The impact of information and communication technology (ict) on bank performance of commercial bank in vietnam

This research endeavours to elucidate the intricate relationship between Information and Communication Technology ICT and the performance of commercial banks operating within the Vietnam

Research Motivation

Along with the strong development trend of the digital economy, innovation and integration, the finance and banking industry is also undergoing a rapid change along with the technological change of the digitalization of the economy (Mention, 2021) Business models with the support of modern technology are creating changes in the operating models of financial institutions, including banks (Schueffel, 2016) With the development of a series of payment technologies, Fintech companies, banks are under increasing pressure to provide better services to customers In particular, the recent development trend of artificial intelligence technology allows banks to rapidly develop services, occupy market share with significantly reduced costs compared to traditional methods Investing in modern technology, including ICT, can help commercial banks quickly change their approach to customers, reduce the cost of managing employees and businesses, increase operational efficiency, capture market share, and ultimately increase profits

The advent of ICT in the banking industry has brought about significant changes in its operations (Gupta et al, 2018) Therefore, this study was conducted with the aim of examining the impact of ICT innovation and application on the profitability of Vietnamese commercial banks from 2009 to 2022 Vietnam has witnessed a significant surge in technological advancements over the past decade This includes the widespread adoption of smartphones, internet penetration, and digital financial services This rapid transformation has profoundly impacted on the banking industry Customers are increasingly demanding convenient, personalized, and digital-first banking experiences Traditional banking models are struggling to keep pace with these changing expectations The Vietnamese banking sector is becoming increasingly competitive, with both domestic and foreign players vying for market share Banks need to innovate to attract and retain customers The State Bank of Vietnam (SBV) has actively encouraged the adoption of technology in the banking sector through various initiatives This further underscore the importance of understanding the impact of technology on bank performance

In Vietnam, ICT - a general term for all types of technologies that allow users to create, access and manipulate information, is a new research indicator and ICT has many different impacts on the system as well as the business efficiency of each bank Therefore, to evaluate and clarify this

2 issue, the authors chose the topic “The impact of Information and Communication Technology

(ICT) on bank performance of commercial bank in Vietnam” to assess the impact of ICT on the operational efficiency of banks From there, the authors propose guidelines to help bank planners and managers understand the role of technology and find the right solutions to do business in the digital age.

Research Objectives

Based on several published studies, the group's research aims to clarify the relationship between building and investing in the development of Information and Communication Technology (ICT) and the banking sector At the same time, based on the research results to evaluate the impact of ICT on the operational efficiency of domestic commercial banks, as well as build a comprehensive view of ICT-related issues in the current era, raise people's awareness of the importance of ICT At the same time, this study also sets out specific objectives including building and systematizing the theoretical basis of ICT and the impact of ICT on the operational efficiency of commercial banks, clarifying the current situation of the impact of ICT on the operational efficiency of commercial banks, develop a model to study and evaluate the impact of ICT on the operational efficiency of commercial banks in Vietnam in the period of 2011 – 2022 and propose recommendations and solutions to promote the application of ICT and improve the operational efficiency of banks

Research Question

In an increasingly competitive landscape, Vietnamese commercial banks are constantly seeking ways to improve performance efficiency The application of information technology is considered one of the key factors However, to comprehensively assess the impact of information technology on performance, it is necessary to consider how it interacts with other factors such as size, capital, and asset quality From there, there are 2 main questions for this research: “What are the components of ICT impact on banking performance of commercial bank in Vietnam?” and

“How does ICT components, size of bank, capital adequacy ratio and loan to deposit ratio affect the preformances of Vietnamese commercial banks?”.

Scope of research

The dissertation research the impact of ICT components, capital adequacy ratio and loan to deposit ratio affect the preformances of 24 Vietnamese commercial banks from 2011 to 2022.

Research method

This research paper has used the data collection method to search and synthesize information related to the topic, including previous research papers to build an overview of the research and the theoretical basis of the topic to provide an overview and see the gaps of the previous research to supplement the research paper of the group The data collected by the research team is the data that has been processed in the ICT Index Report of the Ministry of Information and Communications which is used as the main ICT measurement index for the research paper ICT data is combined with data from the performance of commercial banks through total asset indexes, net profit ratio to total assets, and cost-to-income ratio over the years of 24 commercial banks to form a data panel The research paper was carried out quantitative research, using a synthesized and standardized dataset to perform regression and validation methods to find the most suitable model for the research paper Regression models are performed to study the relationships between variables in the model This paper uses Pooled OLS regression methods and uses FEM fixed impact model, REM random impact, FGLS estimation to overcome the problems of the model Especially, in order to build the most complete model to solve defects and endogenous phenomena, the team researched and used the GMM regression model Using a variety of methods, synthesizing the research dataset to analyze the model results to see the relationships between the variables, thereby showing the correlation of information and communication technology with the operational efficiency of commercial banks in Vietnam

This research makes a novel contribution to the theoretical and empirical literature on the impact of information technology on the performance of Vietnamese commercial banks By employing a large dataset of 24 banks over the period 2011-2022, the study provides a comprehensive overview of the evolving relationship between ICT and bank performance during a period of significant change The application of a variety of econometric models, including Pooled OLS, FEM, REM, and especially GMM, addresses endogeneity issues and ensures the robustness of the estimation results The study identifies specific ICT components that significantly impact bank performance and clarifies the mediating roles of other factors such as

4 size, capital, and asset quality The findings have important implications for the strategic development of commercial banks and the formulation of policies to support the development of digital banking in Vietnam.

Dissertation Structure

Chapter 1: Introduction: Introducing the rationale of the dissertation, the research objectives and research questions, the scope of research, new contribution and reasearch method

Chapter 2: Literature review: Mentioning the concept and the measurements of bank performance, ICT components, and reviewing the results of previous research papers about the factors that affect the performance of Vietnam commercial bank

Chapter 3: Research Methology: Mentioning the sources of data and the research method of the dissertation

Chapter 4: Estimation results and major findings: Presenting the estimation results, performing tests to have the most accurate results and interpreting these results

Chapter 5: Recommendation and Conclusion: Recommending to help Vietnamese commercial banks operate safely and efficiently

Overview of ICT

Definition of ICT

Information and communications technology is a combination of information technology and communication technology According to Alabi (2005), ICT merges computer technology with high-speed communication, audio, and video linked data networks Laudon (2001) argues that ICT involves collecting, storing, manipulating and transmitting information by electronic means Communication technology refers to physical devices and software that link components of various computer hardware and transmit data from one physical location to another According to Rafi Ashrafi (2008), it is critical that the word ICT include a wide spectrum of computerized information and communication technologies Desktops, laptops, mobile devices, wired or wireless intranets, business productivity applications such as word processors and spreadsheets, enterprise software, data storage and security, cybersecurity, and other items are examples of these technologies.

ICT application in commercial banking

In the world of technology, which has an increasingly great influence on today's life, the interaction between devices, systems, software and people is increasing markedly The banking industry is one of the leading industries in the application of information technology in operation and management to meet the changing trend of the world Technological applications have crept into every corner of life ICT has been gradually becoming the foundation for promoting economic growth, participating in solving most of the difficult problems of the country in general and the banking industry in particular, towards national digital transformation, digital economy, and building a prosperous Vietnam The application of ICT in banks is no longer a necessary condition, but it is a mandatory condition if banks want to develop faster and more sustainably ICT brings many benefits to banks, automating production, control, data storage, and online transactions quickly and at low cost, creating many new and innovative products to meet market demand Information and Communication Technology is an extremely important factor in expanding, improving, enhancing performance, reaching more customers and enhancing the bank's competitive advantage compared to other competitors The application of information technology to bank operations has helped the communication process become more convenient, promoting

6 exchanges between banks, businesses between many countries and the process of world economic integration.

Components of ICT System

According to research by the Ministry of Information and Communications, the report on the readiness index for the development and application of information and communication technology in Vietnam in 2020, the ICT Index system of commercial banks includes: technical infrastructure, human resource infrastructure, internal banking applications, etc online services of banks The research team uses the results of the indicators in the ICT Index report over the years to calculate and apply to the model

The ICT index system of commercial banks is measured and calculated independently based on statistics of the Ministry of Information and Communications The component indicators have a reciprocal relationship, which is the basis for calculating the results of ICT indicators

The technical infrastructure index constitutes a comprehensive assessment of a commercial bank's information technology capabilities and the efficacy of its digital applications in underpinning core business functions This metric serves as a pivotal benchmark for gauging a bank's competitive position within the market, as it reflects the institution's capacity to leverage technology to drive operational efficiency, enhance customer experience, and innovate new products and services In particular, the technical infrastructure index of commercial banks includes 5 indicators:

Firstly, server and workstation infrastructure: Includes server and workstation systems used to store, process and manage bank data It also supports the implementation of information security and data safety solutions to ensure the security of the bank's important data

+ Ratio of Virtual Servers/Total Servers (Physical Servers + Virtualized Servers)

∑ Servers + Workstation ratio (PC/Laptop) within the last 3 years / Total workstations

∑ Workstations equipped in the last 3 years

Secondly, communication infrastructure: Includes network systems and communication equipment used to connect the bank's departments and branches with each other and with customers In addition, it also supports the implementation of electronic payment systems such as POS and ATM

+ Percentage of workstations running licensed and manufacturer-supported operating systems Formula:

∑ Workstations running licensed and manufacturer − supported operating systems

∑ Workstations + Internet Bandwidth Ratio for Internet Banking Services/Total Internet Banking Customers Formula:

+ Percentage of Internet bandwidth provided to internal users/Total number of computers connected to the Internet

∑ Connection bandwidth for internal users

∑ Number of computers connected to the Internet + Broadband Ratio/Total Terminals

∑ Bandwidth of a wide area network

Thirdly, ATM and POS infrastructure: Includes ATM and POS systems used to make payment and withdrawal transactions for customers To ensure the safety of these transactions,

8 information security and data safety solutions are also implemented to prevent fraudulent activities and unauthorized access

+ ATM Ratio/Total Payment Cards

∑ Payment Cards + Chip card acceptance rate/Total number of ATMs

∑ ATMs that accept chip cards

∑ Number of ATMs + Ratio of ATMs with recharge function/Total number of ATMs

∑ Máy ATM có chức năng nạp tiền

∑ Number of ATMs + POS Machine Ratio/Total Payment Cards

∑ Payment Cards + Ratio (mPOS+ Wireless POS)/Total POS

Fourth, implement information security and data safety solutions: Including information security and data safety solutions implemented to ensure the security and reliability of systems Formula: Anti-virus software installation rate + Information security + Data safety

Anti − virus software installation rate = ∑ Computer with antivirus software installed

Information security = ((∑ Affiliated units deploy firewalls) x 5 + ∑ Affiliated units deploy spam filtering software + ∑ Affiliated units deploy anti-virus software + ∑ Affiliated units deploy access warning software + (∑ Affiliated units deploy the solution other information security) x 0,5)/(∑ Affiliated units)

Data safety = (∑ Affiliated units install magnetic tape + ∑ Affiliated units install disk cabinets + (∑ Affiliated units installation SAN) x 5 + (∑ Affiliated units installation NAS) x 4 + (∑ Affiliated units installation DAS) x 3 + (∑ Affiliated units installation other storage devices) x 0,5)/ (∑ Affiliated units)

Finally, data centers and disaster backup centers: Data centers and disaster backup centers are built to ensure stable operations and avoid risks when an incident occurs A data center helps banks store and manage data efficiently, while a disaster backup center helps minimize damage when something goes wrong

The human resource infrastructure index of commercial banks includes 3 indicators:

Firstly, the proportion of full-time staff in information technology: This index measures the proportion of officials in the total number of employees of the bank with expertise in information technology This is an important indicator to ensure that the bank has enough human resources to develop and maintain information technology systems

∑ Officer in charge of IT

Secondly, the proportion of full-time staff in charge of information security: This index measures the proportion of full-time officers in charge of information security in the total number

10 of employees of the bank This is an important indicator to ensure that the bank has enough manpower to protect customer information and prevent cyberattacks

∑ Officer in charge of Information security

Finally, the ratio of full-time information technology officers with international certificates in information technology/total number of full-time information technology officers: This index measures the proportion of full-time information technology officers with international certificates in information technology in the total number of full-time information technology officers of the bank This is an important indicator to ensure that the bank's IT staff is qualified and professional enough to meet the needs of the job

∑ IT officers have international certificates specializing in IT

∑ Officer in charge of IT

Therefore, these indicators help the bank evaluate and improve its human infrastructure in the field of information technology and information security and strengthen the ability to develop and maintain safe and effective information technology systems

2.1.3.3.Application of Internal Information Technology

The bank's internal information technology application index includes 3 indicators:

Firstly, implement core banking: Core Banking is the bank's core financial and transaction management system, including customer account management, transfers, borrowing, payments and other financial services This index assesses the level of implementation of the bank's Core Banking system

Formula: Total number of deployed Core bank Modules + Total number of connections to Core bank and other systems (ERP, ATM/POS, Internet Banking, SWIFT, CITAD, Reporting Systems ) + Connection methods between Core bank and other systems (1: file interface, 2: Database, 3: Message Queue, 4: ESB integration axis, 5: Other forms) + Level of automation when processing transactions between Core bank system and other systems (0: not automatic, 1: semi-

11 automatic, 2: automatic) + Handling data reconciliation between Core Bank and other systems (0: no reconciliation, 1: manual reconciliation , 2: with partial automatic collation, 3 with full automatic collation)

Performance of commercial banks

Concept of business efficiency of enterprises

According to British economist Adam Smith: “Efficiency - The result achieved in economic activity, is the revenue from the consumption of goods” This view does not distinguish between efficiency and business results without considering the cost factor to achieve that business efficiency Business results are what an enterprise achieves after a certain period quantified by expenditures such as consumption, revenue, and market share while business performance reflects the level of use of resources, which is calculated by the ratio between the results and the waste spent to achieve that result

From a financial perspective, the maximum goal of an enterprise is to maximize the value of the enterprise based on respecting the law and well implementing social responsibility Therefore, an enterprise with high business efficiency will create an increase in enterprise value in the long term based on effective use of the resources of the enterprise To achieve the goal of maximizing business value, businesses must constantly improve business efficiency

Thus, business efficiency is also reflected in the rate of return on operating capital in relation to the cost of using capital: This means that on a capital spent, the enterprise tries to obtain the highest profit based on considering acceptable risks Therefore, an enterprise with good business performance is an enterprise that generates a return on capital that is greater than the cost of using capital

Performance efficiency of commercial banks

Measuring the performance of a bank is a complex operation The researchers used different approaches to evaluate the performance of banks in different countries at different times Laying the foundation for later studies, the operational efficiency of commercial banks is shown through the relationship between revenue and operating costs (Berger & Mester, 1997) which is simply understood as the fact that banks generate the largest revenue with the smallest costs and input resources

However, Sinkey (1992) argued that return on assets is a comprehensive measure of the overall performance of a bank from an accounting perspective and a key indicator of management efficiency because it indicates how the bank's management ability has converted the bank's assets into net income Meanwhile, Akintoye (2004) offers additional indicators that can be used to represent the performance of organizations, businesses, and banks, namely net profit margin (Net profit margin), return on capital used (ROCE) and return on assets (ROA) The use of different measures to measure performance will complicate the research process when comparing the results of the studies but will provide a better overview and increase the robustness of the model

According to a study by Hughes & Mester (2008), performance is measured in two ways: a structural approach and a non-structural approach With a structural approach, financial indicators ROA, ROE, ROS are commonly used; while the cost, revenue, and profit functions are used to analyse the effectiveness of the unstructured approach.

Methods of measuring the bank's performance

The ROA (Return on Assets) index is an important indicator in evaluating the business performance of a bank This metric is calculated by dividing the bank's total net income (total income minus expenses) by the bank's total assets The formula for calculating ROA is as follows:

Total Asset ∗ 100 The ROA indicator indicates the bank's ability to generate profits from its assets A high ROA indicates that the bank is likely to generate high profits from its assets, while a low ROA may indicate that the bank is using its assets inefficiently or is experiencing profitability issues

ROA is one of the important indicators for investors, regulators and experts in the banking industry to evaluate a bank's business activities

Return on Equity (ROE) is a measure of a company's annual profit (net income) divided by the total value of shareholders' equity, expressed as a percentage The formula for calculating ROA is as follows:

Total Equity ∗ 100 ROE provides a simple measure of return on investment By comparing a company's ROE to the industry average, it is possible to determine something about the company's competitive advantage ROE can also provide insight into how company management is using equity financing to grow the business Sustained ROE and increased over time can mean that a company is good at creating value for shareholders because it knows how to reinvest earnings wisely to increase productivity and profits Conversely, a decrease in ROE can mean that management is making bad decisions to reinvest capital in unprofitable assets (CFA Team, 2024)

The NIM (Net Interest Margin) index is also known as the net interest income ratio This is an important financial indicator in the banking industry NIM is simply understood as the difference between the organization's net interest income and the budget that that financial institution (bank) must pay NIM signifies a financial institution's ability to generate profits from the interest it earns from credit and investment activities The formula for calculating NIM is as follows:

Total assets profitable from average interest*100

In which, Net Interest Income = Interest Income and Similar Income – Interest Expense and Similar Expenses/Average gross profitable assets = Deposits at the SBV + Re-deposits of other credit institutions (excluding risk provisions) + Investment securities (excluding depreciation provisions) + Customer loans (excluding risk provisions); Purchase debt (excluding risk provisions) Value is the value that has not yet been set aside for provision NIM is a measure of the effectiveness as well as profitability of a bank's credit activities If NIM is positive, that is, this bank operates profitably, the greater the positive number, the higher the profitability, the greater

15 the ability to invest profitably If NIM is negative, it means that this bank is operating at a loss, investment activities are ineffective.

Literature Review

The impact of ICT on the performance of businesses in developed countries has been found in many of the above studies Recently, studies on the impact of ICT on emerging and developing economies have gained attention and become more vibrant Lessons learned from successful countries such as the United States and Europe have set a greater impetus for other emerging economies (McCarthy et al, 2014), but not every company succeeds without a strategy to build the capabilities needed to benefit from innovation and technology adoption (Prud'homme, 2015) Research in many companies in emerging countries shows that the adoption of ICT has had a positive impact on their performance

Technological changes have been happening rapidly and directly affecting the bank's development strategies ICT is one of the ways for banks to use modern technology and the advantages of the internet to increase their operational efficiency Innovative ideas may include the use of ICT to create new markets and gain a competitive advantage through greater interactivity, cheaper transactions, and direct communication with partners and customers (Zhu & Andersen, 2021) Changes in the modern business environment force commercial banks that want to develop to rely on ICT to achieve and maintain competitiveness and improve productivity However, although organizations in different sectors are widely adopting ICT, several survey reports have found that many businesses do not progress during the stages of the business lifecycle (Amankwah-Amoah, 2019)

The research objective of Aliyu & Tasmin (2012) is to determine whether ICT improves the performance of commercial banks in Nigeria This study uses both fixed regression (FEM) and random regression (REM) models The results of the study show that ICT has a positive impact on banks' profitability through interest rates, exchange rates, ROA, and ROE Sharing the same view, it is found that the creation of ICT channels has a more beneficial impact on banks that only start the Internet than the transformation of traditional conventional banks In addition, due to a lack of awareness or age factors, and a host of other reasons why ICT do not seem to be significantly feasible or accepted by consumers warmly or quickly Therefore, the strengthening of ICT in the

16 banking industry is imperative for the rapidly changing market, as the ICT revolution sets the stage for an exceptional increase in global financial activity

With a sample of 31 Vietnamese joint stock commercial banks and data on the establishment of Fintech companies in the period 2006 - 2018, Thu Hong & Huu Tuan (2021) found that the increase in Fintech companies has a negative impact on the indicators Standards measuring the performance of banks include ROA, ROE and NIM with correlation coefficient magnitudes of 0.0046 respectively; 0.04 and 0.009 and 0.009 In addition, according to Minh Ngoc and Tien Thanh (2021), the application of core technology has improved the reputation of Viettinbank, specifically in the last 5 years, Vietinbank has always ranked second in the top 10 most prestigious banks Vietnam and is considered the least scandalous bank in Vietnam The fourth industrial revolution and digital transformation are becoming an indispensable part, contributing to socio- economic development and bringing a new, more optimal life to people Banking is one of the key sectors of a country, this will be one of the leading sectors in digital transformation and is the foundation for other industries and sectors to move towards digital transformation (Thu Huong & Yen, 2023) The application of technology in banking business areas has become an extremely necessary and indispensable factor These technologies help the banking system automate processes, improve operational efficiency, enhance customer experience, and contribute to increased revenue (Anh Ngoc & Diem, 2022)

From synthesizing previous research, the research team found that in Vietnam there are not many studies on the impact of ICT on the performance of commercial banks like other countries in the world Studies focus on evaluating common factors ATM, POS, Mobile Banking to see the relationship between information communication technology and bank performance Therefore, the research team decided to use the ICT index combined with many other variables to evaluate the performance of commercial banks in Vietnam

Ben Naceur & Goaied (2008), as well as Chhaidar et al (2022), claimed that bank size affects digital investments and profitability As the bank's size increases, the impact becomes more substantial The cost of operations decreases when digital techniques are adopted ICT can help larger banks maximize their investments by lowering per-employee costs Initial ICT costs might be greater, and the advantages may not be immediately apparent Previous years' investments continue to have an impact on later years, which must be considered when performing panel data

17 analysis The GMM model can solve this problem by eliminating endogeneity from the data In GMM, the variable's previous value is removed from its value for the next period Chhaidar et al.'s

(2022) article employed the Least Square Dummy Variable model, which few economists agree is inconsistent for limited panel data, and therefore recommend GMM instead Adane et al (2021) found that ATM installations have a substantial beneficial influence on the profitability of Ethiopian commercial banks, utilizing panel data from 13 Ethiopian banks from 2012 to 2016 using a random effect model This finding demonstrates that technological adoption without any engagement with ICT may have a major influence on the bank's earnings Eliminating this noise while investigating the link between technology and a bank's profitability is critical for producing reliable results In this study, ICT has very little or no influence Sanga and Aziakpono (2022) found from their African study that ICT has a substantial influence on bank deposits and loans to the private sector, resulting in increased revenue links for African banks Furthermore, we believe that earnings will increase in proportion to sales, however formal research is necessary to precisely estimate this According to Saksonova (2014), NIM outweighs ROA in terms of the bank's performance and stability in the research in the United States, Baltic Countries, and Europe Does this remain true if ICT interventions increase asset quality? This question must be answered with new evidence on the relationship between ICT and NIM

Research hypothesis

Previous studies on the investment and application of information and communication technology in the banking sector have had an impact on business results in terms of return on total assets (ROA), return on equity (ROE), profit after tax (Aliyu & Tasmin, 2012) Several studies have investigated this relationship, however, there are differences in the impact of ICT on bank performance Activities related to information and communication technology are an important part of the development process of banks and contribute to improving the competitiveness of commercial banks in the current period Researchers have applied a variety of methods and approaches to understand the nature of the relationship between ICT and performance efficiency Measuring the application of ICT through the effective use of internet banking, mobile banking, ATM, POS (Mahboub, 2018; Abebe, 2016); the level of application of e-banking at banks, the cost of investment in ICT (Ernest et al, 2016); collect information on product and service evaluation of bank employees (Bethuel, 2012)

Bank size, branch count, assets management ratio, operational efficiency, and leverage ratio are the most important bank-specific characteristics determining the profitability of Indian commercial banks, according to ROA The liquidity ratio, asset management ratio, and asset quality ratio have all been proven to considerably boost ROE (Almaqtari et al., 2019) Countless research has been conducted on bank profitability and technology improvements This section elaborates on a few key studies from the literature

In the context that Vietnam is actively developing and investing in information technology, banks will have more opportunities to increase access and innovation in technology The growth of information technology has contributed significantly to the growth of banks, achieving the expected results through operational improvement and innovation Based on the above arguments, the authors believe that investment and application of information and communication technology have a positive impact on the operational efficiency of Vietnamese commercial banks with 4 specific hypotheses as follows:

Hypothesis 1: The ICT index has a positive impact on the performance of banks (ICT)

Del Gaudio et al (2021) examined 28 European Union banks Using data from 1995 to

2015, the study assessed the influence of ICT on bank earnings The findings are consistent with those of the previous research stated above, indicating that ICT boosts the bank's profitability Their findings also indicate that ICT, in conjunction with IT and FinTech, improves bank financial stability However, does a stable bank mean more revenue? Banks' portfolios always include assets with varied risks Doing a customer profile evaluation based on real data and continuing to invest in riskier goods as directed by legislation can dramatically enhance revenues for banks while having no influence on bank stability This appraisal procedure involves the inherent risk of assets turning non-performing

Financial organizations must invest in ICT to increase earnings According to a study of banks in Oman, understanding about clients is critical to a bank's profitability because it allows bank officials to make better decisions regarding avoiding money laundering Offering goods to customers necessitates a thorough review by the bank to reduce the risk of default and noncompliance Bank workers at all levels should be competent This competency and knowledge of changing legislation may be conveyed to employees through continual grooming and the provision of learning opportunities inside the firm This training can be delivered more effectively using ICT

Hypothesis 2: The technical infrastructure index has a positive impact on the bank's operational efficiency (ITI)

The transformation process in Revolution 4.0 has been bringing many practical benefits to different industries and fields, and the banking industry is no exception To survive in this digital transformation, banks are always trying to innovate products and services These products and services can bring significant operational efficiencies to banks, costs for banks, profits can be increased, and risks can be minimized compared to traditional banking products (Tunay et al,

2018) Determining the level of use of IT infrastructure and its impact on customer service always determines the growth of the customer (Luka et al, 2012)

Modern banking now offers financial services thanks to advancements in IT infrastructure Banks in recent times aid farmers by offering simple loans Banks thus play an important role in rural development Thus, the banking industry contributes to the nation's economic development (Kanoujiya et al, 2021; Balasubramaniam, 2012) Instead, the emphasis is on the IT infrastructure,

20 without providing personnel with development opportunities in proportion to IT growth According to Barpanda & Athira (2022), learning opportunities and a positive work environment might reduce attrition Higher attrition results in knowledge loss, the extent of which is determined by whether the organization is process-dependent or people-dependent

Hypothesis 3: The information technology application index has a positive impact on the bank's operational efficiency (OP)

Obviously, the risk of a crisis causing banks to run out of capital is always a latent thing According to Otoo et al (2021), the operational efficiency of global banks will be improved depending on the effectiveness and effectiveness of internal controls, identifying appropriate objectives for the organization will significantly reduce risks Advanced information technology supported by an effective control mechanism is essential to ensure that IT overcomes the gaps in the necessary processes If the management of the banks specifically identifies the individual responsible for coordinating various activities in the unit, the employees understand the concept and importance of internal control, including the division of responsibilities, communication within the banks to help evaluate the effectiveness of the organization's guidelines and policies, and the reporting system on the organizational structure clarifies all the responsibilities of each department/unit in the organization, the operational efficiency of banks will increase (Otoo et al,

2021) Therefore, advanced information systems supported by higher-level control mechanisms are needed to ensure that the necessary processes are implemented through Information Technology

Hypothesis 4: The bank size index has a positive impact on the bank's operational efficiency (SIZE)

Ben Naceur & Goaied (2008), as well as Chhaidar et al (2022), claimed that bank size influences digital investments and profitability The influence rises in proportion to the bank's size The cost of using digital means in operations decreases For larger banks, having ICT can help maximize investments by lowering per-employee expenses The initial cost of ICT might be higher, and the advantages may not be reflected immediately Previous years' investments continue to have an impact on later years, which must be considered when doing panel data analysis

Hypothesis 5: The capital adequacy ratio has a negative impact on the bank's operational efficiency (CAR)

Hassan (2008) uses industry data from 43 Islamic banks around the world using tabular data analysis to analyse the characteristics of Islamic banks and how the overall financial environment affected their operations between 1994 and 2001 The results show that a larger ratio of equity to total assets affects higher profit margins in Islamic banks The author argues that mandatory reserve requirements do not have a significant impact on profitability measures In addition, the results also show that there is little impact between the capital adequacy ratio (CAR) of Islamic banks on the performance of these banks On the other hand, the positive correlation between the capital adequacy ratio and the bank's performance is shown in the market model and regulatory model in the study by Bаrrios & Blаnco (2003) This study classifies West Bank banks into a different group when determining factors that correlate with capital changes The first group of banks with a capital ratio that exceeds the required ratio is possible to achieve the optimal capital structure (resulting from the market model) by adjusting for unique variables such as bank size, account balance, operating expenses, etc ROА, credit, and account risk In contrast, banks with a capital adequacy ratio that is more than the legally required requirement cannot achieve this optimal level More than women, the degree of influence of capital on other factors also differs according to the bank's conditions

Hypothesis 6: The loan to deposit ratio has a positive impact on the bank's operational efficiency

The loan-to-deposit ratio (LDR) is the ratio between the amount of money that the bank provides credit to the amount that the bank raises, an index used to measure the level of lending as well as show the liquidity of the bank The higher the LDR ratio, the greater the third-party capital used for lending, which means that the bank has been able to perform its financial intermediary function well The higher the bank's credit expansion, the higher the chance of profit because the interest earned from credit extensions will increase the bank's profit The higher the LDR coefficient, the higher the bank's profitability Harun (2016), Eng (2013) and Sudiyatno & Suroso

(2010) show that LDR has a positive and significant effect on bank profitability In addition, Dat et al (2019) showed that the ratio of outstanding debt to customer deposits has a positive impact on ROE.

Research Data

The study has used a secondary and unbalanced panel data collected Many sources were mobilized to collect all the data needed for the empirical study However, the data for performance, investment in computer software and control variables were extracted from banks’ annual reports and Bank scope database The data was collected from the Vietnam ICT index report (Communication of Information) and finally, the internet banking data was extracted from the websites of the various banks

The data of the research paper was synthesized and statistically compiled by the authors from table data, due to a combination of time and space data, the sample included 24 commercial banks in the period from 2011 to 2022 The data is collected from audited financial statements and annual reports of Vietnamese commercial banks In addition, we also use secondary data from the Information and Communications Technology Readiness Index Report published by the Ministry of Information and Communications to measure the information and communication technology indicators of banks The data the team used in the study was a table of 288 observations from 24 commercial banks.

Data Research Methods

The study used both descriptive and inferential statistics in data analysis The analysis is carried out with the help of Eview12 and Stata14 Firstly, the collected data is cleaned, sorted, and collated The data will then be entered into the analysis software, after which the analysis will be performed The project uses quantitative research methods with multivariate regression models applying many different regression methods such as Pooled OLS, FEM, REM and performing related tests in combination with theoretical research methods including classification and theoretical systematization, analysis - synthesis method, comparison method and the most complete model building In addition, this study uses the GMM (Generalized Method of Moments) regression model to estimate the impact of the capital adequacy ratio on the bank's performance The advantage of GMM estimation is that it overcomes endogenous and gives accurate results

The first is to use the Pooled OLS least squares regression method to understand the relationship between independent variables and dependent variables Improve the efficiency of group model regression by implementing the Fixed Effect Model (FEM) and the Random Effects Model (REM) to select the appropriate model Pooled OLS estimation for the stochastic impact

23 model will give the estimation parameters without deviation but is ineffective because it ignores the autocorrelation in the error component

The impact of technological innovation on the performance of the Vietnamese banking sector was tested by using a quantitative method This technique includes both descriptive statistics and multiple regression analysis

• “ROA”, “ROE” and “NIM”: performance in bank “i” for period “t”

• “β”: coefficient of variable where “i” ranges from 1 to 6

• “ICTi,t”: ICT index “i” for period “t”

• “ITIi,t”: information technology infrastructure index in bank “i” for period “t”

• “OPi,t”: Online paying in bank “i” for period “t”

• “SIZEi,t”: bank size in bank “i” for period “t”

• “CARi,t”: capital adequacy ratio in bank “i” for period “t”

• “LDRi,t”: loan-to-deposit ratio in bank “i” for period “t”

Hypothesis Testing

Table 1: Explanation and references of variables Dependent variables Explanation References

ROA Return on total assets Financial Statements

ROE Return on Equity Financial Statements

NIM Net Interest Margin Financial Statements

ICT Information and communication technology index of commercial banks

ITI information technology infrastructure index in bank

OP Online paying in bank (+) ICT Index Report

Control variables Explanation Expected sign References

Annual Assets of Commercial Banks

Data Description

Statistics describing the variables used in the study are shown in the Table 2 In which, the mean values of the independent variables of ICT, ITI and OP are 0.3950705, 0.3950705 and 2.05, respectively This means that 39.50% of the banks in the observation sample have invested in ICT, 33.38% of the banks in the observation sample have invested in technology infrastructure, and 205% of the banks in the observation sample have invested in online paying

Variable Mean Std Dev Min Max

The ROA-dependent variable shows the average return on total assets of the study sample with an average value of 0.0095968, a standard deviation of 0.0094657, equivalent to the fluctuation in the profitability of banks in the study data is negligible and stable with a range ranging from -0.0551175 to 0.0537847 This shows that the average value has a slight fluctuation and the difference in data between banks is small The ROE dependent variable shows the average net return on equity of the research sample with an average value of 0.1109432, a standard deviation

26 of 0.1082903, the fluctuation in the profit of banks in the research data at ROE is not small with a range ranging from -0.8200211 to 0.6592532 The last dependent variable is NIM, we find Nim’s mean value is 0.0349985, a standard deviation of 0.0130979, the fluctuation in the net interest margin of banks in the research data at NIM is small with a range ranging from 0.005724 to 0.0953 This shows that the data is very different and the data for each year of banks for this variable is also very different.

Correlation Matrix

Correlation coefficient is a statistical index that measures the extent of a linear relationship between two variables Based on the results of table above, the correlation coefficients of the independent variables are lower than 70%, so the independent variables have low correlations and are suitable for regression

ROA ROE NIM ICT ITI OP SIZE CAR LDR

Analysis of Regression Results

Table 4: Estimation results from 3 models OLS, FEM, and REM (ROA)

Note: The value in square brackets [] is the corresponding t-statistics

*, **, and *** correspond to statistical significance levels of 10%, 5%, and 1%, respectively

According to the estimation results from 3 Pooled OLS, FEM and REM models for the ROA model, the ICT independent variable has no statistical value in all 3 models This shows that the ICT variable has no impact on the growth nor the sliding of the ROA-dependent variable In contrast, when using the Pooled OLS, REM, and FEM regression models, the ITI (0.00991, 0.00756, and 0.00798) and OP (-0.00110, -0.000971, and -0.000973) variables are both statistically significant at the 5% significance level Meanwhile, the SIZE variable was statistically significant at 1% for the Pooled OLS model of 0.000634 but not at the FEM and REM models Finally, the CAR variables (-0.000945, -0.000488 and -0.000526) and LDR (0.0153, 0.0190 and 0.0180) were statistically significant at 1% for all 3 Pooled OLS models FEM and REM

Table 5: Estimation results from 3 models OLS, FEM, and REM (ROE)

Source: Author’s Calculation Note: The value in square brackets [] is the corresponding t-statistics

According to the estimation results from 3 Pooled OLS, FEM and REM models for the ROE model, the independent variables of ICT and OP have no statistical value in all 3 models This shows that the ICT variable has no impact on the growth nor the sliding of the ROE-dependent variable For knowing that ITI is statistically significant at 5% (0.102) for the Pooled OLS model, 10% (0.0808) for the REM model and no statistical significance for the FEM model In particular, the SIZE variable is statistically significant for all 3 Pooled OLS, FEM and REM models, of which it is at 1% for the Pooled OLS (0.0142) and REM (0.00788) models and 5% for tissue FEM figure (0.00500) Along with that, when using the Pooled OLS, REM and FEM regression models, the LDR variable has statistical significance at the 1% significance level of 0.141, 0.227 and 0.194, respectively Finally, the CAR variable was statistically significant at a significance level of 1% for the Pooled OLS model of -0.00690 and 5% for the REM model of -0.00406 but not statistically significant for the FEM model

Table 6: Estimation results from 3 models OLS, FEM, and REM (NIM)

According to the estimation results from 3 Pooled OLS, FEM and REM models for the NIM model, the independent variables of ICT, ITI and OP have no statistical value in all 3 models This shows that these 3 variables have no impact on the growth as well as the sliding of the NIM- dependent variable For the SIZE variable, the statistical significance is only at 10% with the Pooled OLS model of 0.000506 In contrast, the CAR variables (-0.00146, 0.000933 and -0.000991) and LDR (0.0334, 0.0287 and 0.0289) were statistically significant at 1% for all 3 Pooled OLS, FEM and REM models.

VIF test for multicollinearity

Table 7: VIF test for multicollinearity

Conducting the verification of the phenomenon of multi-line through the VIF digital relationship From the above data table, we can see that the ITI, ICT and OP variables have a VIF coefficient of 4.80, 3.16 and 3.13, because the dataset used by the group belongs to the financial industry, so the comparison coefficient will be 2 So, the ITI, ICT and OP variables are multi- linear, and the LDR (1.12), CAR (1.08), SIZE (1.04) variables are not multi-linear The VIF mean value of 1< 2.39 0.05 is consistent with the H1 hypothesis, so the REM model is more suitable than the FEM Thus, the more suitable model used in the study is the random impact model (REM)

Table 9: Breusch-Pagan Lagrangian and Hausman tests for ROE model

Breusch-Pagan Lagrangian 0.0000 The Pooled OLS is not suitable

Hausman 0.0000 The FEM model is more suitable than the REM model

Breusch-Pagan Lagrangian and Hausman tests were conducted to find the model that best fits the study The inspection results are shown in the data table above With the Breusch-Pagan Lagrangian test, the p-value = 0.0000 < 0.05, which is consistent with the H0 hypothesis, so the Pooled OLS model is not suitable With the Hausman test, the p-value = 0.0000 < 0.05 is consistent with the H0 hypothesis, so the FEM model is more suitable than REM Thus, the more suitable model used in the study is the FEM model

Table 10: Breusch-Pagan Lagrangian and Hausman tests for NIM model

Breusch-Pagan Lagrangian 0.0000 The Pooled OLS is not suitable

Hausman 0.0759 The REM model is more suitable than the FEM model

Breusch-Pagan Lagrangian and Hausman tests were conducted to find the model that best fits the study The inspection results are shown in the data table above With the Breusch-Pagan Lagrangian test, the p-value = 0.0000 < 0.05 is consistent with the H0 hypothesis, so the Pooled OLS model is not suitable With the Hausman test, the p-value = 0.0759 > 0.05 is consistent with the H1 hypothesis, so the REM model is more suitable than the FEM Thus, the more suitable model used in the study is the random impact model (REM).

White and Wooldridge tests

After selecting the REM model that is suitable for the ROA and NIM models and the FEM model that is suitable for the ROE model, the autocorrelation test and the variance of variation were tested to see if the model has a defect The data table below describes the test results of the model

Table 11: White and Wooldridge tests for ROA model (REM)

White 0.0000 The model has a variable error variance

Wooldridge 0.0000 The model has a self- correlation phenomenon

According to the results of the White test, p-value = 0.0000 < 0.05 so it can be concluded that the model has a variable error variance Similarly, the model has a self-correlation phenomenon according to the results of the Wooldridge test when p-value = 0.0000 < 0.05 is consistent with the H0 hypothesis

Table 12: White and Wooldridge tests for ROE model (FEM)

Wooldridge 0.0000 The model has a self- correlation phenomenon

According to the results of the White test, p-value = 0.0000 < 0.05 so it can be concluded that the model has a variable error variance Similarly, the model has a self-correlation phenomenon according to the results of the Wooldridge test when p-value = 0.0000 < 0.05 is consistent with the H0 hypothesis

Table 13: White and Wooldridge tests for NIM model (REM)

Wooldridge 0.0004 The model has a self-

According to the results of the White test, p-value = 0.0000 < 0.05 so it can be concluded that the model has a variable error variance Similarly, the model has a self-correlation phenomenon according to the results of the Wooldridge test when p-value = 0.0004 < 0.05 is consistent with the H0 hypothesis.

FGLS Regression Results

To overcome the autocorrelation phenomenon and variable error variance, the research team conducted regression according to FGLS The FGLS model is often used to overcome inaccurate estimates due to autocorrelation and variable error variance because FGLS allows the model to exist with variable, self-correlated variance without compromising the accuracy of the estimates (Wooldridge, 2010; Romano & Wolf, 2017) The final regression estimation results of the model are presented in the data table below

Table 14: FGLS Regression Results with ROA Model

The above table shows the regression results showing that the FGLS model with Prob > Chi2 = 0.000 < 0.05 shows that the regression model is suitable, in other words, the model has overcome the autocorrelation and variance variance of variance Therefore, we can see that the final model of ROA is:

According to the regression table above, most of the variables are statistically significant to varying degrees Except for the OP variable which has no statistical significance, the effect of investing in online payment applications on the bank's operational efficiency is unclear This is not consistent with the Hypothesis 3 originally posed However, investing in online payment more products means that banks must increase the convenience of products, monitoring activities and ensuring security factors, especially in the era of advanced technology Therefore, investing heavily in online payment applications will increase the difficulty and complexity of monitoring and ensuring safety and security (Abebe et al, 2016) From there, it can increase costs and reduce the profit margin that the bank earns from this activity Therefore, the impact of investment in OP may not be clear

The regression weight of the ICT variable has a negative mark, indicating the negative impact of investment in ICT on the bank's performance at a meaningful level of 5% (-0.00287), in other words, investment in ICT does not contribute to increasing the bank's operational efficiency This is not consistent with the H1 hypothesis, but it is consistent with previous studies by Bethuel

(2012) Because of the strict state regulations on the finance and banking sector, sometimes investment decisions in ICT are made to circumvent regulations or political interests that may be brought Therefore, investments in ICT can cause unnecessary increases in costs, making business results at banks not very positive In Vietnam, 0 VND banks are compulsorily acquired and assigned by the State Bank to Vietcombank and Vietinbank to manage and operate is a specific example of this case The management and operation do not bring benefits to both banks because of the negative equity and very large bad debts of the 0 VND banks In addition, Vietcombank and Vietinbank will have to restructure the organization to ensure that there are enough departments and resources to manage and operate these 0-dong banks, thereby increasing costs and reducing the bank's profits

The regression weight of the investment in technical infrastructure (ITI) variable has a positive mark, showing the same relationship of investment in ITI and the bank's performance at a meaningful level of 5% (0.00309) This is consistent with the Hypothesis 2 In fact, banks tend to focus on improving, upgrading and developing more data storage systems, software to support loan management, credit risk control, early warning software, and improving the credit appraisal process by data digitization measures This helps to increase labour productivity significantly; help banks closely control data, make forecasts about business activities and possible risks; contributing to minimizing costs and improving the bank's business efficiency

The regression weights of the SIZE and LDR variables are positive, showing the same relationship between SIZE, LDR and the bank's performance at the meaningful levels of 5% and 1% at 0.000294 and 0.0115, respectively As previously reported by Ben Naceur and Goaied

(2008) and Chhaidar et al (2022) suggest that bank size plays a role in digital investments and their profitability This impact is even more significant as the size of the bank increases Costs are reduced with the adoption of digital means in operations For larger bank sizes, the use of information technology can optimize investment by reducing the cost per employee The results of many previous studies have also shown that the larger the size of the bank, the higher the bank's profitability Therefore, banks need to diversify capital sources as well as increase the size of banks safely, in which increasing shareholders' own capital is the safest and most effective solution For the LDR variable, in Vietnam, the difference between lending interest rates and deposit interest rates is quite large, leading to the more banks lend, the higher the profits Therefore, the positive

37 relationship between the ratio of outstanding credit to mobilized capital of banks and the profit margin of Vietnamese commercial banks is reasonable (Phan Anh, 2023)

The regression weighting of the CAR variable has a negative mark, showing the adverse effect of the bank's capital ratio compared to the bank's risk-weighted assets and short-term liabilities on the bank's performance at a meaningful level of 5% (-0.00287) This represents the opposite relationship and is consistent with the H5 hypothesis set out earlier The results can be explained by previous studies as follows, the results of an empirical study from commercial banks in the UK and Australia in the period 2010-2019 by Le et al (2020) indicate that the capital adequacy ratio under the Basel III Treaty has a negative impact on the profitability and efficiency of banks Dao and Nguyen (2020) conducted a study on 16 commercial banks in Vietnam between

2010 and 2017, using simultaneously the processes estimated by the smallest cylinder model (OLS) The results of the estimate show a negative relationship between capital (measured by capital adequacy ratio) and bank performance (measured by ROE) The author concludes that it is not always good to keep a large amount of capital or only approve loans that are satisfactory This method generates relatively high CAR for banks, but the experimental results show that, in this case, the bank's profits will be lower

Table 15: FGLS Regression Results with ROE Model

The above table shows the regression results showing that the FGLS model with Prob > Chi2 = 0.000 < 0.05 shows that the regression model is suitable, in other words, the model has overcome the autocorrelation and variance variance of variance Therefore, we can see that the final model of ROE is:

For this statistic, the SIZE and LDR variables are statistically significant at 1% with regression coefficients of 0.00734 and 0.0933, respectively This result shows that it is consistent with the Hypothesis 4 and Hypothesis 6 proposed at the beginning In addition, both SIZE and LDR variables have a positive impact on ROE, so if the growth in bank size and the ratio of loans to total deposits increases by 1 unit, the bank's ROE will increase to a greater extent with 0.00734 and 0.0933 units The remaining variables in the model are not statistically significant The regression coefficients of the ICT, ITI, OP, and CAR variables that are not statistically significant are all contrary to the initial hypothesis

Table 16: FGLS Regression Results with NIM Model

The above table shows the regression results showing that the FGLS model with Prob > Chi2 = 0.000 < 0.05 shows that the regression model is suitable, in other words, the model has overcome the autocorrelation and variance variance of variance Therefore, we can see that the final model of NIM is:

For this statistic, the CAR and LDR variables are statistically significant at 1% with regression coefficients of -0.00105 and 0.0162, respectively This result shows that the CAR variable has an inverse effect on NIM and is consistent with the initial Hypothesis 5 That is, when CAR increases

40 to 1 unit, the bank's NIM will decrease by 0.00105 units In contrast, the LDR variable has a positive impact on NIM and is consistent with the Hypothesis 6 set out at the beginning Therefore, when the ratio of loans to total deposits grows to 1 unit, the bank's NIM will increase to a greater extent with 0.0162 units The remaining variables in the model are not statistically significant The regression coefficients of the variables ICT, ITI, OP, and SIZE are not statistically significant, all contrary to the original hypothesis.

GMM Regression Results

With the GMM regression estimation method in the table data, the problems of endogenousness variance, and autocorrelation have been overcome The results are shown in 3 tables of data below corresponding to 3 models of ROA, ROE and NIM

Table 17: GMM Regression Results of ROA model

The ITI and OP variables have a positive impact on the performance of commercial banks (ROA) during this period The increase in the value of these variables in the model will increase the value of the rate of return on the total assets of the commercial bank The ITI and OP variables were statistically significant at 10% and 1% with coefficients of 0.007091 and 0.001237,

41 respectively On the contrary, the ICT and CAR variables have a negative impact on the performance of commercial banks, while other factors remain unchanged, the decrease of this variable in the model will increase the value of ROA, specifically as follows:

The statistically significant ICT variable with a significance of 1% and a coefficient of - 0.015612 indicates that the bank's operating efficiency (ROA) and ICT are inversely related, if the ICT coefficient increases by 1%, it leads to a decrease of 0.016% in the bank's operational efficiency A bank with a high ICT index will negatively affect the bank's operational efficiency The above results are not consistent with the H1 hypothesis mentioned above but can be explained

Ky et al (2019) argue that the application of ICT as well as the implementation of Fintech technology products in banks increases bank profits and efficiency and enhances customer interaction and develops new customer segments The application of ICT in banking technology also facilitates banks' risk-taking behaviour, thereby attracting and retaining customers by providing quality and timely services, as well as reducing customer costs and increasing bank profits (Wang et al, 2020) However, in addition to some of the positive points mentioned above, intensive investment in ICT and Fintech products also requires the cost of staff training, maintenance costs, upgrades as well as possible risks due to failure (Alt et al, 2018), leading to a significant decline in bank profits Most of the experimental results show that the investment and application of ICT in the financial sector, especially in the banking industry, significantly increases the operational efficiency and profitability of commercial banks in India (Gupta et al, 2018), Europe (Del Gaudio et al, 2021), United States (Pierri & Timmer, 2022) In fact, there are very few documents on the impact of ICT on the profit margins of Vietnamese commercial banks There has not been any research to analyse in detail the impact of each sub-index constituting the ICT composite index on the profit margin of Vietnamese commercial banks

Similar to the FGLS model of ROA, the CAR variable running GMM is statistically significant at 1% with a coefficient of -0.001786 If the CAR ratio increases by 1%, it will lead to a decrease of 0.001786% in the bank's operational efficiency This is consistent with the Hypothesis 5, but this can be explained by the same studies Regulations on capital adequacy as well as the level of capital adequacy of commercial banks have many direct impacts on the credit activities of commercial banks (Van den Heuvel, 2008; Noss & Toffano, 2016) Thus, changes in the capital adequacy ratio (CAR), will affect the credit growth of commercial banks, thereby

42 affecting economic growth, especially for countries with economies that are heavily dependent on bank credit such as Vietnam (Beck & Levines, 2004) In addition, the study by Trung et al (2023) published in the banking journal also showed the impact of the CAR coefficient on the credit growth of Vietnamese commercial banks in the period 2005 - 2021 based on a research sample of

26 commercial banks with different characteristics in size, nature of ownership and business model The main result of this study is that it has shown that the increase in the CAR ratio helps Vietnamese commercial banks expand credit activities faster

In contrast to the FGLS model of ROA, the SIZE variable and LDR running GMM have no statistical significance at the coefficients of -0.000566 and 0.007847, respectively This indicates that when the bank holds high liquidity, the ratio of net income to total assets of the bank will decrease This result is consistent with Bordeleau & Graham's (2010) study of Canadian banks Banks with traditional business models rely primarily on mobilizing deposits and credit for profit-optimized banks with a higher level of liquidity The above view is even more accurate in the case of a capital market with good liquidity, banks only need to hold a certain amount of liquid assets to maximize profits In addition, the negative effect of scale on financial performance represents an increase in size but not an improvement in bank profits, and small banks operate more efficiently than large banks, indicating the offset of economic advantage by scope (Chi,

For the ROA(-1) variable, the statistical significance is at 5% This implies that when other conditions remain constant, if the ROA(-1) increases, the ROA rate for the current year of the enterprises increases This means that the bank's good business results in the previous period will be the premise and create conditions for the bank's activities in the next year, leading to an increase in the bank's performance in the next year

Table 18: GMM Regression Results of ROE model

ICT and LDR variables have a positive impact on the performance of commercial banks (ROA) during this period The increase in the value of these variables in the model will increase the value of the return on equity of the commercial bank The ICT and LDR variables were statistically significant at 1% with coefficients of 0.165738 and 0.225059, respectively On the contrary, the ITI and OP variables have a negative impact on the operational efficiency of commercial banks, the decrease of this variable in the model will increase the value of ROE, specifically as follows:

The statistically significant ICT variable with a significance of 1% and a coefficient of 0.165738 indicates that the bank's operating efficiency (ROE) and ICT are related in the same direction, if the ICT coefficient increases by 1%, it leads to an increase of 0.166% in the bank's operational efficiency This is in line with the H1 hypothesis originally proposed The statistically significant LDR variable with a significance of 1% and a coefficient of 0.225059 indicates that the bank's operating efficiency (ROE) and LDR are related in the same direction, if the LDR coefficient increases by 1%, it leads to an increase of 0.225% in the bank's operational efficiency

In contrast, the variables ITI and OP are statistically significant with a significance of 1% and 5% coefficients of -0.149385 and -0.006580, respectively, indicating that the bank's operating efficiency (ROE) and ITI and OP are inversely related, if the ITI ratio increases by 1%, it leads to a decrease of 0.149% in the bank's operating efficiency and if the OP ratio increases by 5%, it leads to a decrease of 0.006% the bank's operational efficiency This result is inconsistent with the Hypothesis 2 and Hypothesis 3

The SIZE and CAR variables are not statistically significant with coefficients of -0.000730 and -0.002146, respectively A bank with a high capital adequacy ratio will not have a clear impact on the bank's operational efficiency The above results are relatively similar to the results of research by Berger & Bonaccorsi (2006), Dao & Nguyen (2020) and Angelini et al (2011) There are several main reasons why when a bank achieves a higher capital adequacy ratio, or holds more capital, it will be detrimental to the bank's business results and cash flow First, creditors have an information advantage over capital contributors due to the existence of many terms in the borrowing market and contribute to the health of financial markets (Leland & Рyle, 1977) There are many studies that emphasize the role of debt management (Hаrt & Moore, 1995), when managers often seek to reduce market discipline by building a neck cushion Second, banks may also have the problem of reducing their ability to generate money by holding too much capital (Diаmond & Rаjаn, 2001) and increasing the marginal expenditure of capital holdings The theory of representations and the theory of trade-off have similar expectations In other models, the CАR estimation results have a variation in sign depending on the estimate used by the author

For the ROE(-1) variable, the statistical significance is at 1% This implies that when other conditions remain constant, if the ROE(-1) increases, the ROE ratio of the enterprises in the current year increases This means that the bank's good business results in the previous period will be the premise and create conditions for the bank's activities in the next year, leading to an increase in the bank's performance in the next year

Table 19: GMM Regression Results of NIM model

Note: The value in square brackets [] is the corresponding t-statistics

Results of testing for GMM

ROA model

Table 20: Arellano Bond test of ROA model m-Statistic P-value

With the results of the J-statistic test, the p-value = 0.4610 > 0.05 (results are shown in Tables 23 and 24 under the Appendix 1) we accept the H0 hypothesis, that is, the model has no endogenous phenomena, which means that the model form is consistent with the surveyed data In addition, based on the p-values of AR(1): 0.0845 < 0.1 and AR(2): 0.9523 > 0.1, we can refute the hypothesis that a model exists with a chain correlation phenomenon in the Arellano-Bond model Therefore, after testing the above three factors, we have reliable results and can use the GMM estimation results in the case of the ROA model.

ROE model

Table 21: Arellano Bond test of ROE model m-Statistic P-value

With the results of the J-statistic test, the p-value = 0.3549 > 0.05 (results are shown in Tables 25 and 26 under the Appendix 1) we accept the H0 hypothesis, that is, the model has no endogenous phenomena, which means that the model form is consistent with the surveyed data In addition, based on the p-values of AR(1): 0.0174 < 0.05 and AR(2): 0.6906 > 0.1, we can refute the hypothesis that a model exists with a chain correlation phenomenon in the Arellano-Bond model Therefore, after testing the above three factors, we have reliable results and can use the GMM estimation results in the case of the ROE model.

NIM model

Table 22: Arellano Bond test of NIM model m-Statistic P-value

With the results of the J-statistic test, the p-value = 0.3658 > 0.05 (results are shown in Tables 27 and 28 under the Appendix 1) we accept the H0 hypothesis, that is, the model has no endogenous phenomena, which means that the model form is consistent with the surveyed data In addition, based on the p-values of AR(1): 0.0025 < 0.01 and AR(2): 0.7335 > 0.1, we can refute the hypothesis that a model exists with a chain correlation in the Arellano-Bond model Therefore, after testing the above three factors, we have reliable results and can use the GMM estimation results in the case of the NIM model

Recommendation

The research paper has evaluated the impact of ICT on the operational efficiency of 24 commercial banks in Vietnam From the results of the study, the team realized the existence of the impact of ICT on commercial banks, but the impact is still insignificant However, it is undeniable that ICT still plays an important role in improving the operation of commercial banks Information and Communication Technologies can help strengthen the bank's ability to manage and monitor operations, while minimizing information system failures and optimizing business processes Therefore, for commercial banks that have applied ICT to their operations, their operational efficiency has been partially improved Advanced technologies help to speed up transactions, increase flexibility, and minimize waiting times for clients In addition, these technologies help to enhance customer outreach, improve customer experience, and increase customer engagement Therefore, the application of ICT to the operation of commercial banks in Vietnam still has the potential to develop and will play an important role in the competitiveness of banks in the future However, the application of this technology needs to be done wisely and strategically, and it is necessary to ensure information security to avoid security risks and protect customer information

The results of the analysis and assessment of the impact of ICT on the operational efficiency of commercial banks in Vietnam suggest that ICT has the ability to stimulate the development of banking industries However, the Government needs to take measures to overcome the shortcomings to enhance the impact of ICT as follows:

Firstly, to increase investment in Information and Communication Technology, train professional human resources, promote the application of new technological solutions and develop Information and Communication Technology infrastructure to create favourable conditions for commercial banks to use advanced technologies Increasing investment in Information and Communication Technology is essential to put new technologies into practical application at commercial banks, helping to enhance the efficiency of business activities Information and Communication Technology can provide tools and solutions to enhance management, optimize workflows, improve customer service, and create new products and services Investing in Information and Communication Technology will help commercial banks make the most of the potential of technology to improve competition and respond quickly to market requirements At

49 the same time, training professional human resources will help ensure the sustainable development of the Information and Communication Technology industry and create a high-quality human resource team to meet the increasing needs of commercial banks Training professional human resources in the field of Information and Communication Technology means providing the necessary knowledge and skills to work with new technologies, develop and manage Information Technology systems, thereby promoting creativity and absorption of advanced technologies In conclusion, the development of Information and Communication Technology infrastructure is very necessary to create favourable conditions for commercial banks to use advanced technologies, in which Information and Communication Technology infrastructure includes network infrastructure, server systems, etc storage systems and other technologies Investing in ICT infrastructure helps ensure the stability, reliability and safety of the system, and creates a development environment for deploying and exploiting new technology solutions

Secondly, the Government needs to promote investment in Information and Communication

Technology infrastructure to provide a favourable business environment and ensure the confidentiality of transaction information of commercial banks Because the application of new technological solutions and the development of Information and Communication Technology infrastructure are also very important to create favourable conditions for commercial banks to use advanced technologies Specifically, investing in this infrastructure helps provide the resources and technology needed to process and store transaction information effectively A strong and stable ICT infrastructure will help enhance transaction processing capabilities, reduce response time and increase system availability At the same time, security is also an important factor, and investing in ICT infrastructure provides advanced security measures to ensure the safety of banks' transaction information In addition, the Government needs to increase innovation in business through the application of ICT which is also an important measure to enhance the impact of ICT on the business activities of commercial banks in Vietnam Advanced technologies such as online payment systems, e-banking networks, and mobile applications can be applied to enhance business innovation and optimize processes For example, commercial banks can implement an online payment system to help banks provide fast and convenient payment services to customers The e- banking network helps to create a system of links between banks and provide online banking services The mobile app provides access to remote banking services and creates a convenient and flexible banking experience for customers At the same time, security is also an important factor,

50 and investing in ICT infrastructure provides advanced security measures to ensure the safety of banks' transaction information

Thirdly, the Government needs to have Information and Communication Technology policies to first create a platform and environment to encourage investment in Information and Communication Technology The ICT investment environment can be enhanced through solutions such as digital transformation for businesses, e-government or current policies such as financial support and trade promotion for Information and Communication Technology enterprises The government can also promote the adoption of e-government, in which public services are provided online and online to create convenience and enhance interaction between the government and businesses Specifically, financial support and trade promotion are an important part of the recommended policy The Government can provide financial support policies such as loans with preferential interest rates or investment capital support for Information and Communications Technology enterprises In addition, trade promotion can be carried out through the organization of events, exhibitions or training to raise awareness and market access of Information and Communication Technology enterprises In addition, the Government needs to strengthen the construction of digital infrastructure, especially the development of communication networks and the Internet Communication networks and the Internet play an important role in creating conditions for businesses to access and use Information and Communication Technology Investing in the development of network infrastructure and the Internet will create a stable and fast connection environment and help reduce access costs and enhance access to ICT technologies and services This is especially important for small and medium-sized businesses, as a well-developed digital infrastructure will facilitate their access to and application of technology to enhance performance and competitiveness

Finally, the Government needs to review and improve regulations and policies related to

Information and Communication Technology, especially in the field of information security and data management recommend proposals to review and improve regulations and policies related to information security This includes implementing measures to protect sensitive information, such as data encryption, user authentication, and access authentication The Government needs to ensure that Information and Communications Technology enterprises comply with regulations on information security and are responsible for protecting customer information In addition, it is

51 necessary to develop clear guidelines and regulations to manage data reliably, consistently, and in compliance with legal regulations Ensuring data integrity and security is important for users and customers to have confidence in using Information and Communication Technology products and services The government may consider introducing regulations on data management, including the collection, storage, processing and sharing of data, to ensure the appropriateness and security of the use of personal information and business data This will help strengthen the confidence of users and customers in the use of ICT products and services and at the same time encourage the use and development of ICT in business and society.

Conclusion

The findings allow readers to better understand the impact of ICT investment and adoption on commercial banking business in the context of an emerging and dynamic economy like Vietnam Our results also suggest to commercial bank leaders about ICT investment strategies to catch up and lead in the era of strong development of Fintech technology and the digital economy set out by the Government in the digital development strategy to 2045 We believe that the findings of this study are not only useful in the context of commercial banks or the Vietnamese context but can also be extended to enterprises in the manufacturing sector or commercial banks in countries in the region and the world In particular, it is possible to expand the survey to securities companies or some specific industries that are strongly affected by innovation in information and communication technology such as consumer goods and retail

Alabi, 2005, ‘Information Technology and E-Marketing: Prospects and Challenges for Marketing Practitioners in an Emerging Economy’, An Unpublished Paper delivered at the 2005 MMN Marketing Educator Conference, University of Lagos, Nigeria, [online] [Accessed on 27 July

Abebe, G., 2016 The Impact of Information and Communication Technology on Performance of Commercial banks in Ethiopia, Addis Ababa: Addis Ababa University [Accessed on 29 Aug

Aliyu & Tasmin, 2012, ‘The Impact of Information and Communication Technology on Banks‟ Performance and Customer Service Delivery in the Banking Industry’, International Journal of Finance, Insurance and Risk Management, 2(1), pp 81-83, [online] [Accessed on 29 Aug 2024]

Al-Busaidi & Al-Muharrami, 2021, ‘Beyond profitability: ICT investments and financial institutions performance measures in developing economies’, Journal of Enterprise Information

Management 34: 900–21, [online] Available at: https://www.emerald.com/insight/content/doi/10.1108/JEIM-09-2019-0250/full/html [Accessed on 29 Aug 2024]

Angelini, P., Neri, S., & Panetta, F (2011) “Monetary and macroprudential policies”, Bank of Italy Temi di Discussione (Working Paper) No, 801[Accessed on 29 Aug 2024]

Almaqtari et al, 2019, ‘The determinants of profitability of Indian commercial banks: A panel data approach.’, International Journal of Finance and Economics 24: 168–85, [online] Available at: The determinants of profitability of Indian commercial banks: A panel data approach - Almaqtari

- 2019 - International Journal of Finance & Economics - Wiley Online Library [Accessed on 29 Aug 2024]

Alt at el, (2018), ‘FinTech and the transformation of the financial industry’, Electronic Markets, 28(3), 235-243 [Accessed on 29 Aug 2024]

Anh Ngoc & Diem, 2022 Trí tuệ nhân tạo và các ứng dụng trong lĩnh vực ngân hàng, Tạp chí Ngân hàng [Accessed on 29 Aug 2024]

Amankwah‐Amoah, J (2019), ‘Technological revolution, sustainability, and development in Africa: Overview, emerging issues, and challenges’, Sustainable Development, 27(5), 910-922 [Accessed on 29 Aug 2024]

Adane et al, 2021 Determinants of automated teller machine deployment in commercial banks of Ethiopia Heliyon 7: e07712, [online] Available at: Redirecting (elsevier.com) [Accessed on 29 Aug 2024]

Barpanda, Saswat, and S Athira, 2022, ‘Cause of Attrition in an Information Technology-Enabled Services Company: A Triangulation Approach.’, International Journal of Human Capital and Information Technology Professionals (IJHCITP) 13: 1–22, [online] Available at: https://www.igi-global.com/gateway/article/296691 [Accessed on 29 Aug 2024]

Ben Naceur et al, 2008 The determinants of commercial bank interest margin and profitability: Evidence from Tunisia.’, Frontiers in Finance and Economics 5: 106–30, [online] Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id6365 [Accessed on 29 Aug 2024]

Berger & Humphrey, 1997, ‘Efficiency of financial institutions: International survey and directions for future research’, European Journal of Operational Research 98, 98(2), pp 175-212, [online] [Accessed on 29 Aug 2024]

Berger, A, & Allen N Berger a,b,*, Emilia Bonaccorsi, E di Рatti (2006) “Caрital structure and firm рerformance: A new aррroach to testing agency theory and an aррlication to the banking industry”., Journal of Banking & Finance 30 (2006) 1065- 1102[Accessed on 29 Aug 2024]

Beck, T., & Levine, R (2004) Stock markets, banks, and growth: Panel evidence Journal of Banking & Finance, 28(3), pages 423-442 [Accessed on 29 Aug 2024]

Balasubramaniam, 2012, ‘Non-performing assets and profitability of commercial banks in India: Assessment and emerging issues.’, Abhinav, National Monthly Refereed Journal of Research in Commerce & Management 1: 41–52, [online] [Accessed on 29 Aug 2024]

Bordeleau, É., & Graham, C (2010) The impact of liquidity on bank profitability (No 2010-38), Bank of Canada [Accessed on 29 Aug 2024]

Chhaidar et al, 2022 The effect of financial technology investment level on european banks’ profitability Journal of the Knowledge Economy, 1–23, [online] Available at: The Effect of Financial Technology Investment Level on European Banks’ Profitability | Journal of the Knowledge Economy (springer.com) [Accessed on 29 Aug 2024]

Chi, 2022, ‘Bằng phương pháp thống kê, phân tích thực trạng quy mô tổng tài sản và tương quan cặp giữa quy mô và hiệu quả tài chính, đại diện bởi tỷ số lợi nhuận ròng trên tài sản (ROA), kết hợp lý thuyết và các nghiên cứu thực nghiệm.’, sbv.gov.vn, [online] Available at: https://sbv.gov.vn/webcenter/portal/vi/menu/fm/ddnhnn/nctd/nctd_chitiet?leftWidth %25&sh owFooterse&showHeaderse&dDocName=SBV530222&rightWidth=0%25¢erWidt h%25&_afrLoopb3886695384774#%40%3F_afrLoop%3D623886695384774%26center Width%3D80%2525%26dDocName%3DSBV530222%26leftWidth%3D20%2525%26rightWid th%3D0%2525%26showFooter%3Dfalse%26showHeader%3Dfalse%26_adf.ctrl- state%3D1dqd2rdd8m_51 [Accessed on 29 Aug 2024]

CFA Team, 2024, ‘Return on Equity (ROE)’, corporatefinanceinstitute.com, [online] Available at: https://corporatefinanceinstitute.com/resources/accounting/what-is-return-on-equity-roe/ [Accessed on 27 July 2024]

Del Gaudio et al, 2021, ‘How do mobile, internet and ICT diffusion affect the banking industry?

An empirical analysis.’, European Management Journal 39: 327–32, [online] Available at: https://www.sciencedirect.com/science/article/abs/pii/S0263237320301122?via%3Dihub

Dao et al, 2020 “Determinants of Profitability in Commercial Banks in Vietnam, Malaysia and Thailand”., Journal of Asian Finance, Economics and Business, 7(4), 133- 143, [online] Available at: httрs://doi.org/10.13106/jafeb.2020.vol7.no4.133 [Accessed on 29 Aug 2024]

Diamond, D, W., & Rajan, R G (2001) “Liquidity risk, liquidity creation, and financial fragility:

A theory of banking” Journal of political Economy, 109(2), 287-327 [Accessed on 29 Aug 2024]

Dat, 2019, “Nguồn vốn ngân hàng ảnh hưởng đến khả năng sinh lời và rủi ro tín dụng của ngân hàng thương mại cổ phần Việt Nam”, Tạp chí Nghiên cứu Tài chính – Marketing, 52, 12-21 [Accessed on 29 Aug 2024]

Gupta at el, (2018), ‘Information technology and profitability: evidence from Indian banking sector’, International Journal of Emerging Markets, 13(5), 1070-1087 [Accessed on 29 Aug 2024]

Hart, O, D., & Moore, J (1994) “Debt and seniority: An analysis of the role of hard claims in constraining management.” [Accessed on 29 Aug 2024]

Hughes and Mester, 2008 Efficiency in Banking: Theory, Practice, and Evidence, Federal Reserve Bank of Philadelphia [Accessed on 29 Aug 2024]

Kanoujiya et al, 2021, ‘Banks in India: A balancing act between profitability, regulation and NPA’,

Vision, 09722629211034417, [online] Available at: https://journals.sagepub.com/doi/10.1177/09722629211034417 [Accessed on 29 Aug 2024]

Ky et al, (2019), ‘Is Fintech Good for Bank Performance? The Case of Mobile Money in the East African Community’, [online] Available at: http://dx.doi.org/10.2139/ssrn.3401930 [Accessed on 29 Aug 2024]

Le, L, Nasir, M and Huynh, T, (2020), “Capital requirements and banks performance under Basel- III: A comparative analysis of Australian and British banks”, The Quarterly Review of Economics and Finance [Accessed on 29 Aug 2024]

Laudon, 2001, ‘Management Information Systems: Organization and Technology in the Network Enterprise (4th ed.)’, Prentice Hall International in USA, [online] [Accessed on 27 July 2024]

Leland, H E., & Pyle, D H (1977) “Informational asymmetries, financial structure, and financial intermediation”., The journal of Finance, 32(2), 371-387 [Accessed on 29 Aug 2024]

Mahboub, R., 2018 The Impact of Information and Communication Technology Investments on the Performance of Lebanese Banks European Research Studies Journal, XXI(4), pp 435-458, [online] [Accessed on 29 Aug 2024]

Mention, A.L (2021), ‘The age of FinTech: Implicatio ns for research, policy and practice’, The Journal of FinTech, 1(01), [online] Available at: https://doi.org/10.1142/S2705109920500029 [Accessed on 29 Aug 2024]

McCarthy et al (2014), ‘Emerging innovation in emerging economies: can institutional reforms help Russia break through its historical barriers?’, Thunderbird International Business Review, 56(3), 243-260 [Accessed on 29 Aug 2024]

Minh Ngoc & Tien Thanh, 2021 Thực trạng ứng dụng công nghệ trong ngân hàng thương mại, nghiên cứu trường hợp Ngân hàng Công Thương Việt Nam Kỷ yếu Hội thảo Khoa học Quốc gia: Quản trị thông minh trong môi trường phức hợp toàn cầu lý luận và thực tiễn, tr 424-433, Đại học Quốc gia Hà Nội [Accessed on 29 Aug 2024]

Noss, J., & Toffano, P (2016) Estimating the impact of changes in aggregate bank capital requirements on lending and growth during an upswing Journal of Banking & Finance, 62, pages 15-27 [Accessed on 29 Aug 2024]

Phan Anh, 2023, ‘Tác động của hiệu quả vốn sử dụng đến kết quả hoạt động của các ngân hàng thương mại Việt Nam’, kinhtevadubao, [online] Available at: https://kinhtevadubao.vn/tac-dong- cua-hieu-qua-von-su-dung-den-ket-qua-hoat-dong-cua-cac-ngan-hang-thuong-mai-viet-nam- 27866.html#:~:text=K%E1%BA%BFt%20qu%E1%BA%A3%20nghi%C3%AAn%20c%E1%B B%A9u%20ch%E1%BB%89,to%C3%A0n%20v%C3%A0%20hi%E1%BB%87u%20qu%E1% BA%A3%20nh%E1%BA%A5t [Accessed on 29 Aug 2024]

Pierri, N & Timmer, Y (2022), ‘The importance of technology in banking during a crisis’, Journal of Monetary Economics, 128, 88-104 [Accessed on 29 Aug 2024]

Prud’homme van Reine, P (2015), ‘A networking culture to benefit from open innovation-a comparison between technology and business services industries in The Netherlands’, Journal of Innovation Management, 3(2), 71-105 [Accessed on 29 Aug 2024]

Schueffel, P (2016), ‘Taming the beast: A scientific definition of fintech’, Journal of Innovation Management, 4(4), 32-54 [Accessed on 29 Aug 2024]

Sanga & Aziakpono, 2022 The impact of technological innovations on financial deepening: Implications for SME financing in Africa African Development Review 34: 429–42, [online] Available at: The impact of technological innovations on financial deepening: Implications for SME financing in Africa - Sanga - 2022 - African Development Review - Wiley Online Library [Accessed on 29 Aug 2024]

Saksonova, 2014 The role of net interest margin in improving banks’ asset structure and assessing the stability and efficiency of their operations Procedia-Social and Behavioral Sciences 150: 132–

41, [online] Available at: The Role of Net Interest Margin in Improving Banks’ Asset Structure and Assessing the Stability and Efficiency of their Operations - ScienceDirect [Accessed on 29 Aug 2024]

Table 23: Testing reasonable instrument variables of ROA model Cross-sections included Instrument rank

Table 24: J-statistic test of ROA model

Table 25: Testing reasonable instrument variables of ROE model Cross-sections included Instrument rank

Table 26: J-statistic test of ROE model

Table 27: Testing reasonable instrument variables of NIM model Cross-sections included Instrument rank

Table 28: J-statistic test of NIM model

Table 29: List of Commercial Banks

No Commercial Bank Name Stock codes on exchanges

1 Bank For Investment and Development of Vietnam BID

2 Bank for Foreign Trade of Vietnam VCB

3 Vietnam Joint Stock Commercial Bank for Industry and

4 Vietnam Prosperity Joint Stock Commercial Bank VPB

5 Kien Long Commercial Joint Stock Bank KLB

6 Ho Chi Minh Development Joint Stock Commercial Bank HDB

7 Military Commercial Joint Stock Bank MBB

8 Tien Phong Commercial Joint Stock Bank TPB

9 Sai Gon Thuong Tin Commercial Joint Stock Bank STB

10 An Binh Commercial Joint Stock Bank ABB

11 Saigon Commercial Joint Stock Bank SCB

12 Vietnam Bank for Agriculture and Rural Development ARG

13 Vietnam Commercial Joint Stock Export Import Bank EIB

14 Nam A Commercial Joint Stock Bank NAB

15 Ban Viet Commercial Joint Stock Bank BVH

16 Viet Capital Commercial Joint Stock Bank BVB

17 Saigon Hanoi Commercial Joint Stock Bank SHB

18 Orient Commercial Joint Stock Bank OCB

19 Vietnam Asia Commercial Joint Stock Bank VAB

20 Vietnam International Commercial Joint Stock Bank VIB

21 Asia Commercial Joint Stock Bank ACB

22 Vietnam Technological and Commercial Joint Stock Bank TCB

23 Southeast Asia Commercial Joint Stock Bank SSB

24 Saigon Bank for Industry and Trade SGB

Variable Obs Mean Std Dev Min Max

sum ROA ROE ICT ITI OP SIZE NIM CAR LDR

ROA ROE NIM ICT ITI OP SIZE

pwcorr ROA ROE NIM ICT ITI OP SIZE CAR LDR,sig

ROA Coef Std Err t P>|t| Beta Total 025715292 287 0000896 Root MSE = 00811

Adj R-squared = 0.2662 Residual 018475718 281 00006575 R-squared = 0.2815 Model 007239574 6 001206596 Prob > F = 0.0000 F(6, 281) = 18.35 Source SS df MS Number of obs = 288

reg ROA ICT ITI OP SIZE CAR LDR,beta

Cameron & Trivedi's decomposition of IM-test

Prob > chi2 = 0.0017 chi2(27) = 53.61 against Ha: unrestricted heteroskedasticity

White's test for Ho: homoskedasticity

Wooldridge test for autocorrelation in panel data

xtserial ROA ICT ITI OP SIZE CAR LDR

F test that all u_i=0: F(24, 257) = 8.09 Prob > F = 0.0000 rho 57480195 (fraction of variance due to u_i) sigma_e 00640007 sigma_u 00744129

ROA Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = 0.1302 Prob > F = 0.0000

F(6,257) = 14.43 overall = 0.2499 max = 12 between = 0.0074 avg = 11.5 within = 0.2520 min = 1

Group variable: firmbanking Number of groups = 25

Fixed-effects (within) regression Number of obs = 288

xtreg ROA ICT ITI OP SIZE CAR LDR,fe rho 53925261 (fraction of variance due to u_i) sigma_e 00640007 sigma_u 00692388

ROA Coef Std Err z P>|z| [95% Conf Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000

Wald chi2(6) = 89.07 overall = 0.2570 max = 12 between = 0.0105 avg = 11.5 within = 0.2516 min = 1

Random-effects GLS regression Number of obs = 288

xtreg ROA ICT ITI OP SIZE CAR LDR,re

Test: Ho: difference in coefficients not systematic

B = inconsistent under Ha, efficient under Ho; obtained from xtreg b = consistent under Ho and Ha; obtained from xtreg

ROA[firmbanking,t] = Xb + u[firmbanking] + e[firmbanking,t]

Breusch and Pagan Lagrangian multiplier test for random effects

xtserial ROA ICT ITI OP SIZE CAR LDR

ROA Coef Std Err z P>|z| [95% Conf Interval]

Wald chi2(6) = 89.92 max = 12 avg = 11.95833 min = 11

Estimated coefficients = 7 Obs per group:

Estimated autocorrelations = 1 Number of groups = 24

Estimated covariances = 24 Number of obs = 287

Correlation: common AR(1) coefficient for all panels (0.5227)

Cross-sectional time-series FGLS regression

(note: 1 observations dropped because only 1 obs in group)

xtgls ROA ICT ITI OP SIZE CAR LDR,panel(h)corr(ar1)

esttab ols fixed random gls,r2 star(* 0.1 ** 0.05 *** 0.01) brackets nogap compress

ROE Coef Std Err t P>|t| Beta Total 3.36558607 287 011726781 Root MSE = 09544

Adj R-squared = 0.2232 Residual 2.55978498 281 009109555 R-squared = 0.2394 Model 805801093 6 134300182 Prob > F = 0.0000 F(6, 281) = 14.74 Source SS df MS Number of obs = 288

reg ROE ICT ITI OP SIZE CAR LDR,beta

xtserial ROE ICT ITI OP SIZE CAR LDR

ROE Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = 0.0238 Prob > F = 0.0000

xtreg ROE ICT ITI OP SIZE CAR LDR,fe rho 20387442 (fraction of variance due to u_i) sigma_e 08118952 sigma_u 04108569

ROE Coef Std Err z P>|z| [95% Conf Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000

xtreg ROE ICT ITI OP SIZE CAR LDR,re

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

Modified Wald test for groupwise heteroskedasticity

xtserial ROE ICT ITI OP SIZE CAR LDR

ROE Coef Std Err z P>|z| [95% Conf Interval]

xtgls ROE ICT ITI OP SIZE CAR LDR,panel(h)corr(ar1)

esttab pool1 fe1 re1 gls1, r2 star(* 0.1 ** 0.05 *** 0.01) brackets nogap compress

NIM Coef Std Err t P>|t| Beta Total 049236337 287 000171555 Root MSE = 01014

Adj R-squared = 0.4009 Residual 028878671 281 000102771 R-squared = 0.4135 Model 020357666 6 003392944 Prob > F = 0.0000 F(6, 281) = 33.01 Source SS df MS Number of obs = 288

reg NIM ICT ITI OP SIZE CAR LDR,beta

xtserial NIM ICT ITI OP SIZE CAR LDR

NIM Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = 0.2489 Prob > F = 0.0000

xtreg NIM ICT ITI OP SIZE CAR LDR,fe rho 53158659 (fraction of variance due to u_i) sigma_e 00697247 sigma_u 00742778

NIM Coef Std Err z P>|z| [95% Conf Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000

xtreg NIM ICT ITI OP SIZE CAR LDR,re

NIM[firmbanking,t] = Xb + u[firmbanking] + e[firmbanking,t]

Breusch and Pagan Lagrangian multiplier test for random effects

xtserial NIM ICT ITI OP SIZE CAR LDR

NIM Coef Std Err z P>|z| [95% Conf Interval]

xtgls NIM ICT ITI OP SIZE CAR LDR,panel(h)corr(ar1)

esttab pool2 fe2 re2 gls2, r2 star(* 0.1 ** 0.05 *** 0.01) brackets nogap compress

Method: Panel Generalized Method of Moments

White period (period correlation) instrument weighting matrix

White period (cross-section cluster) standard errors & covariance (d.f. corrected)

Standard error and t-statistic probabilities adjusted for clustering

Constant added to instrument list

Variable Coefficient Std Error t-Statistic Prob

Effects Specification Cross-section fixed (first differences)

Mean dependent var 0.000848 S.D dependent var 0.003408 S.E of regression 0.004721 Sum squared resid 0.004636

Arellano-Bond Serial Correlation Test

Test order m-Statistic rho SE(rho) Prob AR(1) -1.725264 -0.000884 0.000512 0.0845AR(2) 0.059838 0.000018 0.000295 0.9523