by and large not significantly different, and the interaction term only has a significant negative effect on the non-performing loan ratio (NPL), and the sign of the coefficient is opposite to that of the interest rate marketization, indicating that the interaction of net profit margin and diversification degree index somewhat attenuates the negative effect of the interest rate marketization on the NPL ratio of the large and medium-sized commercial banks. For the sample of urban commercial banks, after adding the interaction term between diversification and net profit margin, there is no significant change in the effect of the main variables on the ROA index compared with the test results of model 1 in Table 4. This suggests that urban commercial banks are influenced more by traditional performance influences than by diversification, which may be a result of their later development than medium and large commercial banks.

In order to verify the robustness of the previous regression results on interest rate marketization on the performance of commercial banks, this paper conducts the robustness test in four main ways, i.e., extreme value treatment, replacing the sample period, replacing the explanatory variables, and mitigating the endogeneity problem between the current variables. Table 7 shows the results of the robustness test, where columns (1) to (4) represent the four methods of robustness test, respectively.

First, extreme value treatment. In order to eliminate the influence of extreme values on the regression results, all variables are regressed again after 1% bilateral shrinkage, and the regression results are shown in Column (1). The results show that the regression coefficients of the core explanatory variable (NET) are still significantly negative, indicating that the negative impact of interest rate marketization on commercial bank performance is more robust. Second, the sample period is replaced. In order to eliminate the effect of jumps in interest rate marketization on the benchmark regression results, the regression is conducted after excluding the samples of 2016 and the previous years with reference to the practice of existing studies. The regression results reported in column (2) show that the regression coefficient of the core explanatory variable (NET) is still significantly negative, implying that the results of the benchmark regression are not entirely caused by the jump in interest rate marketization. Third, the explanatory variables are replaced. Replacing the explanatory variable (ROA) with an important measure of banks’ total asset profitability (ROE) for the robustness test, the regression results in Column (3) show that the regression coefficient of the core explanatory variable (NET) is still significantly negative, which expands from the level of shareholders’ equity to the level of overall assets, providing a broader and more solid evidentiary support for the negative role of interest rate marketization in enhancing banks’ performance. It further validates the robustness and reliability of the model. Fourth, to address the endogeneity problem among current variables. Interest rate marketization generates certain competition costs, this factor may make the impact of interest rate marketization on commercial bank performance not fully reflected in the current period, considering the time lag effect in causality, lagging both explanatory variables and control variables by one period, column (4) demonstrates the regression results. From the regression results, it can be seen that the regression coefficient of the core explanatory variable (NET) is still significantly negative, proving once again the negative and negative impact of interest rate marketization on commercial bank performance, and the conclusions drawn in the previous section have good robustness.