Factors affecting financial performance of commercial banks in vietnam

INTRODUCTION

REASONABLE FOR RESEARCH

The banking industry plays a vital role in a country's development, and since Vietnamese banks joined the World Trade Organization (WTO) in 2017, their weaknesses have become more evident Issues such as low financial capacity and heightened competition highlight the challenges faced by this relatively young banking system, which suffers from limited management experience and small scale operations Additionally, ongoing risks threaten the stability of the Vietnamese banking sector, as evidenced by the slower growth rate of Vietnam compared to other markets during the global economic crisis of 2008.

To ensure the stability of the banking system and maintain public trust, the restructuring of credit institutions in Vietnam was implemented swiftly between 2011 and 2015, guided by Decision No 734/QD-State Bank This initiative aimed to enhance profitability while ensuring safe operations amidst market fluctuations Commercial banks focused on key financial performance indicators, including capital structure, asset quality, profitability, liquidity, and operational efficiency.

As a result, it is critical to evaluate and identify the current issues affecting the financial performance of commercial banks in Vietnam Because it will aid

Assessing the financial performance of commercial banks in Vietnam is crucial for managers, policymakers, bankers, and investors, especially during the integration phase This evaluation serves as a foundation for developing a sound policy framework to regulate bank operations It enables effective restructuring of the banking system, providing a scientific basis for decisions related to mergers and consolidations.

This research investigates the "Factors Affecting the Financial Performance of Commercial Banks in Vietnam," aiming to identify key influences on banks' financial outcomes and propose strategies for enhancing their performance.

OBJECTIVES OF STUDY

The general objective of this study is to study the factors affecting the financial performance of commercial banks in Vietnam

 Build models based on previous studies

 Verify the impact of these factors on the financial performance of Vietnam Joint Stock Commercial Bank

 Check the direction of impact

To enhance the financial performance of Vietnam's joint stock commercial banks, it is crucial to implement targeted solutions and recommendations These measures should focus on optimizing operational efficiency, strengthening risk management practices, and leveraging technology to streamline processes Additionally, fostering a culture of innovation and continuous improvement will help mitigate negative impacts on financial performance, ensuring sustainable growth and stability within the banking sector.

RESEARCH QUESTION

This thesis is carried with the expectation to find answers for four main question listed here in below:

 What are the determining factors affecting the financial performance of commercial banks in Vietnam?

 What model and method to measure the financial performance of commercial banks in Vietnam?

 How is the impact of the financial performance of commercial banks in Vietnam?

 Based on the research results, what solutions to improve the financial performance of commercial banks in Vietnam?

SUBJECT AND SCOPE OF THE STUDY

The object of this research is financial performance of commercial banks, the factors affecting the financial performance of commercial banks in Vietnam

Data research was carried out on 31 Vietnam Joint Stock Commercial Banks The study used data collected from 2010-2020.

RESEARCH METHODOGY

To enhance the reliability of research findings and address the limitations of individual methods, this study employs a mixed-methods approach, integrating both qualitative and quantitative techniques The quantitative method is utilized to identify relationships and correlations among variables, while the qualitative method serves to validate the results obtained from data analysis.

Effective data collection methods involve developing research models, designing research samples, and gathering data In this study, the author utilized secondary data collection techniques by sourcing published data to support the research objectives.

4 websites of commercial banks such as annual reports, cash flow statements, etc currency, business results, in the period 2010-2020

This study utilizes a quantitative research method, supported by Stata software, to perform regression analysis on acquired panel data The author applies various techniques, including the least squares method (POOLED OLS), random effects method (REM), and fixed effects approach (FEM), to construct an appropriate model To determine the best model for panel data regression, tests such as Preusch and Pagan and the Hausman test are employed Additionally, the study addresses issues like variable error variation and autocorrelation using the feasible generalized least squares (FGLS) approach To resolve endogenous issues, the author implements the S-GMM strategy.

Qualitative method: used to compare results from empirical analysis with results from previous studies to explain research objectives and research questions.

CONTRIBUTIONS

The findings of this thesis serve as a valuable resource for administrators, policymakers, and scholars, aiming to enhance the operational efficiency of banks and advance research and governance within the banking sector.

STRUCTURE OF RESEARCH

This chapter outlines the research framework, detailing the rationale behind the chosen topic, the specific research problem, objectives, and questions It also defines the research object and scope, highlights the significance of the study, and presents the overall structure of the research.

LITERATURE

OVERVIEW OF FINANCIAL PERFORMANCE OF JOINT STOCK

Financial performance, often referred to as business efficiency, measures the outcomes of an enterprise's activities through the profit or loss generated It assesses how well financial goals are being met and involves calculating the monetary value of a company's policies and actions This metric evaluates a company's overall financial health over a specific timeframe and allows for comparisons between similar businesses across various industries Due to its direct nature, financial performance can be easily analyzed and understood.

Financial performance in the banking industry encompasses the outcomes of a bank's commercial and investment activities, along with its internal factors within the economic landscape.

Previous empirical studies such as San and Heng (2013) Ongore & Kusa

In 2013, various estimation methods were employed to assess the factors influencing the financial performance of commercial banks The evaluation of financial performance was based on three key metrics: Return on Equity (ROE), Return on Assets (ROA), and Net Interest Margin.

Financial performance of commercial banks is a category of economic - financial performance that reflects the quality of commercial banks' business

8 activities; it is the commercial bank's ability to achieve business goals by establishing, organizing, and operating business strategies, policies, and programs

2.1.2 The indicators reflect the financial performance of commercial banks a Group of criteria on equity

 Capital adequacy ratio determined by the formula:

Minimum capital adequacy ratio (CAR) = Own capital × 100%

As per Circular 13/2010/TT-NHNN, credit institutions in Vietnam are required to maintain a Capital Adequacy Ratio (CAR) of at least 9% This requirement will be adjusted to a minimum of 8% by the year 2020, in accordance with Circular 41/2016/TT-NHNN.

Credit quality is assessed through the ratio of bad debt to overdue debt relative to total outstanding loans A low ratio indicates strong credit quality and a robust financial position for the bank, while a high ratio suggests poor credit management and potential financial concerns Additionally, the criteria regarding the size and quality of assets play a crucial role in evaluating a bank's overall financial health.

Assessing property condition is crucial for determining asset quality, a key indicator of a commercial bank's financial sustainability and management effectiveness Key criteria for evaluating asset quality include total outstanding loans, growth rates of total assets and loans, bad debt ratios, overdue ratios, and the proportion of outstanding loan debt to total assets.

Assessing the capital situation is crucial for a bank's operations, as mobilized capital serves as a key indicator An increase in mobilized capital allows the bank to expand its loan offerings, reinforcing its role as a borrower Consequently, it is essential for the bank to consistently monitor the scale and structure of mobilized capital, analyzing it by term to ensure effective financial management.

The mobilization of capital involves various objects, including economic organizations and individuals, and is conducted in both VND and foreign currencies This process establishes the structure of each component within the mobilized capital, allowing for a thorough review and evaluation of the sources of mobilized capital to implement appropriate adjustments Furthermore, it is essential to monitor the growth rate of mobilized capital Key criteria for assessing the size and quality of these capital sources include capital structure, average mobilizing interest rates, total capital sources, and capital growth rate, along with a focus on profitability metrics.

Profitability serves as a crucial indicator of a bank's operational size, quality, efficiency, business orientation, and competitiveness It is evaluated through key metrics such as Return on Equity (ROE), Return on Assets (ROA), and Net Interest Margin (NIM).

Profitability is a measure to assess the business situation of commercial banks Profitability is analyzed through the following parameters:

ROE: Return on equity measures how much profit a bank's equity will generate over a given time period (usually one year)

Equity × 100 ROA: shows profitability on total assets - a measure of a bank's management, showing a bank's ability to convert assets into net income (Return on Assets – ROA)

A high Return on Assets (ROA) or Return on Equity (ROE) signifies that a bank is valued by both customers and investors, indicating strong profitability This financial strength not only enhances the bank's competitiveness but also serves as a key indicator of its overall performance in the commercial banking sector.

Net Interest Margin (NIM) reflects the percentage difference between a bank's total returns and total interest expenses, highlighting the profit banks earn from the disparity in interest rates between deposits and investments.

NIM(%) =Total interest income − total interest expenses

Total earning assets × 100 d Group of criteria for liquidity

Ensure liquidity: The ratio of credit to mobilized capital On November 15,

In 2019, the State Bank of Vietnam introduced Circular No 22/2019/TT-NHNN, which established the limits and ratios for ensuring safety in the operations of banks and foreign bank branches.

The Loan-to-Deposit Ratio (LDR), defined as the ratio of outstanding loans to total deposits, indicates the financial health of banks In joint stock commercial banks, the LDR has reached its peak, demonstrating that private banks are increasingly willing to balance liquidity risk against profitability This trend highlights the strategic choices made by these banks in managing their financial operations.

Cash position index: this index has good liquidity

Cash position index = Cash + Deposits with other credit institutions

Total assets The cash position index has good liquidity This shows that if the cash position index is high, the liquidity status of commercial banks is higher

2.1.3 Indicators reflecting the financial performance of joint stock commercial banks according to the safety framework CAMEL

In the 1970s, the United States implemented the CAMEL analytical framework through three federal banking regulators: the Federal Reserve, the Federal Deposit Insurance Corporation, and the Office of the Comptroller of the Currency This framework, part of the "Uniform Financial Institutions Rating System," evaluates the overall health of banking and financial institutions based on five key components: Capital Adequacy, Asset Quality, Management, Earnings, and Liquidity.

The Capital Accord, established by the Basel Committee on Banking Supervision, aims to standardize capital adequacy criteria across banks Basel I set the initial capital adequacy standards for G-10 countries, while Basel II enhanced these requirements by increasing the responsiveness of capital to significant banking risks Basel III further strengthens the banking sector's resilience against financial and economic shocks, thereby minimizing the potential spillover effects on the real economy A robust capital foundation not only instills confidence in depositors but also enables financially strong companies to seize commercial opportunities and navigate unforeseen challenges, ultimately leading to greater profitability.

FACTORS AFFECTING THE FINANCIAL PERFORMANCE OF

 Domestic and international political and social environments:

The stability of the socio-political environment, both domestically and internationally, plays a crucial role in enhancing a bank's financial performance A secure socio-political climate instills confidence in individuals when depositing their money, thereby facilitating the attraction of both domestic and international investment capital This stability also encourages individuals and businesses to invest in the medium and long term, further bolstering the bank's growth and financial health.

ROE ROE = Profit after tax

Net Interest Income + Non − Interest Income

Customer deposits to total assets

Total loan to customer deposit (LTD)

Scaling up operations refers to the process of expanding a company's activities to enhance growth However, an unstable socio-political environment can negatively impact the confidence of individuals and organizations in depositing their money, ultimately affecting business performance This instability can lead to deteriorating credit quality and challenges in capital mobilization, presenting significant obstacles for banks.

A stable and sustainable economy requires a robust legal framework tailored to current conditions The operations of commercial banks create diverse economic relationships governed by legal provisions; therefore, inadequate legal regulations can pose risks to businesses and adversely affect the financial performance of commercial banks.

Vietnam's rapid monetization has prompted the need for new legislation and updates to outdated laws to align with the current economic landscape The newly established law provides a robust legal framework aimed at effectively addressing conflicts and complaints arising from both business and social activities.

 Economic policy of the government

The government plays a vital role in the growth of the banking sector by acting as the central authority and supervisor of commercial banks Through the Central Bank, the government influences the industry as both a major debtor and creditor, shaping the macroeconomic environment that directly affects banks' financial performance Additionally, state regulations, including legal capital requirements and capital adequacy ratios, govern the operations of commercial banks, ensuring stability and compliance within the financial system.

18 information transparency standards, among others Commercial banks must adhere to appropriate development directions as a result of these restrictions

A well-developed financial market featuring a variety of banks and non-bank financial institutions fosters competition among entities This heightened competition influences bank profitability, prompting institutions to formulate effective development strategies, innovate new products, and enhance customer service quality to sustain and boost their financial performance.

Financial risks have an impact on commercial banks' financial performance If bad debt continues to rise, commercial banks must increase their risk provisioning

Effective management and governance significantly impact a bank's financial performance, as strong administrative capabilities can reduce operating costs and enhance resource efficiency, ultimately leading to increased profitability for commercial banks.

To thrive in today's rapidly evolving landscape, banks must embrace advanced technology and invest in skilled personnel By integrating innovative technology, they can create modern banking products and services that cater to the needs of a developed port society.

 Qualification and quality of employees:

The success of a commercial bank heavily relies on the human factor, as developed societies demand innovative and high-quality banking services To meet these demands, banks must enhance their human resources to swiftly adapt to market and societal changes Employing professionals with strong ethics and expertise fosters customer loyalty, mitigates business and investment risks, and reduces operating costs Moreover, prioritizing human resource development alongside new technologies is essential for ongoing growth and efficiency in the banking sector.

RESEARCH METHODS

DATA COLLECTION

Collect panel data through an observational sample of 31 commercial banks in Vietnam from 2010-2020 These Bank figures are collected from the financial statements

Source: Compiled by the author

Table 3.1 List of Commercial banks in Vietnam

No Full Name No Full Name

1 An Binh Commercial Joint Stock Bank

17 Petrolimex Group Commercial Joint Stock Bank (PG Bank)

2 Asia Commercial Joint Stock Bank

3 Bac A Commercial Joint Stock Bank

19 SaiGon Joint Stock Commercial Bank (SCB)

4 Joint Stock Commercial Bank For

20 Southeast Asia Commercial Joint Stock Bank (SeABank)

5 Bao Viet Joint Stock Commercial Bank

21 SaiGon Bank For Industry and Trade (SAIGONBANK)

6 Vietnam Joint Stock Commercial Bank

For Industry and Trade (VietinBank)

22 SaiGon – HaNoi Commercial Joint Stock Bank (SHB)

RESEARCH MODELS

7 Dong A Commercial Joint Stock Bank

23 SaiGon Thuong Tin Commercial Joint Stock Bank (Sacombank)

8 Viet nam Export Import Commercial

24 Vietnam Technological And Commercial Joint Stock Bank (Techcombank)

9 Ho Chi Minh City Development Joint

25 Tien Phong Commercial Joint Stock Bank (TPBank)

10 Kien Long Commercial Joint Stock

26 Vietnam Asia Commercial Joint Stock Bank (VietABank)

11 Lien Viet Post Joint Stock Commercial

27 Vietnam Joint Stock Commercial Bank (Vietbank)

12 Military Commercial Joint Stock Bank

28 JointStock Commercial Bank For Foreign Trade Of Vietnam

29 Vietnam International Commercial Joint Stock Bank (VIB)

14 Nam A Commercial Joint Stock Bank

30 Vietnam Prosperity Joint Stock Commercial Bank (VPBank)

15 National Citizen Bank (NCB) 31 Viet Capital Commercial Joint Stock

16 Orient Commercial Joint Stock Bank

The dependent variable on financial performance is taken based on previous studies: Ongore and Kusa (2013), Gul and Zaman (2011), Bảo (2016)

The research model is built as follows:

FIP it = β 1 + β 2 EA it + β 3 SIZE it + β 4 FATA it + β 5 LLR it + β 6 CIR it + β 7 NPL it + β 8 LOAN it + β 9 LIQ it + β 10 GDP t + β 11 CPI t +  + i

FIP it : is the financial performance of bank i at time t as expressed by ROA, ROE, NIM

EA it = Capital adequacy of bank i at time t

SIZE it = Bank size of bank i at time t

FATA it = Asset quality of Bank i at time t

LLR it = Loan loss reserves of bank i at time t

CIR it = Management quality of bank i at time t

NPL it = Bad dept ratio of bank i at time t

LIQ it = Liquidity of bank i at time t

GDP t = Economic growth (GDP) at time t

CPI t = Inflation rate at time t

 it = Error term where i is cross sectional and t time identifier

DESCRIPTION VARIABLE AND RESEARCH HYPOTHESIS

Return on Assets (ROA) is a key financial metric that measures a bank's profitability by dividing net profit after tax by total assets This ratio is derived from the financial statements of commercial banks in Vietnam and has been widely utilized in academic studies to assess banking performance, including research conducted by San and Heng (2013), Ongore and Kusa (2013), Gul and Zaman (2011), and Robin and Bloch (2018).

(2016), Cuong (2017) ROA is measured according to the following formula:

Return on Equity (ROE) assesses the profitability generated from shareholders' equity, indicating how effectively a company utilizes its investment funds to drive earnings growth By evaluating ROE, investors can gauge a firm's efficiency in converting every unit of equity into profits, as highlighted in various research studies, including those by San and Heng.

(2013), Ongore and Kusa (2013), Gul and Zaman (2011), Robin and Bloch (2018) Bảo (2016, Cường (2017) ROE is measured according to the following formula:

Net interest margin (NIM) is a key metric that reflects the net return on a bank's earning assets, encompassing investment securities, loans, and leases Defined as the ratio of interest income to total assets, NIM serves as an important indicator in various research studies, including those by San and Heng (2013), Ongore & Kusa (2013), Gul and Zaman (2011), Robin and Bloch (2018), Bảo (2016), Cường (2017), and Duong and Nguyen (2021).

Capital adequacy of a bank is measured by Equity to Asset ratio (EA)

Equity to Total Assets (EA) measures a bank's capital adequacy, indicating its capacity to absorb financial shocks A high EA signifies a bank's robust ability to withstand financial risks, minimizing reliance on external funding and enhancing profitability Additionally, well-capitalized banks can seize more business opportunities and demonstrate effective risk management, which lowers insolvency risks and reduces borrowing needs Research by San and Heng (2013), Ongore and Kusa (2013), Bảo (2016), and Duong et al (2020) confirms that a strong EA positively influences financial performance.

H1: Capital adequacy has a positive impact on financial performance

The variable bank size is chosen by many scholars to include in the research model, which is measured by taking the natural logarithm of total assets

The size of a bank is measured by the logarithm of its total assets, with growth in total assets signaling an expansion phase However, this increase often leads to a higher proportion of risky assets, which can decrease the capital adequacy ratio Empirical studies have consistently demonstrated a positive correlation between bank size and performance.

37 effect on financial performance as the study of Gul and Zaman (2011), San and Heng (2013), Duong & Nguyen (2021) , Bảo (2016) Based on that, the author formulates the following hypothesis:

H2: Bank size has a positive impact on financial performance

Asset quality is crucial for banks as it enables them to gauge the risk levels associated with their customer disclosures By evaluating asset performance, banks strive to minimize non-performing loans, as these can significantly harm profitability Additionally, asset quality reflects the proportion of fixed assets held by a bank relative to its total assets, providing insight into its financial health.

Total assets Dembel (2020) studied show that inflation has a negative impact on financial performance Based on that, the author formulated the research hypothesis:

H3: Asset quality has a negative impact on financial performance

Loan Loss Reserves reflect the credit risk levels faced by banks, highlighting the historical vulnerability of loan quality that has often led to bank failures.

Total assets San and Heng (2013), Phan Thu Bao (2016) studied shows a possitive effect on financial performance Based on that, the author develops the following hypothesis:

H4: Asset quality has a possitive impact on financial performance

Management quality is calculated as Operating Expense divided by Net Income plus Non – Interest income

Net Interest Income combined with Non-Interest Income is crucial for assessing financial performance Research by Tomuleasa and Cocris (2014) and Gul and Zaman (2011) indicates that a high bad debt ratio adversely affects financial outcomes Consequently, this leads to the formulation of the research hypothesis.

H5: Management quality has a possitive impact on financial performance

NPL shows the ratio of bad debt to total outstanding debt

Total outstanding debt Usman & Lestari (2019), Getahun (2015) studied show that bad dept ratio has a negative impact on financial performance Based on that, the author formulated the research hypothesis:

H6: Bad dept ratio has a negative impact on financial performance

The ratio of total loans to total assets, referred to as LOAN, has been shown to positively influence financial performance, according to a study by Gul and Zaman (2011) This finding led the author to develop a research hypothesis centered on the relationship between LOAN and financial outcomes.

H7: LOAN has a positive impact on financial performance

Liquidity is defined as the ratio of total deposits to total assets, encompassing both demand and time deposits It serves as a key indicator of a bank's performance, reflecting its capacity to meet obligations primarily to depositors.

Experimental studies on liquidity management yield mixed results Owoputi et al (2014) discovered that liquidity management negatively impacts Return on Assets (ROA), Return on Equity (ROE), and Net Interest Margin (NIM) Consequently, the fourth hypothesis is established.

H8: Liquidity has a negative impact on financial performance

Inflation refers to the rate at which the overall prices of goods and services increase in an economy over time As inflation rises, consumer purchasing power diminishes, leading to a situation where individuals can acquire fewer goods and services for each unit of currency.

In general, inflation is calculated by calculating the inflation rate of a price index, such as the consumer price index (CPI)

Gul and Zaman (2011), San and Heng (2013), Bảo (2016) studied show that inflation has a positive impact on financial performance But Ongore and Kusa

(2013) research shows that inflation has a negative impact on financial performance Therefore, the fourth hypothesis is formulated as follows:

H9: Inflation has a positive impact on financial performance

H9: Inflation has a negative impact on financial performance

The GDP growth rate variable is calculated as the GDPG growth rate of the year of observation

Ongore and Kusa (2013) research shows that GDP has a negative effect on ROA and NIM and has a positive effect on ROE Meanwhile Robin and Bloch

(2018), Phan Thu Bao (2016) GDP has a negative effect on financial performance The opinions from scholars are different, but based on the fact of Vietnam, the author hypothesizes as follows:

H10: Economic growth has a positive impact on financial performance

Table 3.2 Summary of research on financial performance

Name Measurement method ES Empirical evidence in previous studies

ROA Return on assets Profit after tax / Total assets

San and Heng (2013); Ongore and Kusa (2013); Bảo (2016); Duong and Nguyen (2021); Gul and Zaman (2011);

Equity Profit after tax)/Equity

NIM Net interest margin Net interest income /total earning assets

+ San and Heng (2013); Ongore and Kusa (2013); Bảo

SIZE Bank size Log (Total asset) + Gul and Zaman (2011); Bảo (2016); San and Heng (2013);

FATA Asset quality Fixed assets

Loan Loss Reserves Total Assets - San, O T., & Heng, T B (2013); Phan Thu Bao (2016)

Net Interest Income + Non − Interest Income + Getahun (2015); Desta (2016); Tomuleasa and Cocris

NPL Bad dept ratio - Usman & Lestari (2019); Getahun (2015)

Bad debt Total outstanding debt

Total Assets + Gul and Zaman (2011); Tibebe (2020)

LIQ Liquidity Total Customer Deposits

Total Assets - Owoputi et al (2014)

San and Heng (2013); Ongore and Kusa (2013); Bảo (2016); Duong & Nguyen (2021); Gul and Zaman (2011);

CPI Inflation rate CPI =CPI 𝑡 − 𝐶𝑃𝐼 𝑡−1

San and Heng (2013); Ongore and Kusa (2013); Bảo

RESEARCH PROCESS

The author conducted a study by utilizing secondary data sourced from commercial banks' websites, including annual reports, balance sheets, and accounting records, covering the years 2010 to 2020, and subsequently recalculated the pertinent indicators.

The author uses STATA 13 software to perform a summary description of the data's features, such as mean, maximum, minimum, and standard deviation, for the dependent variable and independent variables

Step 3: Analyze the correlation matrix between variables

One of the assumptions of linear regression is that the independent variables

4 • Model analysis by Pooled OLS, FEM, REM methods

5 • Check for suitable model selection

6 • Testing and handling model defects

To investigate the relationships among independent variables, a correlation matrix is utilized, revealing that 45 variables exhibit no connection Additionally, the article employs the variance inflation factor (VIF) to identify group correlations, which are indicative of multicollinearity.

Step 4: Test the model by Pooled OLS, FEM, REM model

The author implements three regression models: the Pooled Ordinary Least Squares (OLS) method, the Fixed Effects Model (FEM), and the Random Effects Model (REM) to analyze the data effectively.

Step 5: Check for suitable model selection

After estimating with three approaches, Pooled OLS, FEM, and REM, the author runs a series of tests, including the F test, Breusch – Pagan, and Hausman, to choose the best model

Step 6: Testing and handling model defects

Model defect tests enhance the reliability and relevance of study findings by addressing common defects in quantitative research Key tests include the Variance Exaggeration Factor (VIF) for detecting multicollinearity, the Modified Wald test for identifying variable variance, and the Wooldridge test for assessing autocorrelation To tackle issues such as variable error variance and autocorrelation in panel data, the author utilizes the Feasible Generalized Least Squares (FGLS) approach Additionally, the author employs the System Generalized Method of Moments (S-GMM) to resolve endogenous issues within the model.

RESEARCH METHODS

The least squares estimation model, also known as the OLS estimation model, is a model for estimating the coefficients of the explanatory variable on the

The mean of the dependent variable is calculated as 46, following the principle of minimizing the sum of the squares of the model's residuals Residuals represent the differences between the actual and predicted values of the dependent variable, influenced by the explanatory factors.

The fixed effects model (FEM) is employed when crossover units are not uniform, allowing for the analysis of the impact of explanatory variables on the dependent variable while considering individual characteristics FEM assumes that the partial regression coefficients remain consistent across cross units, although the regression intercepts vary.

The REM model estimates distinct intercepts for each cross-unit while also assessing the overall impact of explanatory variables Each unit's cross-intercept is based on a common, time-invariant intercept combined with a random variable that varies by subject and time, reflecting the unique error component associated with each subject.

3.5.4 Feasible Generalized Least Square (FGLS)

Feasible Generalized Least Squares (FGLS) is an effective technique for estimating models that exhibit variable variance or autocorrelation Rather than assuming a specific structure for variable variance, FGLS utilizes Ordinary Least Squares (OLS) to estimate the model parameters, even in the presence of these issues By calculating the variance-covariance matrix of the model errors, FGLS transforms the original variables, allowing for accurate parameter estimation despite the complexities introduced by variable variance and autocorrelation.

3.5.5 System Generalized Model of Moments (S-GMM)

Antoniou et al (2006) highlight the effectiveness of System GMM for estimating dynamic models, emphasizing its ability to address endogeneity concerns By incorporating lagged variables, System GMM delivers robust estimates even when faced with issues of variable variance or autocorrelation It is recommended to utilize the first degree of the dependent variable as the explanatory variable in this model.

The author relies on four conditions to ensure that the System GMM estimates are appropriate:

The Hansen test evaluates the appropriateness of the instrumental variable under the null hypothesis (H0), which asserts that no endogenous phenomenon exists For the instrumental variable included in the model to be deemed statistically significant, the P-value of the test should exceed 10%.

Hansen Sargan's test with the hypothesis H0: the instrumental variable is exogenous with the condition that the P-value is greater than 10%

The Arellano-Bond test (AR(2)) is utilized to assess autocorrelation at various levels, with the null hypothesis (H0) indicating no autocorrelation A p-value exceeding 0.1 suggests that the null hypothesis cannot be rejected, indicating the absence of second-order autocorrelation.

In addition, it is necessary to ensure that the number of instrument variables cannot exceed the number of units to be studied

3.5.6 Check for suitable model selection

 F-test to choose Pool OLS or FEM model:

H0: There is no difference between different subjects or time points

H1: There is a difference between objects or different times

If p-value ≤ 𝛼 with (𝛼 = 5%) then reject H0, FEM model is selected, otherwise if p-value ≥ 𝛼 the OLS model is selected

 Hausman test is performed to choose between FEM and REM models:

H0: Error of estimate does not include inter-subject deviations

H1: The error of the estimate includes deviations between objects

If p-value ≤ 𝛼 with (𝛼 = 5%) then reject H0, FEM model is selected, otherwise if p-value ≥ 𝛼 then the REM model is selected

 Breusch & Pagan test to choose OLS and REM:

H0: Error of estimate does not include inter-subject deviations

H1: The error of the estimate includes deviations between objects

If p-value ≤ 𝛼 with (𝛼 = 5%) then reject H0, REM model is selected, otherwise if p-value ≥ 𝛼 then the OLS model is selected

Upon choosing the suitable model, we utilize the Random Effects Model (REM) to analyze the outcomes, employing the Modified Ward test and assessing autocorrelation through the Wooldridge test.

 In the FEM model, the Modified Ward test is used to test the phenomenon of variance:

In statistical analysis, if the p-value is less than or equal to the significance level (α = 5%), we accept the alternative hypothesis (H1) and reject the null hypothesis (H0), indicating that the regression model exhibits variable variance Conversely, if the p-value is greater than or equal to α, we accept H0, suggesting that the regression model does not demonstrate changes in variance.

 The test for autocorrelation in the FEM model is the Wooldridge test with the hypothesis:

H0: There is no autocorrelation in the model

H1: There is autocorrelation in the model

If the p-value is less than or equal to 𝛼 (with 𝛼 set at 5%), we accept the alternative hypothesis (H1) and reject the null hypothesis (H0), indicating the presence of autocorrelation in the model When autocorrelation and variance changes are detected, the Feasible Generalized Least Squares (FGLS) model effectively addresses these issues by controlling for both autocorrelation and variable variance.

SGMM addresses the endogeneity issue of certain explanatory variables by utilizing instrumental variables To ensure the validity of these instrumental variables, the Sargan test or Hansen test can be employed to assess their over-identifying properties.

 Hansen test is used to test the over-identifying of instrumental variables (to determine whether there is a correlation between instrumental variables and residuals in the model) with the hypothesis:

H0: Instrumental variables are suitable (satisfactory over-identifying)

H1: Instrumental variables are not appropriate (not satisfactory over- identifying)

In statistical analysis, if the p-value exceeds 10%, we accept the null hypothesis (H0), indicating that the instrumental variables used in the model are appropriate Conversely, a p-value below 10% leads to the rejection of H0 and the acceptance of the alternative hypothesis (H1), suggesting that the instrumental variables are unsuitable for the model.

 The second order autocorrelation test (AR2) to test the second order correlation of residuals in the model, with the hypothesis:

H0: There is no quadratic correlation of the residuals

H1: There is a quadratic correlation of the residuals

If the p-value is greater than 10%, we accept the null hypothesis (H0), indicating that the model's residuals do not exhibit second-order autocorrelation and that the model meets the necessary criteria Conversely, if the p-value is less than 10%, we reject the null hypothesis, suggesting that the model's residuals do show second-order autocorrelation, making the model unsatisfactory.

In Chapter 3, the author outlines the research model and variables, detailing the data collection process and the methodologies employed, including OLS, FEM, REM, FGLS, and S-GMM This comprehensive approach ensures that the results align with the intended objectives and facilitates the identification of the most effective model for analysis.

RESEARCH RESULTS AND DISCUSSION

DESCRIPTIVE STATISTICAL

Table 4.1 Summary of Descriptive statistics

Variable Obs Mean Std Dev Min Max

Source: Calculation results from Stata software

Table 4.1 presents descriptive statistics for the variables analyzed in this study, which encompasses data from 31 commercial banks in Vietnam, including both state-owned and private joint-stock banks The statistics include key metrics such as return on total assets, return on equity, net profit margin, equity to asset ratio, asset size, fixed assets to total assets, provision ratio, liquidity, inflation rate, and total domestic assets.

Return on total assets (ROA) of 31 commercial banks in Vietnam from 2010-

In 2020, the average value was 0.79% with a standard deviation of 0.63% The lowest recorded value was 0.01% from the National Citizen Bank (NCB) in 2012, while the highest value reached 4.72% from SaiGon Bank For Industry and Trade (SAIGONBANK) in 2010.

The average return on equity (ROE) is 8.95%, with a standard deviation of 6.46% The lowest recorded ROE was 0.06% for the National Citizen Bank (NCB) in 2012, while the highest was 26.82% for Asia Commercial Joint Stock Bank (ACB) in 2011.

The net interest margin (NIM) in the banking sector has an average of 2.87% and a standard deviation of 1.29% Notably, Vietnam Public Bank (PVcomBank) recorded a peak NIM of 8.836% in 2013, while Vietnam Prosperity Joint Stock Commercial Bank (VPBank) reported the lowest NIM at 0.102% in 2019.

Capital adequacy (EA) averages 9.35% with a standard deviation of 4.05% The lowest recorded EA is 2.93%, noted at Sai Gon Bank for Industry and Trade (SAIGONBANK) in 2019, while the highest EA is 25.53%, achieved by Kien Long Commercial Joint Stock Bank (Kienlongbank) in 2010.

The size of banks, indicated by the natural logarithm of total assets, averages 8.013 with a standard deviation of 0.494 Viet Capital Commercial Joint Stock Bank (VCA/CAB) recorded the smallest size at 6.915 in 2010, while the largest size was noted at 9.1808 by the Joint Stock Commercial Bank for Investment and Development of Vietnam (BIDV) in 2020.

The asset quality (FATA) has an average of 1.29% and a standard deviation of 1.19% The lowest recorded value is 0.07%, attributed to Bao Viet Joint Stock Commercial Bank (BVB) in 2017, while the highest value of 6.01% was noted at Sai Gon Bank for Industry and Trade (SGB) in 2014.

Bad dept (LLR) has a mean of 1.05%, with a standard deviation of 0.86%

The minimum value is -1.01% of SaiGon – HaNoi Commercial Joint Stock Bank (SHB) in 2012 and the maximum value is 5.40% of Vietnam Prosperity Joint Stock Commercial Bank (VPBank) in 2019

The management quality, measured by the CIR, has an average score of 48.38% and a standard deviation of 12.90% The lowest recorded value was 21.36% for Viet Capital Commercial Joint Stock Bank (VCA/CAB) in 2013, while the highest was 87.07% for National Citizen Bank (NCB) in 2014.

The average non-performing loan (NPL) ratio stands at 2.21%, with a standard deviation of 1.75% The lowest recorded NPL ratio is 0.02%, attributed to Tien Phong Commercial Joint Stock Bank (TPBank) in 2010, while the highest is 21.40%, reported by Southeast Asia Commercial Joint Stock Bank (SeABank) in the same year.

The average loan percentage stands at 54.73%, with a standard deviation of 12.44% The lowest recorded loan percentage is 19.10%, attributed to Southeast Asia Commercial Joint Stock Bank (SeABank) in 2011, while the highest percentage of 78.80% was noted at the Joint Stock Commercial Bank for Investment and Development of Vietnam (BIDV) in 2020.

Liquidity (LIQ) has an average of 63.87% and a standard deviation of 12.83% The lowest recorded liquidity was 25.08% at Tien Phong Commercial Joint Stock Bank (TPBank) in 2012, while the highest was 89.37% at SaiGon Thuong Tin Commercial Joint Stock Bank (STB) in 2015.

The macroeconomic model incorporates GDP and CPI as key variables, highlighting significant fluctuations in the inflation rate Notably, the standard deviation of the Consumer Price Index (CPI) reached its peak in 2011 and 2015, indicating considerable variability in inflation during those years.

CORRELATION ANALYSIS

Table 4.2 Correlation between ROA and independent variables

ROA EA SIZE FATA LLR CIR NPL LOAN LIQ CPI GDP

Table 4.2 presents the correlation coefficients for independent variables, revealing that EA, LLR, and CPI have coefficients of 0.3290, 0.1552, and 0.2747, respectively, all significant at the 1% level Additionally, SIZE shows a correlation coefficient of 0.0983 at the 5% significance level, while LOAN has a coefficient of 0.1164 at the 10% significance level.

The analysis reveals that 55 variables positively correlate with the dependent variable, Return on Assets (ROA) In contrast, the Current Income Ratio (CIR) and Liquidity (LIQ) exhibit negative correlations with ROA, with coefficients of -0.6287 and -0.254, respectively, both significant at the 1% level Additionally, Gross Domestic Product (GDP) shows a negative relationship with ROA, reflected by a correlation coefficient of -0.1054, significant at the 10% level.

Table 4.3 Correlation between ROE and independent variables

ROE EA SIZE FATA LLR CIR NPL LOAN LIQ CPI GDP

Table 4.3 shows that the independent variables SIZE, LOAN, and CPI have correlation coefficients of 0.4592, 0.1853, and 0.1597, respectively, at the 1% significance level, and LLR has a correlation coefficient of 0.1083 at the

At a 10% significance level, all examined variables show a positive relationship with the dependent variable, Return on Equity (ROE) However, correlation coefficients for Economic Activity (EA), Fixed Assets to Total Assets (FATA), Capital Investment Ratio (CIR), Non-Performing Loans (NPL), and Liquidity (LIQ) are -0.1650, -0.2419, -0.613, -0.1948, and -0.1184, respectively, indicating a negative correlation with ROE at a 1% significance level Additionally, Gross Domestic Product (GDP) does not demonstrate statistical significance in relation to ROE.

Table 4.4 Correlation between NIM and independent variables

NIM EA SIZE FATA LLR CIR NPL LOAN LIQ CPI GDP

Table 4.4 presents correlation coefficients for independent variables EA, FATA, LLR, LOAN, and CPI, which are 0.4209, 0.2136, 0.5276, 0.2628, and 0.2271, respectively, all showing a positive relationship with the dependent variable NIM at the 1% significance level In contrast, CIR and GDP exhibit negative correlations with NIM, with coefficients of -0.4006 and -0.1110, also at the 1% significance level Additionally, SIZE and LIQ do not demonstrate statistical significance in relation to NIM.

MULTICOLLINEARITY TEST

Multicollinearity occurs when independent variables in a model exhibit linear correlation with one another To investigate this phenomenon, the study utilized the Variance Inflation Factor (VIF) criterion to test the hypothesis of no multicollinearity, with the results detailed in the accompanying table.

Table 4.5 Multicollinearity VIF Variable VIF 1/VIF

VIF of all independent variables is less than 4, so multicollinearity in the model is assessed as not serious

ESTIMATED THE THE POOLED OLS, FEM, REM MODELS

This study analyzes the correlation coefficient to explore the relationship between model variables, followed by a regression analysis aimed at measuring the impact of independent variables on the dependent variable Various methods, including Pooled OLS, FEM, and REM, are employed, along with tests to determine the most suitable regression approach.

Table 4.6 Estimated of Model 1 (Pooled OLS, FEM and REM)

Variable POOLED OLS FEM REM

Coef P- value Coef P- value Coef P- value

Selection OLS and FEM FEM and REM OLS and REM

There is no difference between different subjects or time points

There is no correlation between the characteristic error between the subjects and the explanatory variables

The error of the estimate does not include the deviations between subjects

P-value Prob>F = 0.000 Prob>chi2 = 0.003 Prob > chi2 =0.000

Conclude Rejected Ho Rejected Ho Rejected Ho

Coef P- value Coef P- value Coef P- value

Coef P-value Coef P-value Coef P-value

Source: Calculation results from Stata software.

SELECTION TEST OF 3 MODELS POOLED OLS AND FEM

 Check the fit between Pooled OLS model and FEM

The author employs the Wald F-test to reassess the compatibility between the Pooled OLS model, a classical linear regression approach, and the Fixed-Effects Model (FEM) at a defined significance level, testing the hypothesis H0: The Pooled OLS model is more appropriate The findings of this analysis are presented in the results.

The results of model 1, model 2, and model 3 give the same results:

We see that Pro>chi2 = 0.0000 < 0.05, so we reject hypothesis H0, accept hypothesis H1, FEM model is more suitable

 Testing the fit between Pooled OLS and REM models

Next, to choose between Pooled OLS and REM models, the research uses Breusch

- Pagan test with hypothesis H0: Pooled Ols model is more suitable The model gives the following results:

The results of model 1, model 2, and model 3 give the same results:

The results show that the coefficient Pro>chibar2 = 0.0000 < 0.05 should reject the hypothesis H0, accept the hypothesis H1 is to use the REM model

 Check the fit between FEM and REM models

After conducting the Wald F-test and Breusch-Pagan test, the findings indicate that the Fixed Effects Model (FEM) and Random Effects Model (REM) are more appropriate than the Ordinary Least Squares (OLS) models To further assess the compatibility of the FEM and REM, the author employs the Hausman test, with the null hypothesis stating that the REM is the better-fitting model The results of this analysis support the preference for either the FEM or REM models.

1, model 2, and model 3 give the same results:

The results of model 1 show that Pro>chi2 = 0.003 < 0.05, model 2 and model

3 givé the same result Pro>chi2 = 0.000 < 0.05, so we reject the hypothesis H0, accept the hypothesis H1 so chooses FEM model

Thus, after performing the tests to select model, model 2 and model 3, the author finds that the FEM model is the most suitable to measure the factors affect financial performance

Model defect tests enhance the reliability and relevance of research findings by addressing three common issues in quantitative research: multicollinearity, variable variance, and autocorrelation The analysis confirms the absence of multicollinearity in Table 4.5 through the variance inflation factor (VIF) test Furthermore, this section employs the Modified Wald test to assess variable variance and the Breusch-Godfrey test to identify autocorrelation.

 Test of variance of variable error

The Modified Wald test was conducted to evaluate the hypothesis H0: there is no variance phenomenon The results, as shown in Tables 4.9, 4.10, and 4.11, indicate that Prob > chi2 = 0.0000, which is less than 5% Therefore, we reject H0 and accept H1, concluding that the model exhibits a variable variance phenomenon.

Table 4.9 Modified Wald Test for ROA

Modified Wald test for groupwise heteroskedasticity in fixed effect regression model

Ho: sigma(i)^2 = sigma^2 for all i chi2 (32) = 6191.75 Prob>chi2 = 0.0000

Table 4.10 Modified Wald Test for ROE

Table 4.11 Modified Wald Test for NIM

The Wooldrige test results, presented in tables 4.12, 4.13, and 4.14, indicate a coefficient of Prob > F = 0.000, which is less than the significance level of 0.05 Consequently, the thesis rejects the null hypothesis (Ho) that suggests the presence of autocorrelation in the model.

Table 4.12 Wooldridge test - ROA autocorrelation test

Wooldridge test for autocorrelation inpaneldata

Table 4.13 Wooldridge test - ROE autocorrelation test

Table 4.14 Wooldridge test - NIM autocorrelation test

ESTIMATED THE FGLS

The fixed effects model often encounters issues such as variable variance and autocorrelation To address these challenges, this study employs the feasible generalized least squares (FGLS) method, which effectively mitigates these phenomena within the model.

DEFECT TESTING OF MODEL

4.7.1 Check the endogeneity of the variables in the model

The endogenous phenomenon of independent variables is common in panel data models, and it affects the topic's analysis results when using the OLS method

The research findings are compromised due to bias and instability, rendering the estimated results ineffective and unreliable This instability stems from endogenous phenomena that arise from measurement errors linked to inaccurate financial statement data.

The Durbin-Wu-Hausman test is employed to assess the endogeneity of the research models, specifically to identify the presence of endogenous variables This test involves a systematic four-step process to ensure accurate evaluation.

 Step 01: Regression of the equation with the dependent variable, the independent variable includes all the variables in the main regression model and the instrumental variable

 Step 2: Get the remainder (r) of the equation you just finished running in step

 Step 3: Regression of the main equation, but adding the residual (r) of step

02 The purpose is to check whether the residual (r) has an impact on the dependent variable of the main equation

 Step 04: Test the coefficient (r) with the hypothesis: H0: The variable is not endogenous (exogenous) and H1: The variable is endogenous

If the p-value of the test (r) is less than the statistical significance level (5%)

=> the conclusion rejects H0 => The variable is endogenous and the model occurs endogenous phenomenon

The following table summarizes the results of testing the endogeneity of each independent variable in the research models with each dependent variable (ROA, ROE and NIM)

Variable P- Value Result P- Value Result P- Value Result

Source: Author's compilation from data analysis results of Stata 13.0 software

 ROA: the endogenous variable is the independent variable SIZE and NPL The rest EA, FATA, LLR, CIR, LOAN, LIQ, CPI and GDP are exdogenous

 ROE: the endogenous variable is the independent variable SIZE and NPL Remaining EA, FATA, LLR, CIR, LOAN, LIQ, CPI and GDP are exdogenous

 NIM: the endogenous variable is the independent variable FATA, NPL and LOAN EA, SIZE, LLR, CIR, LIQ, CPI and GDP are exdogenous

4.7.2 Check the suitability of the GMM model

Sargan test, Hansen test, and second order series correlation AR (2) are important tests to use when using GMM estimation to check the appropriateness of instrumental variables in the model

The Hansen and Sargan tests evaluate the hypothesis H0, which posits that the instrumental variable is exogenous, meaning it is not correlated with the error term of the model A higher p-value for the Hansen statistic indicates a more favorable outcome.

The AR (2) test assesses the presence of second-order correlations in the residuals of a research model, with the null hypothesis (H0) stating that no chain correlation exists A higher p-value indicates a more favorable outcome in this analysis.

Table 4.17 Sargan Hansen and Arellano-Bond Test

Source: Author's compilation from data analysis results of Stata 13.0 software

The results in Table 4.17 show that: The validity and the instrumental variables of the model are shown in the number of instrumental variables: ROA is

23, ROE is 27, NIM is 28, all of which are less than or equal to the number of important groups.

Hansen and Sargan test have p-values greater than 10% in all models ROA, ROE, NIM, showing that the instrumental variables in the regression model are suitable

The AR (2) test has p-values greater than 10% in all models ROA, ROE, and NIM Therefore, it can be concluded that the GMM estimates in all three models are efficient

In dynamic panel data models, the first-order lag of the dependent variable is included as an independent variable alongside other factors The System GMM method is commonly used for estimating these models due to its effective handling of endogeneity The findings from this research are presented as follows:

Table 4.18 S-GMM regression results of Model 1 Dependent Variable ROA Coef Std Err t P>t

The regression analysis reveals that the variables EA, SIZE, FATA, CIR, LOAN, LIQ, CPI, and GDP are statistically significant at the 1% level, while other variables do not show significance Additionally, a negative correlation was identified between FATA and both LIQ and ROA.

The variables EA, SIZE, CIR, LOAN, CPI, and GDP positively influence ROA, as indicated by the positive sign of the regression coefficient Additionally, the first-order lag variable of ROA shows a significant P-value of 0.014 Consequently, the regression model derived from this study is established based on these findings.

ROA = -0.0853 + 0.4943L1.ROA + 0.1309EA + 0.0083SIZE + 0.1945FATA + 0.0198LLR + 0.0183CIR – 0.0415NPL + 0.0183LOAN – 0.0155LIQ + 0.0543CPI + 0.0504GDP

Table 4.19 S-GMM regression results of Model 2 Dependent Variable ROE Coef Std Err t P>t

The regression analysis revealed that variables such as EA, SIZE, FATA, CIR, NPL, LOAN, CPI, and GDP are statistically significant at the 1% level, while other variables lack significance Notably, a negative correlation was identified between FATA and both NPL and ROE, whereas EA, SIZE, CIR, LOAN, CPI, and GDP positively influence ROE Additionally, the first-order lag variable of ROE exhibited a P-value of 0.001, confirming a positive regression coefficient Thus, the regression model is effectively established based on these findings.

ROE = -0.7070 + 0.5580L1.ROE + 0.8301EA + 0.0578SIZE – 2.1355FATA + 0.5621LLR + 0.2069CIR – 1.6159NPL + 0.1742LOAN – 0.0289LIQ + 1.1073CPI + 0.7098GDP

Table 4.20 S-GMM regression results of Model 3 Dependent Variable NIM Coef Std Err t P>t

The regression analysis reveals that the variables EA, SIZE, LLR, CIR, NPL, LOAN, LIQ, and CPI are statistically significant at the 1% level, while other variables lack significance Notably, a negative correlation exists between LIQ and both CPI and NIM, whereas EA, SIZE, LLR, CIR, NPL, and LOAN positively influence NIM Additionally, the first-order lag variable of NIM shows a P-value of 0.002, indicating a positive regression coefficient Thus, the study's regression model is established based on these findings.

NIM = -0.0955 + 0.2877L1.NIM + 0.1819EA + 0.0096SIZE – 0.1597FATA + 0.2736LLR + 0.0281CIR + 0.1408NPL + 0.0438LOAN – 0.0315LIQ - 0.0760CPI + 0.0261GDP

CONCLUSIONS AND RECOMMENDATIONS

CONCLUSION

The financial performance of commercial banks is a crucial factor in their operations, especially in the context of integration and globalization In Vietnam, commercial banks face significant challenges such as increased competitiveness and business risks Therefore, improving financial efficiency is vital for these banks to thrive in the international economic landscape, necessitating the formulation of targeted business strategies.

The study investigates the influence of seven key parameters—equity to asset ratio (EA), total asset size (SIZE), asset quality (FATA), loan loss reserves (LLR), and liquidity (LIQ)—on various dependent variables, including the same metrics It incorporates macroeconomic factors such as GDP (economic growth) and inflation (INF) using panel data from 31 commercial banks in Vietnam spanning 2010 to 2020 To address model imperfections, the research employs the FGLS regression estimation method for panel data and the S-GMM method to resolve endogenous issues The findings reveal significant insights into the relationships between these financial metrics and macroeconomic indicators.

To begin with, the thesis identifies the elements that influence the financial performance of commercial banks in Vietnam At the 1% level of significance, the

The analysis reveals that factors such as EA, SIZE, CIR, LOAN, CPI, and GDP positively influence Return on Assets (ROA), while FATA and LIQ negatively impact it For Return on Equity (ROE), EA, SIZE, CIR, LOAN, CPI, and GDP also show a positive effect, whereas FATA and NPL adversely affect ROE, with LLR and LIQ being statistically insignificant In terms of Net Interest Margin (NIM), EA, SIZE, LLR, CIR, NPL, and LOAN positively contribute, while LIQ and CPI negatively affect it, and both FATA and GDP lack statistical significance.

Second, the impact of these factors on the financial performance of commercial banks in Vietnam was assessed in this study

The author will provide recommendations to enhance the financial efficiency of commercial banks, based on the research findings, which will be elaborated on in the subsequent section.

PROPOSE RECOMMENDATIONS

5.2.1 Solutions for improving capital adequacy

Chapter 4 reveals that capital adequacy significantly enhances the financial performance of commercial banks in Vietnam, indicating that higher equity leads to better financial outcomes To improve capital adequacy, banks can issue more domestic and foreign shares, increase contributions from strategic shareholders, and retain profits while limiting debt lending It is essential for banks to enhance capital efficiency, manage resources effectively, and eliminate unproductive excess capital Implementing these strategies will not only boost capital adequacy but also improve the overall efficiency of capital utilization.

The study reveals that larger banks tend to perform better due to enhanced operational efficiency Increased bank size facilitates diversification in financial activities and the ability to offer a broader range of products and services, leading to a competitive edge Joint-stock commercial banks can expand their size by boosting loan mobilization through strategies such as deposit mobilization, issuing securities, and borrowing from other financial institutions.

5.2.3 The solution to the bad debt problem

Banks need to enhance their bad debt management strategies by minimizing bad debts and reducing risk provisions while collaborating closely with debt settlement centers to streamline loan recovery processes New customer screening is essential, prioritizing credit approval for those with high credit ratings, as evidenced by CIC's lower outstanding loans at other institutions In cases where customers show signs of insolvency, staff should directly assess the reasons behind cash flow issues and the debtor's future repayment capabilities, promptly reporting findings to management for timely intervention Management should also conduct regular checks on borrowers' capital use to ensure compliance with loan purposes, issuing warnings if funds are misallocated Furthermore, the State Bank and individual banks must implement specific regulations to effectively manage the overall bad debt landscape, focusing on gross bad debt ratios.

To maintain liquidity safety, strictly adhere to the State Bank's restrictions on the ratio of loans to mobilized capital at a suitable level A greater emphasis is

95 being placed on mobilizing and growing corporate capital Diversify your lending products and avoid high-risk lending

Credit operations play a vital role in the revenue generation of commercial banks, making it essential to enhance credit quality and reduce bad debt percentages for maximizing profits Implementing a comprehensive internal credit rating system is crucial for assessing risk levels and facilitating swift, informed lending decisions Before approving loans, it is important to meticulously review professional procedures, ensuring that customer evaluations are accurate, documentation is complete and valid, and any discrepancies are promptly addressed to mitigate potential future losses Continuous monitoring throughout the lending process is necessary to ensure that loan capital is utilized effectively, while regular assessments of ongoing loans help identify problematic accounts quickly and develop proactive strategies for resolution.

LIMITATIONS OF THE TOPIC AND RESEARCH

This study acknowledges a limitation regarding the independent variable, highlighting that besides the variables examined, there are additional factors that can influence the financial performance of commercial banks.

This study highlights that students did not incorporate the S - Sensitivity to Market Risk factor from the CAMELS model to assess how fluctuations in interest rates and exchange rates affect the financial efficiency of the banking system.

This chapter evaluates the research findings, limitations, and future development directions related to joint stock commercial banks in Vietnam It provides recommendations aimed at enhancing the financial performance of these banks while acknowledging existing limitations that require resolution in future research endeavors.

1 Tomuleasa, I I., & Cocriş, V (2014) Measuring the financial performance of the European systemically important banks Financial studies, 18(4)

2 Said, R M., & Tumin, M H (2011) Performance and financial ratios of commercial banks in alaysia and China International Review of Business Research Papers, 7(2), 157-169

3 Desta, T S (2016) Financial performance of “The best African banks”: A comparative analysis through CAMEL rating Journal of accounting and management, 6(1), 1-20

4 Tibebe, A N (2020) Determinants of Financial Performance of Private Commercial Banks in Ethiopia: CAMEL Approach

5 Getahun, M (2015) Analyzing financial performance of commercial banks in Ethiopia: CAMEL Approach Unpublished Thesis, Addis Ababa University

6 Usman, B., & Lestari, H S (2019) Determinants of bank performance in Indonesia Jurnal Minds: Manajemen Ide Dan Inspirasi, 6(2), 11-28

7 Ongore, V O., & Kusa, G B (2013) Determinants of financial performance of commercial banks in Kenya International journal of economics and financial issues, 3(1), 237

8 San, O T., & Heng, T B (2013) Factors affecting the profitability of Malaysian commercial banks African Journal of Business Management, 7(8), 649-660

9 Gul, S., Irshad, F., & Zaman, K (2011) Factors Affecting Bank Profitability in Pakistan Romanian Economic Journal, 14(39)

10 Dembel, M B (2020) Factors Affecting the Performance of Commercial Banks (A Case Study on Commercial Banks in Ethiopia): CAMEL Ratings

Research Journal of Finance and Accounting, 11(5) DOI: 10.7176/RJFA/11-

11 Musyoki, D., & Kadubo, A S (2012) The impact of credit risk management on the financial performance of banks in Kenya for the period International

Journal of Business and Public Management, 2(2), 72-80

12 Robin, I., Salim, R., & Bloch, H (2018) Financial performance of commercial banks in the post-reform era: Further evidence from Bangladesh Economic Analysis and Policy, 58, 43-54

13 Palamalai, S., & Britto, J (2017) Analysis of financial performance of selected commercial banks in India Srinivasan, Palamalai and Britto, John

(2017),“Analysis of Financial Performance of Selected Commercial Banks in India”, Theoretical Economics Letters, 7(7), 2134-2151

14 Owoputi, J A., Olawale, F K., & Adeyefa, F A (2014) Bank specific, industry specific and macroeconomic determinants of bank profitability in Nigeria European scientific journal, 10(25)

15 Sangmi, M U D., & Nazir, T (2010) Analyzing financial performance of commercial banks in India: Application of CAMEL model Pakistan Journal of Commerce and Social Sciences (PJCSS), 4(1), 40-55

16 Kaneza, C (2016) Factors affecting the financial performance of commercial banks listed on the Nairobi Securities Exchange (Doctoral dissertation,

United States International University-Africa)

17 Ahsan, M K (2016) Measuring financial performance based on CAMEL: A study on selected Islamic banks in Bangladesh Asian Business Review, 6(1), 7-56

18 Kouser, R., & Saba, I (2012) Gauging the financial performance of banking sector using CAMEL model: comparison of conventional, mixed and pure

Islamic banks in Pakistan International Research Journal of Finance and Economics, 82(1), 67-88

19 Alharthi, M (2017) Financial performance and stability in Islamic banks: Evidence from GCC countries Corporate Ownership and Control, 14(4),

20 Said, R M., & Tumin, M H (2011) Performance and financial ratios of commercial banks in Malaysia and China International Review of Business

21 Annual reports, financial statements of banks for the period 2010 - 2020

22 DUONG, T T N., PHAN, H T., HOANG, T N., & VO, T T T (2020) The effect of financial restructuring on the overall financial performance of the commercial banks in Vietnam The Journal of Asian Finance, Economics, and

23 Bảo, P T (2016) Phân tích các yếu tố ảnh hưởng đến hiệu quả tài chính của các Ngân hàng thương mại cổ phần Việt Nam Luận văn thạc sĩ kinh tế Đại học Kinh tế Thành Phố Hồ Chí Minh

24 Trung, N H (2013) Các yếu tố ảnh hưởng đến năng lực tài chính của các ngân hàng thương mại Việt Nam Luận văn thạc sĩ kinh tế Đại học Kinh tế Thành Phố Hồ Chí Minh

25 Cường, N Đ (2017) Phân tích nhân tố ảnh hưởng đến năng lực tài chính của các ngân hàng thương mại Việt Nam Luận văn thạc sĩ kinh tế Đại học Kinh tế Thành Phố Hồ Chí Minh

26 TP, T., MINH, H., & NGA, P T H NĂNG LỰC TÀI CHÍNH CỦA CÁC NGÂN HÀNG THƯƠNG MẠI VIỆT NAM Luận án tiến sĩ kinh tế Đại Học Ngân Hàng Thành phố Hồ Chí Minh

27 Nguyen, D T., Ta, H T., & Nguyen, H T D (2018) What determines the profitability of Vietnam commercial banks International Business Research, 11(2), 231-245

28 DUONG, T T N., & NGUYEN, H Q (2021) Bank Restructuring and Financial Performance: A Case Study of Commercial Banks in Vietnam The

Journal of Asian Finance, Economics and Business, 8(10), 327-339

29 Luu, H N., Nguyen, L Q T., & Vu, Q H (2019) Income diversification and financial performance of commercial banks in Vietnam: do experience and ownership structure matter? Review of Behavioral Finance

30 Hoang, C., Nguyen, L., Tran, M., & Hoang, T (2020) The impact of income diversification on liquidity creation and financial performance of Vietnamese Commercial Banks Accounting, 6(4), 553-568

APPENDIX Appendix 1: Summary of Descriptive statistics

Appendix 2: Correlation between ROA and independent variables gdp 313 0608068 0103052 0291 0707579 cpi 313 0583466 0479276 006312 1867773 liq 313 6387669 1283139 2508404 8937174 loan 313 5473468 1244773 1910427 7880604 npl 313 0221875 0175436 0002 214 cir 313 4837603 1289911 2135882 8706558 llr 313 0105874 0086222 -.0101404 0540799 fata 313 0129649 0119722 0007003 0601816 size 313 8.012878 4940276 6.915157 9.180896 ea 313 093511 0405302 0293141 2553888 nim 313 0287164 0129721 0010243 088359 roe 313 0895443 0646337 0006826 2682345 roa 313 0079426 0063235 0001007 0472891

Variable Obs Mean Std Dev Min Max gdp -0.1054 -0.1157 -0.0335 -0.0226 -0.0331 0.1042 -0.0636 0.0006 0.0141 -0.0860 1.0000 cpi 0.2747 0.2554 -0.2554 0.0775 -0.1386 -0.2519 0.2166 -0.3329 -0.5492 1.0000 liq -0.2543 -0.2794 0.3425 0.1448 0.0501 0.2318 -0.0953 0.5912 1.0000 loan 0.1164 -0.0676 0.2909 0.2054 0.0929 -0.0792 -0.1185 1.0000 npl -0.0811 0.1599 -0.2113 0.0744 0.0546 0.0469 1.0000 cir -0.6287 -0.1244 -0.2684 0.1682 -0.3258 1.0000 llr 0.1552 0.1163 0.1696 -0.0981 1.0000 fata 0.0000 0.4943 -0.3636 1.0000 size 0.0983 -0.6583 1.0000 ea 0.3290 1.0000 roa 1.0000 roa ea size fata llr cir npl loan liq cpi gdp

Appendix 3: Correlation between ROE and independent variables

Appendix 4: Correlation between NIM and independent variables gdp -0.0561 -0.1157 -0.0335 -0.0226 -0.0331 0.1042 -0.0636 0.0006 0.0141 -0.0860 1.0000 cpi 0.1597 0.2554 -0.2554 0.0775 -0.1386 -0.2519 0.2166 -0.3329 -0.5492 1.0000 liq -0.1184 -0.2794 0.3425 0.1448 0.0501 0.2318 -0.0953 0.5912 1.0000 loan 0.1853 -0.0676 0.2909 0.2054 0.0929 -0.0792 -0.1185 1.0000 npl -0.1948 0.1599 -0.2113 0.0744 0.0546 0.0469 1.0000 cir -0.6130 -0.1244 -0.2684 0.1682 -0.3258 1.0000 llr 0.1083 0.1163 0.1696 -0.0981 1.0000 fata -0.2419 0.4943 -0.3636 1.0000 size 0.4592 -0.6583 1.0000 ea -0.1650 1.0000 roe 1.0000 roe ea size fata llr cir npl loan liq cpi gdp gdp -0.1110 -0.1157 -0.0335 -0.0226 -0.0331 0.1042 -0.0636 0.0006 0.0141 -0.0860 1.0000 cpi 0.2271 0.2554 -0.2554 0.0775 -0.1386 -0.2519 0.2166 -0.3329 -0.5492 1.0000 liq -0.0928 -0.2794 0.3425 0.1448 0.0501 0.2318 -0.0953 0.5912 1.0000 loan 0.2628 -0.0676 0.2909 0.2054 0.0929 -0.0792 -0.1185 1.0000 npl 0.0011 0.1599 -0.2113 0.0744 0.0546 0.0469 1.0000 cir -0.4006 -0.1244 -0.2684 0.1682 -0.3258 1.0000 llr 0.5276 0.1163 0.1696 -0.0981 1.0000 fata 0.2136 0.4943 -0.3636 1.0000 size -0.0497 -0.6583 1.0000 ea 0.4209 1.0000 nim 1.0000 nim ea size fata llr cir npl loan liq cpi gdp

Appendix 6: Estimated Pooled OLS of ROA

Mean VIF 1.86 gdp 1.05 0.952162 npl 1.11 0.897606 llr 1.29 0.774240 fata 1.70 0.587131 cpi 1.70 0.586883 cir 1.78 0.562146 loan 1.79 0.558489 liq 2.37 0.421452 size 2.83 0.352970 ea 2.92 0.341908 Variable VIF 1/VIF

_cons -.0182232 0081756 -2.23 0.027 -.0343116 -.0021347 gdp 0029101 0246947 0.12 0.906 -.0456854 0515057 cpi 0135415 0067632 2.00 0.046 0002324 0268505 liq -.008115 002981 -2.72 0.007 -.0139813 -.0022487 loan 0088211 0026694 3.30 0.001 003568 0140741 npl -.0290569 0149401 -1.94 0.053 -.0584569 000343 cir -.0207717 0025676 -8.09 0.000 -.0258244 -.015719 llr -.0622392 032731 -1.90 0.058 -.1266489 0021705 fata -.0402341 0270692 -1.49 0.138 -.0935022 0130339 size 0038103 0008461 4.50 0.000 0021454 0054752 ea 0737536 0104781 7.04 0.000 0531342 094373 roa Coef Std Err t P>|t| [95% Conf Interval]

Total 012475835 312 000039987 Root MSE = 00439 Adj R-squared = 0.5189 Residual 005810251 302 000019239 R-squared = 0.5343 Model 006665584 10 000666558 Prob > F = 0.0000 F( 10, 302) = 34.65 Source SS df MS Number of obs = 313

Appendix 7: Estimated FEM of ROA

F test that all u_i=0: F(30, 272) = 6.84 Prob > F = 0.0000 rho 5569362 (fraction of variance due to u_i) sigma_e 00348925 sigma_u 00391202

The fixed-effects regression analysis reveals significant findings regarding various financial indicators The coefficient for GDP shows a positive relationship, but it is not statistically significant (p = 0.308) In contrast, the Consumer Price Index (CPI) exhibits a significant positive effect (p = 0.000), while liquidity (liq) has a negative and significant impact (p = 0.005) Loans demonstrate a positive and significant relationship (p = 0.001), whereas non-performing loans (NPL) show no significant effect (p = 0.721) The capital adequacy ratio (CIR) and loan loss reserve (LLR) both have significant negative coefficients (p = 0.000), indicating potential risks Additionally, firm size and earnings assets (EA) present significant positive correlations (p = 0.000) The overall model is robust, with an F-statistic of 39.16 (p = 0.0000) and an R-squared value of 0.5901, indicating a strong explanatory power The analysis includes 313 observations across 31 banks, highlighting the importance of these financial metrics in understanding bank performance.

Appendix 8: Estimated REM of ROA rho 2717327 (fraction of variance due to u_i) sigma_e 00348925 sigma_u 00213136

The analysis reveals significant coefficients in a random-effects GLS regression, highlighting key relationships among various financial metrics Notably, the variable 'liquidity' (liq) shows a negative impact with a coefficient of -0.0098 (p = 0.001), while 'loan' (loan) demonstrates a positive effect at 0.0126 (p < 0.000) The 'capital adequacy ratio' (cir) is also significant with a coefficient of -0.0196 (p < 0.000) Furthermore, 'return on assets' (roa) and 'equity' (ea) exhibit strong positive associations, with coefficients of 0.0745 (p < 0.000) and 0.0056 (p < 0.000), respectively The overall model fit is robust, indicated by a Wald chi-squared statistic of 378.25 and an R-squared value of 0.5756, suggesting a considerable explanatory power for the variables examined across 31 bank groups and 313 observations.

Appendix 9: Estimated Preuch – Pagan of ROA

Apppendix 10: Estimated Hausman of ROA

Estimated results: roa[bank,t] = Xb + u[bank] + e[bank,t]

Breusch and Pagan Lagrangian multiplier test for random effects

Test: Ho: difference in coefficients not systematic

The analysis reveals that the variable B exhibits inconsistency under the alternative hypothesis (Ha) but demonstrates efficiency under the null hypothesis (Ho) In contrast, the results from the xtreg model show that the variable b is consistent under both Ho and Ha The coefficients for GDP, CPI, liquidity, loans, non-performing loans (NPL), capital adequacy ratio (CAR), loan loss reserves (LLR), financial assistance (FATA), firm size, and earnings announcements (EA) are presented, highlighting their respective impacts and standard errors Specifically, GDP shows a positive coefficient of 0.0207, while CPI has a coefficient of 0.0240 Conversely, liquidity and non-performing loans exhibit negative coefficients, indicating potential challenges in these areas The findings underscore the importance of these variables in financial performance assessment, providing valuable insights for stakeholders.

Apppendix 11: Estimated of model 1 (Pooled OLS, FEM and REM)

Apppendix 12: Modified Wald Test for ROA

Apppendix 13: Wooldridge test - ROA autocorrelation test

H0: sigma(i)^2 = sigma^2 for all i in fixed effect regression model

Modified Wald test for groupwise heteroskedasticity

Wooldridge test for autocorrelation in panel data

Apppendix 14: Estimated the FGLS of ROA

_cons -.0113145 0072902 -1.55 0.121 -.025603 002974 gdp 0061973 0127711 0.49 0.627 -.0188335 0312281 cpi 0090781 0041525 2.19 0.029 0009393 0172169 liq -.0046593 0023049 -2.02 0.043 -.0091769 -.0001417 loan 0056046 0024819 2.26 0.024 0007402 010469 npl 0016555 0088051 0.19 0.851 -.0156021 0189132 cir -.0216565 001929 -11.23 0.000 -.0254373 -.0178756 llr -.1843407 0254394 -7.25 0.000 -.2342011 -.1344803 fata -.0770753 0299568 -2.57 0.010 -.1357895 -.018361 size 0032147 0008306 3.87 0.000 0015866 0048427 ea 0623148 0088803 7.02 0.000 0449097 0797199 roa Coef Std Err z P>|z| [95% Conf Interval]

Prob > chi2 = 0.0000 Wald chi2(10) = 343.36 max = 11 avg = 10.09677 Estimated coefficients = 11 Obs per group: min = 4 Estimated autocorrelations = 1 Number of groups = 31 Estimated covariances = 31 Number of obs = 313

Correlation: common AR(1) coefficient for all panels (0.6077)

Cross-sectional time-series FGLS regression

Apppendix 15: S-GMM regression results of Model 1

Difference (null H = exogenous): chi2(8) = 11.62 Prob > chi2 = 0.169 Hansen test excluding group: chi2(3) = 1.56 Prob > chi2 = 0.667 iv(L4.ea L3.fata L4.llr L2.cir L3.loan L3.liq L3.cpi L2.gdp)

Difference (null H = exogenous): chi2(2) = 5.75 Prob > chi2 = 0.056 Hansen test excluding group: chi2(9) = 7.44 Prob > chi2 = 0.592 GMM instruments for levels

Difference-in-Hansen tests of exogeneity of instrument subsets:

(Robust, but can be weakened by many instruments.)

Hansen test of overid restrictions: chi2(11) = 13.19 Prob > chi2 = 0.281 (Not robust, but not weakened by many instruments.)

Sargan test of overid restrictions: chi2(11) = 13.19 Prob > chi2 = 0.281 Arellano-Bond test for AR(2) in first differences: z = -1.08 Pr > z = 0.281 Arellano-Bond test for AR(1) in first differences: z = -1.82 Pr > z = 0.069 D.(L6.size L2.npl) collapsed

GMM-type (missing=0, separate instruments for each period unless collapsed) L4.ea L3.fata L4.llr L2.cir L3.loan L3.liq L3.cpi L2.gdp

GMM-type (missing=0, separate instruments for each period unless collapsed) D.(L4.ea L3.fata L4.llr L2.cir L3.loan L3.liq L3.cpi L2.gdp)

Instruments for first differences equation

Warning: Uncorrected two-step standard errors are unreliable.

_cons -.0853323 0168686 -5.06 0.000 -.1198326 -.0508321 gdp 0504434 0130937 3.85 0.001 0236637 077223 cpi 0542762 0181973 2.98 0.006 0170585 0914939 liq -.0154556 0048967 -3.16 0.004 -.0254704 -.0054408 loan 0182608 0055928 3.27 0.003 0068222 0296994 npl -.0414748 0770108 -0.54 0.594 -.1989795 11603 cir 0182995 0064811 2.82 0.008 0050443 0315548 llr 0197772 0698317 0.28 0.779 -.1230446 1625991 fata -.1944904 061667 -3.15 0.004 -.3206136 -.0683673 size 0083303 0019087 4.36 0.000 0044266 012234 ea 1309161 0299866 4.37 0.000 0695867 1922455

The dynamic panel-data estimation using a two-step system GMM reveals a coefficient of 0.4942864 for the return on assets (ROA), with a standard error of 0.1886786, resulting in a t-value of 2.62 and a p-value of 0.014, indicating statistical significance The F-statistic is 35.05 with a probability greater than F of 0.000, suggesting a strong model fit The analysis includes 181 observations across 30 groups, with a minimum of 0 and an average of 6.03 observations per group, utilizing 23 instruments over the time variable of years.

Appendix 16: Estimated Pooled OLS of ROE

_cons -.1419261 0833202 -1.70 0.090 -.3058879 0220356 gdp 0128246 251671 0.05 0.959 -.4824262 5080754 cpi 1611375 0689261 2.34 0.020 0255012 2967739 liq -.0945645 0303807 -3.11 0.002 -.1543492 -.0347799 loan 1098593 0272049 4.04 0.000 0563241 1633946 npl -.4081831 152259 -2.68 0.008 -.707806 -.1085602 cir -.223266 0261674 -8.53 0.000 -.2747596 -.1717725 llr -.653055 3335707 -1.96 0.051 -1.309472 0033621 fata -.3638417 2758693 -1.32 0.188 -.9067111 1790278 size 0438432 0086223 5.08 0.000 0268758 0608107 ea -.0112607 1067854 -0.11 0.916 -.2213983 1988769 roe Coef Std Err t P>|t| [95% Conf Interval]

Total 1.30338515 312 004177517 Root MSE = 0447 Adj R-squared = 0.5217 Residual 603465165 302 001998229 R-squared = 0.5370 Model 69991999 10 069991999 Prob > F = 0.0000 F( 10, 302) = 35.03 Source SS df MS Number of obs = 313

Appendix 17: Estimated FEM of ROE

The fixed-effects regression analysis, conducted with 313 observations across 31 banks, reveals significant insights into various financial indicators Notably, the coefficient for consumer price index (CPI) is 0.249, indicating a strong positive relationship with the dependent variable, supported by a t-value of 3.82 and a p-value of 0.000 Additionally, loans show a positive impact with a coefficient of 0.130 and a t-value of 4.22, also statistically significant Conversely, liquidity (liq) has a negative coefficient of -0.089, while the capital adequacy ratio (cir) displays a significant negative effect with a coefficient of -0.196 The results emphasize the importance of CPI and loans in influencing financial performance, while highlighting the detrimental effects of liquidity and capital ratios The overall model demonstrates a robust fit, with an R-squared value of 0.5294, indicating the explanatory power of the included variables.

Appendix 18: Estimated REM of ROE rho 2949676 (fraction of variance due to u_i) sigma_e 03343235 sigma_u 02162467

The analysis reveals significant findings from a random-effects GLS regression involving various economic indicators Notably, the loan variable shows a strong positive coefficient of 0.139, with a p-value of 0.000, indicating a robust relationship with the dependent variable Conversely, the liquidity (liq) variable has a negative coefficient of -0.100, significant at the 0.001 level The capital adequacy ratio (cir) also demonstrates a substantial negative impact with a coefficient of -0.205, while the loan loss reserve (llr) exhibits a significant negative coefficient of -1.481 Other variables such as GDP and non-performing loans (npl) did not show statistically significant results The overall model fit is indicated by a Wald chi-squared statistic of 320.94, suggesting strong explanatory power with an R-squared value of 0.5194 The analysis included 313 observations across 31 banking groups, providing a comprehensive overview of the relationships among the variables studied.

Appendix 19: Estimated Preuch – Pagan of ROE

Appendix 20: Estimated Hausman of ROE

Estimated results: roe[bank,t] = Xb + u[bank] + e[bank,t]

Breusch and Pagan Lagrangian multiplier test for random effects

Test: Ho: difference in coefficients not systematic

The analysis reveals that the variable B is inconsistent under the alternative hypothesis (Ha) but efficient under the null hypothesis (Ho), while the results obtained from the xtreg model indicate that b is consistent under both hypotheses The findings show that GDP has a coefficient of 0.2358, CPI 0.2493, and a negative liquidity effect of -0.0888 Additionally, loan growth is positively associated with a coefficient of 0.1302, whereas non-performing loans (NPL) show a slight positive effect of 0.0309 The capital adequacy ratio (CIR) demonstrates a negative impact of -0.1962, while the loan loss reserve (LLR) significantly declines by -1.8568 Furthermore, firm size has a positive coefficient of 0.0850, and the effect of earnings assets (EA) is slightly negative at -0.0146 These results underscore the varying impacts of financial variables in the regression analysis.

Appendix 21: Estimated of model 2 (Pooled OLS, FEM and REM)

Appendix 22: Modified Wald Test for ROE

Appendix 23: Wooldridge test - ROE autocorrelation test

H0: sigma(i)^2 = sigma^2 for all i in fixed effect regression model

Modified Wald test for groupwise heteroskedasticity

Wooldridge test for autocorrelation in panel data

Appendix 24: Estimated the FGLS of ROE

_cons -.1856108 0771405 -2.41 0.016 -.3368035 -.0344182 gdp 1035134 1573419 0.66 0.511 -.2048711 411898 cpi 1185462 0449177 2.64 0.008 0305092 2065832 liq -.0539521 0253376 -2.13 0.033 -.1036128 -.0042914 loan 0750785 0270511 2.78 0.006 0220594 1280976 npl -.0145784 1041055 -0.14 0.889 -.2186215 1894647 cir -.2210381 0198731 -11.12 0.000 -.2599888 -.1820875 llr -1.811316 2834354 -6.39 0.000 -2.366839 -1.255793 fata -.4077743 2817255 -1.45 0.148 -.9599463 1443976 size 0486516 0087231 5.58 0.000 0315546 0657486 ea -.0302866 0797624 -0.38 0.704 -.1866182 1260449 roe Coef Std Err z P>|z| [95% Conf Interval]

Prob > chi2 = 0.0000 Wald chi2(10) = 310.70 max = 11 avg = 10.09677 Estimated coefficients = 11 Obs per group: min = 4 Estimated autocorrelations = 1 Number of groups = 31 Estimated covariances = 31 Number of obs = 313

Correlation: common AR(1) coefficient for all panels (0.5761)

Cross-sectional time-series FGLS regression

Appendix 25: S-GMM regression results of Model 2

Difference (null H = exogenous): chi2(11) = 13.87 Prob > chi2 = 0.240

Hansen test excluding group: chi2(4) = 2.77 Prob > chi2 = 0.596 iv(L4.ea L2.fata L3.fata L2.llr L3.cir L3.loan L2.loan L2.liq L3.liq L3.cpi L3.gdp) Difference (null H = exogenous): chi2(15) = 16.65 Prob > chi2 = 0.340

Hansen test excluding group: chi2(0) = 0.00 Prob > chi2 = gmm(L6.size L.npl, collapse lag(1 ))

Difference (null H = exogenous): chi2(2) = 3.09 Prob > chi2 = 0.213

Hansen test excluding group: chi2(13) = 13.56 Prob > chi2 = 0.406

Difference-in-Hansen tests of exogeneity of instrument subsets:

(Robust, but can be weakened by many instruments.)

Hansen test of overid restrictions: chi2(15) = 16.65 Prob > chi2 = 0.340

(Not robust, but not weakened by many instruments.)

Sargan test of overid restrictions: chi2(15) = 13.67 Prob > chi2 = 0.550

Arellano-Bond test for AR(2) in first differences: z = -1.26 Pr > z = 0.207

Arellano-Bond test for AR(1) in first differences: z = -2.65 Pr > z = 0.008

GMM-type (missing=0, separate instruments for each period unless collapsed)

L4.ea L2.fata L3.fata L2.llr L3.cir L3.loan L2.loan L2.liq L3.liq L3.cpi

GMM-type (missing=0, separate instruments for each period unless collapsed)

D.(L4.ea L2.fata L3.fata L2.llr L3.cir L3.loan L2.loan L2.liq L3.liq

Instruments for first differences equation

Warning: Uncorrected two-step standard errors are unreliable.

_cons -.7070091 1290745 -5.48 0.000 -.9709962 -.4430221 gdp 7098222 1044962 6.79 0.000 4961036 9235409 cpi 1.107343 108646 10.19 0.000 8851371 1.329549 liq -.0289039 0453928 -0.64 0.529 -.1217426 0639347 loan 1741612 0510972 3.41 0.002 0696556 2786668 npl -1.615946 4550678 -3.55 0.001 -2.546664 -.6852276 cir 2068634 0488957 4.23 0.000 1068605 3068663 llr 562061 732885 0.77 0.449 -.9368571 2.060979 fata -2.135496 4811878 -4.44 0.000 -3.119636 -1.151357 size 0578476 0180573 3.20 0.003 0209162 094779 ea 8300599 204343 4.06 0.000 4121314 1.247988

L1 .5579845 1452458 3.84 0.001 2609236 8550454 roe roe Coef Std Err t P>|t| [95% Conf Interval]

Number of instruments = 27 Obs per group: min = 0

Time variable : year Number of groups = 30

Group variable: bank Number of obs = 181

Dynamic panel-data estimation, two-step system GMM

Appendix 26: Estimated Pooled OLS of NIM

_cons -.0079404 0157605 -0.50 0.615 -.0389547 0230738 gdp -.0398537 0476049 -0.84 0.403 -.133533 0538257 cpi 0769231 0130378 5.90 0.000 0512667 1025794 liq -.0108458 0057467 -1.89 0.060 -.0221543 0004628 loan 0360615 005146 7.01 0.000 025935 046188 npl -.0704149 0288006 -2.44 0.015 -.1270902 -.0137396 cir -.0077605 0049497 -1.57 0.118 -.0175007 0019798 llr 7329887 0630967 11.62 0.000 6088239 8571536 fata 0903806 0521822 1.73 0.084 -.0123061 1930673 size 0012334 001631 0.76 0.450 -.0019761 0044429 ea 0885758 020199 4.39 0.000 0488271 1283245 nim Coef Std Err t P>|t| [95% Conf Interval]

Total 052501995 312 000168276 Root MSE = 00846 Adj R-squared = 0.5751 Residual 021591844 302 000071496 R-squared = 0.5887 Model 030910151 10 003091015 Prob > F = 0.0000 F( 10, 302) = 43.23 Source SS df MS Number of obs = 313

Appendix 27: Estimated FEM of NIM

Tiêu đề	Factors Affecting Financial Performance of Commercial Banks in Vietnam
Tác giả	Kim Thi Que Chi
Người hướng dẫn	Dr. Lê Hà Diễm Chi
Trường học	Banking University of Ho Chi Minh City
Chuyên ngành	Finance - Banking
Thể loại	Graduation Thesis
Năm xuất bản	2021
Thành phố	Ho Chi Minh City

Định dạng
Số trang	155
Dung lượng	2,87 MB