The purpose of this essay was to determine the primary elements that contribute to the level of liquidity risk posed by joint stock commercial banks in Vietnam.. The Pooled OLS, FEM, and
INTRODUCTION
Background of the study
Liquidity risk is a critical concern for banks, necessitating continuous measurement and analysis by management to make timely adjustments Additionally, it is essential for the State Bank of Vietnam (SBV) authorities to understand liquidity risk to implement policies that stabilize the liquidity of the entire banking system.
Vietnam, a developing economy, faces challenges in its banking sector due to incomplete macro-management policies and management levels compared to developed nations The influx of foreign banks has intensified competition, impacting the capital mobilization of domestic banks, particularly smaller institutions, and consequently increasing credit risk.
Numerous studies have been conducted on bank credit risk, both domestically and internationally, including notable works by Truong Quang Thong (2013), Dang Van Dan (2015), and Chung (2009) These studies typically analyze the credit risk of individual banks or groups of banks within a specific region over a defined time frame However, it is important to recognize that each bank and banking system possesses unique characteristics influenced by varying economic conditions, leading to differences in liquidity among the research subjects.
The Covid-19 pandemic, which has been causing acute respiratory infections since the end of
In 2019, COVID-19 was first identified in Wuhan, China, and has since spread to 215 countries, prompting governments worldwide to implement strict measures to curb infection rates, including halting travel and closing schools and businesses This has disrupted global trade, leading to decreased economic activity and rising unemployment, threatening financial stability in many nations The pandemic has particularly impacted supply chains, especially for imported goods from China, causing production slowdowns and shortages Service industries, including tourism and events, have been heavily affected Furthermore, the pandemic has significantly decreased credit demand due to reduced economic activity, increased delinquent debt as businesses and households face financial strain, and heightened the need for digital, non-cash transactions as consumer behavior shifts.
The state has adjusted its monetary policy to manage inflation, real estate, securities, and bonds, tightening credit and affecting commercial banks' lending activities The Federal Reserve's decision to raise interest rates consistently has compelled the state to follow suit to combat inflation and maintain the exchange rate Liquidity, defined as a bank's ability to fund asset growth and meet cash and collateral obligations, has become a critical concern, especially amid the pandemic's economic impact, increasing the liquidity risk for commercial banks and raising the likelihood of their failure This has prompted both investors and policymakers to be wary of the liquidity risks posed by these banks Recognizing the importance of liquidity risk, the author chose to research "The determinants of liquidity risk of commercial banks in Vietnam."
Research objectives
Conduct research into the elements that contribute to the liquidity risk of Vietnamese joint stock commercial banks, with the end goal of formulating policies to reduce this risk
- Identifying the various elements that contribute to the liquidity risk faced by commercial banks
- Determine how much of an impact various factors have on the liquidity risk posed by commercial banks in Vietnam
To mitigate liquidity risks, commercial banks in Vietnam should implement robust liquidity management policies that include maintaining adequate cash reserves, diversifying funding sources, and enhancing asset-liability matching Additionally, banks should adopt stress testing and scenario analysis to identify potential liquidity shortfalls and develop contingency funding plans Strengthening regulatory compliance and fostering transparent communication with stakeholders will further enhance banks' resilience against liquidity challenges By prioritizing these strategies, Vietnamese banks can effectively limit their exposure to liquidity risks while ensuring sustainable operations.
Research questions
What factors are the liquidity risks of Vietnamese commercial banks affected by?
How do these factors affect the liquidity risk of Vietnamese commercial banks?
How did the Covid 19 epidemic affect the liquidity of commercial banks?
What policies aim to limit liquidity risks during operations for commercial banks in Vietnam?
Research contribution
This research contributes to understanding the factors that impact the liquidity risk of commercial banks by examining the influence of internal elements and macroeconomic factors The author adapts the research model based on a solid theoretical foundation and relevant practical studies.
The study's findings provide valuable empirical insights that can help commercial banks choose effective strategies to mitigate liquidity risks, enhance liquidity safety, and improve operational efficiency.
Thesis structure
The thesis has five main chapters:
Chapter II: Literature review and hypothesis development Chapter III: Research methods
LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT
The theory of liquidity risk
2.1.1 Concept of liquidity and liquidity risk
Liquidity, as defined by Rudolf Duttweiler, is the ability to meet financial commitments when due, primarily tied to cash flows In the banking sector, liquidity refers to a bank's capacity to fulfill obligations such as deposit and loan payments swiftly Liquidity risk occurs when there are insufficient funds available, often due to the complexities of banks acting as financial intermediaries, as noted by Bonfim and Kim (2012) Banks typically rely on short-term deposits to fund longer-term loans, creating a maturity mismatch that heightens liquidity risk (Diamond & Dybvig, 1983) To manage this risk, banks can maintain a liquidity buffer to align asset and liability maturities, although holding excess liquid assets can negatively impact profitability due to opportunity costs Therefore, achieving liquidity safety in operational management remains a challenge for banks, despite the incentives to maintain liquid asset buffers, as highlighted in Bonfim and Kim's research.
Liquidity risk refers to the potential inability of a bank to meet its financial obligations as they come due, which can jeopardize both its profits and reserves (Comptroller of the Currency, 2001) This risk arises when a bank cannot maintain adequate funds to satisfy the needs of lenders and depositors at reasonable interest rates Defined by the SBP, liquidity risk encompasses the threat of loss due to the failure to meet obligations or fund asset growth It also highlights the importance of being able to liquidate assets quickly and at fair market value (Muranaga and Ohsawa, 2002) If financial institutions struggle to sell assets at reasonable prices, they risk depleting their cash reserves, especially during volatile sales conditions, leading to potential financial losses and reduced profits Various factors contribute to liquidity risk, including significant deposit withdrawals (Jeanne and Svensson, 2007; Kumar, 2008) and challenges stemming from long-term lending commitments, which can further strain a bank's liquidity (Kashyap et al., 2002).
When substantial obligations arise, banks are required to fulfill them completely Moreover, in times of severe liquidity crises, banks holding a significant amount of long-term loans may struggle to sell these loans effectively.
Liquidity risk, as defined by Basel (2008), refers to the inability of a financial institution to secure adequate capital to fulfill its obligations without disrupting daily operations or compromising its financial stability This risk can lead to significant losses, as banks may be compelled to acquire additional capital at elevated costs or liquidate assets at unfavorable prices when actual liquidity demands surpass expectations In situations where solvency is jeopardized, commercial banks often resort to sourcing loans at exorbitant rates, particularly in underdeveloped money markets.
The inability of a bank to liquidate its assets for capital rotation can result in a severe liquidity crisis, as it struggles to convert assets into cash and is unable to mobilize or borrow funds This situation ultimately hinders the bank's capacity to meet its financial obligations and commitments.
In the role of a payment intermediary within the economy, banks are highly sensitive to market fluctuations, which can expose them to increased credit risk compared to other financial institutions.
Liquidity risk in banking arises from two key factors: the maturity transformation of a bank's liabilities and assets, and the liquidity of its assets, which refers to how easily they can be sold without significant loss These elements are interdependent; banks with liquid assets can overlook maturity transformation, while those with shorter-maturing assets may require fewer liquid holdings Economic recessions exacerbate liquidity risk due to decreased resource generation and heightened depositor demand This situation can lead to a domino effect, potentially resulting in the collapse of individual banks or the financial sector as a whole While past research has focused on the liability side of banks' balance sheets, asset-side risks, such as delayed borrower payments or project cancellations, also pose significant threats The banking structure itself, along with macroeconomic variables and internal financing rules, contributes to liquidity risk Severe liquidity crises can trigger major financial turmoil, leading to bankruptcies and mass withdrawals Reports indicate that Pakistani banks faced increased liquidity strain due to high credit demand from public sector companies and delayed loan repayments, compounded by excessive government borrowing, making monetary control challenging for the State Bank of Pakistan.
The business outcomes of the bank are impacted by the risk of liquidity
When liquidity risk arises, a bank must quickly raise substantial capital to meet its payment obligations, often leading to the liquidation of assets at prices below market value This discrepancy can result in significant losses for the bank If such losses happen frequently, they can adversely affect the bank's overall financial performance, potentially leading to sustained losses.
The risk of the bank's liquidity leading to losses in trust from its clients
A reputable bank prioritizes exceptional service, product quality, and a strong public image When liquidity risk occurs, it can erode client trust, negatively impacting the bank's performance and hindering its ability to attract capital This loss of trust can create a downward spiral, further diminishing the bank's liquidity.
Liquidity risk from one bank can affect the entire banking system
Many financial institutions maintain funds in the accounts of other institutions to facilitate broader payment acceptance Liquidity risk arises when these institutions risk losing access to customer deposits, posing a threat to the entire financial system This risk can disrupt liquidity and negatively impact the economy.
Liquidity ratios and the funding gap method are two effective approaches for assessing a bank's liquidity (Vodová, 2011) These methods can be utilized in combination to provide a comprehensive evaluation of a bank's financial health.
This method assesses a bank's liquidity by analyzing liquidity ratios derived from balance sheet items, often compared to industry averages Commonly used indicators in this evaluation include various liquidity metrics.
The L1 index measures the proportion of a bank's total assets that are liquid, reflecting its liquidity status A high L1 index suggests strong liquidity but may also indicate an excess of unprofitable reserves Consequently, banks should evaluate their operational strategies and consider necessary adjustments to optimize asset management.
The L2 index indicates a bank's capacity to maintain solvency during unforeseen customer deposit withdrawals A L2 value of 100% or higher signifies that the bank is well-equipped to handle unexpected withdrawal requests from its clients.
The L3 ratio is a measure that indicates what proportion of total assets are comprised of loans The lower the ratio, the higher the bank's liquidity as compared to other institutions
The L4 index measures the short-term comparison of deposits to loans, with values above 100% indicating low bank liquidity supported by external capital sources, while values below 100% suggest self-financed loans However, Poorman and Blake (2005) argued that liquidity ratios alone are inadequate for assessing bank liquidity, citing the example of Southeast Bank of Miami in 1991, which had a high liquidity ratio yet still failed To address this issue, Saunders and Cornett (2006) recommended using the funding gap as a more effective measure of liquidity risk.
Determinants of liquidity risk of commercial banks and hypothesis development
Large banks often depend on government support as a last resort rather than actively maintaining high liquidity levels, a viewpoint supported by Vodová (2011) This aligns with the concept of "too big to fail," highlighted during the US government's interventions to rescue banks during the 2008 financial crisis Empirical studies by Bonfim and Kim (2011) and Vu Thi Hong (2015) further substantiate this claim, demonstrating that a bank's liquidity tends to decrease as its asset size increases.
Large-scale banks benefit from economies of scale, enabling them to effectively mobilize capital due to their extensive geographical presence and access to the interbank market Consequently, these larger institutions typically experience lower liquidity risk Empirical research by Dang Van Dan (2015) further supports this, indicating that a bank's asset size negatively influences its exposure to risk.
In 2017, an analysis of data from all 24 commercial banks in Vietnam revealed that several major banks, including the Vietnam Joint Stock Commercial Bank for Industry and Trade, exhibited unusually high funding gaps Notably, the Petrolimex Petroleum Joint Stock Commercial Bank was also identified among those with significant discrepancies in their funding.
2017 and Kien Long Commercial Joint Stock Commercial Bank in 2017 are examples of smaller banks with a significantly smaller asset size and a much smaller financing gap Vodová's (2011)
The author hypothesizes that LSIZE, an explanatory variable, will positively correlate with the bank's liquidity risk However, the study aims to demonstrate that LSIZE actually has a negative impact on liquidity risk.
Bank size is determined by the logarithm of total assets
The Loan-to-Deposit Ratio (LDR) serves as a vital tool for commercial banks to manage their liquidity by balancing stable capital and liquid assets, as noted by the ESRB (2013) Unlike other liquidity ratios that focus primarily on either assets or capital, LDR provides a more holistic view of a bank's liquidity, acknowledging the interdependent nature of capital and asset liquidity Scholars such as Bonfim and Kim (2012) and Moussa (2015) have utilized LDR to assess the liquidity risk faced by commercial banks, highlighting its significance in financial analysis.
A negative relationship exists between the ratio of loans to short-term mobilized capital and the liquidity of Vietnamese commercial banks, assuming other factors remain constant This finding aligns with predictions that when a bank's mobilized capital is largely short-term and it engages heavily in lending, its liquid assets will be less available, thereby decreasing liquidity Furthermore, if lending levels funded by short-term capital sources remain high, the bank's liquidity risk will increase To mitigate this risk, banks should maintain a reasonable Loan-to-Deposit Ratio (LDR) Moreover, banks with a higher proportion of long-term deposits and the same LDR will experience less liquidity pressure compared to those with high short-term deposits and substantial medium to long-term loans.
The study conducted in 2012 aligns with previous research, suggesting that LDR (Liquidity Coverage Ratio) positively influences liquidity risk The author expects the research findings to confirm this beneficial relationship.
Banks maintain a sufficient level of liquid assets to safeguard against unexpected withdrawals, although this practice incurs an opportunity cost, as extremely liquid assets typically yield little profit Consequently, a higher allocation to liquid assets reduces liquidity risk but also diminishes potential earnings While strong liquidity is indicated by having more cash on hand than customer deposits, it limits the bank's capacity for profitable investments A sudden surge in client withdrawals can significantly heighten a bank's liquidity risk, regardless of its liquid asset levels.
The primary objective of financial institutions is to maximize profits by effectively utilizing mobilized capital while minimizing liquid asset reserves Research conducted by Valla and Escorbiac (2006), Vodová (2011), and Bonfim and Kim (2011) indicates a negative correlation between banks' profitability ratios and their liquidity The author expects that the research findings will reveal a detrimental effect of Return on Equity (ROE) on liquidity risk.
Return on Assets (ROA) is a key financial ratio that assesses a company's profitability in relation to its total assets It serves as an important metric for management, analysts, and investors, reflecting how efficiently a business converts its assets into profits.
Profit after tax as a percentage of average assets is a key indicator of a company's financial health A higher Return on Assets (ROA) signifies effective balance sheet management and profit generation, while a lower ROA indicates potential for growth and improvement.
Return on assets (ROA) is a key metric that assesses a bank's profitability in relation to its asset value, highlighting how effectively the bank utilizes its resources to generate profits A higher ROA indicates greater capital efficiency, making it an essential indicator for bank administrators to monitor The research suggests that ROA may negatively influence liquidity risk, underscoring its importance in financial management.
An increase in nonperforming loans (NPLs) heightens the risk of liquidity disruptions, potentially leading to bankruptcy for companies Consequently, commercial banks face liquidity risks stemming from poor loan quality Research, including studies by Iqbal (2012) and Vong and Chan (2009), has established a positive correlation between NPLs and liquidity risk Additionally, studies such as those by Cai and Thakor (2008) and Vu Thi Hong (2012) in Vietnam indicate an inverse relationship between asset quality and liquidity risk.
According to Truong Quang Thong's risk management theory (2012), a bank's equity capital acts as its ultimate defense and buffer against potential risks The safety of a bank can be assessed through its equity ratio; a low equity ratio indicates high financial leverage and increased risk.
(2013) empirical research contradicts the author's hypothesis, and the findings he reports have no
Research indicates a positive correlation between a bank's equity and its liquidity risk Empirical studies by Vodová (2011) and Vu Thi Hong (2015) further support this theory, showing that a higher equity ratio enhances a bank's liquidity Consequently, the author of this study predicts an inverse relationship between the equity-to-total capital ratio (ETA) and liquidity risk.
The current state of liquidity risk and its management
Rising inflation in Vietnam significantly impacts bank liquidity, as consumers increasingly prefer investing in gold and foreign currency for better returns rather than keeping their money in banks This behavior leads to reduced deposits and increased cash withdrawals, ultimately diminishing banks' reserves and compromising their liquidity.
The Vietnamese monetary market remains underdeveloped, characterized by a flawed bank asset structure that struggles to meet liquidity demands This sector is overly dependent on market loan capital, suffers from a fragile financial foundation, and faces unreasonable debt burdens.
Macro factors such as exchange rates, interest rates, and inflation significantly influence the liquidity risk of commercial banks in Vietnam Additionally, geopolitical risks play a crucial role in affecting this liquidity risk In 2022, the sharp increase in the USD/VND exchange rate, coupled with rising interbank interest rates and elevated borrowing costs, has created substantial challenges for commercial banks in the latter half of the year, thereby considerably heightening their liquidity risk.
Vietnamese commercial banks are transitioning to a centrally managed liquidity risk model, moving away from the traditional branch-based approach This shift aligns with Basel's guidelines for effective liquidity risk management, emphasizing the importance of mobilizing capital growth Additionally, banks are prioritizing stringent management of system-wide capital resources to ensure financial stability and compliance.
Taming credit expansion enhances liquidity safety, reflecting a shift from rigid state bank regulations to the implementation of internal monitoring and reporting standards This evolution emphasizes the importance of establishing internal limits to ensure effective liquidity management.
Banks are implementing significant measures to reduce bad debt and recover lost funds while ensuring compliance with state regulations on ratios and liquidity safety indicators Effective risk management is essential for commercial banks, necessitating both preventative strategies and provisions to mitigate liquidity risks.
RESEARCH METHODS
Research model
This study aims to analyze the factors influencing bank liquidity risk, focusing specifically on the effect of the Loan-to-Deposit Ratio (LDR) The "Funding Gap" method proposed by Saunders and Cornett (2006) is utilized as a proxy for assessing liquidity risk, chosen for its relevance among various liquidity approaches Additionally, the author employs Chung's (2009) model to evaluate the impact of different factors on liquidity risk, providing a structured framework for the analysis.
𝐹𝐺𝐴𝑃 𝑖𝑡 : Bank funding gap (i) in year (t)
𝐿𝐷𝑅 𝑖𝑡 : Ratio of loans to total short-term deposits of bank (i) in year (t)
𝑁𝐿𝑃 𝑖𝑡 : Non-performing loans ratio of bank (i) in year (t)
𝐸𝑇𝐴 𝑖𝑡 : The ratio of equity capital to total capital of the bank (i) year (t)
𝐿𝑆𝐼𝑍𝐸 𝑖𝑡 : Logarithm of total assets of the bank (i) year (t)
𝑅𝑂𝐸 𝑖𝑡 : Bank's rate of return on equity (i) year (t)
𝑅𝑂𝐴 𝑖𝑡 : Bank's rate of return on asset (i) year (t)
𝐺𝐷𝑃 𝑡 : Gross domestic products in year (t)
𝐶𝑂𝑉𝐼𝐷 𝑡 : Covid 19 epidemic in the year (t) εit: The residual is not observable
Funding gap FGAP Loans - Depossits
Return on assets ROA Return/Assets
Loan-to-deposit LDR Loans/Depossits
Non-performing loan NLP Total non-performing loans/total loans
Equity to total capital ETA Equity/ total assets
Bank size LSIZE Logarithm of total assets
Return on equity ROE Return/ Equity
Research methods
This article explores the application of quantitative research methods and regression techniques, including Pooled Ordinary Least Squares (OLS), Fixed Effects Model (FEM), Random Effects Model (REM), and Generalized Least Squares (GLS) It emphasizes the importance of conducting tests to identify the most effective model while addressing potential deficiencies within the modeling process.
Panel data offers enhanced reliability in estimating model parameters due to its significant variability, which leads to more accurate results It effectively addresses individual heterogeneity issues, minimizes multicollinearity and bias, and captures the dynamics of relationships between explanatory and response variables over time (Hsiao, 2014) The author opts for a panel data-based estimation approach because of its advantages over cross-sectional and time series data Common regression models in panel data analysis include pooled least squares (OLS), fixed effects model (FEM), and random effects model (REM) While OLS is the simplest method, it overlooks assumptions related to multicollinearity, autocorrelation, and heteroskedasticity This research utilizes FEM and REM to address the limitations of OLS estimation, tailoring the choice of regression models to the study's objectives and characteristics.
The pooled OLS model, utilized for panel data without distinguishing between time and objects, functions similarly to a standard OLS model This approach allows for the analysis of accumulated data, and it is adequate to conduct tests typical of conventional OLS models, provided that the dataset meets the necessary criteria for OLS application, which includes having more than two fixed effects models (FEM) and random effects models (REM).
The model must adhere to the following conditions, as stated by Gujarati (2008), for the OLS estimator's parameters to be free of bias
Assumption of normal distribution: Error has a normal distribution 𝜺𝒊 = 𝑵 (, 𝜺)
Homoscedasticity assumption: The variance of random errors is constant: 𝟐 = 𝟐
Assumption of independent relationship between dependent variables: 𝒄𝒐𝒗 (𝑿𝒊, 𝑿𝒋) = 𝟎
The dependent variables must not be completely collinear with each other
Serial correlation: Errors are not correlated with each other 𝒄𝒐𝒗 (𝑼𝒊, 𝑼𝒋) = 𝟎
Assumption about independence between dependent variable and error: There is no correlation between x and u 𝒄𝒐𝒗 (𝑿𝒊, 𝑼𝒋) = 𝟎
Linearity assumption: The relationship between x and y is linear
The above assumptions will be tested using tools on Stata 13 for panel data
In order to address limitations of the overall OLS regression model, this research employs two distinct estimate approaches for panel data: FEM and REM (Greene, 2007)
A fixed effects model (FEM) maintains constant effects for each factor, similar to an ordinary least squares model that utilizes dummy variables for fixed cross units, time, or both This model is particularly suitable when there is a correlation between an individual's intercept and one or more independent variables However, the presence of numerous variables heightens the risk of multicollinearity, significantly reducing the data's degrees of freedom.
The random effects model (REM) accounts for variations in experimental conditions among units as a result of chance, addressing individual changes over time that impact research outcomes While using this model, it is crucial to consider autocorrelation, but it offers a more effective means of avoiding heteroskedasticity compared to other modeling approaches.
The essay is evaluated in a number of ways to determine which research model will best fit the data
✓ The F-test is used to choose whether the FEM fixed-effects model performs better than the pooled OLS model
H0: OLS model is better than FEM model
H1: The OLS model is not as good as the FEM model
If the value, P-value < 0.05 = > FEM model is better and vice versa
✓ Use the Breusch-Pagan test to choose whether the random effects model (REM) is better than the Pooled OLS model
H0: OLS model is better than REM model
H1: The OLS model is not as good as the REM model
If value, P-value < 0.05 = > REM model is better and vice versa
✓ Use Hausman test to select FEM fixed-effects model and REM random-effect model
H0: The REM model is better than the FEM model
H1: REM model is not as good as FEM model
If P-value > 0.05 => REM model is better and vice versa
Following the selection of an acceptable panel data regression model, the following Stata 13 commands will be used to conduct hypothesis testing:
To assess multicollinearity in a model, the variance inflation factor (VIF) is a widely utilized metric The VIFj value indicates the extent of variance inflation among the independent variables A VIFj value of less than 10 suggests that multicollinearity is not present among the independent variables, while values exceeding this threshold indicate a potential multicollinearity issue.
Heteroskedasticity can be analyzed using the xttest0 test with the Random Effects Model (REM) and the xttest3 test with the Fixed Effects Model (FEM), both of which operate under the same two test hypotheses to evaluate the variance of the study.
H0: There is no heteroscedasticity phenomenon
H1: There is a phenomenon of heteroskedasticity
If P-value > 5% accept H0: The regression model does not have heteroscedasticity
On the contrary, if P-value < 5%, the model has heteroscedasticity
✓ Autocorrelation: The model's autocorrelation can be examined with xtserial
H0: There is no autocorrelation phenomenon
H1: There is an autocorrelation phenomenon
If P-value > 5% accept H0: The regression model does not have autocorrelation
When P-value < 5% accepts H1, the regression model has autocorrelation
While REM and FEM estimating methods improve upon some limitations of OLS, empirical evidence shows that OLS still struggles with certain assumption violations that result in less accurate outcomes, particularly autocorrelation and heteroskedasticity The FGLS (Feasible Generalized Least Squares) regression model effectively addresses these issues.
The steps when implementing the research model include:
Step 1: Statistics that are descriptive of the model's variables, including the minimum value, the maximum value, the mean value, and the standard deviation of the variables
Step 2: Run the results of regression models: Pooled OLS, FEM, REM using STATA 13 software
- Use the coefficient Prob>F in the Pooled OLS model to choose between the Pooled OLS and FEM models
- Use Hausman test to choose between FEM and REM models
- Use Breusch-Pagan test to choose between Pooled OLS and REM models
Step 4: Perform tests to check the defects of the selection model such as: autocorrelation test, change variance test
Step 5: Fix model defects using FGLS regression method
Step 6: After regression model by FGLS method, the author will focus on analyzing research results, and at the same time compare research results with initial expectations If there is a difference between expectations and research results, the author will try to give appropriate explanations based on theory and reality.
Data
The author analyzed consolidated financial statements, annual reports, and internal documents from 25 Vietnamese commercial banks spanning 2005 to 2022 Data was sourced from securities websites and the Vietnam Joint Stock Commercial Bank's site, while macroeconomic indicators like inflation and economic growth rates were obtained from the General Statistics Office of Vietnam.
Between 2020 and 2022, Vietnam's market economy experienced significant fluctuations, particularly influenced by the Covid-19 pandemic, which affected the operations of Joint Stock Commercial Banks This period highlighted the broader economic impacts and specific liquidity challenges within the banking sector A comprehensive study covering 2005 to 2022 aims to accurately assess the credit risk faced by these banks during various volatile periods The research incorporates updated financial statements from all banks through 2022 to provide a clear picture of the current liquidity risk landscape However, as of August 2023, only 25 Vietnamese joint stock commercial banks provided sufficient data for analysis, resulting in a sample of 432 observations categorized as disproportionate panel data, ensuring a balanced representation of small, medium, and large banks.
EMPIRICAL RESULT
Data Descriptive Statistics
Variable Obs Mean Std.Dev Min Max
The funding gap (FGAP) is a critical indicator of a bank's liquidity risk, with a smaller value suggesting lower risk Statistical analysis reveals an average funding gap of -10.75215, with values ranging from a minimum of -211.5681 to a maximum of 121.511, and a standard deviation of 31.86235 This data indicates significant variability in funding gap values among banks, influenced not only by inter-bank differences but also by each bank's fluctuations over time Consequently, banks exhibit a funding gap that consistently varies from low to high throughout the period from 2005 onward.
Return on Equity (ROE) ranges from a minimum of -0.8200214 to a maximum of 0.4509542, with a mean of 0.1267611 and a standard deviation of 0.0931898 Analyzing the merged data reveals that variations in business performance are primarily driven by the individual capabilities of banks rather than overarching economic conditions.
Vietnam's economic growth, measured by GDP, exhibits stability with a minimum of 0.839 and a maximum of 9.513, closely aligning with an average of 3.825 Additionally, the low average volatility of 2.285 underscores the consistency of the country's economic performance.
The logarithmic variable of total asset size (LSIZE) reveals that despite significant differences in asset sizes among banks at any given time and substantial growth in their assets over time, the logarithmic transformation narrows the range of values The smallest observed value is -0.8390485, while the largest is 3.326461, resulting in a mean of 1.860019 and a standard deviation of 0.6911043 Consequently, the variation among observations relative to the mean is relatively low.
Non-performing loan ratio (NPL): the minimum NPL ratio is 0.0001832, the maximum value is 0.2790931, the mean and standard deviation are 0.0198993 and 0.0218795, respectively
Loan-to-deposit ratio (LDR): The minimum LDR ratio is 0.2349639, the maximum value is 2.507711, the mean and standard deviation are 0.905188 and 0.2440103 respectively
The Equity to Total Capital (ETA) ratio ranges from a minimum of 0.0267 to a maximum of 0.7121, with a mean of 0.1094 and a standard deviation of 0.0770 Newly established banks often experience limited access to mobilized capital, resulting in a higher ETA ratio.
Return on Assets (ROA) varies significantly among banks, with a minimum value of -0.551175 and a maximum of 0.0797726 The mean ROA is 0.126156, accompanied by a standard deviation of 0.0110021, indicating a notable disparity in how effectively banks utilize their assets.
Between 2005 and 2022, Vietnam experienced significant fluctuations in its inflation rate, with a minimum of 0.0019 and a maximum of 0.199 The average inflation rate during this period was 0.0666, accompanied by a high standard deviation of 0.5509, indicating considerable variability in inflation trends.
Testing the assumptions of the regression model
Correlation matrix between independent variables in the model
The author used Stata 13 software to calculate the correlation between variables at 5% level, giving the following results:
FGAP NPL ROE ETA ROA LDR LSIZE
The correlation matrix indicates that the independent and dependent variables identified for the regression model analyzing factors influencing liquidity risk show significant correlations with FGAP at the 5% level Specifically, the independent variables ROA, LDR, INF, and COVID exhibit a positive correlation with FGAP, aligning with initial predictions and existing scholarly research Conversely, the independent variables NPL, ROE, ETA, Lsize, and GDP demonstrate a negative correlation with FGAP.
Independent variables such as NPL, ROE, GDP are not statistically significant at 5% level This demonstrates that nonperforming loans, return on equity, and gross domestic product are not
20 acceptable for evaluation as potential factors affecting the liquidity risk of commercial banks This also aligns with the findings of numerous earlier research
Regression results according to Pooled OLS, FEM, REM
Pooled OLS model regression results
Source SS df MS Number of obsC2
FGAP Coef Std Err t P>|t| Beta
The regression analysis reveals that the variables ROE, ETA, ROA, LDR, LSIZE, and INF significantly influence the changes in FGAP The average coefficient of determination (R²) is 0.4425, indicating that these independent variables account for 44.25% of the variation in the dependent variable Additionally, the author employs multiple Fixed Effects Model (FEM) and Random Effects Model (REM) tests to identify the most suitable model for their analysis.
Regression results of fixed effects model (FEM)
The study implements the fixed-effects analysis model FEM with the collected panel data, and the results are as follows:
Fixed-effects (within) regression Number of obs = 432
Group variable: Bank Number of groups = 24
R-sq: within = 0.4305 Obs per group: min = 18 between = 0.3613 avg overall = 0.4530 max
FGAP Coef Std Err t P>|t| [95% Conf Interval]
The regression findings from the fixed effects method (FEM) indicate that the coefficients ETA, ROA, LDR, LSIZE, and INF significantly explain variations in FGAP In contrast, the variables ROE, NPL, and GDP do not show a statistically significant impact on FGAP Additionally, the Pooled OLS analysis model demonstrates an explanatory power that exceeds the FEM model by 43.05%.
Regression results of random effects model REM
The author implements a random-effects regression model with panel data of variables FGAP, NPL, ROE, ETA, ROA, LDR, LSIZE, GDP, INF, COVID with the following results:
Random-effects GLS regression Number of obsC2
Group variable: Bank Number of groups$
R-sq: within = 0.4296 Obs per group: min between = 0.3766 avg overall = 0.3960 max
FGAP Coef Std Err z P>|z| [95% Conf Interval]
The random effects regression analysis reveals that the coefficients ETA, ROA, LDR, LSIZE, INF, and COVID significantly influence the variation in FGAP, while ROE, NPL, and GDP do not show a statistically significant impact The explanatory power of the random effects model stands at 42.96%, which is lower than both the Pooled OLS and fixed effects models, although the differences between the FEM and REM results are minimal The author is tasked with conducting the model's defect test.
Examine the model for any defects
The author employs the Hausman test to identify the most suitable model for analysis, evaluating both the Fixed Effects Model (FEM) and the Random Effects Model (REM) as potential options The null hypothesis (H0) posits that there is no correlation between the explanatory variables and the random component of the model.
Covid 6.479452 5.827857 0.651596 0.3550133 b = consistent under Ho and Ha; obtained from xtreg
B = inconsistent under Ha, efficient under Ho obtained from xtreg
Test: Ho: difference in coefficients not systematic chi2(9) = (b-B)'[(V_b-V_B) ^ (-1) = 5.93
From the Hausman test results, H0 is rejected, meaning there exists a correlation between the explanatory variables and random components, so the FEM fixed effects model is appropriate
Test heteroscedasticity with the hypothesis H0: Homoskedasticity
Table 9: Heteroscedasticity test White's test for Ho: homoskedasticity against Ha: unrestricted heteroskedasticity chi2(53) = 194.38
The results show P-value = 0 |z| 95% Conf Interval
The coefficients in the results table are largely consistent with expectations, displaying the correct signs and achieving statistical significance at the 5% level Additionally, the signs of these coefficients align with the theoretical analysis discussed earlier.
Results and discussion
The analysis of liquidity risk factors reveals that a regression model utilizing the GLS approach identified seven statistically significant explanatory variables affecting FGAP at a 5% significance level: ETA, ROA, LDR, LSIZE, GDP, INF, and COVID Conversely, two additional variables showed no statistical significance, indicating a lack of correlation with FGAP.
The research project employed Pooled OLS, FEM, and REM models to analyze the factors influencing liquidity risk in Vietnamese commercial banks from 2005 to 2022, while the FGLS method was used to address issues of heteroskedasticity and autocorrelation in the FEM model The study identified key determinants of liquidity risk, including bank size, LDR ratio, ROA ratio, ETA, GDP, inflation, and the impact of COVID-19 Notably, there was insufficient evidence to conclude that NPL and ROE significantly affect liquidity risk The results indicated a negative relationship between liquidity risk and ROE, ETA, ROA, bank size, and GDP, while NPL, LDR, inflation, and COVID-19 were found to positively impact liquidity risk These findings align with the proposed research hypothesis and are consistent with previous studies in the field.
The liquidity risk is positively influenced by factors such as LDR, NPL, INF, and COVID, while it is negatively affected by variables including ETA, LSIZE, GDP, ROE, and ROA.
The FGAP ratio at the 5% significance level indicates that a bank's size negatively influences its liquidity risk This outcome aligns with the author's initial research hypothesis and corroborates the findings from previous studies by Aspachs et al (2003) and Lucchetta.
A study conducted by Vodová (2011) indicates that banks in the sample experienced a faster growth in total assets, including liquid assets, compared to credit growth Consequently, as these banks expand their operations, the FGAP ratio declines, thereby decreasing the likelihood of encountering liquidity risk.
At a significance level of 5%, the LDR ratio positively influences liquidity risk, supporting the author's research hypothesis These findings are consistent with previous studies conducted by Delechat et al (2014) and Sudirman.
A higher Loan-to-Deposit Ratio (LDR) can enhance profitability, yet it also raises the risk of cash shortages When the LDR exceeds optimal levels, it suggests that a bank is extending more loans than it is receiving in deposits, heightening the chances of encountering liquidity problems.
The Return on Assets (ROA) ratio negatively influences liquidity risk, with a significance level of 5% This aligns with the research findings of Vong and Chan (2009) and Iqbal (2012) A higher ROA indicates that a bank generates more returns on its assets, reducing the likelihood of cash shortages By effectively leveraging its assets for profit, a bank secures the financial resources needed to cover expenses and debts Furthermore, such banks tend to invest in a substantial amount of liquid assets, thereby minimizing their exposure to liquidity risk.
The return on equity (ROE) ratio in this study is deemed statistically insignificant due to inaccuracies in credit provision expenses, which ultimately affect the accuracy of the bank's profit calculations.
At a 5% significance level, the ratio of shareholders' equity to total assets (ETA) negatively influences the liquidity risk of businesses This finding aligns with the author's research and corroborates studies by Aspachs et al (2003), Bunda (2003), and Vodová (2011) A high capital structure indicated by ETA suggests that banks are more risk-averse, resulting in cautious lending practices that help mitigate liquidity issues and prevent excessive loan expansion.
The current year's GDP growth rate negatively impacts bank liquidity risk at a significance level of 5%, aligning with findings from previous studies by Mousa (2015), Saxegaard (2006), and Vodova (2013) This correlation suggests that a slowing economy hampers enterprise consumption of manufactured goods, resulting in reduced cash flow and difficulties in repaying existing bank loans, thereby exposing banks to increased liquidity risks associated with credit.
NPL: In this particular study, the non-performing loan ratio (NPL) is not statistically significant
At a 5% significance level, the FGAP ratio reveals that inflation positively impacts liquidity risk, aligning with the author's hypothesis and supporting findings from Vodova (2011) and Bonfim and Kim (2011) This suggests that inflation contributes to various challenges in business operations, potentially resulting in delayed loan repayments or even bankruptcies Consequently, this situation can elevate a bank's bad debt ratio, heightening the risk of capital losses and significantly diminishing the bank's liquidity.
COVID: Covid 19 has a big impact on the statistics, with a significance level of 5% When
The onset of Covid-19 significantly elevates liquidity risk for banks, as the pandemic disrupts the economy, causing businesses to halt operations and leading to economic stagnation This situation results in firms and customers struggling to meet their debt obligations, potentially facing bankruptcy and loss of repayment capacity Consequently, the likelihood of banks facing liquidity shortages increases during such crises.
DISCUSSSION AND CONCLUSION
Conclusion
The author has systematically established the theoretical foundation for liquidity risk theory and examined empirical studies at both domestic and international levels This research led to the proposal of a comprehensive model that identifies nine factors influencing liquidity risk in commercial banks Among these, six internal factors are highlighted: bank size (LSIZE), non-performing loan ratio (NPL), loan-to-deposit ratio (LDR), equity to total capital (ETA), return on assets (ROA), and return on equity (ROE) Additionally, two macroeconomic factors, economic growth (GDP) and inflation rate (INF), along with the control variable of Covid-19 (COVID), are considered in this analysis.
To evaluate the impact of various factors on the liquidity risk of Vietnamese commercial banks, the author employs quantitative research methods, specifically multivariable regression analysis using balanced panel data This approach facilitates an assessment of how these factors influence liquidity risk The study utilizes three distinct estimation methods: the Pooled OLS least squares model, the Fixed Effect Model (FEM), and the Random Effect Model (REM) for the panel data analysis Subsequently, Breusch-Pagan Lagrangian tests are conducted to differentiate between Pooled OLS and REM, along with a Hausman test to select between REM and FEM, ensuring the model accurately fits the research data Finally, the STATA statistical package is used to validate essential regression assumptions, leading to significant findings from the regression analysis.
Based on the Hausman and Breusch-Pagan Lagrangian tests, the case study author determined that the Fixed Effects Model (FEM) would yield the best results among the three models analyzed However, regression assumptions indicated the presence of autocorrelation and heteroskedasticity To address these issues and ensure accurate estimations, the author opted to use the Feasible Generalized Least Squares (FGLS) method for the model estimation.
The study's regression analysis reveals that liquidity risk in Vietnamese commercial banks is significantly influenced by factors such as bank size, LDR ratio, ROA ratio, ETA, GDP, inflation, and the impact of COVID-19 However, there is insufficient evidence to conclude that non-performing loans (NPL) and return on equity (ROE) affect liquidity risk The results indicate that ROE, ROA, LDR, GDP, inflation, and COVID-19 exert a similar influence on liquidity risk, while ETA, NPL, and bank size negatively impact the likelihood of cash shortages These findings align with the author's research hypothesis and corroborate the results of previous studies.
Recommendations for hedging against liquidity risks
Smaller banks face greater risks related to liquidity issues compared to larger institutions To ensure operational safety, these financial entities should implement policies that minimize high-risk assets and maintain sufficient highly liquid assets Moreover, it is crucial for these institutions to consistently monitor their asset utilization to align with their current operational status.
To foster economic growth in line with planned targets, the State Bank and macro-management agencies must implement effective policies Confidence in anticipated economic expansion allows banks to prepare for future lending without compromising liquidity by adjusting their capital structure Additionally, it is crucial to consider the banking system's response to new regulations during the formulation and execution of economic policies For example, in 2008, the State Bank of Vietnam utilized interest rate tools and mandatory reserves to adopt a restrictive monetary policy aimed at curbing inflation and stimulating the economy, which inadvertently created liquidity challenges for commercial banks.
Bank Capitalization and Equity Policy
Vietnam's commercial banks are predominantly small institutions facing challenges such as limited resources, inadequate liquidity, and inexperienced management To overcome these issues, it is crucial for these banks to expand their capital scale Prioritizing increases in bank equity and total assets is essential, but this must be coupled with a focus on each institution's unique strengths and weaknesses Effective and secure allocation and utilization of assets are vital for achieving the goal of increasing total assets.
In the banking sector, commercial banks must optimize capital utilization to align with their operational nature, focusing on the production and real economy During market volatility, diversifying investment portfolios by acquiring state bank bills and appropriate government bonds is crucial to maintaining stable business operations.
Strategy for Using Mobilized Capital
Banks must enhance their capital mobilization capabilities and diversify their sources by assessing their strengths and weaknesses compared to competitors Implementing effective capital mobilization policies is essential to reduce the risk of liquidity crises and maximize earnings To achieve this, banks should consistently mobilize adequate corporate capital to meet their operational requirements.
Every bank must establish a dedicated division responsible for the careful and systematic management of its diverse capital sources This division should ensure ongoing and thorough oversight of the departments tasked with capital mobilization and utilization, promoting effective financial practices within the institution.
29 capital, as well as the coordination of the operations of these various divisions so that there is neither an excess nor a shortage of liquidity
When the deposit mobilization department expects substantial deposits soon, it is crucial to promptly inform the liquidity management department This communication enables effective coordination of available capital, allowing for proactive planning to address potential surpluses or deficits in liquidity efficiently.
Banks must regularly review their strategies for building and maintaining relationships with capital owners while diversifying their capital sources to ensure a liquidity cushion during financial difficulties This diversification is crucial for effective liquidity management policies, as relying on a limited number of funding sources increases liquidity risk Therefore, the liquidity management department should continuously monitor and analyze various capital sources, weighing their benefits and drawbacks to make informed decisions Examples of potential funding sources to meet liquidity needs include:
Mature assets and those not yet at maturity that can be sold, along with short-term investment instruments easily convertible to cash, are key components of a liquid investment strategy.
- The issuing of certificates of deposit for extended periods of time helped to mobilize deposits
- Credit limits that other financial institutions have promised to make available to this financial institution
- A cap on discounts that is imposed by the central bank
- Cash in a foreign currency that was brought in from banks located in other countries
To quickly access funds, banks can utilize asset mortgaging and engage in repurchase agreements (repos) with other financial institutions A repo is a contract where the buyer acquires government bonds or other financial assets, agreeing to repurchase them from the seller at a predetermined price and date in the future.
To effectively address the significant issue of bad debt in banks, a comprehensive evaluation of loan quality and debt recovery potential is essential This involves utilizing specific statistics to assess the value of bad debts By doing so, banks can accurately identify delinquent accounts and develop appropriate and efficient solutions to mitigate the impact of outstanding debts.
Commercial banks must actively utilize their financial resources to sell collateral and collect debts to address bad debts swiftly They are also allowed to resell unrecoverable debts at market prices and offer security for debt trading organizations This practice enables banks to leverage external resources, thereby accelerating the management of bad debts from previous years.
Banks can utilize funds allocated for credit risk provisioning to offset capital losses from unmanageable bad debts, thereby swiftly enhancing capital available for loans This approach facilitates the repatriation of bad debt, ultimately benefiting the economy's production and operational efficiency.
Commercial banks must classify debt according to the criteria established by the state bank, enabling them to make informed credit risk provisions This classification is essential for effectively monitoring overdue and bad debts, which is crucial for efficient debt collection practices.
Banks are mandated by the State Bank to allocate funds for each category of bad loans, ensuring transparency in their provisioning practices They must not hide their provisioning status or manipulate levels to enhance financial statements or artificially inflate earnings, as this is crucial for maintaining their reputation and trust among shareholders and customers.
In accordance with the circular issued by the state bank, provisions for credit risk should be made in their whole and on a consistent basis once every three months.
Limitations of the study and future research directions
The thesis, while achieving remarkable successes, is inevitably bound by certain limitations that must be examined to enhance and advance its development.
To begin, the author is only able to collect data from 24 commercial banks in the period between
The research data set on liquidity risk conditions of commercial banks in Vietnam remains limited and small due to the restricted time available for research between 2005 and 2022 Consequently, the findings cannot be generalized to accurately represent the broader liquidity risk landscape in the country.
The study exclusively assesses the liquidity risk of commercial banks using the FGAP ratio, which may limit its ability to provide a comprehensive view of the overall liquidity risk landscape for these banks.
The author highlights the importance of examining both internal and macroeconomic factors that influence liquidity risks within the commercial banking system, aiming to provide recommendations for managing these risks effectively However, the analysis lacks a detailed exploration of the objective and subjective causes of liquidity risk in banks, which is essential for implementing preventive and mitigating strategies.
The author intends to extend this research model to include data from various banking groups, aiming to assess the similarities and differences in liquidity risk levels among domestic banks By utilizing comprehensive data from the entire commercial banking system, the study seeks to uncover the overarching characteristics of liquidity risk impacting the Vietnamese banking sector.
To thoroughly assess a bank's liquidity risk, it is essential to utilize a diverse range of approaches for a comprehensive evaluation The author plans to enhance the study by increasing the data volume and exploring alternative models to better analyze the liquidity risk faced by banks in the near future.
Diamond, D.W and Rajan, R.G (1998) ‘Liquidity risk, liquidity creation and financial fragility:
A theory of banking’, SSRN Electronic Journal [Preprint] doi:10.2139/ssrn.112473
Ayu Effendi, K and Malinda, S (2018) ‘Liquidity risk of Islamic banking in Islamic and non Islamic countries’, Proceedings of the 4th Sriwijaya Economics, Accounting, and Business
Cont, R., Kotlicki, A and Valderrama, L (2020) ‘Liquidity at risk: Joint stress testing of solvency and liquidity’, Journal of Banking & Finance, 118, p 105871 doi:10.1016/j.jbankfin.2020.105871
Bonfim, D and Kim, M (2012) ‘Liquidity risk in banking: Is there herding?’, SSRN Electronic Journal [Preprint] doi:10.2139/ssrn.2163547
Diamond, D and Rajan, R (1999) Liquidity risk, liquidity creation and financial fragility: A theory of banking [Preprint] doi:10.3386/w7430
Chollete, L., Naes, R and Skjeltorp, J.A (2007) ‘What captures liquidity risk? order based versus trade based liquidity measures’, SSRN Electronic Journal [Preprint] doi:10.2139/ssrn.967598
Vodová, P (2013) ‘Determinants of commercial banks’ liquidity in Hungary’, Acta academica karviniensia, 13(1), pp 180–188 doi:10.25142/aak.2013.016
T T Tran, T et al (2019) ‘The determinants of liquidity risk of commercial banks in Vietnam’, Banks and Bank Systems, 14(1), pp 94–110 doi:10.21511/bbs.14(1).2019.09
Agbada & Osuji (2013), “The Efficacy of Liquidity Management and Banking Performance in Nigeria”, International Review of Management and Business Reseach Vol.2 Issue.1 (March
Pavla Vodová (2011), “Liquidity of Czech Commercial Banks and its Determinants”,
International journal Mathematical Model anh Methods in Applied Sciences, pp.1060-1067
Giannoti, Gibilaro & Mattarocci (2010), “ Liquidity Risk exposure for Specialised and unspecialized Real Estate Banks: Envidence form the Italian market”, Journal of Propery
Trương Quang Thông (2012), “Quản trị ngân hàng thương mại”, Nhà xuất bản Kinh tế thành phố
Valla and Escorbiac (2006), “Bank liquity and finacial stability Banque de France financial stability review”, pp.89-104
Chung-Hua Shen (2009), “Bank Liquidity Risk and Performance”, Working paper
Dermine, J (1986), “Deposit rates, creadit rates and bank capital: The Klein_ Monti model revisited”, Journal of Banking & Finance, 10(1), 99-114
Bhati, S., De Zoysa, A and Jitaree, W (2019) ‘Factors affecting the liquidity of commercial banks in India: A longitudinal analysis’, Banks and Bank Systems, 14(4), pp 78–88 doi:10.21511/bbs.14(4).2019.08
Drobotya, Ya., Doroshenko, O and Yaremenko, A (2022) ‘Liquidity risk of a commercial bank’, Efektyvna ekonomika [Preprint], (2) doi:10.32702/2307-2105-2022.2.81
Wuryandani, G (2012) ‘The determinants of bank liquidity’, SSRN Electronic Journal [Preprint] doi:10.2139/ssrn.2242754
Shin, H.S (2019) Risk and liquidity [Preprint] doi:10.1093/oso/9780198847069.001.0001
Bello, N., Hasan, A and Saiti, B (2017) ‘The mitigation of liquidity risk in Islamic banking operations’, Banks and Bank Systems, 12(3), pp 154–165 doi:10.21511/bbs.12(3-1).2017.01
Buch, C and Goldberg, L (2014) International Banking and Liquidity Risk Transmission lessons from across countries [Preprint] doi:10.3386/w20286
Li, L (2016) ‘Liquidity risk’, Commercial Banking Risk Management, pp 103–119 doi:10.1057/978-1-137-59442-6_5.