THE THEORETICAL BASIS
Commercial banks
Commercial banks are financial institutions that primarily offer financial services to individuals, businesses, and organizations They engage in various activities such as accepting deposits, providing loans, offering payment services, asset management, and other financial services.
Profit
Profit is the positive amount of money or financial value that an individual or organization earns from business activities after deducting costs, interest expenses, taxes, and other payable amounts It represents the difference between revenue and expenses or production costs associated with the creation and sale of goods or services Profit reflects business efficiency and the ability to generate financial value from business operations It can be used to measure the success and sustainability of an organization or investment decision.
Profit on equity or Return on Equity (ROE)
ROE is a ratio calculated by dividing net profit after tax by the total value of equity based on the balance sheet and financial statements at the end of a specific period (such as the first 6 months or the last 6 months of the year)
In theory, a higher ROE indicates more efficient capital utilization Stocks with higher ROE are often favored by investors
CAP: Measuring the capital adequacy ratio of a bank is essential Shareholders’ capital plays a crucial role in ensuring the continuity of activities related to the bank Therefore, it is considered a tool to help the bank manage all risks arising from its operations, ensuring payment capability in loss situations and minimizing losses to the maximum extent possible Additionally, it enables the expansion of credit operations, which are the primary profit-generating activities for the bank This variable is expected to move in the same direction as profitability
Hypothesis 1 (H1): The capital adequacy ratio has a positive impact on the profitability of the bank
LOAN: Measuring the loan-to-assets ratio of a bank is important Within the overall asset structure of a bank, the loan-to-assets ratio always holds a significant proportion A higher ratio indicates that the bank is actively expanding its lending activities to generate higher profits However, as the loan portfolio grows rapidly, the corresponding increase in non-performing loans can lead to a decrease in profitability Following the risk and return trade-off principle, the bank’s lending growth is equivalent to an increase in credit risk, which can negatively impact the bank’s profitability
Hypothesis 2 (H2): The loan-to-assets ratio has an inverse impact on the profitability
ME: Measuring the effectiveness of bank management poses a significant challenge, particularly in cost management This indicator assesses the level of cost incurred by the bank to invest each unit of its assets When the cost ratio increases, it indicates that the bank is utilizing a higher amount of expenses for investments This suggests that the bank may not be effectively implementing cost-saving measures, consequently negatively impacting profitability
Hypothesis 3 (H3): Management efficiency has an inverse impact on the profitability
TANG: Measuring the fixed asset ratio of a bank is crucial as it helps evaluate the utilization and management of the company’s fixed assets Analyzing this ratio can determine the effectiveness of the company’s fixed asset management Furthermore, using this ratio in decision-making regarding investment opportunities, reliability, and overall financial health is highly beneficial and insightful Fixed assets are considered significant investments for many companies, highlighting the importance of this ratio They are also essential components in the company’s operations and production
Efficiently managing and effectively utilizing these assets can significantly impact the company’s profitability and financial performance
Hypothesis 4 (H4): The fixed asset ratio has a positive impact on the profitability of the bank
GDP: Measuring the GDP growth rate of an economy is extremely important GDP growth represents the level of expansion of the domestic economy in a year A stable economy helps businesses operate efficiently in production and business, leading to increased borrowing demand and reduced likelihood of non-performing loans This indirectly brings higher profits to banks
Hypothesis 5 (H5): The GDP growth rate has a positive impact on the profitability of the bank
Figure I-1: Research model on factors impacting BIDV’s ROE
Table 1: Description of Variables in the Model
Variable Symbol Formula Expection symbol
Debt-to-equity ratio LOAN 𝑇𝑜𝑡𝑎𝑙 𝐷𝑒𝑏𝑡
GDP Growth GDP Average GDP Growth Rate per Capital +
Return in Equity (ROE) Evaluation Criterion
Return in Equity (ROE) evaluation criterion is the company's efficiency in using equity capital to create profits ROE shows the ratio between profit after tax and equity Criteria to evaluate ROE include:
Relative ROE: Comparing ROE with industry peers to evaluate a company's relative performance helps determine whether a company's ROE is better or worse than its competitors
Compare ROE with company goals: Determine the company's ROE target based on factors such as growth goals, investment levels, and risks Compare the actual ROE with the company's target to evaluate whether it has met its target
ROE Growth: Track ROE changes and growth over time This allows to evaluation of the company's financial performance in terms of increasing ROE from performance improvement measures or business expansion
Compare ROE to return on equity: Evaluate ROE by comparing it to the return on equity required to beat a similar investment If ROE exceeds the return on equity, this indicates that investment in the company achieved a return higher than the minimum return on investment
Financial leverage: Compare the ratio of debt to equity and see if the company using financial leverage brings higher profits or not
Operating performance and asset management: Evaluate the company's financial performance through indicators such as gross profit ratio, net profit ratio, total asset ratio, and company asset ratio
The above evaluation criteria can be obtained from financial reports, research, and reports of financial consulting companies, portfolios, and reports of investment funds.
RESEARCH METHODOLOGY AND DATA
Research Model and Methodology
Based on previous research, this article suggests a research model that includes two groups of variables: macroeconomic factors and internal aspects of a bank The internal bank variables consist of fixed assets ratio (TANG), capital adequacy ratio (CAP), debt-to-equity ratio (LOAN), and management efficiency (ME) The macroeconomic variable used in the model is the growth of the gross domestic product (GDP)
The study employed quantitative research using descriptive statistical techniques and regression analysis Regression analysis was conducted using a panel dataset collected for the study The estimation method used was Ordinary Least Squares (OLS) Once the model was established, various tests were conducted, including p-value tests to examine multicollinearity, the Breusch-Pagan test (Breusch and Pagan, 1980) to test for heteroscedasticity, the Breusch-Godfrey test to test for autocorrelation In case the research model encountered issues such as autocorrelation, multicollinearity, and/or heteroscedasticity, the Feasible Generalized Least Squares (FGLS) method would be employed after addressing these issues
The empirical analysis in this study employed two sets of data: macroeconomic factors and internal factors of the Bank for Investment and Development of Vietnam (BIDV) Internal factors were sourced from the audited consolidated financial reports of BIDV, while macroeconomic data, specifically the economic growth rate, was obtained from the International Monetary Fund (IMF) All data utilized in the study was collected between 2006 and 2019, ensuring a comprehensive analysis of the factors influencing BIDV's profitability during this period.
RESEARCH RESULTS
The current situation
In 2020, BIDV's pre-tax profit reached 9.026 trillion VND, exceeding the State Bank of Vietnam's financial plan (106%), but still decreased by 15.9% compared to
2019 This was because BIDV proactively reduced its income by over 6.4 trillion VND to restructure debt and waive interest fees for customers affected by Covid-19, as directed by the State Bank of Vietnam As a result, BIDV's ROE in this year was 9.18%
In 2021, BIDV's ROE reached approximately 13%, while other banks achieved above 20% A low ROE means that the retained earnings for capital replenishment are also reduced, putting pressure on the capital adequacy ratio This is particularly concerning when credit growth is high Therefore, BIDV is facing the pressure to strengthen capital mobilization from external sources to ensure capital adequacy ratio (CAR)
Meanwhile, in 2022, BIDV ranked in the top 10 banks with the highest ROE, with 19.34% This was mainly due to BIDV recording a 70% growth in after-tax profit compared to 2021, reaching over 18 trillion VND This demonstrates that the bank is balancing shareholder capital with borrowed capital in a harmonious manner.
Descriptive Statics
Based on the data collected from BIDV's consolidated financial reports for each year from 2010 to 2022, the group presented a descriptive statistical table of the variables used in the study The values in the table include: Mean, Median, Maximum, Minimum, and Standard Deviation of the 5 variables
Table 2: Descriptive statics between observed variables
Variable CAP ME LOAN TANG GDP
Source: The group compiled the data using Eviews 10
Capital adequacy ratio: The average rate is 5.15% The minimum value was
4.06% in 2017, and the maximum value was 6.61% in 2010 BIDV achieved a value of 6.61% because the bank successfully met its business objectives and targets Specifically, the total assets as of December 31, 2010, reached approximately 366,268 billion VND; mobilized capital also increased to 251,924 billion VND; and pre-tax profit simultaneously reached 4,626 billion VND
Figure III-1: Capital adequacy ratio from 2010 to 2022
Management efficiency: The average rate is -1.26% The minimum value was -
1.64% in 2012, and the maximum value was -0.94% in 2011
Figure III-2: Management efficiency from 2010 to 2022
- Debt-to-equity ratio: According to the descriptive statistics, LOAN has an average value of 71.50%, a median of 70.91%, a minimum value of 67.51% in 2014, and a maximum value of 78.81% in 2020 During the Covid-19 pandemic, BIDV not only actively issued various credit packages to support customers in this difficult period but also reduced lending interest rates for individual customers since August 2020, with the "Connect - Reach Further" loan package, which has a total scale of up to 30,000 billion VND
Fixed Assets Ratio: The average value is 0.84%, the minimum value is 0.5% in
2022, and the maximum value is 1.03% in 2014
Figure III-4: Fixed Assets Ratio from 2010 to 2022
GDP Growth: The average value is 9.3%, with a minimum value of 1.68% in
2015 and a maximum value of 19.62% in 2011 During the period from 2011 to 2015, global investment flows continuously declined Specifically, according to the State Bank of Vietnam, these flows decreased from 11.8 trillion USD (20% of global GDP) to 2 trillion USD within a span of three years (2007-2009) Subsequently, they began to grow again and reached 6.1 trillion USD in 2010, but then declined to 5.3 trillion USD and 4.6 trillion USD in 2011 and 2012, respectively, only amounting to one-third compared to 2007, which is equivalent to 6% of global GDP
Figure III-5: GDP Growth from 2010 to 2022
Testing some limitations of the estimation method
In econometrics, multicollinearity is an important issue in regression models It occurs when the independent variables in the model are highly correlated with each other The presence of multicollinearity can lead to biased indicators in the model, resulting in quantitative analysis results that are not highly meaningful Multicollinearity is a violation of the initial assumption of linear regression models, which states that the independent variables must be linearly independent from each other There are two main causes of multicollinearity:
Firstly, nature of variables: Some variables have similar characteristics and show little variation For example, income and salary can be considered similar as they both measure the financial aspect of individuals Similarly, preferences and interests can be closely related Age and experience can also lead to multicollinearity as they often increase simultaneously
Secondly, survey environment characteristics: In some cases, differences in the survey environment can cause two variables to become multicollinear For example, two conducting research in environment 2, it is necessary to adjust the survey design to avoid multicollinearity
To test for multicollinearity in a model, the following methods can be used:
The correlation coefficient quantifies the linear association between two variables To detect multicollinearity among independent variables, the correlation coefficient between each pair is calculated A high correlation coefficient (typically exceeding 0.8) indicates a strong linear relationship, suggesting multicollinearity and the presence of redundant information in the independent variable set.
Variance Inflation Factors (VIF): VIF is a statistic calculated based on the regression model It measures the degree of multicollinearity for each independent variable in the model If the VIF value of a variable exceeds a threshold (usually 10), we can conclude that the variable has high multicollinearity
An auxiliary regression model is employed to gauge the relationship between independent variables When this model reveals a correlation among these variables, it indicates the presence of multicollinearity.
In this research study, the testing of multicollinearity was conducted using two methods: Variance Inflation Factors (VIF) and auxiliary regression models
Testing for multicollinearity helps determine the extent of its impact in a regression model If multicollinearity is detected, it is necessary to consider methods to reduce multicollinearity, such as removing highly correlated variables, using principal component analysis, or employing non-linear regression methods like Ridge regression or Lasso regression This ensures the accuracy and significance of quantitative analysis results in econometrics
Firstly, the group conducted a test using Variance Inflation Factors (VIF) When the VIF values of two independent variables are greater than 10, it indicates that the regression model definitely exhibits multicollinearity If the VIF value is less than 2, it suggests that the regression model does not have multicollinearity If the VIF value falls between 2 and 10, it indicates that the regression model lacks sufficient data to draw a conclusion
Table 3: Multicollinearity coefficients of observed variables (VIF)
Source: The group compiled the data using Eviews 10.
The table above shows the Variance Inflation Factor (VIF) coefficients of the independent variables, and all of them are less than 10 However, the VIF coefficients of the ME and TANG variables are greater than 2, so we cannot immediately conclude
To determine if a research model exhibits multicollinearity, auxiliary regression models were employed The presence of a significant auxiliary regression model indicates multicollinearity, while the absence of suitable models suggests no significant correlation among independent variables and, therefore, no multicollinearity The auxiliary regression models were estimated using specific hypotheses.
- H: The auxiliary regression model is inappropriate
- K: The auxiliary regression model is appropriate
Figure III-6: Running auxiliary regression with the dependent variable as CAP
Source: The group compiled the data using Eviews 10
Figure III-7: Running auxiliary regression with the dependent variable as ME
Source: The group compiled the data using Eviews 10
Figure III-8: : Running auxiliary regression with the dependent variable as LOAN
Source: The group compiled the data using Eviews 10
Figure III-9: Running auxiliary regression with the dependent variable as TANG
Source: The group compiled the data using Eviews 10
Figure III-10: Running auxiliary regression with the dependent variable as GDP
Source: The group compiled the data using Eviews 10
With a significance level of 5%, the respective p-values are 0.59, 0.18, 0.25, 0.10, and 0.56, all of which are greater than the significance level of 0.05 Therefore, we accept the null hypothesis (H) and can conclude that the auxiliary regression models are not appropriate, meaning that the independent variables are linearly independent from each other Thus, we can infer that the research model does not exhibit multicollinearity
Heteroscedasticity, a significant concern in econometrics, arises when error variance in a regression model is not constant This deviation from the assumption of homoscedasticity, where error variance is uniform, leads to observations with varying error magnitudes Heteroscedasticity adversely affects the estimation accuracy of variance, making the estimated coefficients inefficient and biased As a result, hypothesis testing power diminishes, limiting the reliability of the regression model.
The main cause of this phenomenon is often the existence of observations in variables with significantly different values compared to the remaining observations, or observations of the same variable measured on different scales For example, when measuring income, it is not appropriate to use the same currency unit, such as thousands of dollars, for individuals with high incomes and millions of dollars for those with low incomes This difference in measurement can result in varying error variances
Additionally, heteroscedasticity can also occur when there are errors or inaccuracies in the data transformation process Independent variables may have a nonlinear relationship with the dependent variable, leading to heterogeneity in variance This can happen when the linear regression model is not suitable for the data, and alternative methods such as nonlinear regression models need to be employed
To test for heteroscedasticity, the group utilized the Breusch-Pagan-Godfrey test This test examines the following hypotheses:
- H: The model's error term has constant variance
- K: The model's error term has varying variance
Figure III-11: Results of the test for heteroscedasticity
Source: The group compiled the data using Eviews 10
With a significance level of 5%, we have a p-value (F-statistic) of 0.8006, which concluded that the research model does not exhibit heteroscedasticity in the error variance
Autocorrelation is a phenomenon in quantitative economics where the value of a variable in a time series depends on its past values This phenomenon commonly occurs in time series data and can significantly impact the results of a model Autocorrelation violates the initial assumption of the regression model, which assumes that the error term at different values of the independent variable is uncorrelated Autocorrelation renders the estimates inefficient, introduces bias in the estimated variances of the OLS estimates, undermines the reliability of t-tests and F-tests, and can lead to inefficiency in the variances and standard errors of predictions
There are several methods to detect autocorrelation in econometric models In this study, the group employed the Breusch-Godfrey test to examine first- and second- order autocorrelation
First, the Breusch-Godfrey test was used to test for first-order autocorrelation The hypotheses tested are as follows:
- H: The errors of the research model do not exhibit first-order autocorrelation
- K: The errors of the research model exhibit first-order autocorrelation
Figure III-12: Results of the test for first-order autocorrelation
Source: The group compiled the data using Eviews 10
Regression Analysis Results
After conducting regression analysis using the Pooled OLS model, the group obtained the following results presented in Table 5:
Table 5: Regression Analysis Results using Pooled OLS Model
No Variable Beta coefficient P-value
Source: The group compiled the data using Eviews 10
Meaning: 89.02% of the fluctuations of Return on Equity (ROE) is caused by the fixed assets ratio (TANG), capital adequacy ratio (CAP), debt-to-equity ratio (LOAN), and the growth of the gross domestic product (GDP) The remaining 10.98% is caused by other factors
4.1 Capital adequacy ratio (CAP) variable:
At a significance level of 0.05, the CAP variable has a p-value of 0.0858, which exceeds the threshold This indicates that the CAP variable is not statistically significant, despite its negative beta coefficient (-1.1296) and inverse relationship with return on equity (ROE).
The ME variable has a p-value of 0.0024 < alpha = 0.05, indicating that this variable has an impact on ROE Specifically, the beta coefficient of ME 𝛽̂ = -12.4476) 2 shows a negative impact If management efficiency increases by 1% while other factors remain constant, the average ROE will decrease by 12.45%
This factor has the strongest impact on ROE The reason could be that banks are focused on expanding their operations and developing new service offerings, which requires investments in employee salaries, construction expenses, costs associated with launching new products, and other types of expenses Banks are using higher costs for investments The bank has not been able to effectively save costs, leading to a decrease in profits
4.3 Debt-to-asset ratio (LOAN) variable:
With a p-value of 0.0002 < alpha = 0.05, the LOAN variable is statistically significant and has a negative correlation with ROE (𝛽̂ = -1.0169) Holding other 3 variables constant, a 1% increase in the debt-to-asset ratio will lead to an average decrease of 1.02% in ROE
When the lending ratio increases, it indicates that the bank is expanding its lending activities to generate higher profits However, generating income from lending activities also requires effective management and reasonable cost control in order to achieve efficiency An increase in bad credit quality and non-performing loans will result in a decrease in profits
4.4 Fixed assets ratio (TANG) variable:
With a significance level of alpha = 0.05, the TANG variable is significant in the model as its p-value is 0.0012 < 0.05 The beta coefficient of TANG 𝛽̂ = -17.7209 4 indicates that the fixed assets ratio has a reverse impact on the return on equity If the fixed assets ratio increases by 1% while other factors remain constant, the average ROE will decrease by 17.72%
The company invested in purchasing machinery and equipment and improving technological capabilities to meet the needs of the company and its customers This creates conditions for increased labor productivity and revenue growth The research results differed from the initial expectations as they had a contrary impact on ROE
4.5 Gross domestic product (GDP) growth rate variable:
The GDP variable has a p-value of 0.0098 < alpha = 0.05, indicating that it has an impact on ROE, specifically a negative impact 𝛽̂ = -0.3354 If the GDP growth rate 5 increases by 1% while other factors remain constant, the average ROE will decrease by 0.34%
A stable economy helps businesses operate efficiently, leading to a greater need for loans and a lower risk of bad debts, indirectly resulting in higher profits for banks.
CONCLUSION AND RECOMMENDATIONS
Research findings identify four factors influencing BIDV's Return on Equity (ROE): Management Efficiency (ME), Loan-to-Asset Ratio (LOAN), Fixed Asset-to-Total Asset Ratio (TANG), and Economic Growth Rate (GDP) Consequently, the study recommends that commercial banks adopt strategies to enhance these factors, thereby boosting their profitability.
Management Efficiency (ME): The research underscores that ME has a strong and inverse impact on BIDV's ROE BIDV should prioritize improving its management efficiency This can be achieved by controlling costs by minimizing unnecessary costs such as advertising costs, management costs and working with technology more effectively; Optimize profits, improving service quality by enhancing employee training Additionally, BIDV should develop new products and services that align with customer needs and market trends to create differentiation and gain a competitive edge over other banks
Loan-to-Asset Ratio (LOAN): The results indicate that the LOAN variable affects BIDV's ROE inversely This suggests that BIDV should reconsider its credit expansion strategy, ensuring the quality and efficiency of its lending activities To mitigate the risk of non-performing loans, BIDV should focus on adjusting its credit strategy towards quality rather than quantity It should concentrate on areas with high growth potential, low default risks, and reduce lending in high-risk sectors such as real estate and construction Additionally, BIDV should enhance monitoring and management of loan portfolios, promptly address non-performing loans, and implement credit risk prevention measures
The Fixed Asset-to-Total Asset Ratio (TANG) inversely impacts BIDV's Return on Equity (ROE) To mitigate this, BIDV should identify underperforming fixed assets and reevaluate their investments to minimize waste and maximize utilization Additionally, the company should explore funding options such as customer deposits and capital raised from the securities market to reduce capital costs and enhance profitability.
Economic Growth Rate (GDP): The research indicates that the GDP growth rate affects BIDV's ROE inversely To balance the impact of GDP and minimize negative effects on ROE, BIDV should adjust its investment portfolio to reduce risk and seek profit opportunities in various markets or industries Furthermore, the bank can diversify its business activities, such as financial services and insurance, or expand strategic partnerships to increase income from multiple sources BIDV should explore ways to strengthen capital management and increase shareholders' equity to meet growth requirements and ensure adequate capital capacity in the face of GDP fluctuations This will contribute to maintaining stable ROE and sustainable development
In conclusion, by implementing these recommendations, BIDV can work towards enhancing its profitability and ensuring long-term success in the competitive banking sector.