Rủi ro tín dụng của hệ thống ngân hàng dưới tác động của đại dịch Covid-19: Nghiên cứu trường hợp tại Việt Nam

Bài viết tiến hành kiểm tra sức chịu đựng vi mô để tìm hiểu ngành ngân hàng Việt Nam có thể chịu được rủi ro tín dụng ngày càng gia tăng dưới ảnh hưởng của đại dịch COVID-19 đang diễn ra trên toàn thế giới hay không. Mời các bạn cùng tham khảo!

INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 CREDIT RISK OF BANKING SYSTEM UNDER THE EFFECT OF COVID-19 EPIDEMIC: THE CASE OF VIETNAM RỦI RO TÍN DỤNG CỦA HỆ THỐNG NGÂN HÀNG DƯỚI TÁC ĐỘNG CỦA ĐẠI DỊCH COVID-19: NGHIÊN CỨU TRƯỜNG HỢP TẠI VIỆT NAM Tien Nhat NGUYEN - Ngoc Quynh Anh LE University of Economics, Hue University Ntnhat@hce.edu.vn Abstract The paper conducts micro stress testing to investigate whether the Vietnamese banking sector can withstand the increasing credit risk under the influence of the COVID-19 pandemic that has been occurring worldwide The study was conducted to build scenarios with levels of high, low, and medium of credit risk shock that Vietnamese commercial banks may encounter in the absence of capital support from the State Bank and interbank market The research results show that all banks stay well in the good normal scenario (Vietnam can control the COVID pandemic) and their capital adequacy ratios (CAR) are all above 9% However, in the worst economic scenario (the epidemic in Vietnam is not able to control), (the shock caused by the reduction in the ratio of collaterals to non-performing loans) – (the shock caused by the increase in the ratio of non-performing loans)/12 banks in the sample are negatively affected by the increase in NPL, which leads to the decline in their CAR below the regulatory level of 9% Keywords: Capital adequacy ratio; Stress Testing; Credit risk; COVID-19, Vietnam Tóm tắt Bài viết tiến hành kiểm tra sức chịu đựng vi mơ để tìm hiểu ngành ngân hàng Việt Nam chịu rủi ro tín dụng ngày gia tăng ảnh hưởng đại dịch COVID-19 diễn toàn giới hay không Nghiên cứu thực nhằm xây dựng kịch với mức độ sốc rủi ro tín dụng cao, thấp trung bình mà ngân hàng thương mại Việt Nam gặp phải trường hợp không hỗ trợ vốn từ Ngân hàng Nhà nước thị trường liên ngân hàng Kết nghiên cứu cho thấy tất ngân hàng kịch bình thường tốt (Việt Nam kiểm sốt đại dịch COVID) hệ số an toàn vốn (CAR) 9% Tuy nhiên, kịch kinh tế xấu (dịch bệnh Việt Nam chưa thể kiểm soát), (cú sốc giảm tỷ lệ tài sản đảm bảo nợ xấu) - (cú sốc tăng tỷ lệ nợ xấu)/ 12 ngân hàng mẫu chịu ảnh hưởng tiêu cực tỷ lệ nợ xấu tăng, dẫn đến hệ số CAR họ giảm xuống mức quy định 9% Từ khóa: Tỷ lệ an tồn vốn; kiểm tra sức chịu đựng; rủi ro tín dụng; covid-19, Việt Nam 1349 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 Introduction The new strain Corona-virus has had negative effects and unpredictable evolution in most of the countries Many experts predicted that the economic shock caused by COVID-19 might be more serious than the 2008 financial crisis Besides, an economist of a consulting firm, namely Rosenberg Research and Associates Inc, claimed that during the 2008 financial crisis, air transport was not stopped, borders were not closed, frozen and quarantined, and people were not afraid to leave their homes, but during COVID-19 we were talking about a tangible fear that causes people to withdraw from economic activity According to the Financial Post, a series of Canadian banks such as Scotiabank, Canadian Imperial Bank of Commerce (CIBC) and Royal Bank of Canada (RBC) forecast the risk of a great financial crisis Mr Mark Carney who works for Bank of England said that the shock from the previous time would be very different At the heart of the 2008 financial crisis, bad debt was the focus of the crisis and limited in the field In the banking sector, it ended when the US Government released USD 700 billion to buy back collateral, and this time the epidemic was a threat from outside, affecting the whole economy Van Luc CAN and the BIDV Research & Training Institute have just published a report assessing the impact of the COVID-19 on the global economy and Vietnam, including the financial sector Regarding the banking system, it claims that the COVID-19 epidemic will have influence in three important aspects The first is the decrease in credit demand due to the lower credit needs of businesses and households, especially in the first and second quarters of the year 2020 Secondly, there might be a potential increase in bad debt, when businesses and households are negatively affected by the epidemic, leading to difficulties in production and business activities Finally, the demand for digital banking transactions and non-cash payments probably increases because of a quarantine Therefore, it is inevitable to assess the credit risk of either each bank or banking system to recognize and take measures to respond to the crisis of COVID-19 Since the banking industry can be considered as a pillar of the economy in the crisis, the State Bank of Vietnam issued Circular 01/2020/TT-NHNN which aims to regulate the financial institutions, rescheduling the term of the loan payment, exempting and reducing interest rates, keeping the debt group in order to support the economics and customers to overcome the crisis Recently, a lot of techniques applied to forecast the credit risk have been developed and implemented by the central bank In which Stress Testing method has been widely used to measure the credit risk tolerance of the banking system In the United States, JP Morgan Chase is one of the largest banks that adopted inspection techniques since the 1990s and performed regular stability checks (daily/ weekly against risk) for risk management purposes of existing portfolios as a basis for business planning The IMF also conducts quarterly and annual stability assessment to measure the credit risk of the whole system under the Financial Sector Assessment Program (FSAP) Up to now, in Vietnam, there are also some typical studies to test the credit risk tolerance of the banking system However, these studies mostly focus on the macro approach, meaning that conducting a regression analysis between a loan quality and macro-economics factor and hardly ever apply a methodology that assumes a direct shock to the loan quality Meanwhile, 1350 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 more than 50% of FSAP (Financial Sector Assessment Program) use the method based on a data of bad debt to measure the credit risk, and 60% of them conduct the micro stress test method, 30% apply macro stress test (Cihak M., 2005) The study will apply the macro stress test technique to assess the credit risk tolerance of Vietnamese commercial banks during the COVID-19 epidemic The study focuses on constructing three scenarios with varying levels of high, low, and medium with either considering a scenario of the IMF applied to assess the banking system in several countries or the situation of epidemic expansion in Vietnam It also takes into account experts’ judgments and the research that predicts the change in the current social economy and the fact of the influence of COVID-19 on developed countries, for example, China, Italy, the US, and Korea that have been heavily affected by the COVID-19 epidemic The objective of the study is to calculate the impact of the current COVID19 epidemic on the increasing rate of bad debts, the decrease in collateral, leading to a change of the Capital Adequacy Ratio (CAR) of commercial banks in Vietnam The research results perhaps are meaningful for bank managers and policymakers to provide instant responses to adverse socio-economic changes during and after the COVID-19 epidemic Literature review In fact, in the United States, JP Morgan Chase is one of the largest banks that has adopted Stress test technique since the 1990s, and coducting regular stability testing (daily/ weekly against market risk) The International Monetary Fund (IMF) also conducts stability checks to assess risks across the finance system, regularly publishing in semi-annual or annual financial stability reports, namely Financial Sector Assessment Program (FSAP) More than 50% of FSAP missions used the NPL data-based method to measure credit risk, of which 60% used the micro stress test method and 30% used the macro stress test method (Cihak , 2007) Later, there were also numerous macro-level studies on credit risk using Stress test method Hoggarth (2005) uses the macro-test technique to analyze output shocks, inflation, exchange rates and short-term interest rates to the non-recoverable NPL ratio of the UK banking system Virolaimen (2004) builds a macro credit risk model for the Finnish business sector, researching a scenario including a severe recession and banking crisis Amediku (2006) estimates the changes in macro variables affecting NPL ratios and the forecast of the quality of loan portfolios of Ghanaian banks Vazquez (2012) tested credit risk tolerance based on 03 models (i) macro model, (ii) micro model, and (iii) VAR model In general, the research results all confirm that adverse changes in the macro environment have an impact on banks’ credit risk, but banks’ ability to withstand shocks depends on the strength and financial stability of the bank In recent years, there are also some studies that aim to test the credit risk tolerance of the banking system in Vietnam However, most of them focus on the macro approach, meaning that based on regression analysis between loan quality and macro variables with little focus on the method of assuming a direct shock on loan quality According to the current context, the study will apply macro-checking method to assess the credit risk tolerance of Vietnamese commercial banks during the Covid-19 pandemic The study will build three credit shocks with different scenarios corresponding to three levels from low, medium, to high; based on the epidemic 1351 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 development situation of Vietnam and the world recently, and judgments of experts regarding to the change in the current social economy and the influence of Covid-19 on some countries before the current Covid-19 pandemic such as China, America, Italy, Korea The study calculates the impact of the Covid-19 on the increasing rate of bad debts, the decrease in collateral, leading to the change in the minimum CARs of commercial banks in Vietnam Methodology 3.1 IMF Stress Test Stress Testing is used to assess the vulnerability of a portfolio due to changes in macroeconomics factors; or an influence of events that are extreme, exceptional, and abnormal but plausible as defined by Basel There are two common approaches to conduct a systemic stress testing for the credit risk regarding the performance of loans (Cihak, 2004) The first one is micro stress test which immerses a shock assumption directly in a loan quality, and the another is macro stress test which investigates a relationship between a loan quality and macro-economics variables Each method has its advantages and disadvantages The advantage of a loan quality-based approach is that monitoring agencies have data of bad debts, lending status However, a drawback of this method is that the NPLs not reflect the current situation For the regression method or VAR (Vector AutoRegression), a main challenge is to have enough data series over time of NPLs, with adjusting the periods where there are major changes in industry structure or economy From the pros and cons of these approaches, the study will apply the micro stress test to assess the credit risk tolerance of Vietnamese commercial banks by linking a direct shock to factors reflecting the lending The implementation mechanism of the method is to reduce the value of the collateral and increase the bad debt ratio 3.2 Research process The study constructs three scenarios in which two rates are adjusted (rate of collateral and bad debt) and keep other assumed rates following the IMF scenario The way of implementing two shocks is as follows: 1st shock: The value of the collateral is decreasesed to an assumed rate Accordingly, the provision portion also increases, thereby affecting the CAR The CAR after the shock is calculated as follows: Step 1: The value of the collateral is reduced according to the assumed rate; Step 2: Reducing the value of collateral for each group of bad debts according to the ratio determined in Step 1; Step 3: Calculating the necessary provision for risks in case of a decrease in the value of collateral for each group of debts according to the rate of provision for each loan interval Step 4: Calculating the allowance for the difference that equals to the current provision level minus the required level of risk provision Step 5: Recalculating the results of equity after shock, Risk Weighted Assets (RWA) after 1352 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 shock, CAR after shock, and CAR difference 2nd shock: The NPL ratio and provisioning are increased by a certain percentage, thereby affecting the CAR Step 1: NPLs are increase according to the assumed rate Step 2: Calculating the incremental value of the risk reserve For the reserve rate increase, the study keeps the IMF’s rate Step 3: Recalculating the results of equity after shock, Risk Weighted Assets (RWA) after shock, CAR after shock, and CAR difference 3.3 Constructing scenarios of credit risk under influence of COVID-19 epidemic The main principles applied to construct the scenarios are: hSelecting the rate of scenario following the IMF’s scenario and several studies (Cihak M., 2004) (Goldstein, 2012); hMatching the scenario with current conditions of Vietnam; hMaking basic assumptions of the scenario; hClarifying a sensitivity of the result under scenario assumptions; hEnsuring a practice of scenario Scenario for shocks 1st Scenario: This is IMF’s scenario that sets out to test for banking system of several countries; 2nd Scenario: Positive scenario This is the scenario where the disease lasts until the end of 2nd quarter of the year 2020, and the disease was under control in Vietnam; the production and business activities will return to operate as normal production 3rd Scenario: Moderate scenario This is the scenario that assumes the epidemic to be controlled but economic activities are still limited 4th Scenario: Negative scenario This is a scenario that assumes the epidemic in Vietnam cannot be controlled or controlled but economic activities are forced to stop for a long time The detail of each shock accompanied with scenarios are described as follows: 1st shock: The decrease of collateral value According to the expert Tom Barrack - US billionaire - chairman and CEO of Colony Capital, “The US mortgage market is in danger of collapsing in the current situation when the US cannot control the epidemic in the second quarter of 2020 This has huge consequences, leading to the borrower default if the Government and banks not help A mass blockade by the US to 1353 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 stop the epidemic of acute pneumonia COVID-19 will cause a series of collateral to be deposited and foreclosed Tenants risk being evicted, borrowers’ default, banks collapse The repayment effect can be dramatically increased At this time, the mortgage market is in danger of collapsing Including real estate investors, landowners, homeowners, travel services, restaurants, and hotels Both the tenant and the employee are working for the landlord” The study will be based on this assumption and assume that in the worst scenario, Vietnam cannot control the pandemic, the risk of affecting collateral is very high and taking the IMF scenario with the rate of asset reduction mortgage on 75% discount loans makes the worst scenario for the Vietnamese economy to test credit risk tolerance However, according to experts, housing prices in Vietnam increased continuously because of the “tradition” of the hoarding property of the people The saving rate of Vietnamese people on average is nearly 50% of income, double that of most other countries In particular, a very large saving rate used to store real estate properties and spending on housing lead to very high house prices here CBRE Vietnam has built scenarios on the real estate market under COVID-19 influence In scenario is an epidemic that lasts until the end of June with a 5% decrease in land prices, 714% of office space, and a 5% increase in apartment prices Since then, the study chooses the best scenario when the epidemic is controlled in the second quarter of 2020, the collateral rate for loans decreases at 10% This is reasonable, while the majority of collateral and collateral is real estate The second scenario that CBRE has given after surveying the real estate market is that if the epidemic lasts until the end of September, the ground price will decrease by 10%, the official price will decrease by 8-10%, down by 5% Thus, the total reduction in real estate value is about 25% if the epidemic is controlled in Q3 / 2020 Based on scenario by CBRE, the study selected the average scenario for the collateral rate of collateral loss for loans at 40% when the disease is controlled but the economy is still limited 2nd shock: The increase of NPL ratio To carry out shock 2, the study will perform the credit risk test under IMF scenario with the rate according to column 1, table Scenario 2: according to the report of the Ministry of Planning and Investment, as a result, the State Bank of Vietnam forecasts that the NPL ratio will increase from 2.9-3.2% if the epidemic is controlled by the end of Q1 / 2020 and the NPL ratio will increase by 4% if the disease is controlled review at the end of quarter 2/2020 (column 3, Table 1) Thus, the best scenario for the Vietnamese bank at the time of implementation is expected that the epidemic will be controlled by the end of Q2 / 2020 and the NPL ratio will increase by 4% Besides, for the medium scenario, based on the SBV’s estimate, the NPL ratio will increase by about 1% for each quarter of the year if the epidemic is controlled and the economy is still restricted by the end of 2020 NPLs will be 6-7% (NPLs increase in Q3 and Q4 is about 23%) (column 4, Table 1) For the worst scenario, according to the report of early March of (Financial Magazine-Finance Ministry Information Agency, 2020), 23 credit institutions estimated that about 926,000 billion VND of outstanding loans were affected affected by the COVID-19 pandemic, accounting for about 14.27% of the total outstanding loans of these 23 credit institutions Based on this number, the study will assume the worst-case scenario for these affected out1354 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 standing loans will turn into bad debt for credit institutions and choose the worst increase for NPL ratio is 15% VND To test the stamina of the banking system (column 5, Table 1) For scenario 2,3,4, the study only changes the NPL increase rate and inherits the other indicators under the IMF scenario Table 1: Credit risk scenarios Shock Rate (%) 1st Collateral 2nd NPLs 2nd 3rd 1st Positive Sce- ModerateSceIMF Scenario nario nario 75 10 25 Research data 40 4th Negative Scenario 75 15 (Source: Authors’ synthesis) The research data is collected from data on the Financial Statements of 12 joint-stock commercial banks of Vietnamese banking system in 2019 These are 12 out of 18 Basel II standard banks on minimum capital adequacy ratio and also the banks that research can collect sufficient data to perform the tolerance test Besides, 12 over 31 Vietnamese commercial banks in the study belong to the group of banks that account for nearly 80% of the total assets and equity of the Vietnamese banking system, so the research results partly reflect the level of risk The credit of the entire banking system The data includes total outstanding loans, safe debt groups (groups 1, 2), bad debt groups (groups 3, 4, 5), insured assets for substandard loans (for groups 3, 5), credit risk provisions, equity, convertible risk assets, and minimum CAR Table Descriptive statistics of data used for Stress test of Vietnamese commercial banks in 2019 Unit: VND billion Indicators Descriptive statistics Mean Standard Deviation Minimum Maximum 377,846 340,021 94,409 1,097,501 631 24,944 Total debt(s) 383,476 1st Group 372,922 Outstanding 2nd Group Bad debt(s) 3rd Group 4th Group 5th Group 4,924 5,630 1,318 939 3,373 344,630 335,325 7,053 95,645 92,480 5,260 1,236 1,139 305 1,681 3,219 1355 218 449 1,116,997 1,072,557 19,496 5,448 4,305 11,356 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 Collateral for un- 10,045 8,230 1,415 29,507 4th Group 1,708 1,756 213 6,450 3rd Group 1,879 5th Group 6,458 Credit-risk proviEquity Risk CAR (%) NPL/Total 5,472 4,836 Weighted 12,946 1,154 8,514 11 17,411 1,208 12,580 2,701 5,646 348 26,720 (NPL – Credit-risk 4,419 469 41,169 3,858 Debt 1,807 80,883 17 -6 11 (Source: Authors’ synthesis) Result The research on liquidity risk testing with two approaches as follows: 1st Shock: Reducing collateral for ineffective loans; 2nd Shock: Increasing NPL ratio and respectively test using four scenarios, namely IMF’s scenario, Positive scenario, Moderate scenario, and Negative scenario (Table 1) for 12 commercial banks in Vietnam Table The result of the credit-risk tolerance test as reducing the ratio of collateral for inefficient loans BANK VCB ACB BID CTG HDB MBB TCB TPB STB VIB VPB LPB 1st and 4th Scenarios (75%) CARs 9.50 6.01 4.60 8.97 8.52 10.68 15.11 10.40 6.18 6.53 6.69 7.56 Δ CAR 0.00 -3.69 -6.10 -0.33 -2.48 0.00 -1.39 0.00 -4.52 -3.17 -4.71 -3.34 2nd Scenario (10%) CARs 9.50 9.70 10.70 9.30 11.00 10.68 16.50 10.40 10.70 9.70 7.25 10.90 1356 Δ CAR 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 -4.15 0.00 Unit: % 3rd Scenario (40%) CARs 9.50 9.70 10.26 9.30 11.00 10.68 16.50 10.40 10.70 9.70 6.28 10.90 Δ CAR 0.00 0.00 -0.44 0.00 0.00 0.00 0.00 0.00 0.00 0.00 -4.42 0.00 (Source: Authors’ calculation) INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 The result of the credit risk tolerance test of 12 commercial joint-stock banks in the Vietnamese banking system following a shocking approach by reducing the ratio of insured assets to ineffective loans (Table 3) (i) For the Positive scenario (with depreciation rate for inefficient loans of 10%) and the medium scenario (with a decrease in mortgage assets for inefficient loans is 40%) However, the study showed that 11/12 banks that were given to test still meet the standards of Basel II with CAR ratio greater than 9% However, this degree of influence still causes the minimum capital adequacy ratios of some banks to decrease slightly compared to 2019 Specifically, BIDV in the good scenario still maintains the CAR ratio but when checked with the average scenario level, although the CAR ratio is guaranteed (greater than 9%), it starts to show a decrease to 10.26% when the property’s property ratio decreases by 40%, corresponding to a decrease of 0.44% compared to the rate Initial CAR (column 6,7 - table 3) It is also worth noting that the CAR of VPB plummeted with 4.15% in the good scenario and 4.42% in the medium scenario (column 5.7 table 3) All CAR indicators of VPB in scenarios not meet the minimum safe capital ratio requirement with 7.25% and 6.28% in the good and medium scenarios (columns 4.6 - table 3) (ii) For the worst scenario, the study using the scenario of the IMF as the worst scenario with the ratio of the collateral for inefficient loans decreased by 75%, it shows that 8/12 banks cannot keep the capital ratio Minimum safety following regulations, accounting for 66.67% of banks are checked Specifically, BID, VPB, STB, ACB, LPB, VIB, HDB, and CTG all showed a decrease in CAR ratio ranked from high to low leading to a decrease of 6.10%, 4.71%, 4.52%, 3.34%, respectively , 3.17%, 2.48%, and 0.33%, and CAR of all banks are below the regulated standards of 4.6%, 6.69%, 6.18%, 6.01%, 7.56%, 6.53%, 8.52%, and 8.97%, respectively In contrast, banks, VCB, MBB, and TPB, still maintained their CAR in all scenarios after implementing the shock of collateral value reduction for ineffective debts This is explained by the fact that these three banks have quite a high provision for credit losses, so when the collateral value for ineffective loans decreases, the difference risk provision is zero Since then, with the rate of decrease in collateral for ineffective loans given, it cannot affect the equity and riskweighted assets as well as the CAR of the three banks this row This result shows that the defense ability of these banks is quite solid in the crisis caused by the corona flu Besides, despite showing a decline of 1.39% CAR after TCB’s shock, the bank’s CAR still ranks the highest in the banking system Through this result, it shows that there are two methods to strengthen banks’ defenses: banks can minimize the decrease in CAR or banks can increase their initial CAR through financial indicators Table The result of the credit-risk tolerance test as increasing the ratio of bad debt Bank VCB ACB 1st Scenario (25%) 2nd Scenario (4%) 3rd Scenario (7%) 4th Scenario (15%) CARs Δ CAR CARs Δ CAR CARs Δ CAR CARs Δ CAR 9.02 -0.68 9.58 -0.12 9.50 -0.20 9.28 -0.42 9.07 -0.43 9.43 -0.07 1357 9.38 -0.12 9.24 -0.26 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 BID 9.02 -1.68 10.39 -0.31 10.17 -0.53 9.63 -1.07 HDB 10.24 -0.76 10.87 -0.13 10.78 -0.22 10.53 -0.47 TCB 15.65 -0.85 16.36 -0.14 16.25 -0.25 15.98 CTG MBB TPB STB VIB VPB LPB 8.53 10.14 9.74 9.23 8.60 6.51 9.72 -0.77 -0.54 -0.66 -1.47 -1.10 -0.74 -1.18 9.17 10.59 10.29 -0.13 -0.09 -0.11 10.44 -0.26 7.12 -0.13 9.51 10.69 -0.19 -0.21 9.07 10.52 10.21 10.25 9.37 7.02 10.54 -0.23 -0.16 -0.19 -0.45 -0.33 -0.22 -0.36 8.82 10.35 9.99 9.77 9.01 6.79 10.16 -0.48 -0.33 -0.52 -0.41 -0.93 -0.69 -0.46 -0.74 (Source: Authors’ calculation) The result of the credit risk tolerance test of the 12 commercial banks in the Vietnamese banking system following the shocking approach by increasing the bad debt ratio showed that the results were almost similar to the results of implementing 1st shock (Table 4): (i) For the best scenario (with a bad debt growth rate of 4%) and the medium scenario (with a bad debt growth rate of 7%) that the study gives, 11/12 Bank is given after testing, it still meets the standards of Basel II with a CAR ratio of more than 9% However, this level of influence still makes the minimum capital adequacy ratios of some banks decrease compared to 2019 Especially for out of 11 banks with satisfactory CAR ratios, BID and STB has a larger difference of CAR before and after the shock than the other banks, with a decrease of 0.31%, 0.26% in the positive scenario, and 0.53%, 0.45% in the moderate scenario In contrast, similar results from a bank shock, VPB failed to pass from the test with the lowest NPL increase scenario with a post-shock CAR of 7.12% (less than 9% as required) Although the ratio of CAR ratio before and after the shock is not large, the initial CAR of VPB is quite low (ii) For the case scenario (with an increase of 15% in bad debt), BID and STB still have the difference of the highest CAR of 1.07% and 0.93%, respectively, but still, maintain the system CAR complies with regulations, respectively 9.63% and 9.77% And VPB bank still shows the highest level of credit risk compared to Vietnamese commercial banks with a CAR of not meeting standard 6.79% In particular, for this scenario, the study also found that CTG, one of the major banks in Vietnam, started to show instability with a CAR of 8.82 below the regulated standards (iii) For the IMF scenario (with a 25% NPL increase), 9/12 banks still show good defensive ability, only banks have CAR ratios after the shock of not meeting the standard CTG, VPB, and VIB with CARs of 8.53%, 6.51%, and 8.60%, respectively 1358 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 dài hạn Ngược lại, khía cạnh tài hỗ trợ cho khía cạnh học hỏi phát triển, quy trình nội thành cơng hơn, tiếp tục góp phần làm nên thành cơng khía cạnh khách hàng 2.2 Kỹ thuật phân tích bao liệu (DEA) Kỹ thuật phân tích biên giới thiệu từ nghiên cứu Farrell (1957), nhiên đến xuất nghiên cứu Charnes cộng (1978) kỹ thuật thu hút nhiều ý giới nghiên cứu Bài viết nhóm nghiên cứu đề xuất việc áp dụng quy hoạch tuyến tính vào việc giải toán Farrell, biết đến với tên gọi Data Envelopment Analysis, hay DEA Trong thập niên gần đây, phương pháp phân tích bao liệu (DEA) xem phương pháp hữu ích đánh giá suất hiệu sản xuất, đặc biệt tổ chức phi sản xuất Ưu trội nó, là: kỹ thuật phi tham số Điều có nghĩa để đo lường hiệu theo phương pháp không cần phải xác định trước dạng hàm Đường biên sản xuất theo phương pháp xác định dựa liệu quan sát Ngoài ra, cịn dễ dàng áp dụng tổ chức sử dụng nhiều đầu vào để tạo nhiều đầu ngân hàng Đó lí mà lĩnh vực ngân hàng lĩnh vực có nhiều nghiên cứu ứng dụng DEA nhất, theo khảo sát Liu cộng (2013) Vận dụng phương pháp phân tích bao liệu, tốn hiệu sử dụng viết viết sau: Trong đó: n: số lượng đơn vị định (Decision Making Unit, viết tắt DMU) xem xét; m: số lượng yếu tố đầu vào; s: số lượng yếu tố đầu ra; yrk: lượng đầu thứ r tạo đơn vị k; xik: lượng đầu vào thứ i sử dụng đơn vị k; λj: trọng số đo lường khả trở thành “ đơn vị chuẩn” (benchmark) DMUj DMU đo lường hiệu (k) Giá trị thu toán quy hoạch tuyến tính ln ≥ 1, giá trị ϕ_k-1 lượng đầu mà DMUk tăng lên mà không thay đổi lượng yếu tố đầu vào Như vậy, giá trị 1⁄ϕk ≤ định nghĩa số hiệu kỹ thuật tương quan DMUk so với đơn vị định khác sở liệu (xem Coelli, 1996) Đây công thức xác định hiệu dựa lập luận theo định hướng đầu ra, nghĩa việc tối đa hóa hiệu đơn vị định xác định sở tối đa hóa đầu thu với yếu tố đầu vào cố 1366 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 định Tác giả lựa chọn mô hình định hướng đầu cho viết theo Ngo Dang Thanh (2012), mơ hình phù hợp với tổ chức phi sản xuất ngân hàng 2.3 Sự kết hợp BSC DEA đánh giá hiệu Phương pháp BSC DEA có nhiều ưu điểm đánh giá hiệu tổ chức hai có hạn chế lớn việc đánh giá Sự kết hợp BSC DEA đánh giá hiệu theo nhiều nghiên cứu phát huy ưu điểm khắc phục nhược điểm hai phương pháp; vậy, hướng nghiên cứu nhiều nhà nghiên cứu lý thuyết thực hành quan tâm Theo Eilat cộng (2008), Najafi cộng (2009) số hạn chế BSC giảm bớt BSC sử dụng kết hợp với DEA Sau số ưu điểm mơ hình kết hợp BSC-DEA theo tổng hợp từ nghiên cứu Nguyễn Thị Hạnh (2017): - BSC đo lường hoạt động tổ chức theo cách tiếp cận cân đối lại chung chung, DEA cung cấp thơng tin rõ ràng hữu ích cho nhà quản trị; - BSC giúp xác định hệ thống mục tiêu phù hợp cho tổ chức DEA giúp cung cấp đánh giá chuyên sâu hoạt động quản trị tổng thể tổ chức dựa đầu vào đầu ra; - Sự kết hợp BSC-DEA cung cấp cách tiếp cận với khả phân tích cải thiện Nó cho phép thực phân tích đồng thời nhiều đầu vào nhiều đầu ra, đồng thời đầu vào nên giảm để đạt mức đầu xác định đầu tăng lên với mức đầu vào cho trước xác định để đạt hiệu quả; - DEA chuyển hóa thước đo hoạt động thành thơng tin quản trị, cịn BSC cung cấp đầu vào phù hợp cho DEA, kết hợp đầu vào đầu khác cho mơ hình đánh giá hiệu mức độ khác Đặc điểm giúp cho cho kết hợp BCS-DEA trở thành công cụ lý tưởng để đánh giá hiệu hoạt động ngân hàng (Chen cộng sự, 2008) Tổng quan nghiên cứu đề xuất mơ hình 3.1 Tổng quan nghiên cứu 3.1.1 Tổng hợp số nghiên cứu sử dụng công cụ kết hợp BSC-DEA đánh giá hiệu Trên giới có nhiều nghiên cứu đề xuất việc sử dụng kết hợp hai phương pháp BSC DEA việc đo lường đánh giá hiệu đơn vị, tổ chức; nhiên nghiên cứu tiếp cận góc độ khác Nghiên cứu Rouse cộng (2002) xem nghiên cứu đề xuất ý tưởng cho kết hợp hai phương pháp để đo lường hiệu hoạt động phận dịch vụ kỹ thuật bảo trì hãng hàng khơng quốc tế giai đoạn từ 1993-1996 Tuy nhiên, nghiên cứu này, hai công cụ đo lường sử dụng cách rời rạc mà chưa có kết hợp chặt chẽ với Đến nghiên cứu Rickards (2003) thật phát triển mô hình đo lường hiệu 1367 INTERNATIONAL CONFERENCE FOR YOUNG RESEARCHERS IN ECONOMICS & BUSINESS 2020 ICYREB 2020 dựa kết hợp hai công cụ BSC DEA Tác giả đề xuất mơ hình DEA để đo lường hiệu 69 đơn vị cơng ty đa quốc gia, tiêu đầu vào đầu lựa chọn từ tiêu đánh giá hiệu theo bốn phương diện mơ hình BSC Và ý tưởng sử dụng mơ hình đo lường hiệu tổng qt kết hợp hai công cụ BSC DEA tiếp tục phát triển hai nghiên cứu Eilat cộng (2006, 2008) ứng dụng việc đánh giá hiệu dự án nghiên cứu phát triển (R&D) Khác với hướng nghiên cứu Rickards (2003), Chen Chen (2007) đề xuất phát triển mơ hình DEA tương ứng với phương diện mơ hình BSC truyền thống tài chính, đổi học hỏi, khách hàng quy trình nội Trên sở so sánh kết mô hình để đánh giá hiệu 30 doanh nghiệp ngành công nghiệp bán dẫn Đài Loan giai đoạn từ 2002-2005 Min cộng (2008) đề xuất mơ hình DEA để đo lường đánh giá hiệu khách sạn hạng sang Hàn Quốc Tuy nhiên mơ hình khơng sử dụng tiêu đầu vào đầu theo phương diện mơ hình BSC mà tác giả lại đề xuất sử dụng số hiệu mơ hình làm thước đo đánh giá phương diện tài mơ hình BSC đề xuất sử dụng cho ngành khách sạn Tương tự, để đánh giá hiệu 15 dự án R&D Bồ Đào Nha lựa chọn từ phận bảo dưỡng công ty đa quốc gia hoạt động lĩnh vực vận chuyển, Amado cộng (2012) đề xuất sử dụng mơ hình DEA tương ứng với phương diện BSC Tuy nhiên, khác với nghiên cứu trước, mơ hình không sử dụng để đo lường tách biệt hiệu phương diện mơ hình BSC mà nhằm xem xét mối quan hệ nhân chúng Cụ thể: tác giả sử dụng đầu mơ hình (đo lường phương diện học hỏi phát triển) làm đầu vào mơ hình (đo lường phương diện quy trình nội bộ), đầu mơ hình lại sử dụng làm đầu vào mơ hình (đo lường phương diện khách hàng) cuối đầu mơ hình sử dụng làm đầu vào mơ hình (đo lường phương diện tài chính) Đối với lĩnh vực ngân hàng, cịn nghiên cứu tìm hiểu mơ hình kết hợp BSC DEA để đo lường đánh giá hiệu đơn vị tổ chức hoạt động lĩnh vực Nghiên cứu Chen cộng (2008) nghiên cứu đề xuất ứng dụng mơ hình kết hợp BSC DEA lĩnh vực ngân hàng để đánh giá hiệu ngân hàng Hualien (Hoa Liên) Đài Loan giai đoạn từ 2001 đến 2006, chi tiết theo 24 quý hoạt động Trong nghiên cứu này, nhóm tác giả đề xuất sử dụng so sánh kết nhiều mơ hình DEA: ngồi mơ hình đại diện cho phương diện truyền thống BSC tài (F), đổi học hỏi (L), khách hàng (C) quy trình nội (I), tác giả đề xuất bổ sung phương diện quản trị rủi ro (R)) Tập hợp biến mà nhóm tác giả giả đề xuất sử dụng cho mơ hình bao gồm: biến đầu (các khoản cho vay ngân hàng, số lượng hộ vay, khoản thu nhập từ phí) biến đầu vào (số lượng nhân viên, tổng tài sản, vốn huy động từ tiền gửi tài sản cố định) Tuy nhiên, sở chọn biến cho mơ hình nghiên cứu chưa thật hợp lý biến lựa chọn chưa thật đại diện cho phương diện mơ hình BSC đề xuất Shahroodi Bahraloloom (2014) đề xuất mơ hình kết hợp BSC DEA để 1368 ... Huy Trường Đại học Kinh tế, Đại học Huế ntthuyenkt@gmail.com Tóm tắt Sự cạnh tranh gia tăng mạnh mẽ hệ thống ngân hàng Việt Nam năm gần đặt yêu cầu cần phải đo lường hiệu hoạt động ngân hàng, ... để đánh giá hiệu hoạt động ngân hàng (Chen cộng sự, 2008) Tổng quan nghiên cứu đề xuất mơ hình 3.1 Tổng quan nghiên cứu 3.1.1 Tổng hợp số nghiên cứu sử dụng công cụ kết hợp BSC-DEA đánh giá hiệu... nhiều nghiên cứu đề xuất việc sử dụng kết hợp hai phương pháp BSC DEA việc đo lường đánh giá hiệu đơn vị, tổ chức; nhiên nghiên cứu tiếp cận góc độ khác Nghiên cứu Rouse cộng (2002) xem nghiên cứu