MINISTRY OF EDUCATION AND TRAINING UNIVERSITY OF ECONOMICS HO CHI MINH CITY - - NGUYEN QUOC ANH IMPACTS OF CREDIT RISK ON VIETNAMESE COMMERCIAL BANKS’ BUSINESS PERFORMANCE SUMMARY OF ECONOMIC DOCTORATE THESIS Faculty: Finance - Banking Code: 62.34.02.01 Academic Instructor: Assoc Prof S.D Nguyễn Đăng Dờn Ho Chi Minh City – 2016 CHAPTER INTRODUCTION 1.1 REASONS FOR RESEARCH The empirical studies in Vietnam and other countries indicate that macroeconomic factors, such as inflation, and GDP growth significantly impact nonperforming loans (NPL) Specific factors of the banks were also tested, the previous research indicated that NPL ratio of the previous year and credit growth ratehave the strongest impact on the NPL ratio of banks Some studies show that there is a relationship between credit risk and business performance of commercial banks through profitability indicator ratios.NicolaePetria (2013), Hasan Ayaydin (2014) use the ROE (Profit after tax / Equity) as the dependent variable and study the impact of credit risk on business performance of commercial banks The results indicate that credit risk has negative impact on business performance of commercial banks Credit risks appears objective in the market economy, especially in international integration trend and financial crisis The consequences of credit risk are the decline in bank profits and the destabilization in the banking and economy system From the above reasons, there is a need to study the impact of credit risk on business performance of commercial banks in Vietnam Research gap While there are several studies on the factors affecting credit risk of commercial banks in Vietnam, there is no study on the effects of credit risk on business performance of commercial banks The authors propose that credit risk may affect business performanc ecommercial banks in Vietnam The authors examine factors affecting credit risk, such as macroeconomic factors and specific factors of commercial banks, in the period of 2005 to 2015 The findings also detect the factors affecting credit risk, indicating that NPL and credit risk management affect the business performance of commercial banks in Vietnam 1.2 RESEARCH PURPOSES Particularly, the study aims to provide insights into the factors affecting credit risk, the impact of credit risk on business performance of commercial banks in Vietnam, and suggest solutions to limit credit risk, in order to enhance business performance of commercial banks in Vietnam To be specified, this research: - Identify the factors affecting credit risk of commercial banks - The impact of credit risk on business performance of commercial banks - Measure factors influencing the level of credit risk, and the impact of credit risk on business performance of commercial banks in Vietnam - Suggest solutions to limit credit risk, and enhance business performance of commercial banks in Vietnam 1.3 RESEARCH SUBJECTS AND SCOPE OF THE STUDY - Research subject: the factors influencing the level of credit risk, and the impact of credit risk on business performance of commercial banks in Vietnam - The study scope: focuses on factors influencing credit risk and business performance of commercial banks in Vietnam The major data sources were collected from 26 commercial banks in Vietnam through Bankscope, macroeconomic data sources were collected from ADB indicators with coverage from 2005- 2015 1.4 RESEARCH METHODOLOGY Quantitative research methods: using The Pooled OLS, Fixed effects method (FEM) & Random effects model (REM) , and the GMM method to examine factors influencing credit risk; using Feasible Generalized Least Squares (FGLS) model to study the impact of credit risk on business performance of commercial banks in Vietnam Furthermore, the authors also use several research methods such as interpretation, synthesis, comparison and analysis to study the factors influencing credit risk of commercial banks; current state of credit risk and business performance of commercial banks in Vietnam 1.5 RESEARCH STRUCTURE The content of this research consists of chapters: Chapter 1: Introduction Chapter 2: Theoretical Foundations Chapter 3: Research Methodology Chapter 4: Research Results Chapter 5: Conclusions and suggested solutions CHAPTER THEORETICAL FOUNDATIONS 2.1 CREDIT RISK After examining multiple perspectives on credit risk, the authors define credit risk as risk that arise during credit activities of banks, when customers not pay their debts or repay debts on time for the banks This is the main risk in the banking business, so credit operations and credit risk will affect the profitability and effectiveness of banks Credit risk can lead to other risks that cause serious consequences and disrupt the balance and stability of banks 2.2 BUSINESS PERFORMANCE OF COMMERCIAL BANKS Commercial banks use resources such as labor, infrastructure, financial resources for core activities: taking deposits, lending and investments This is the basis for determining the level of efficiency and factors affecting the efficiency of commercial banks Similarly, in the study of banking activities, some authors use the production approach with a view to banks as production units (Benston, 1965; Ferrier et al, 1990; Shaffnit et al, 1997 ; Zenios et al, 1999); some authors use intermediate approach, which views banks as financial intermediaries (Sealey and Lindley, 1977; Maudos and Pastor, 2003; Casu et al, 2003); and some other authors use modern approach that banks play both roles (Frexias and Rochet, 1997; Denizer et al, 2000;.Athanassopoulos and Giokas, 2000) 2.3 IMPACT OF CREDIT RISK ON BUSINESS PERFORMANCE OF COMMERCIAL BANKS 2.3.1 Credit risk impacts profitability and risks of commercial banks When credit risks arise, NPL increases, which results in lower sales and leads to a loss Furthermore, when NPLs incur, the costs will increase significantly, for example: the interest payment costs, NPL management cost, provisioning costs and other credit-related expenses The increase in costs results in lower profit than the initial estimate NicolaePetria (2013) indicates credit risk negatively affects business performance of banks (as measured by ROE, ROA), which have a direct impact on and degrade business performance of banks (Hasan Ayaydin, 2014) 2.3.2 Credit risk lead to bank instability, thereby affect the business performance of commercial banks NPL results in losses on bank assets.When high level of NPL is not limited, it will lead to a series of serious effects Normal losses occur in the lending amount, increase in operating costs, decrease in profit, decrease in the value of assets, etc which lead to bank reputation loss 2.3.3 Credit risk affects macroeconomic factors Highlevels ofcredit risk may impose systemic risk in the banking system, which then damage general economic conditions of a country, the macroeconomic factors (Vania Andriani1, SudarsoKaderiWiryono, 2015) 2.4 RESEARCH OVERVIEW ABOUT IMPACT OF CREDIT RISK ON BUSINESS PERFORMANCE OF COMMERCIAL BANKS 2.4.1 The impact factors of credit risk Table 2.1: Summary of the impact factors of credit risk Author(s) Rajan and Dhal (2003) Research subject(s) Non performing loans Variables Model Findings Dependent variable: Panel data, Bank size has the NPL ratio FEM, REM negative impact on NPLs GDP ofcommercial Independent banks in India variables: loan (2003 -2008) growth, loan loss reserve, GDP growth rate, unemployment rate, interest rate growth rate has positive effect on NPLs In good business environment, NPLs decrease Berge andBoye (2007) Problem loans of Dependent variable: GMM Northern Europe problem loans banks Independent (1993 –2005) variables: GDP growth rate, unemployment, real interest rates, inflation Problem loans impact nominal interest rates and unemployment rates Salas andSaurina (2002) Macroeconomic and microeconomic variables impact NPLs of Spanish banks (19851997) Dependent variable: Panel problem loans data,FEM, REM NPL Independent ratio variables: GDP growth rate, bank size, efficiency, marginal income ratio, leverage ratio, market power Bank size has negative impact on credit risk GDP growth rates has positive impact on problem loans ZribiandBoujelbène (2011) Examine the macroeconomic and microeconomic variables that impact credit risk of ten commercial banks in Tunisia Dependent variable: Panel data, Ownership credit risk REM, FEM structure, profitability and Independent the variables: the macroeconomic ownership indicators (GDP structure, the growth rates, prudential inflation, regulation of exchange rates, capital, profits, and interest rates GDP, inflation, affect credit risk exchange rates, and interest rates Louzis et al (2012) The macroeconomic factors and bank’s variables that affect NPL in Greece bank system Dependent variable: Dynamic the NPL ratio Panel Data, GMM Independent variables: real GDP growth rate, unemployment rates, interest rates and public debt, ROE, liquidity ratio, noneffectiveness ratio, bank size Problem loans due to macroeconomic variables (real GDP growth, unemployment, interest rates and public debt) Messai Study factors Dependent variable: Panel data, affecting NPLs the NPL ratio FEM, REM of 85 banks in three countries Independent The GDP growth rate, and ROA has negative impactson NPLs; (2003-2009) Ahlem (2013) Selma (Italy, Greece variables: GDP, and Spain) unemployment, (2004-2008) interest rates, growth in outstanding loans, loan loss reserve MarijanaCurak, Sandra Examine factors PepurandKlimePoposki impacting NPLs (2013) of the banking system in Southeast Europe (2003-2010) Bucuret al, (2014) The influence of macroeconomic conditions on credit risk in Romanian (2008-2013) Tehuluet al (2014) Dependent variable: Panel data of the NPL ratio 69 banks in 10 countries, Independent GMM variable: GDP, unemployment rates, interest rates, growth in outstanding loans, ROA, inflation There is a negative relationship between bank size and NPL ratio Dependent variable: Multivariate credit scores regression, SPSS Independent variable: GDP, inflation, money supply, unemployment rate The growth rate of money supply and exchange rate have a negative relationship with credit risk The unemployment rate has a positive relationship with credit risk Examine the Dependent variable: Panel bank-specific loan loss reserve GLS determinants of credit risk in Independent variable: credit Ethiopian growth, bank size, commercial ownership, banks operating (2007 - 2011) inefficiency, bank liquidity, profitability HasnaChaibiandZiedFtiti Credit risk (2015) determinants: Evidence from a cross-country study (2005-2011) unemployment and interest rates has a positive impact on NPLs data, Credit growth and bank size have negative impact on credit risk Dependent variable: Dynamic loan loss reserve Panel Data Independent variable: inflation rate, GDP, interest rates, unemployment, exchange rate, efficiency, leverage, size, profitability, loan loss reserve Operating inefficiency and ownership have positive impact on credit risk All examined macroeconomic variables affect the NPL ratio Đào Thị Thanh Bình and Factorsaffecting Dependent variable: Panel data of Bank size has NPLs of 14 positive impact Đỗ VânAnh (2013) commercial banks in Vietnam (2008-2012) the NPL ratio commercial on NPLs ROE banks in has negative Independent Vietnam, impact on NPLs variable: bank size, FEM, REM ROE, GDP, inflation Đỗ Quỳnh Anh, Nguyễn Factors affecting Dependent variable: Panel data, Inflation and Đức Hùng (2013) NPLs of the NPL ratio REM, FEM, GDP growth commercial GMM have impact banks in Vietnam Independent onNPLs variable: GDP, (2005 -2011) inflation, credit NPLs growth, bank size affectfollowing year’s NPLs Bank size has positive relationships with NPLs Võ Thị Quý and Bùi Examine factors Ngọc Toản (2014) affecting credit risk of commercial banks (2009 – 2012) Nguyễn Thị Ngọc Diệp Define bank and Nguyễn Minh Kiều characteristics (2015), affecting credit risk in commercial banks in Vietnam (2010- 2013) Dependent variable: GMM loan loss reserve model, panel data Independent of 26 variable: credit commercial growth, bank size, banks GDP growth Credit risk, credit growth, GDP growth rate, the impact of negative loss ratio in the year on credit risk Dependent variable: Data panel loan loss reserve of 32 commercial Independent banks in variable: credit Vietnam growth, loan size, with and operational cost regression / income loans least squares ratio (OLS) Credit growth, loan size, and the operational cost / income loans ratio impact credit risk 2.4.2 The impact of credit risk on business performance of commercial banks Table 2.2: Summary of studies about the impact of credit risk on business performance Author(s) Research subjects Athanasolouet al (2006) The factors affecting the profitability of Greek banks (1985 – 2001) Hassan Sanchez (2007) and The factors determining the efficiency of commercial banks in Latin America Variable Model Findings Dependent variable: Panel data, The bank-specific factors, ROE, ROA, FEM, REM, such as: loan loss reserve, 3GLS Equity-to-Asset ratio, Independent variable: operating costs have an loan loss reserve, impact on bank profits Equity-to-Asset ratio, operating costs Dependent variable: DEA model bank efficiency Independent variables: capitalization profitability, level, loan Loan loss reserve has negative relationship with business performance Capitalization level, and profitability have positive relationship with business (1996-2003) Aremu Mukaila Ayanda (2013) loss reserve, labor, credit balance performance The effectiveness Dependent variable: ECM Model of the bank in ROE, ROA, NIM Nigeria Independent variable: (1980-2010) loan loss reserve, total Loans-to-Total Assets Ratio, Equityto-Total Asset Ratio, bank size, GDP Nicolae Petria Examine (2013) determinants of banks' profitability of EU 27 banking systems Loan loss reserve, total Loans-to-Total Assets Ratio, Equity-to-Total Asset Ratio have negative effect on business performance Bank size has positive effect on business performance Dependent variable: Panel data, Credit risk has negative ROE, ROA REM, FEM impact on business performance (ROE) of Independent variable: commercial banks size, credit risk, cost efficiency, liquidity, HHI, GDP, inflation (2004-2011) Hasan Ayaydin (2014) Zou et (2014) Factors affecting Dependent variable: Panel the capital and NIM, ROE, ROA, GMM profits of Turkish Independent variable: banks loan loss reserve, (2003-2011) capital ratio, foreign ownership, HHI, liquidity, inflation, GDP al Examine the relationship between credit risk and profitability of commercial banks in Europe (2007-2012) Alshatti (2015) Examine the impact of credit risk on financial efficiency of commercial banks in Jordan data, Loan loss reserve negatively impact bank performance, measured through ROE variable Dependent variable: OLS ROE and ROA Independent variable: the NPL ratio, CAR, bank size Dependent variable: Panel data ROA and ROE Independent variable: CAR, leverage ratio, NPL ratio, loan loss reserve Credit risk has no positive impact on the profitability of commercial banks NPL ratio has a significant impact on ROE and ROA Credit risk impacts business performance of commercial banks (2005-2013) Samuel (2015) The influence of Dependent variable: OLS credit risk on the ROA profitability of banks in Nigeria Independent variable : Lending rate has negative relationship with profitability NPL / credit balance, outstanding loans / total deposits Gizawet al The influence of Dependent variable: Panel data and NPL ratio, loan loss (2015) credit risk on the ROE, ROA multivariate business regression performance of Independent variable: analysis banks in CAR, NPL ratio, loan loss reserve Ethiopia reserve impact business performance of banks in Ethiopia (2003-2004) Kodithuwakku The influence of (2015) credit risk on business performance of commercial banks in Sri Lanka Dependent variable: Multivariate ROA regression, Eview Independent variable: reserve / total loans ratio; loan loss reserve / NPLs ratio; loan loss reserve / total asset ratio, NPLs / loans ratio Nguyễn Việt Analysis factors The inputs and DEA Hùng (2008) affecting business outputs cost factors performance of 32 commercial banks in Vietnam (2001– 2005) Loans and regulations affect the profitability of banks NPL ratio, Total loans/Total assets ratio, deposits / total loans ratio, total costs / total revenues ratio, income from interest rate / income from operations ratio have negative impact on the business performance of banks Market shares, equity to total assets ratio has positive impact on the business performance of banks Trịnh Quốc Trung (2013) Nguyễn Văn Sang Analysis factors affecting business performance of 39 commercial banks in Vietnam (2005-2013) Dependent variable: Tobit ROE, ROA regression model Independent variable: cost / revenue ratio, deposit / loan ratio, capital / total assets ratio, market share, loan / total assets, ratio of overdue debts and total residual debt The higher the NPL ratio, the lower the operating efficiency The higher theratio of loans / total assets,the higher operational efficiency Total operating expenses / revenue ratio negatively correlated with ROE; The higher theSelf-financing ratio, the lower the ROE Source: Compiled from relevant research SUMMARY OF CHAPTER In this chapter, the authors introduce the theoretical foundations, and measurement method of credit risk in commercial banks The authors also analyze the causes and effects of credit risk on general economy, business performance of commercial banks In general, previous researches in Vietnam and abroad on the factors influencing credit risk and the impact of credit risk on business performance were also reviewed by the authors It's also the foundation for research in the following chapters CHAPTER METHODOLOGY 3.1 RESEARCH METHODOLOGY Based on the theoretical foundation of credit risk and business performance of commercial banks, the authors select and identify key research issues, using quantitative research methods To be specific, the authors use multivariate regression models through Pooled regression model, Fixed effect, Random effect and use the GMM method to solve the endogenous regression on panel data In addition, the authors also use the methodological interpretation, synthesis, comparison and analysis method to achieve research objectives 3.2 FACTORS AFFECTING CREDIT RISK MODEL Recent studies on this issueuse dynamic tabular data (Dynamic Panel Data), for example, Cheng and Kwan (2000); Calderon and Chong (2001); Salas and Saurina(2002); Beck and Levine ( 2004), Santos-Paulino and Thirlwall (2004); Carstensen and Toubal (2004); Athanasoglou et al (2009); and Merkl and Stolz (2009).The authorsselect multivariate regression model which is consistent with previous studies, our model, which is based on the model of Hasna Chaibi and Zied Ftiti (2015), identifies factors affecting the commercial banks RRTD: NPLit = α+ γNPLi,t-1 + βjXi,t + vi + εi,t (1) Where: α: is the intercept NPLi,t-1: the NPL ratio of bank iin year t NPL ratio was used to measure the degree credit risk (Vania Andriani, Sudarso Kaderi Wiryono, 2015) γ : is the impact of negative loss ratio in the year t Xi,t : the vector of independent variables, including macroeconomic variables and bank specific variables Bank specific variables:ETAi,t , LEVi,t , SIZEi,t , EFFi,t , ROEi,t , NIIi,t , PLLi,t; macroeconomic variables: GGDPt , INRt , INFt , UNRt , EXRt βj: the impact of independent variables on NPL ratio vi: unobserved characteristic among banks εi,t: is the accumulation of the structure The lagged variable of the dependent variable - NPL – has a correlation with v Therefore, if we apply the smallest quadratic OLS methodology, an unbalanced and unstable estimation will be likely to occur The (1) regression equation will be stabilized if it is estimated by GMM (Generalized Method of Moments) introduced by Arellano and Bond (1991) * Bank internal variables: (1) Loan loss reserve (LLR ) = Loan loss reserve/Total loans Hypothesis 1: There is a positive correlation between Loan loss reserve and the NPL ratio (2) EFF- Operating inefficiency = Operating expenses/Operating income Hypothesis 2: There is a positive correlation between operating inefficiency and NPL ratio (3) Leverage (LEV) = Total liabilities/Total assets Hypothesis 3: There is a positive correlation between leverage and NPL ratio (4) Non-interest income (NII ) = Non-interest income/Total income Hypothesis 4: There is an inverse correlation between non-interest income and NPL ratio (5) Bank size (SIZE ) = Natural log of total assets Hypothesis 5: There is a positive correlation between the bank size and NPL ratio (6) Return on Equity (ROE) = Net income/Total equity Hypothesis 6: There is an inverse correlation between ROE and the NPL ratio * Macroeconomic variables (7) Inflation (INF) = Inflation rate Hypothesis 7: There is a positive correlation between the inflation rate and NPL ratio (8) GDP Growth (GGDP) = GDP Growth rate Hypothesis 8: There is an inverse correlation between GGDP and NPL ratio (9) Nominal interest rate (INR) = Real interest rate Hypothesis 9: There is a positive correlation between the real interest rate and NPL ratio (10) Unemployment rate (UNR) Hypothesis 9: There is a positive correlation between the unemployment rate and NPL ratio (11) Exchange rate (EXR) Hypothesis 11: There is an inverse correlation between the exchange rates and NPL ratio Table 3.1: Description of Model 1’s variables Variable Calculation Expectation Dependent variable Problem loan / Total loan measuring credit risk: NPL ratio (NPL) Independent variable Bank’s internal variables Loan loss reserve (LLR) EFFinefficiency Loan loss reserve/Total loans Operating Operating expenses/Operating income + + Leverage (LEV) Total liabilities/Total assets + Non-interest income (NII) Non-interest income/Total income - Bank size (SIZE) Natural log of total assets + Return on Equity (ROE) Net income/Total equity - Inflation (INF) Inflation rate + GDP Growth (GGDP) GDP Growth rate - Nominal interest rate (INR) Nominal interest rate + Unemployment rate (UNR) Unemployment rate + Exchange rate (EXR) VND/USD rate - Macroeconomic Variables Source: Compiled from relevant studies 3.3 IMPACT OF CREDIT RISK ON BUSINESS PERFORMANCE OF COMMERCIAL BANK MODEL Nicolae Petria (2013), Hasan Ayaydin (2014), Aremu Mukaila Ayanda (2013) study the factors influencing business performance of commercial banks, their studies conclude: the NPL ratio and loan loss reserve’s impact on business performance of commercial banks The authors use ROE and ROA as dependent variables; credit risk is represented by NPL ratio (NPL) ratio and provision for loan losses as PLL; other control variables were included in the model through vector X The authors use the multivariate regression model, used in the studies of Athanasolou et al (2006), Aremu Mukaila Ayanda (2013), Hasan Ayaydin (2014), Alshatti (2015) Our model is as followed: (ROEit, ROAit) = α+ β1NPLi,t + β2PLLi,t + βjXi,t + vi + εi,t (2) Where: α: the intercept β1 and β2: the impact of NPL and PLL on ROE, ROA Xi,t : variable vectors: bank internal variables includes: EFFi,t, LEVi,t, NIIi,t, SIZEi,t , and macroeconomic variables include: GGDPt , INRt , INFt , UNRt , EXRt βj: the impact of independent variable i on ROE, ROA vi : Unobserved characteristics among commercial banks εi,t : is the accumulation of the structure The authors use four models - Pooed OLS, Fixed Effects, Random Effect and FGLS - on panel data to estimate and examine the impact of credit risk on business performance of commercial banks Hypothesis 12: There is an inverse correlation between NPL, LLP and ROE, ROA Table 3.2: Description of Model 2’s variables Variable Calculation Return on Equity (ROE) Net income/Total equity Return On Assets (ROA) Net income/Total asset Expectation Non-performing loan ratio (NPL) Non-performing loan /Total loan - Loan loss reserve (LLR) Loan loss reserve /Total loan - Leverage (LEV) Total liabilities/Total asset - Non-interest income (NII ) Non-interest income/Total income + Bank size (SIZE) Natural log of total assets + Operation inefficiency (EFF) Operating expenses/Operating income - Macroeconomic Variables Inflation (INF) Inflation rate GDP Growth (GGDP) GDP Growth rate + Nominal interest rate (INR) Nominal interest rate - Unemployment rate (UNR) Unemployment rate - Exchange rate (EXR) VND/USD rate Source: Compiled from relevant researches +/- +/- 3.3 DATA SOURCES Internal bank data sources were obtained from Bank-scope and audited financial statements of 26 Vietnamese commercial banks with coverage from 2005 to 2015 The authors used major data sources from 26 commercial banks, whose total assets account for over 75% total assets of Vietnamese commercial banks Therefore, the data sources are representative of Vietnamese commercial banks Macroeconomic data was extracted from ADB Indicators with coverage from 2005 to 2015 CHAPTER 3: SUMMARY This chapter analyzes and decides which regression model to be appropriate for our research goal The research measures the factors affecting credit risk of commercial banks in Vietnam, using dynamic panel data; the dependent variable as NPL ratio representing credit risk The authors study the impact of credit risk on business performance Macroeconomic variables and bank internal variables inherent in the bank were analyzed and selected The hypotheses are presented in details in order to determine the impact direction of variables CHAPTER RESULTS AND DISCUSSION 4.1 MACROECONOMIC CONDITIONS IMPACT CREDIT RISK AND BUSINESS PERFORMANCE OF COMMERCIAL BANKS According to the empirical studies, the macroeconomic factors affect the credit risk and business performance of commercial banks In recent years, Vietnamese economy is unstable because of the global financial crisis The changes in GDP growth, inflation, interest rates, exchange rates will change the macroeconomic conditions The changes in monetary policy, and interest rates will affect the bank lending channel, NPLs, and impact business performance of commercial banks 4.2 CREDIT RISK AND BUSINESS PERFORMANCE OF COMMERCIAL BANKS IN VIETNAM 4.2.1 Credit risk Credit outstanding balance accounts for large proportion of bank portfolio: Credit activities still account for about 60-80% of the total assets of commercial banks, thus the income from credit activities accounts for a large proportion of the total income of banks Credit average growth rate reaches 19.15% during 2008 – 2015 NPL ratio and loan loss reserve expenses increase: credit activities of commercial banks in Vietnam increase in the direction of increasing the size and growth rate, but does not focus on improving credit quality Furthermore, there is unfavourable change in the economy Thus, the credit quality decreases significantly In 2012, the NPL ratio was 4,08%; in 2013 and 2014, the NPL ratio decreased and it was reduced to 2.55% in 2015 4.2.2 Business Performance ROA and ROE increased in 2008-2010 However, in the period of 2008 – 2015, both ROA and ROE indicators reduced sharply in 2012 (ROA: 43.12%, ROE: 46.8%) In 2013 and 2014, the profitability of commercial banks increased compared to 2012, but it was only equal of 50% of the average during 2009-2011 Table 4.1: Profitability of commercial banks in Vietnam Year 2008 2009 2010 2011 2012 2013 2014 2015 ROA (%) 1,29 1,01 1,29 1,09 0,62 0,49 0,51 0,4 ROE (%) 14,56 10,42 14,56 11,88 6,31 5,56 5,49 5,7 Source: Annual Report of the State Bank of Vietnam from 2008-2015 4.3 IMPACT OF CREDIT RISKS ON BUSINESS PERFORMANCE OF COMMERCIAL BANKS IN VIETNAM 4.3.1 Credit risk decreases profits 4.3.2 The increase in loan loss reserve lead to the decrease in profits 4.3.3 Bank restructuring to limit credit risk and improve business performance 4.4 FINDINGS 4.4.1 Factors affecting the credit risk Table 4.2 Descriptive statistics of model Variable Obs Mean Std Dev Min Max NPL LLR EFF LEV NII SIZE ROE GGDP INF UNR EXR INR 233 271 262 276 264 276 275 286 286 286 286 286 0.022471 0.015822 0.01115 0.006623 0.487958 0.190311 0.869828 0.11084 0.160688 0.271395 17.34343 1.619804 0.114088 0.074759 6.246387 0.742069 9.280675 6.03656 2.206564 0.262572 18932.05 2319.773 9.820909 2.178888 Source: Computation using STATA 13 0.001 0.000129 0.079532 0.015271 -2.00369 11.88353 0.000749 5.247367 0.63 1.8 15916 7.62 0.1246 0.037018 2.0527 1.129474 0.785564 20.56153 0.444905 7.547248 23.11632 2.6 22380.54 13.46 The authors tested the correlation and multicollinearity model The authors estimates all models: Pooled model, FEM model and REM model However, as discussed above, due to the endogenous phenomenon in the model, the authors use GMM regression ((Ahmad and Ariff (2007), Podpiera and Weill (2008), Louzis et al (2012), Hasna Chaibi and Zied Ftiti (2015)) using panel data The final analytical results are based on the results of GMM regression Table 4.3 Regression result of model Variable L.NPL LLR EFF LEV NII SIZE ROE GGDP INF UNR EXR INR _cons Pooled NPL FEM NPL REM NPL GMM NPL 0.172*** [2.69] 1.269*** [7.90] 0.01 [1.31] -0.00281 [-0.31] 0.0140** [2.57] -0.00332*** [-3.28] -0.0192 [-1.16] 0.000113 [0.06] -0.000289 [-0.93] -0.00458 [-1.09] 8.57E-07 [1.11] 0.00194** [2.07] 0.0347 [1.24] 0.0773 [1.15] 1.829*** [9.06] 0.00754 [0.84] -0.00789 [-0.74] 0.0138** [2.24] -0.00652** [-2.17] -0.00874 [-0.45] 0.000943 [0.51] -0.000264 [-0.86] -0.00604 [-1.48] 1.23E-06 [1.07] 0.00127 [1.35] 0.0890** [2.17] 0.172*** [2.69] 1.269*** [7.90] 0.01 [1.31] -0.00281 [-0.31] 0.0140** [2.57] -0.00332*** [-3.28] -0.0192 [-1.16] 0.000113 [0.06] -0.000289 [-0.93] -0.00458 [-1.09] 8.57E-07 [1.11] 0.00194** [2.07] 0.0347 [1.24] 0.0868* [1.79] 2.151*** [8.60] 0.00175 [0.27] -0.00302 [-0.53] 0.0120*** [3.75] -0.00774*** [-4.27] -0.000479 [-0.03] -0.000662 [-0.75] -0.000146 [-1.28] -0.00421** [-2.37] 0.00000246*** [4.63] 0.000844* [1.88] 0.0886*** [3.95] Chow test (p-value) Hausman test (p-value) Bresh-Pagan test (pvalue) Sargan test (p-value) TTQ test (p-value) N R-sq T * 0.0108 0.00 204 204 204 0.452 0.469 statistics in brackets p