Tóm tắt Luận án The profitability of Vietnamese commercial banks in the context of the world economic crisis

. Government needs a different approach to responding to potential economic crises in the future. 1.6 THESIS STRUCTURE The content of the thesis includes 5 chapters: Chapter 1: Introduction Chapter 2: Theoretical basis of the impact of the economic crisis on the profitability of commercial bank Chapter 3: Research methodology and data Chương 4: Research results and discussions Chương 5: Conclusion and policy suggestions for Vietnamese commercial banks to achieve reasonable profitability in the context of the world economic crisis. 12 CHAPTER 2 THEORETICAL BACKGROUND OF THE IMPACTS OF ECONOMIC CRISIS ON COMMERCIAL BANK'S PROFITABILITY 2.1 THEORETICAL BACKGROUND OF THE ECONOMIC CYCLE AND ECONOMIC CRISIS 2.1.1 Concept of economic cycle and economic crisis Samuelson and Nordhaus (2007) stated that the economic cycle is the volatility of real GDP in a three-phase order of recession, recovery and prosperity. In which, the recession phase is the decline 13 of real gross domestic product (GDP) for two or more consecutive quarters (the negative economic growth rate in two consecutive quarters). Gordon (1994) argued that the economic cycle consisted of periods of expansion that occurred almost simultaneously in many economic activities, and that followed by periods of crisis and periods of recovery integrated into the expansion phase of the next cycle. The US National Agency for Economic Research (NBER) considered that an economic slowdown is a decline in economic activity across the country, lasting months to more than a year. An economic slowdown can be related to a simultaneous decline in many economic indicators such as employment, investment, and corporate profits, and can be associated with inflation or deflation. A severe and long-term recession is called an economic crisis (NBER, 2010).. 2.1.2 Theories of economic cycles and economic crisis According to the Keynesian (1936) school, stemming from the existence of indelible uncertainty in a monetary economy, as a result, the economy experienced periods of rapid growth and recession, then there is the crisis. One of the implications of this view is that the economy does not regulate itself, and the government must act to stimulate demand when the economy weakens, and to cool it when it gets too hot. If there is not enough money flowing into the economy to support the refinancing process, the recession will prolong 14 and the result is economic crisis. When the economic crisis occurs, the tightening of government spending, and the fear of lending of banks, make the crisis more serious. Therefore, the government needs to boost spending to increase aggregate demand in times of crisis and to tighten spending during hot growth. For the financial market, in a period of crisis, the central bank needs to act quickly to create market liquidity, prevent cash hoarding, and clear capital flows in the economy. In addition, too high savings can negatively affect the economy because households reduce consumption, businesses do not invest to produce more goods due to fear of not having enough demand to consume goods they produce and are uncertain about the prospects of the economy. According to the post-Keynesian school, typically Minsky (1975), there was a time where the financial system changed from steady state to state of crisis (often called the Minsky moment). This theory argues that financial shocks and investor misbehavior are the cause of the financial crisis. Accordingly, the Minsky moment was when investors, financing their portfolios by over-borrowing, were forced to sell their best assets to repay the loan, leading to a sharp decline in both value and liquidity in the financial market. Minsky's view of the financial crisis is summarized in three phases as follows: - Stage 1: This is the period when the economy recovers and develops stably, following a shock or previous financial crisis. During this period, investors expressed excitement and optimistic assessment of 15 the economy. Investors' psychology is quite stable, tends to choose to invest in safe portfolios or assets. From their own capital or borrowing from a bank, they have stable and safe profits but no sudden changes and can afford to pay principal and interest to the bank. From such early successes, they started to invest in assets that could bring more returns. - Stage 2: Based on the successes from stage 1 investing, investors begin to expand their portfolios. A business sector can have many investors participating and one investor participating in many different sectors. Due to the rapidly increasing investment demand, investors began to increase their debt. Banks lend more leading to high credit growth, more profits but also increased risks. Credit capital poured more into different sectors, creating price bubbles of assets. Increased risks erode investors' profits. Since then, they can only hold on to paying interest to banks, while principal debt becomes more difficult and needs more time to pay. - Stage 3: The financial bubble really exploded, causing the economy to fall into crisis. Banks restrict lending, enhance lending terms, or stop lending due to perceived high risks. At that time, speculation on assets, especially high-risk assets such as securities, real estate was frozen. Investors who are unable to pay their banks, are financially depleted, have to sell off their assets at increasingly lower prices to repay the banks and try to withdraw from the investment portfolio. As a result, they run the risk of losing their capital and falling into 16 bankruptcy. Banks themselves also fell into liquidity, unable to meet the surging demand for loans of investors as well as depositors to increase cash withdrawals from banks. After that, the economy started a new cycle, returned to a stable state, prioritizing safety for phase 1. Investors limit using debt as financial leverage, prioritizing first choice to invest in areas of high safety, while credit growth of banks is also reduced in line with reduced demand for loans of investors. The Keynesian and the Post Keynesian (Minsky) schools play a central role, the foundation of the theory in the study of this topic. Economic crisis is cyclical, considered as a phase of economic cycle, it is called cyclical crisis. The economic crisis that occurs locally in a particular field is a specific crisis, such as financial crisis, credit crisis, business crisis... Approach to economic crisis as a constituent phase of the economic cycle is supported by the majority of well-known economists. Therefore, in the next content, the thesis will apply in combination these theories with profitability theories to analyze the relationship between economic crisis and the profitability of commercial banks. In the scope of the thesis, it is limited in the context of the financial crisis. 2.2 THEORETICAL BACKGROUND OF THE PROFITABILITY OF COMMERCIAL BANK 2.2.1 The views on performance and profitability of commercial 17 bank Commercial bank’s performance can be assessed by many methods and the profitability of commercial bank is one of the methods to evaluate performance of commercial bank (ECB, 2010; Yuanita, 2019). The ability to create profits is seen as a way to evaluate the performance of commercial bank (ECB, 2010). 2.2.2 The theories of commercial bank profitability Profitability of a bank is the indicator measuring the efficiency of banking operations. The studies of the bank's profitability or the bank's performance are basically based on two theories: market power theory (MP) and the theory of efficient structure (ES). Market power theory The theory of market power (MP - market power) has two main approaches: Structure-Conduct-Performance (SCP) theory, proposed by Chamberlin (1933) and Robinson (1933), and Relative market power (RMP) theory, proposed by Smirlock (1985). According to the SCP theory, the more centralized banks are, the more able to manipulate the market by imposing high interest rates on loans and low deposit rates due to lower levels of competition. According to the Relative market power (RMP) theory, banks with large market shares and differentiated products and services will be able to control 18 the market and achieve higher profits. The theory of efficient structure Demsetz (1973) was the first to study efficient structure theory (ES). This theory holds that the most efficient banks get both the profit and the higher share; banks increase profitability as an indirect result of improved bank governance, not the power of market benefits. Thus, it can be seen that the market power theory (MP - market power) states that the profitability of banks is a function of market factors, while efficient structure theory (ES) argues that bank performance is influenced by internal performance and governance decisions, i.e. internal factors. Accordingly, many researchers have relied on the above theory to introduce some useful variables to the model of bank profitability and most admit that the bank's profitability is one function according to both internal and external factors (Olweny & Shipho, 2011). 2.2.3 The indicators reflect the profitability of commercial bank Two basic ratios commonly used to evaluate profitability of commercial bank are ROA and ROE (Nguyen Minh Kieu, 2009). - Return on Assets (ROA). ROA is a ratio measuring the ability of commercial banks to manage and use financial resources to create 19 profits. ROA is calculated according to the formula: 𝑅𝑂𝐴 = 𝑝𝑟𝑜𝑓𝑖𝑡 𝑎𝑓𝑡𝑒𝑟 𝑡𝑎𝑥 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠 - Return on Equity (ROE). ROE is a ratio reflecting the efficiency of equity, that is, the bank's return from equity. ROE is calculated by the formula: 𝑅𝑂𝐸 = 𝑝𝑟𝑜𝑓𝑖𝑡 𝑎𝑓𝑡𝑒𝑟 𝑡𝑎𝑥 𝐸𝑞𝑢𝑖𝑡𝑦 In the scope of the thesis, the author limits the use of two criteria, ROA and ROE, to evaluate profitability of Vietnamese commercial banks. 2.3 THE IMPACT OF ECONOMIC CRISIS ON COMMERCIAL BANK'S PROFITABILITY 2.3.1 The economic crisis affects the macro economy 2.3.2 Economic crisis increases credit risk 2.3.3 Economic crisis increases liquidity risk 2.4 EMPIRICAL STUDIES 2.4.1 Factors affecting the profitability of commercial bank In the world In Vietnam 2.4.2 An empirical studies of the impact of the economic crisis on 20 the profitability of commercial bank In the world In Vietnam 2.5 DISCUSSION OF PREVIOUS STUDIES First, regarding research objectives, studies have proved costly in the close relationship between profitability of commercial banks and economic crisis in recent decades (Asian financial crisis in 1997 and the global financial crisis in 2008). These two crises, especially the global financial crisis in 2008, have severely affected the Vietnamese economy in general and businesses in particular. Although previous studies have proven the impact of financial crisis (positive or negative), there is no research that can explain why these impacts are so closely and scientifically based on the theory and specific evidence. According to the author's review, since these two crises, in Vietnam, there have been a number of studies on the profitability of commercial banks in the context of the world economic crisis, but with the different scope and approach. Therefore, this thesis is designed to study profitability of commercial banks in the context of world economic crisis, and solve the following problems: (i) identify factors affecting the profitability of commercial banks in the context of world economic crisis and recovery period, (ii) study the differences in commercial banks' profitability in the crisis and recovery period. (iii) study the relationship between micro and macro factors to the 21 profitability of commercial banks, how these factors affect the profitability of commercial banks in each period. At the same time, research will explain the mechanism of action based on the Keynesian view (1936). Second, In terms of the research methodology, previous studies used traditional estimation methods (also known as frequentist) such as Pooled OLS, FEM, REM, GMM. These methods allow studying the impact of micro and macro factors on the profitability of commercial banks, but there are still many disadvantages. Specifically, the scientific conclusions in the frequentist statistics are based on the data set, regardless of known information (Nguyen Ngoc Thach, 2019). In frequentist statistics, parameters are considered to be constant but unknown. But for time series data, these parameters will change, thus assuming constant parameters is no longer suitable. Therefore, more broadly, in Bayesian statistics, the parameters are assumed as random variables and follow a distribution law (van de Schoot & Depaoli, 2014; Bolstad & Curran, 2016). Bayesian conclusions based on a prior information combined with collected data sets should have higher accuracy. For frequentist statistics, a sufficiently large data set is required to draw conclusions. Whereas for Bayesian statistics, making conclusions regardless of the size of the data (Baldwin & Fellingham, 2013; Depaoli & van de Schoot, 2016; Doron & Gaudreau, 2014), overcomes the drawback of frequentist statistics. Recently, the Bayesian method has been interested in using by researchers around 22 the world as well as in Vietnam because of its high accuracy and predictability in economic issues. The research on the profitability of commercial banks in Vietnam using the Bayesian method has not been performed by any author. Therefore, to achieve the goal is to study the specific factors that affect the profitability of commercial banks in different stages of the economic cycle, and evaluate fluctuations of the profitability of commercial banks under the impact of macro shocks, the author uses the Bayesian method to achieve the research objectives and is a new approach compared to previous studies. Besides, with a sample size of 30 banks and a relatively short period of time in the context of economic crisis (2007 - 2011), the Bayesian regression method is appropriate. Therefore, in terms of the research methodology, the thesis ensures the novelty and does not overlap with the previous studies. Third, considering the scope of space and time, the thesis ensures the inconsistency with previous studies. Most of the researches on the profitability of commercial banks associated with the context of economic crisis in different countries. Specifically, Andries et al (2016) studied the relationship between the financial crisis and the profitability of Eastern and Western European commercial banks from 2004 to 2013 through time dummy variables (Dummy = 1 in crisis period 2009 - 2013). In Vietnam, there are a number of works assessing the impact of the financial crisis in 2008 to the profitability of commercial banks such as Le (2017), related to the topic, with a 23 sample of 40 banks from 2005 to 2015 and dummy variable (D = 1) refers to the crisis in 2008 and 2009. Or the study of Nguyen Anh Tu and Pham Tri Nghia (2018), with a sample of 27 banks from 2005 to 2017 using dummy variables (D = 1) to indicate the crisis in 2008 and 2009, ie is the bottom of the crisis. According to the author's review, up to now, no research in Vietnam has been done associated with both the crisis and the recovery period to fully assess the impact of the factors in each stage on the profitability of commercial banks. Therefore, the scope of research in addition to finding out the impact of the entire crisis cycle on commercial banks' profitability. The thesis also studies the impact of factors on profitability in the entire period (2007 to 2018) and divided into two specific phases corresponding to two phases of the world economic cycle - the crisis period ( 2007 to 2011) and the post-crisis period (2012 to 2018). Therefore, the research scope of the thesis includes the scope of space and time which does not overlap with previous studies. The above analysis proves that the thesis has different subjects, objectives, methods, and scope of the research compared to previous studies. In addition, the thesis also ensures the novelty, scientific and highly applicable, especially in the context that the Vietnamese economy in general and commercial banks in particular still have many potential risks, uncertainty, especially in the current world economic instability. CONCLUSION OF CHAPTER 2 24 In chapter 2, the author presents the theoretical background of the impact of economic crisis on the profitability of commercial banks. Besides, the author systematizes relevant domestic and foreign research projects to find scientific gaps in research. CHAPTER 3 RESEARCH METHOD AND DATA 3.1 RESEARCH DESIGN 3.2 RESEARCH METHOD 3.2.1 Bayesian method To analyze the profitability of Vietnamese commercial banks in the context of the world economic crisis, Bayesian method applied. From conditional probabilities: p(A|B) = p(A, B) p(B) We have Bayes' theorem: p(A|B) = p(B|A)p(A) p(B) Where: p(A|B): Posterior probability, it is necessary to find the hypothetical probability A given by the data collected 25 p(B|A): The likelihood of the data, probability of data collected under the correct hypothesis A (data collected) p(A): Prior probability: Probability of the hypothesis A that we believe has occurred (true) before the data is collected. p(B): Constant, probability of the data A, B are random vectors 3.2.2 The superiority of the Bayesian method over the frequentist method First, Bayesian analysis is a powerful analytical tool for statistical modeling, interpretation of results, and prediction of data. Second, the universality of the Bayesian approach can be viewed as a methodological advantage over the traditional approach.. Third, in Bayesian analysis, we can use prior information, either belief or empirical evidence, in the data model to obtain a more balanced outcome for a particular problem. Fourth, by using the knowledge of the entire posterior distribution of the model parameters, Bayesian inference is much more comprehensive and flexible than traditional inference.. Fifth, Bayesian method can be used to simulate many models, including complex functions with arbitrary precision.. 26 Sixth, Bayesian inference provides a simpler and more intuitive explanation of outcomes in terms of probabilities. Seventh, the Bayesian model satisfies the probability principle, assuming that information in a sample is fully represented by the probability function (Berger và Wolpert 1988). Finally, as mentioned briefly earlier, the estimated accuracy in Bayesian analysis is not limited by sample size - Bayesian simulation methods can provide an arbitrary level of accuracy and are not affected by limitations such as aucorrelation, endogenous, heteroskedasticity that the frequentist method encounters. 3.3 RESEARCH MODEL 3.3.1 Research process 3.3.2 Research model proposed From the overall theoretical background presented and inherited the research model of Sufian (2011), and the previous studies are reviewed in chapter 2, the author proposes research models in the thesis, as follows: Table 3.1. Proposed models in the thesis , Model 1 ROA = ∝0 + ∝1SIZE + ∝2LOAN + ∝3LLP + ∝4DEP + ∝5LIQUI 27 + ∝6INT + ∝7OPE + ∝8CAP + ∝9INF+ ∝10GGDP + ε1 Stage: 2007 - 2011 Model 2 ROE = β0 + β1SIZE + β2LOAN + β3LLP + β4DEP + β5LIQUI + β6INT + β 7OPE + β8CAP + β9INF+ β10GGDP + u1 Model 3 ROA = ∝0 + ∝1SIZE + ∝2LOAN + ∝3LLP + ∝4DEP + ∝5LIQUI + ∝6INT + ∝7OPE + ∝8CAP + ∝9INF+ ∝10GGDP + ε2 Stage : 2012 - 2018 Model 4 ROE = β0 + β1SIZE + β2LOAN + β3LLP + β4DEP + β5LIQUI + β6INT + β 7OPE + β8CAP + β9INF+ β10GGDP + u2 Model 5 ROA = ∝0 + ∝1SIZE + ∝2LOAN + ∝3LLP + ∝4DEP + ∝5LIQUI + ∝6INT + ∝7OPE + ∝8CAP + ∝9INF+ ∝10GGDP + ∝11DUMMY+ ε3 Stage : 2007 - 2018 Model 6 ROE = β0 + β1SIZE + β2LOAN + β3LLP + β4DEP + β5LIQUI + β6INT + β 7OPE + β8CAP + β9INF+ β10GGDP + β 11DUMMY + u3 28 Source: author’s synthesis Table 3.2. Measure the variables used in the model variable Formula for calculation Dependent variable ROA Profit after tax/total assets ROE Profit after tax/equity Independent variable SIZE Logarithm of bank total assets CAP Equity/total assets LOAN Total loan/total assets LLP Loan loss provision/total loans LIQUI Liquid assets/total assets DEP deposite/total assets INT Interest expense/total liabilities OPE Operating expense/total assets INFL Inflation GGDP GDP growth DUMMY Get 1 for the crisis period (2007 - 2011) and 0 for the post-crisis period Source: author’s synthesis 3.3.3 Research hypothesis 29 On the basis of systematizing the fundamental theories and empirical studies presented by the author in Chapter 2, the research hypotheses on profitability of Vietnamese commercial banks in the context of the economic crisis. gender, aggregated as follows: Table 3.3. Summary variables used in the research model Independent variable Notation Previous studies Expected Bank size SIZE Sufian (2011), Alexiou and Sofoklis (2009), Kosmidou et al. (2007), Nguyen Pham Nha Truc and Nguyen Pham Thien Thanh (2016). + Bank capital CAP Tran Viet Dung (2014), Alexiou and Sofoklis (2009), Petria et al. (2013), Rahman et al. (2015), Gyulai and + 30 Szucs (2017), Berger (1995). Loan loss provision LLP Iacobelli (2016), Sufian (2011), Anathasoglou et al. (2008), Alexiou and Sofoklis (2009), Tran Viet Dung (2014). - Total loans Sufian (2011), Le (2017), Zhang and Dong (2011), Rahman et al. (2015). + Liquid assets LIQUI Kohlscheen et al. (2018), Kosmidou et al. (2005), Ndoka et al. (2016), Bassey and Moses (2015), Tran Thi Thanh Nga (2018). + Deposite DEP Sufian (2011), Lim and Randhawa + 31 (2005), Zhang and Dong (2011). Interest expense INT Yao et al. (2018), Islam vand Nishiyama (2016), Dietrich and Wanzenried (2011). - Operating expense OPE Batten and Vo (2019), Sufian (2011), Kosmidou (2008), (2016) - Inflation INFLAT Rahman et al. (2015), Le (2017), Tran Viet Dung (2014). - GDP growth GGDP Kohlscheen et al.(2018), Athanasoglou et al. (2008), Trujillo- Ponce (2013), Nguyen Pham Nha Truc and Nguyen Pham Thien Thanh + 32 (2016), Le (2017), Tran Viet Dung (2014). World economic crisis DUMMY Chronopoulos et al. (2015), Lindblom and Willesson (2010), Kamarudin et al. (2016). + Notes: (-) negative correlation, (+) positive correlation Source: author’s synthesis 3.4 DATA 3.4.1 Research data description The thesis uses unbalanced panel data from 2007 to 2018. Data collected from financial statements of 30 Vietnamese commercial banks, macroeconomic data of the General Statistics Office, State Bank of Vietnam in the period from 2007 to 2018. 3.4.2 Model test In order to ensure that Bayes inference based on MCMC model simulation is reasonable, the study is based on the convergence diagnosis of MCMC chain. 33 CONCLUSION OF CHAPTER 3 In chapter 3, the author presented in detail the research methodology, research model, research data and model test. Chapter 4 presents the results of the study and discusses the results obtained. CHAPTER 4 RESULTS AND DISCUSSION 4.1 THE PROFITABILITY OF VIETNAMES COMMERCIAL BANKS DURING THE WORLD ECONOMIC CRISIS 4.1.1 Descriptive statistics and matrix correlation 4.1.2 Return on assets (ROA) Table 4.3. Summary of regression results Mean coefficient Std. dev MCSE SIZE 0,0028 0,0238 0,0002 LLP -0,2124 3,1773 0,0307 LOAN 0,0059 0,2537 0,0025 34 CAP 0,0986 0,3433 0,0974 DEP 0,0157 0,1555 0,0015 LIQUI 0,0026 0,2176 0,0021 INT -0,0114 0,9827 0,0021 OPE -0,3451 3,5132 0,0035 INFLAT -0,0136 0,4035 0,0040 GGDP 0,1753 1,8488 0,0184 _cons -0,0932 0,8183 0,0081 var 0,0399 0,0051 0,0000 Source: author’s calculation 4.1.3 Return on equity (ROE) Table 4.5. Summary of regression results Mean coefficient Std. dev MCSE SIZE 0,0453 0,0252 0,0002 LLP -4,1234 3,3891 0,0344 35 LOAN 0,1965 0,2747 0,0027 CAP 0,3312 0,3634 0,0036 DEP 0,1119 0,1653 0,0016 LIQUI 0,0684 0,2348 0,0023 INT 0,0287 1,0515