(Luận văn thạc sĩ) bank restructuring and bank efficiency the case of vietnam

INTRODUCTION

OVERVIEW

The banking system plays a vital role in regulating capital flows within the economy and fostering both economic and social development in Vietnam Over the past sixty years, the country has seen the establishment and growth of four state-owned commercial banks, thirty joint-stock commercial banks, joint-venture banks, and banks with full foreign ownership, leading to the gradual evolution of a robust banking structure.

In 2006, Vietnam's accession to the WTO marked the beginning of its market economy transformation, necessitating a restructuring of the banking system to address economic challenges and public demands The global financial crisis of 2008 exacerbated risks within the banking sector, leading to deteriorating performance characterized by profit declines, low credit growth, and increasing bad debts In response, the central bank and commercial banks have been implementing measures to enhance operational efficiency and navigate these difficulties This research aims to assess the recent restructuring of Vietnam's banking system, drawing on previous studies from other countries to inform the chosen model and methodology Through quantitative analysis, the study seeks to provide recommendations for sustainable improvements in the banking sector's restructuring process.

Bank restructuring is closely linked to efficiency, as demonstrated by various studies across global economies The impact of restructuring measures on bank efficiency can yield both positive and negative outcomes These results are influenced by the specific economic conditions of each country, including its current status and the challenges it faces.

Numerous studies indicate a positive relationship between bank restructuring and efficiency, with significant improvements observed in Turkey post-restructuring (Zaim, 1995) Privatization and mergers and acquisitions are identified as effective strategies for enhancing banking efficiency Conversely, some researchers argue that restructuring may not influence operational efficiency and could even lead to a decline, as evidenced by Elyasiani and Mehdian (1995), who found no performance improvement in U.S banks following restructuring This thesis aims to explore the connection between bank restructuring and efficiency in Vietnam, an emerging economy, to determine whether this relationship is positive or negative.

This thesis analyzes the restructuring of the Vietnamese banking system from 2007 to 2015, focusing on the financial structure before and during this period It evaluates the impact of restructuring measures on the performance and operational efficiency of commercial banks Additionally, the study investigates how reforms have influenced Vietnam's banking structure and performance By comparing the banking system's structure from 2001 to 2006 and during the restructuring phase, it tests various hypotheses, including quiet-life, efficient structure, and market structure conduct, to assess their outcomes.

MOTIVATIONS AND RESEARCH QUESTIONS

This section reviews existing literature to pinpoint gaps that inform the objectives and research questions of the study It is organized into three parts: the methods of bank restructuring (Section 1.2.1) and the relationship between structure and performance within the banking system (Section 1.2.2).

1.2.1 The method of bank restructuring

A restructuring program is essential for addressing the challenges faced by distressed banks (McComb, Gruben, & Welch, 2002) Key methods of restructuring include merging domestic banks, allowing foreign bank entries, state intervention, and privatizing state-owned commercial banks Research by Hawkin and Turner (1999) suggests that domestic mergers are often the most cost-effective solution for restructuring the banking system Additionally, Berger et al (1999) found that mergers can enhance efficiency by improving the risk-return trade-offs through greater diversification A study by Krishnasamy et al (2004) highlighted the importance of technology in boosting productivity in Malaysia’s post-merger banks during 2000-2001 In Asia, Unite and Sullivan (2003) noted that foreign banks can easily expand their operations Furthermore, Daniel (1997) emphasized that governments can enhance banks' effectiveness through various strategies, such as Indonesia's recapitalization program, where owners contributed 20% of the capital shortfall (Fane and McLeod, 2002), and Korea's government intervention in purchasing non-performing loans and supporting private banks' recapitalization efforts.

(2000) pointed out that the inefficient operation of state-commercial banks accelerated the process of recession in Brazil h

Tran et al (2015) explore the effects of restructuring on the efficiency of commercial banks in Vietnam, revealing fluctuations in banking efficiency However, their study does not distinguish between the influences of bank restructuring and environmental factors on overall bank efficiency.

Research on bank restructuring methods has been conducted across various countries with diverse economies, yet there is a lack of studies focusing on their impact in smaller transitioning economies like Vietnam Previous research often fails to distinguish between the effects of bank restructuring and environmental factors, such as financial crises and a weak domestic economy, on bank efficiency This gap in the literature inspires the formulation of Research Question 1 (RQ1).

The restructuring measures in Vietnam, including government intervention, mergers and acquisitions of commercial banks, and the privatization of state-owned banks, significantly impact the performance of commercial banks during the study period These strategic actions aim to enhance operational efficiency, improve financial stability, and foster competitive banking environments By analyzing these factors, we can better understand how such restructuring efforts influence the overall effectiveness and growth of commercial banks in Vietnam.

1.2.2 The structure and performance relationship in the banking system

A study by Chen, Skully, and Brown (2005) found that large state-owned banks and smaller banks in China operated more effectively than average banks Similarly, Fu and Heffernan (2009) reported that from 1985 to 2002, the Chinese banking system adhered to the Quiet Life Hypothesis This aligns with Berger and Hannan's (1997) findings in the United States, which also supported the Quiet Life Hypothesis In Europe, Molyneux and Forbes (1995) indicated that banks followed traditional Structure-Conduct-Performance models, while Goldberg and Rai (1996) suggested that banks in high-concentration countries were influenced by the Market Power Hypothesis and X-Efficiency Hypothesis Additionally, Ho and Baxter (2011) analyzed the Vietnamese banking sector, revealing that reforms included restructuring the banking system, gradually opening to foreign investment, partially privatizing state-owned banks, and enhancing the capitalization of Vietnamese banks.

This thesis seeks to explore the impact of first-stage reforms on bank structure and performance in Vietnam, addressing a notable gap in existing studies that have not utilized hypotheses to analyze this relationship This inquiry is driven by the need to understand how these reforms affect the banking sector, leading to the formulation of Research Question Two (RQ2).

RQ2: What are the effects of reform on Vietnam's commercial bank structure and performance?

RESEARCH OBJECTIVES

Previous studies have explored banking system restructuring, but none have specifically addressed Vietnam's unique economic context Vietnam operates as a mixed economy, recognized by the World Trade Organization as a low-level, transitioning economy To enhance integration into regional and global markets, the development of its monetary and banking sectors is essential Despite a complicated legal framework that is gradually improving, efforts are underway to promote the credit and banking markets, diversify banking products and services, and reduce bad debt The State Bank of Vietnam has implemented various measures aimed at enhancing the banking system, including restructuring initiatives This research seeks to evaluate the positive effects of this restructuring on the banking system, focusing on its impact on banking performance and the overall structure of Vietnam's commercial banks.

DATA AND METHODOLOGY

This section summarizes the data and methodology used to test the developed hypotheses to help answering the two research questions and the empirical findings of this thesis

The research data were sourced from reputable institutions such as Orbis Bank Focus and the IMF The initial phase involved identifying the research data sources and methodologies for investigation Subsequently, raw data was meticulously collected, processed, and analyzed through counting, synthesis, selection, editing, encoding, and evaluation Visual tools like tables and graphs were employed to enhance the content analysis, ensuring the reliability and clarity of the research findings.

This thesis evaluates the performance of commercial banks using Data Envelopment Analysis (DEA) and Stochastic Frontier Analysis (SFA) methods To address the research question, it employs Random Effects and Three-Stage DEA, incorporating environmental factors to test the hypotheses of Research Question 1 Additionally, the study utilizes panel data alongside DEA, Generalized Least Squares (GLS), and Ordinary Least Squares (OLS) methodologies to derive comprehensive results.

4 General Least Squares h approaches were applied to estimate the equations in RQ2 According to Greene

(2003), using panel data would allow us to know the difference among behaviour, individuals or time periods Finally, a Hausman test was used to identify the optimal model.

STRUCTURE OF THE THESIS

The thesis is organized into several key chapters: Chapter 2 reviews existing literature on restructuring, which aids in formulating testable hypotheses for the two research questions Chapter 3 focuses on the banking case in Vietnam during the study period Chapter 4 details the sample data and methodology employed to test the hypotheses related to the three research questions Finally, Chapter 5 presents the empirical results from the hypothesis testing of these questions.

LITERATURE REVIEW

INTRODUCTION

Chapter 1 provided an outline of this thesis and identified its two research questions In Chapter 2, the literature review is presented This chapter reviews the academic literature related to these three thesis research questions and develops hypotheses to test and answer them Section 2.2 reviews the literature relating to factors that impact the bank efficiency The theoretical issues and literature pertaining to bank restructuring and efficiency are discussed in Section 2.3 Next, Section 2.4 overviews the previous literature on method of bank restructuring Finally, Section 2.5 summarizes these hypotheses and their corresponding research questions In particular, the research is based on the theory of structured finance bank (Waxman, 1998; Dziobek and Ceyla, 1998).

BACKGROUND

The definition of restructuring commercial-banking system

According to Waxman (1998), restructuring the banking system addressed issues within specific components, particularly for banks at risk of bankruptcy, while the overall banking system continued to function effectively This process involves a comprehensive overhaul of all system components, including central banks, commercial banking institutions, banking-related social policies, development banks, and micro-credit organizations.

Pazarbasioglu and Dziobek (1997) emphasized that restructuring the banking system is essential for restoring solvency and profitability, enhancing the system's capacity, and improving operational efficiency This process involves financial restructuring, activity adjustments, and safety monitoring, all aimed at achieving these three objectives Similarly, Alexander (1997) noted that a comprehensive restructuring program typically encompasses macroeconomic measures, institutional reforms, and legal adjustments, significantly impacting macroeconomic stability, monetary policy, and balance of payments Such initiatives also enhance the efficiency and transparency of public policy and influence future financial market activities.

In conclusion, the restructuring of the banking system is a gradual process involving various legal measures and financial institutions aimed at addressing structural weaknesses This initiative seeks to enhance the operational efficiency of the banking system, improve access to banking services for labor and small to medium enterprises, and establish a solid foundation for the healthy development of both the banking sector and the economy The restructuring encompasses all types of banks, including the State Bank of Vietnam (SBV), commercial banks (CBs), domestically invested banks, and foreign-invested banks.

Features of the restructuring of commercial banks

Restructuring the commercial banking system differs significantly from restructuring other sectors such as trade, services, and telecommunications, primarily due to the extensive impact commercial banks have on the broader economy Unlike other sectors, which may not affect all areas uniformly, commercial banks engage in specialized commodity and currency trading, linking their activities closely with various sectors and society at large Studies indicate that the restructuring of commercial banks is characterized by two key features that highlight its unique role in economic stability and growth.

The influence of commercial banks is extensive, and a weak banking system can jeopardize various sectors, leading to a widespread risk of bankruptcy akin to a domino effect Consequently, it is crucial to implement drastic and radical restructuring of the banking system to prevent systemic failures.

To effectively restructure the national bank's status, it's essential to recognize that the banking system is interconnected with all economic sectors Therefore, achieving the goals of banking system reform requires the active participation, collaboration, and efficient management of various state agencies.

The causality of the restructuring

Kithinji, Mwangi, and Ogutu (2017) argue that a healthy economy relies on a robust banking system They emphasize that prior to the global financial crisis, it was essential to restructure the banking system to mitigate risks A banking system in crisis poses threats not only to economic stability but also to social well-being, as the failure of a major bank can have widespread repercussions Consequently, without addressing these inefficiencies and instabilities, it becomes challenging for a country to establish a stable and healthy banking environment conducive to economic growth.

The restructuring of the banking system is essential due to numerous structural issues that have led to inefficiencies, weak capital, liquidity shortages, and deteriorating asset quality, putting banks at risk of disintegration and bankruptcy These challenges arise from internal weaknesses, including unclear strategies, ineffective management, unsuitable financial structures, and inadequate risk control tools Additionally, poor human resource management and irrational coordination hinder effective operations Rapid expansion often overlooks critical details, leading to vulnerabilities such as fraud and financial misreporting Therefore, it is imperative for the government to initiate a comprehensive restructuring of banks and financial institutions to prevent potential disruptions in the mass credit system.

Inefficient and predictable macroeconomic regulatory policies can lead to misleading regulations, ultimately resulting in an ineffective banking system For instance, in a rapidly growing economy, if the government urges banks to increase credit without appropriate restraint, it can lead to lax lending practices This scenario may cause a rise in non-performing loan (NPL) ratios and inefficient debt management, resulting in liquidity loss for banks and potentially jeopardizing the stability of the entire financial system.

The objective of the restructuring

In light of the financial crisis, Alexander (1997) emphasizes the urgent need for restructuring the banking system to restore its functionality and ensure the consistent provision of banking services to the economy.

In order to do that, the necessary tasks are: h

- Maintaining the stability of the banking system to ensure the liquidity, the payments and the operation of the banking system and the whole economy

- Finding solutions for problems in a timely manner to prevent them from spreading

Restoring public financial intermediaries is a fundamental goal of restructuring, aimed at ensuring stability and confidence in the banking system By implementing these changes, the overall liquidity of the system can be stabilized, leading to improved bank credit levels and enhancing the confidence of all economic sectors in the banking sector.

Restructuring aims to strengthen the banking system while minimizing costs for the central bank, deposit insurance, and government The goal is to create the most efficient restructuring process, ultimately reducing financial burdens associated with these institutions.

Intervention in credit organizations is essential when their activities stray from fundamental economic functions, when emerging issues pose significant threats, or when there is a risk of systemic collapse.

Restructuring should be approached from multiple perspectives, establishing clear criteria that address the necessity for change It is essential to identify which aspects require restructuring and outline how banks will function post-restructuring Additionally, considerations must be made regarding the management of bad loans, ensuring that the process is conducted voluntarily, supported by a comprehensive roadmap and specific actionable steps.

To effectively manage banking crises, a robust legal framework is essential for government intervention in restructuring efforts, ensuring that the deposit insurance system is adequately capitalized for rapid and efficient response This intervention raises various legal challenges concerning existing shareholders, necessitating adjustments to relevant laws Additionally, it is crucial that any mergers promote the sustainable development of banks without disrupting the financial system Daniel suggested that the government could enhance bank recovery by acquiring debts from financially weakened institutions.

THE STRUCTURE AND PERFORMANCE RELATIONSHIP

This article explores the relationship between structure and performance in banking by examining three key hypotheses: the market power hypothesis, the efficiency structure hypothesis, and the Quiet-life hypothesis We will analyze existing literature to test these theories and understand their implications for the banking sector.

The market power hypothesis highlights a firm's capacity to influence product pricing through the manipulation of supply, demand, or both Firms with significant market power can control market prices, manage profit margins, and create barriers for new competitors entering the market.

Early empirical research in banking, particularly by Phil Molyneux and William Forbes in 1995, examined the traditional structure-conduct-performance (SCP) and relative efficiency (RE) hypotheses Their findings indicated that the traditional SCP paradigm effectively explains the market behavior of European banks.

In the 1993-2000period, the cost, the technical and allocative efficiency of 43 Chinese banks were examined by Xiaogang Chen, Michael Skully and Kym Brown

In 2005, an analysis was conducted to assess the impact of the Chinese government's 1995 deregulation program on bank efficiency The findings revealed that large state-owned banks and smaller banks outperformed medium-sized banks in terms of efficiency Furthermore, technical efficiency consistently surpassed allocative efficiency among Chinese banks Overall, the financial deregulation initiated in 1995 significantly enhanced cost efficiency levels, encompassing both technical and allocative efficiency.

Fu and Heffernan (2009) explored the link between market structure and performance in China's banking system from 1985 to 2002, during a time of significant reform They employed panel data estimation techniques to test both the market-power and efficient-structure hypotheses Additionally, their model was expanded to assess the effects of bank size and ownership, as well as to investigate whether the big four banks experience a "quiet life."

Ye, Xu, and Fang (2012) examined five hypotheses regarding the relationship between market structure, profitability, and efficiency using data envelopment analysis Their study focused on a panel of the 14 largest nationwide banks in China from 1998 to 2007, assessing the impact of competition on bank performance and efficiency The findings indicated that neither the structure-conduct-performance hypothesis nor the efficient-structure hypothesis is applicable in the context of China's banking sector.

Tan (2013) investigated the factors influencing bank profitability in China from 2003 to 2009, focusing on inflation, GDP growth, and stock market volatility The study found that the Chinese banking sector operated under monopolistic competition, as indicated by Panzar-Rosse’s H statistic Additionally, the Lerner index revealed that joint-stock commercial banks exhibited the highest level of competition during this period.

Chinese banks exhibit varying levels of efficiency, with state-owned commercial banks demonstrating the highest technical efficiency, followed by joint-stock commercial banks, while city commercial banks rank lowest Additionally, scale efficiency plays a more significant role than pure technical efficiency in the overall performance of these banks, which are also experiencing a misallocation of inputs and outputs in their operations Throughout the analyzed period, the productivity of state-owned, joint-stock, and city commercial banks remained stable, with notable productivity growth observed in 2005 and 2009.

In their 1997 study, Berger and Hannan analyzed U.S bank data from 1985, revealing stronger support for the structure-conduct-performance paradigm compared to the relative market power and efficient-structure hypotheses Their research addressed gaps in prior studies by exploring the intricate relationships among market structure, profits, prices, and firm efficiency By replicating four existing analytical approaches and introducing new innovations, they confirmed that the structure-conduct-performance hypothesis was more applicable, even though none of the theories fully aligned with the data Additionally, their findings supported Hick's quiet-life hypothesis, suggesting that firms with significant market power may prioritize less rigorous efficiency maximization.

Goldberg and Rai (1996) analyzed bank data from 11 European countries between 1988 and 1991, finding evidence supporting market power for most banks, except in countries with low concentration ratios, where the X-efficiency hypothesis was supported While the relationship between market structure and performance has been extensively studied in U.S banks, research on European banks remains limited Two primary explanations for the correlation between profitability and concentration are the traditional Structure-Conduct-Performance (SCP) hypothesis and efficiency considerations Previous empirical tests of these hypotheses have yielded mixed results due to the lack of direct effects in the models This study employs a random amplitude cost approach, as proposed by Aigner et al (1977), to address issues of X-inefficiency and inefficient scale, assuming normally distributed errors Additionally, ineffective measures are directly integrated into the analysis, following the methodology suggested by Berger and Hannan (1993).

Berger and Mester (1997) analyzed the efficiency and productivity growth of the U.S banking industry using extensive data from the late 1980s to the early 1990s Their findings revealed a slight decrease in cost efficiency during this period, alongside a significant decline in profit efficiency among large banks.

Finally, the Quiet-life hypothesis (QLH) is considered Hicks (1935) argued that

In his 'quiet life' hypothesis, the notion that "the best of monopoly profits is a quiet life" suggests that firms with greater market power tend to foster a more relaxed environment This results in reduced emphasis on maximizing cost efficiency, allowing these companies to operate with less pressure and effort.

Numerous researchers have examined the QLH hypothesis within their national banking systems For instance, Koetter and Vins (2008) investigated German savings banks using bank-specific Lerner indices to assess competition and its impact on cost and profit efficiency from 1996 to 2006 Their findings indicated a decline in both market power and average revenues among these banks Similarly, Koetter et al (2012) applied Lerner indices in their study of U.S commercial banks, which revealed a state of quiet life within this sector.

The initial restructuring of the Vietnamese banking system is poised to transform its overall structure and influence the performance of banks This investigation examines the implications of these structural changes and the effects of reforms on banking performance, aiming to address gaps in restructuring theory within the context of a small transitioning economy and emerging market.

The collusive behavior of dominant firms in the industry significantly impacts the market's price-setting process, enabling these firms to achieve higher profits compared to their competitors This establishes a positive correlation between market concentration and firm performance.

CHAPTER SUMMARY

This chapter explores the theoretical foundations and existing literature on the relationship between bank restructuring and efficiency, as well as the impact of bank reform on structure and performance Findings indicate varying results for restructuring methods across different economies, with several studies highlighting that financial reform enhances efficiency in certain nations Additionally, the relationship between banking structure and performance often aligns with either the market power hypothesis or the efficiency structure hypothesis While previous research benefits from extensive and reliable data sources, they fail to distinguish the effects of restructuring methods and environmental factors on banking efficiency This gap is addressed in this thesis.

To sum up, the results of research questions and hypotheses are summarized in Table 2.1

Table 2.1: Summary of the two research questions and their hypotheses

This table summarizes the hypotheses associated with RQ1 and RQ2 developed in previous sections

The restructuring measures implemented as government interventions, including the merger and acquisition of commercial banks and the privatization of state-owned banks, significantly impacted the performance of commercial banks during the study period These strategic actions aimed to enhance efficiency, competitiveness, and financial stability within the banking sector, ultimately leading to improved operational outcomes and profitability for the banks involved.

H1.1: Government intervention increases the effective of commercial banks' performance

H1.2: Merger and acquisition increases the effective of commercial banks' performance

H1.3: Privatization of state-owned commercial banks increases the effective of commercial banks' performance

RQ2: What are the effects of reform onVietnam's commercial bank structure and performance?

The SCP hypothesis posits that the collusive behavior of dominant firms in an industry significantly impacts the market's price-setting process, enabling these firms to achieve higher profits compared to their competitors This theory suggests a positive correlation between market concentration and firm performance.

Firms with larger market shares and well-differentiated products possess greater market power, enabling them to achieve higher profits compared to typical firms This hypothesis, grounded in RMP, highlights the correlation between market share, product differentiation, and profitability.

H2.3: This hypothesis is based on ES theory: low costs of production by relatively efficient firms enable them to compete more aggressively, capture a bigger market share and earn higher profits

H2.4: Quiet-life hypothesis: Banks that have higher market share have lower efficiency h

BANK RESTRUCTURING - THE CASE OF

INTRODUCTION OF THE BANKING SYSTEM- THE CASE

In the mid-1990s, Vietnam saw the establishment of numerous banks and financial institutions, prompting significant reforms in its financial system To effectively stimulate domestic investment, it became essential to provide conducive services and an environment for both foreign and domestic entities Consequently, bank credit growth surged from 18% of GDP in 1991 to approximately 21% in 1997, highlighting the evolving landscape of Vietnam's financial sector.

In Vietnam, the gross domestic saving rate increased significantly from 2.9% in

Between 1990 and 1997, the percentage of private banks in the financial sector rose from 16.2% to 18.0%, reflecting a substantial enhancement in financial services due to the influx of private, foreign, and mixed ownership banks (source: GSO 1998) However, this growth is still modest when compared to the rapid advancements seen in China and other emerging economies like Malaysia and Thailand (Harvie and Hoa).

In 1997, Vietnam's banking sector experienced limited growth, with many new banks being small and overshadowed by state-owned commercial banks (SOCBs) As a result, the Vietnamese banking system struggled with competitiveness and innovation, leading to the classification of Vietnam as an "under-banked country."

7 General statistics office of Vietnam report 1998 https://www.gso.gov.vn/default.aspx?tabidq5

Since its expansion in 2000, the Vietnamese economy has encountered significant challenges, including a high rate of non-performing loans, which reached 8.6% in 2012, and issues of cross ownership among banks The banking crisis, triggered by the global financial crisis and a fragile domestic economy, resulted in the defaults of several banks, including De Nhat, Tin Nghia, and Habu banks In response, the government implemented bank-bailout policies aimed at restructuring the banking sector and enhancing its efficiency.

This thesis explores the relationship between government restructuring measures and the efficiency of banks, while also examining the transition costs associated with the restructuring process in the Vietnamese banking system.

3.1.2 Introduction of the Bank Restructuring in Vietnam

Since 2000, the Vietnamese government initiated the privatization of State Commercial Banks to align them with regional and international standards However, significant restructuring efforts only began in 2005.

Between 2007 and 2008, Vietcombank and Vietinbank underwent equitization, resulting in significant growth and improved financial indicators, including the lowest non-performing loan (NPL) ratios in the banking system Post-equitization, these banks have become pivotal in steering the financial market, currently boasting the largest charter capital among Vietnamese banks, with total assets increasing several times since 2005.

Additionally, the transformation from rural commercial banks to urban ones conformed to the market economy during the period of 2005-2007 The “upgrade” in

The State Bank of Vietnam's 2012 annual report highlights significant advancements in the banking sector, emphasizing the importance of increasing the charter capital of commercial banks This move not only strengthens the financial stability of these institutions but also aligns their management mechanisms with international standards, marking a crucial step towards modernization and improved governance in the banking system.

Between 2011 and 2013, struggling banks were encouraged to undergo restructuring, with oversight and support from the government and the State Bank This restructuring process included various strategies such as recapitalization, mergers and acquisitions (M&A), privatization, assistance from the State Bank, and the establishment of the Asset Management Company (VAMC).

In recent years, mergers and acquisitions (M&A) have become the primary strategy for restructuring banks in Vietnam, with nine transactions completed to date These restructurings have involved transforming operational models and enhancing capital Addressing the weaknesses within the banking system has often been a top priority, leading to the essential establishment of asset management companies.

Table 3.1: Types of bank in 2003-2007

Rural commercial Bank Commercial Join Stock Bank

Source: Annual report of commercial banks from 2003 to 2007

The WAMC was established by the State Bank Chairman through Decision 1459/QĐ-NHNN on June 27, 2013, as a government tool to swiftly address the bad debt issue, mitigate credit risks for banks and enterprises, and promote sustainable credit growth With a charter capital of 500 billion VND, VAMC operates as a non-profit entity focused on reducing credit costs and resolving bad debt By December 31, 2013, VAMC had successfully acquired 38.9 trillion VND in bad debts.

35 credit institutions, equivalent to 32,400 billion VND which was the value of the special bond

To attract foreign investment, the government has implemented significant changes, increasing foreign ownership limits from 5% in 2007 to 30% in 2014 In specific cases aimed at restructuring weak institutions, this limit can exceed 30%, with decisions made at the discretion of the Minister.

3.1.3.Types of bank restructuring in Vietnam

3.1.3.1 The change of bank ownership

In the early 2000s, Vietnamese government had a privatization plan for state commercial banks to put Vietnam's banking and financial sector on a par with other countries in the region

During the period of 2003-2007, all the first rural commercial banks were transformed into urban commercial banks to be in line with the Vietnamese market economy

On 11 June 2006, the government promulgated decree No.141/2006/NĐ-CP on the list of legal capital of credit institutions Then, the regulation on the grant of licenses for the establishment and operation of joint-stock commercial banks was promulgated by the Decree No.24/2007/QĐ-NHNN, which aimed at increasing the h minimum charter capital from 732.70 billion VND to 1000 billion VND In this way, rural commercial banks are transformed into urban commercial banks

The transformation of banks was crucial, leading to an increase in authorized capital and a restructured administration system that aligns with international development trends.

In August, 2007, there were 13 requirements for establishing new banks with an authorised capital of over VND 1000 billion

Table 3.2: Joint stock commercial bank after 2007

No Joint stock commercial bank Authorised capital

7 Dong Duong Thuong Tin Commercial Joint Stock

9 Viet Nam urban development 1,000 Ha Noi

11 Foreign trade of Asia 1,000 Ha Noi

Source: State bank of Vietnam, Annual report of commercial banks 2008

The inefficient operation of State Enterprises (SOE) made it more difficult for the economy to develop Therefore, the equitization was inevitable, especially in the financial crisis in 2008 h

Between 2007 and 2013, Vietnam's banking sector was dominated by five major commercial banks: Vietcombank, Vietinbank, the Joint Stock Commercial Bank for Investment and Development of Vietnam, Agribank, and the Mekong Delta Housing Bank (MHB).

Table 3.3: The operational status of state bank before and after privatization

2014 45,142 3,062 1,116 0.31 3.65 NA NA Source: State bank of Vietnam, Vietinbank, BIDV, Vietcombank, MHB Annual report from 2008 to 2015

CHAPTER SUMMARY

This chapter provides an overview of the restructuring process in Vietnam's banking sector during its early stages, focusing on the period from 2011 to 2015 It analyzes various restructuring methods employed by state and commercial banks, including self-restructuring, mergers and acquisitions (M&A), the establishment of Asset Management Companies, and the facilitation of foreign investor participation The subsequent chapters will delve into the effectiveness of these restructuring efforts, presenting research design and empirical findings in chapters 4 and 5.

DATA AND METHODOLOGY

INTRODUCTION

The fourth chapter outlines the data and methodology employed in researching the restructuring of Vietnamese banks, addressing literature gaps and focusing on hypothesis testing The study utilizes DEA and SFA methods to assess bank efficiency and examine the effects of the transition period and financial crisis on this efficiency Additionally, the Three-Stage DEA method investigates the influence of restructuring methods, including mergers and acquisitions, privatization, and state intervention, on bank performance The analysis further explores the structure of Vietnam's banking system during the restructuring phase by testing the structure-efficiency hypothesis Data is sourced from reputable platforms such as Vietnamese Orbis Bank Focus and IMF-published economic variables Following the methodology, the chapter presents variables related to two key research questions and discusses robustness tests.

DATA AND SAMPLE

This thesis analyzes the financial statements of 26 banks from the Orbis Bank Focus database, covering the period from 2001 to 2015 The selected banks were impacted by the 2008 banking crisis and subsequently adopted restructuring strategies Annual data is utilized to assess technical efficiencies, with a focus on the changes in banking operations before, during, and after the restructuring implementation.

This article examines 12 mergers and acquisitions, focusing on the restructuring of key financial metrics such as deposits, interest expenses, non-interest expenses, total loans, interest revenue, and non-interest revenue For the first research question (RQ1), six country-specific factors—GDP, real interest rates, fiscal surplus as a percentage of GDP, trade, inflation, and credit growth—are sourced from the International Monetary Fund (IMF).

The research categorizes banks into two groups: those without restructuring and those implementing restructuring measures In Vietnam, the restructuring process varies over time, with banks designated as restructured in the year they meet specific criteria Table 4.1 presents the number of banks in each category from 2007 to 2015, indicating that the initial restructuring results began to emerge in 2007.

Table 4.1: Restructuring measures of Vietnamese banks

No Date Before restructure Measure After restructure

1 15-12-2011 De Nhat joint-stock commercial bank

Merger SaiGon joint-stock commercial bank

Tin Nghia joint-stock commercial bank

SaiGon joint-stock commercial bank

2 28-08-2012 Ha Noi joint-stock commercial bank

Merger SaiGon Ha Noi joint-stock commercial bank

SaiGon Ha Noi joint- stock commercial bank h

3 03-10-2013 Laoviet bank Acquisition Joint Stock

Commercial Bank for Investment and development

Joint Stock Commercial Bank for Investment

4 2013 Ho Chi Minh Joint Stock

Merger Ho Chi Minh Joint

Stock Commercial Bank for development

5 01-10-2015 Southern joint-stock commercial bank

Merger SaiGonThuong Tin joint-stock commercial bank

SaiGonThuong Tin joint- stock commercial bank

6 23-04-2013 Maybank Merger An Binh joint-stock

An Binh joint-stock commercial bank

7 12-08-2015 Me Kong joint-stock commercial bank for development

8 01-05-2015 Me Kong housing bank Merger Joint Stock

Commercial Bank for Investment and development Bank for Investment

9 22-05-2015 PG Bank Merger Joint Stock

Commercial Bank for industry and trade

State bank purchase all shares with 0 Dong

Vietnam Construction Commercial One Member Limited Liability Bank

State bank purchase all shares with 0 Dong

Ocean Commercial One Member Limited Liability Bank

State bank purchase all shares with 0 Dong

One Member Limited Liability Global Petroleum Bank

Bank for Investment and Development of Viet Nam

State-owned commercial bank privatization

Joint stock Commercial Bank for Investment and Development of Viet Nam

Bank for Foreign Trade of Vietnam

State-owned commercial bank privatization

Joint Stock Commercial Bank for Foreign Trade of Vietnam

Commercial Bank for Industry and Trade - Main Operation Center

State-owned commercial bank privatization

Vietnam Joint Stock Commercial Bank for Industry and Trade - Trade Finance Center h

State bank intervention-VAMC 13 purchases bad debt

State bank intervention-VAMC purchases bad debt

18 04-10-2013 Saigon Commercial Bank State bank intervention-VAMC purchases bad debt

State bank intervention-VAMC purchases bad debt

State bank intervention-VAMC purchases bad debt

21 11-10-2013 SaiGon joint-stock commercial bank

State bank intervention-VAMC purchases bad debt

State bank intervention-VAMC purchases bad debt

23 2014 Saigon Commercial Bank State bank intervention-VAMC purchases bad debt

State bank intervention-VAMC purchases bad debt

State bank intervention-VAMC purchases bad debt

26 2014 Vietnam Prosperity commercial joint-stock bank

State bank intervention-VAMC purchases bad debt

27 2014 Vietnam Technological and Commercial Bank

State bank intervention-VAMC purchases bad debt

28 2014 Bank for Foreign Trade of

State bank intervention-VAMC purchases bad debt

Commercial Bank intervention-VAMC purchases bad debt

State bank intervention-VAMC purchases bad debt

31 2014 Lien Viet joint-stock commercial bank

State bank intervention-VAMC purchases bad debt Source: State bank of Vietnam, financial statement of commercial banks from 2007 to

Table 4.1 outlines the banks that have adopted restructuring methods, detailing nine mergers and acquisitions, three privatization cases, and three instances where the State Bank of Vietnam acquired assets for zero VND Additionally, several commercial banks have sold bad debts to the Vietnam Asset Management Company (VAMC) as a state intervention strategy for debt restructuring A comprehensive summary of the restructuring efforts can be found in the Appendix.

METHODOLOGY FOR RQ1, RQ2

Section 4.3 describes the estimation methods for RQ1 and RQ2 In this section, DEA, Three-Stage DEA 17 and GLS are presented in details For RQ1, the Three-Stage DEA was applied because variables which were not affected and then affected by environment factors were considered For RQ2, since there were not any environment variables, thus, DEA was selected to use

Charnes et al (1978) pioneered the study of Data Envelopment Analysis (DEA), a programming technique designed to evaluate the efficiency of decision-making units (DMUs) on the efficient frontier DEA establishes frontiers through linear combinations of best practice observations, enabling the identification of inefficiencies among the sample By analyzing data errors in individual units, DEA can effectively measure performance variations This method not only highlights DMUs that operate on the frontier but also reveals inefficiencies relative to other units in the dataset.

Data Envelopment Analysis (DEA) is closely linked to economic production theory, but it also serves as a valuable tool for evaluating management activities This set of measures is designed to assess the performance of both production and service operations In this context, benchmarking Decision-Making Units (DMUs) through DEA may not establish a traditional "production line," yet it effectively highlights efficiency standards.

"best execution frontier" (Cook, Tone and Zhu, 2014) DEA by Sherman and Zhu

In 2013, it was noted that the calculation parameters for the "non-balance method" of interest do not conform to a specific functional shape However, a common correlation output relationship remains unspecified This method aims to estimate production frontier parameters, as discussed by Lovell and Schmidt in 1988 These parameters are determined by defining a specific function related to the head, which can also be derived from a combination of relatively intense methods in each hybrid, as suggested by Tofallis.

2001), in the specified unit of the frontier, first by DEA and then on the surface of a

18 Decision making unit h smooth installation This allows the relationship between the estimated best multi input and multi output

The framework, adapted from the production function, emphasizes that increased input leads to greater output and is applicable across various fields The Data Envelopment Analysis (DEA) establishes a functional form based on the most efficient producers, distinguishing itself from Ordinary Least Squares (OLS) which focuses on average manufacturers Unlike Stochastic Frontier Analysis (SFA), DEA identifies a "frontier" approach, positing that if one company can achieve a specific output with certain inputs, other companies of similar scale should also be capable of doing so This method enables the identification of the most efficient producers, facilitating the calculation of effective solutions for varying input and output levels.

Data Envelopment Analysis (DEA) is utilized in Research Question 2 (RQ2) to assess the X-efficiency of Vietnamese banks, focusing on key efficiency indicators such as X-efficiency and Scale-efficiency X-efficiency is defined as the ratio of the predicted minimum costs, which would be incurred if a bank operated at the efficiency level of the best-practice bank within the same sample, to the actual predicted costs, with adjustments for random error As outlined by Berger and Mester (1997), a bank-specific measure of X-efficiency is calculated to provide a comprehensive evaluation of banking efficiency.

XEFF 21 = 𝐶̂ 𝑚𝑖𝑛 𝐶̂ 𝑖 = exp [𝑓 (𝑤 exp [𝑓 (𝑤 ̂ 𝑖 ,𝑦 𝑖 )× exp (ln𝑢̂ 𝑚𝑖𝑛 )

𝑖 ,𝑦 𝑖 ) ̂ ×exp (ln𝑢̂ 𝑚 ) = 𝑢̂ 𝑚𝑖𝑛 𝑢̂ 𝑖 (4.1) where 𝐶̂ 𝑚𝑖𝑛 : the predicted minimum costs as used by the best-practice bank

𝑢̂ 𝑚𝑖𝑛 : the minimum of the U, across all banks in the sample

𝑢̂ 𝑖 :the predicted actual cost inefficiency of a specific bank

The DEA method is employed to assess Scale-efficiency, which determines if banks with comparable management and production technologies are achieving optimal economies of scale Scale economies are quantified through specific calculations.

C(Y) is the cost function and, Y: a vector of outputs;

Yp represents various products produced, p: indices of different products

SINEFF = SCALE – 1 if SCALE> 1 SINEFF = 1 – SCALE if SCALE< 1

When the SCALE is below 1, a bank is functioning beyond its optimal capacity and needs to reduce its output to achieve the best resource allocation On the other hand, if SCALE exceeds 1, the bank should enhance its output levels to lower costs The bank attains SINEFF at SCALE equal to 1, indicating that SINEFF equals 0.

In addressing RQ2, we first utilized Data Envelopment Analysis (DEA) to evaluate bank efficiency, followed by the application of Generalized Least Squares (GLS) for the equations, while managing heteroskedasticity within the Ordinary Least Squares (OLS) framework If the variance-covariance matrix of the error term is known, we can transform a heteroskedasticity model into a homoskedastic model.

4.3.1.3 The Three-Stage Data Envelopment Analysis

The Three-Stage DEA is an expanded version of Data Envelopment Analysis (DEA) that addresses the limitations of traditional methods by incorporating external elements and environmental impacts in its calculations Research by Berger and Humphrey (1992) and Wheelock and Wilson (1995) highlights the sensitivity of banks to macroeconomic conditions, indicating that banking crises can significantly influence efficiency This methodology, as discussed by Avkiran and Rowlands (2008), is particularly useful for examining the effects of various economic factors on banking efficiency For research question RQ1, the Three-Stage DEA was deemed suitable to analyze the impact of these economic factors on bank performance.

Step 1: Adding minimized inputs and maximized outputs simultaneously, then using the non-oriented variable returns to scale

Step 2: Separating results from stage 1 into environmental effects, statistical noise and managerial inefficiency Then adjusted these factors to see how their impact on the bank operation

Step 3: Repeating step 1 with the data adjusted at step 2

Based on the literature review and the objectives of the thesis, appropriate research methods were carefully selected and combined to effectively address each research question For Research Question 1, a Three-stage Data Envelopment Analysis (DEA) was utilized to assess the influence of environmental variables on dependent variables In addressing Research Question 2, the study examined the impact of environmental variables by analyzing how input and output variables were affected, employing DEA alongside Fixed Effect, Random Effect, and Ordinary Least Squares (OLS) methods.

4.3.2.1 RQ1: The effect of restructuring methods on banking performance

The Three-Stage DEA model

In this study, stage 1 utilizes three inputs—deposits (TD), interest expense (IE), and non-interest expense (NE)—to assess the role of a bank in mobilizing funds at the lowest possible cost.

The three bank outputs capturing both traditional and non-traditional activities of banks consist of total loans (TL), Interest revenue (IR) and Non-interest revenue (NR)

The total deposit amount serves as the primary input for this study While purchased funds may be relevant, their analysis is often limited due to insufficient data from many banks, especially smaller institutions Additionally, physical capital is assessed through other operating expenses.

To discover the impact of environmental factors on the validity of the initial efficiency analysis, three bank restructuring measures, six country-specific factors are included in step 2

Step 3 applies DEA method introduced in step 1 but replaces with the adjusted inputs and outputs obtained in step 2

In step 2, the test is carried out by the following models:

This article analyzes the impact of three restructuring methods on three inputs—total deposits (TD), interest expense (IE), and non-interest expense (NE)—and three outputs—total loans (TL), interest revenue (IR), and non-interest revenue (NR) The analysis extends to six country-specific factors, as detailed in Table 4.2 Subsequently, six equations are established to assess the influence of these control variables on the input and output variables.

Finally, step 3 is similar to stage 1 using DEA method but replaces three inputs and three outputs adjusted in step 2

4.3.2.2 RQ2: The effects of reform to structure and performance

The ideas of this thesis are based on the love study of Berg (1995) and Fu

In 2009, a study was conducted to evaluate the Quiet-Life Hypothesis (QLH) and its effectiveness during a period of system restructuring The research aimed to determine the impact of restructuring on the system's overall structure and assess the degree of its effectiveness.

First, according to Berger (1995) and Goldberg and Rai (1996), Eq (4.17) is used to test the validity of the hypotheses:

ROBUSTNESS TEST

This section examines the robustness test aimed at addressing methodological concerns regarding various proxies used to assess the impact of restructuring methods on the performance and efficiency of commercial banks Ultimately, the study evaluates the effects of these reforms on the structure and performance of Vietnam's commercial banking sector.

The thesis employs Stochastic Frontier Analysis (SFA) regression to assess the robustness of Data Envelopment Analysis (DEA) regression in Research Questions 1 and 2 SFA is an economic modeling technique that originated from the production model at a random frontier, introduced by Meeusen and Van Den Broeck in 1977 The production frontier model, excluding random components, can be expressed in a specific mathematical format.

In the equation \( y_i = f(x_i, \beta) \times TE_i \), \( y_i \) represents the observed scalar output of manufacturer \( i \), where \( i \) ranges from 1 to \( I \) The vector \( x_i \) consists of \( N \) inputs utilized by manufacturer \( i \), while \( f(x_i, \beta) \) defines the production boundary, incorporating a vector of estimated technological parameters.

Technical efficiency (TEi) is defined as the ratio of observed output to the maximum feasible output A TEi value of 1 indicates that a company is operating at its maximum potential, while a TEi value less than 1 reflects a shortfall in observed output compared to what is achievable.

Random shocks represent unpredictable elements in the production process that are not directly linked to product-specific issues or fundamental technology Each manufacturer experiences unique shocks, which are considered random and can be characterized by a general distribution These shocks, denoted by mining {v_i}, highlight the variability faced by different producers in their operations.

The stochastic production frontier will become:

TEi is also a random variable, with a specific distribution function, common to all manufacturers

An exponential TE_ (i) = exp {〖- u〗 _i}, where ui 0, because we require TEi

Now, they assume that f (xi, β) has Cobb Douglas log-linear form, the model can be written as:

The noise component, denoted as 𝑣𝑖, is typically regarded as a normal two-sided distribution variable, while ui represents a non-negative ineffective component Together, these elements create a synthetic error term characterized by a defined distribution, leading to the term "successful synthetic error model."

In the composed error model, the error term consists of two components: the noise component (𝑣 𝑖), typically treated as a two-sided normally distributed variable, and the non-negative technical inefficiency component (ui) This combination creates a compound error term with a distribution that needs to be specified, which is why this model is commonly referred to as the "composed error model."

Stochastic Frontier Analysis (SFA) evaluates the efficiency of "cost" and "profit" (Kumbhakar & Lovel, 2003), focusing on "Marginal Costs" to minimize overall company expenses effectively Unlike traditional random h specifications, it incorporates non-negative cost inefficiencies The "Analysis of Margins" emphasizes profit maximization for manufacturers, considering both output and input decisions, rather than merely reducing costs with exogenous output levels This approach aligns with the "production frontier" model Additionally, random amplitude analysis has been utilized in microdata to assess consumer demand and segment customers In a two-phase methodology, a stochastic frontier model was first estimated, followed by a regression of consumer characteristics based on deviations from the frontier (Baltas, 2005).

To test the robustness of random effect and fixed effect method, Hausman test is used

The Hausman Test, also known as the Hausman specification test, identifies endogenous variables in a regression model Endogenous variables are those whose values are influenced by other variables within the system The presence of these endogenous variables can undermine the effectiveness of ordinary least squares (OLS) regression, as one of the key assumptions of OLS is the absence of correlation between predictor variables and the error term.

In the linear model represented by y = bX + e, y is the dependent variable, X is a vector of regressors, b denotes the vector of coefficients, and e signifies the error term There are two estimators for the coefficients: b0 and b1 Both estimators are consistent under the null hypothesis, but b1 is considered efficient due to its smaller asymptotic variance within the group of estimators that includes b0 Conversely, under the alternative hypothesis, b0 remains consistent while b1 does not.

Then, the Wu–Hausman statistic is as follows: h

The statistic, denoted by the Moore–Penrose pseudo inverse (†), follows an asymptotic chi-squared distribution under the null hypothesis The degrees of freedom for this distribution are equal to the rank of the matrix formed by the difference between the variances, specifically Var(b0) - Var(b1).

Rejecting the null hypothesis indicates inconsistency with b1 and allows for the examination of endogenous variables by comparing specific variable estimates through Instrumental Variable (IV) estimation and Ordinary Least Squares (OLS) This approach also helps validate additional tools by comparing IV estimates derived from a tool set Z against those from a proper subset of Z It is essential to ensure the validity of this subset and possess the necessary tools to determine its parameters Hausman emphasized that the covariance between estimates can effectively distinguish between effective and ineffective estimates.

CHAPTER SUMMARY

This chapter outlines the data and sample used to test the hypotheses from Chapter 2, addressing four research questions identified in Chapter 1, along with robustness tests Section 4.2 discusses data sources, database characteristics, and the necessity of differentiating between revolving and term loans, as well as sampling procedures and composition The study utilizes multiple data sources, including Orbis Bank Focus, IMF, the General Statistics Office of Vietnam, and the State Bank of Vietnam Given the complexity of variables for the three models, a meticulous approach was taken to ensure accuracy, with careful examination of the variables conducted multiple times Based on existing literature, the sampling procedures were articulated, leading to the determination of the final sample.

This section details the methodology for Research Questions 1 and 2, including model specification and estimation methods A restructuring model is introduced to evaluate the hypotheses for RQ1, while a system of one and five simultaneous equations is developed for RQ2 Additionally, the chapter outlines the selection and measurement of variables, supported by descriptive statistics, all sourced from reliable references The variable selection process is discussed in Sections 4.5-4.5.1, focusing on the impact of restructuring methods on banking efficiency, which involves three inputs (total deposits, interest expense, and non-interest expense) and three outputs (total loans, interest revenue, and non-interest revenue) Lastly, the robustness of the findings from the main analysis is examined in Section 4.6.

This chapter examines the performance differences of banks before and during restructuring using descriptive statistics The findings indicate that in stage 2, competitiveness rises, leading banks to encounter increased competition and greater challenges.

RESULTS AND DISCUSSION OF THE RESULTS

INTRODUCTION

Chapter 4 detailed the data, sample, model specifications, estimation methods, and variable selections used to test the hypotheses related to the three research questions from Chapter 2 Due to the distinct characteristics of revolving loans compared to term loans, they were analyzed separately The test results are organized into sections: Section 5.2 and Section 5.3 address the hypotheses for RQ1 and RQ2, respectively, while Section 5.4 provides a summary of the chapter.

EMPIRICAL RESULTS FOR RQ1

Between 2001 and 2011, the average operating efficiency of commercial banks in Vietnam increased from 85% to over 95%, but then declined to below 95% by 2015 due to the banking crisis and government restructuring policies, along with associated transition costs (Nakhun and Necmi, 2009; Hsing-Chin et al, 2010) The initial decline in efficiency during the restructuring period was attributed to the costs incurred while shifting from the old banking model to a new one, a finding consistent with research on Taiwan's banking system (Hsing et al, 2010) To further assess the impact of restructuring and environmental variables on the operations of Vietnamese commercial banks, stage 2 of the DEA analysis was conducted, incorporating inputs and outputs to evaluate their influence on bank inefficiency through regression SFA.

Figure 5.1 Average banking efficiency scores in step 1 and step 3 DEA

Figure 1-a CRS_TE in step 1 and

CRS_TE scores estimated from the observed data (Stage 1)

CRS_TE scores estimated from the data adjusted for country-specific factors

CRS_TE scores estimated from the data adjusted for country-specific factors, and restructuring (Stage 3)

Figure 1-b VRS_TE in step 1 DEA and Step 3 DEA

VRS_TE scores estimated from the observed data (Stage 1)

VRS_TE scores estimated from the data adjusted for country-specific factors

VRS_TE scores estimated from the data adjusted for country-specific factors, and restructuring (Stage 3)

Figure 1-c SCALE in step 1 DEA and Step 3 DEA

SCALE scores estimated from the observed data (Stage 1)

SCALE scores estimated from the data adjusted for country-specific factors

SCALE scores estimated from the data adjusted for country-specific factors, and restructuring (Stage 3)

Note: CRS_TE: constant variable to scale of technical efficiency; VRS_TE: variable return to scale technical efficiency; SCALE: Scale efficiency Source: Author’s calculation h

A comparison of bank performances from 2001 to 2008 reveals that banks that implemented restructuring outperformed those that did not During this period, restructuring banks focused on profit maximization and relaxed lending regulations, which initially provided competitive advantages However, this approach exposed weaknesses that were detrimental during the financial crisis, leading to the necessity for mandatory restructuring in the subsequent period.

Non-restructured banks Restructured banks Sample average

Non-restructured banks Restructured banks Sample average Source: Author’s calculation h

Figure 5.2-c Comparison of average SCALE between non-restructured banks and restructured banks in DEA step 1

Non-restructured banks Restructured banks Sample average

Note: CRS_TE: constant variable to scale of technical efficiency; VRS_TE: variable return to scale technical efficiency; SCALE: Scale efficiency

After 2009, banks that underwent restructuring experienced significantly lower effectiveness compared to those that did not, primarily due to the global economic crisis and internal challenges, resulting in increased bad debts This heightened liquidity risk diminished operational efficiency, necessitating state-mandated restructuring measures Furthermore, during the initial restructuring phase from 2011 to 2015, banks that implemented these measures continued to see declining performance, attributed to ineffective restructuring strategies and the transition costs incurred in the early stages, which increased operational expenses This finding aligns with previous research, such as that by Hsiao et al.

Research from Thoraneenitiyan (2009) and other studies indicated a decline in bank efficiency during restructuring, attributed to transition costs However, this investigation reveals that the decrease in efficiency was also influenced by external factors, including financial crises and a struggling domestic economy This will be further illustrated in the subsequent sections.

Notes Estimation based on the sample of banks during the first stage of banking reform (2007-

2015) SI: State intervention; MA: Merger and acquisition; COP: State-owned commercial bank privatization Symboy *, ** and *** have meaning that a variable is significant at 0.1, 0.05 and 0.01, respectively

Table 5.1 presents the regression results from SFA's step 2, revealing a positive correlation between mergers (MER) and the input and output slacks of most banks This suggests that merged banks exhibit decreased resource efficiency in the post-restructuring phase, as the acquiring banks inherit the inefficiencies associated with the weaker merged entities, including credit and liquidity risks.

The coefficients for the Cost of Production (COP) and Service Index (SI) are both positive and statistically significant, indicating a positive relationship between SI and bank inefficiencies This finding aligns with previous research, including studies by Hawkin and Turner (1999) and Thoraneenitiyan and Necmi.

In 2009, research indicated that the rate significantly affects the System Input (SI) variables more than the output variable This suggests that government-supported banks, particularly those receiving recapitalization, incur higher marginal costs in revenue generation, resulting in increased inefficiency This scenario is particularly evident in Vietnam, where state-owned commercial banks that received government assistance often allocate substantial resources to support governmental and community initiatives.

The privatization of state-owned commercial banks has been linked to an increase in both input and output variables, which paradoxically correlates with greater bank inefficiency This finding contradicts existing literature that suggests private banks outperform state-owned banks in efficiency within emerging markets The results indicate that privatization may be nominal, as the government retains a significant shareholding in these banks post-privatization, ultimately failing to enhance efficiency.

The performance of banks is influenced not only by restructuring variables but also by environmental factors such as economic growth, interest rates, trade, credit growth, inflation, and fiscal policies Economic growth negatively affects the inefficiency index, as higher demand for loans and improved business efficiency lead to better bank performance Conversely, rising interest rates contribute positively to the inefficiency index, increasing costs for enterprises and hindering their access to capital, often due to stringent government monetary policies Additionally, inflation diminishes the operational efficiency of commercial banks by raising input costs, which can further exacerbate non-performance indicators in their results.

5.2.2 Stage 3: DEA results on adjusted data

Figure 5.1 illustrates the efficiency scores from DEA's step 3, highlighting the significant improvement in scores after adjusting for operating environment variables and restructuring factors The findings indicate that banks' efficiencies vary based on their operating environments, influenced by restructuring programs and six country-specific external factors A comparison between DEA's step 1 and step 3 reveals that efficiency scores rise after removing the effects of these country-specific factors, suggesting their detrimental impact on estimated bank efficiency Specifically, the adjustments lead to an average increase in efficiency scores of 6-15% Ultimately, the analysis confirms that while country-specific factors exert the greatest influence on bank efficiency, the effect of restructuring measures is comparatively minimal.

The performance of banks is significantly influenced by both restructuring and environmental variables, with the latter having a more profound impact During the restructuring phase, banks' efficiency is primarily affected by external factors like economic crises and inherent weaknesses in the economy, rather than the restructuring initiatives themselves This indicates that early-stage restructuring measures may be less effective due to time lags and the overall efficacy of the policies implemented.

Besides, the results also show that in the first phase of the restructuring from

From 2011 to 2015, the performance of banks undergoing government restructuring declined due to ineffective measures and high transition costs in the initial stages, which increased operational expenses and hindered their ability to generate output.

This study highlights how government restructuring measures affect bank efficiency, revealing that the privatization of state-owned commercial banks is ineffective due to the government retaining majority ownership, which minimally alters the operational structure of these banks.

Previous studies indicate a positive correlation between state intervention and bank inefficiency Empirical evidence demonstrates that state-owned banks, which receive government support through recapitalization, tend to incur higher marginal costs in generating revenue, resulting in increased inefficiency.

The consolidation and merger of banks can yield positive outcomes; however, the effectiveness of these processes may be diminished by the risks associated with bad debt and the liquidity challenges faced by struggling banks.

EMPIRICAL RESULTS FOR RQ2

Research question 1 examined how various bank restructuring methods influence bank efficiency at the individual bank level Following this, research question 2 will explore the effects of bank restructuring on the overall structure and performance of the banking system.

Table 5.2: Condition numbers - testing for multicollinearity

Notes CR4: four-bank concentration ratio, HERF: Herfindahl-Hirschman index, MS: market share, XINEFF: X-inefficiency, SINEFF: scale inefficiency

Before conducting estimations, a correlation matrix is created to assess the relationships among the model's variables The findings indicate that from 2001 to 2006, both the SINEFF and XINEFF indices exhibit a negative correlation with MS In the subsequent stage, only the XINEFF index maintains a negative correlation with MS, as shown in Table 5.2.

Table 5.3 presents the P-values from the Hausman test, indicating that fixed effects should be applied to equations with P-values below 0.05 Consequently, fixed effects are deemed optimal for eight cases across both stage 1 and stage 2, while random effects are utilized for the remaining cases The findings are detailed in the tables below.

Table 5.4: Market-power vs efficient-structure

Dependent variable = ROA Dependent variable = ROE

1 st stage 2 nd stage 1 st stage 2 nd stage

The analysis is divided into two stages: the first stage covers the period from 2001 to 2006, focusing on the initial banking reform, while the second stage spans from 2007 to 2015, reflecting the subsequent reform phase Key financial metrics include Return on Assets (ROA) and Return on Equity (ROE), with the intercept denoted as INT The Herfindahl-Hirschman index (HERF) measures market concentration, while market share (MS) and inefficiency indicators such as X-inefficiency (XINEFF) and scale inefficiency (SINEFF) are also considered Additional factors include average income per person (AIP), the loan to asset ratio (LA), and the equity to asset ratio (KA) Statistical significance is indicated by symbols *, **, and ***, representing significance levels of 0.1, 0.05, and 0.01, respectively.

During Stage 1 (2001-2006), the analysis reveals a significant correlation between market power and Return on Assets (ROA), indicating that a higher market share positively influences ROA Conversely, both X-inefficiency and Scale-inefficiency negatively affect ROA, suggesting that increased efficiency in these areas correlates with improved financial performance The findings support the existence of X-efficiency and Scale-efficiency hypotheses during this period The dominance of the big four state banks, which benefited from state incentives and advantages over private banks, contributed to their higher market share and profitability Thus, the observed relationships align with the realities of the banking sector in Stage 1.

During stage 2 (2007-2015), the market share index positively influenced banks' profitability, as evidenced by the return on equity (ROE) Additionally, there was a statistically significant positive relationship between X-inefficiency and both return on assets (ROA) and ROE during this period.

In 2015, a notable trend emerged where banks with higher X-efficiency exhibited lower Return on Assets (ROA) and Return on Equity (ROE), while Scale-inefficiency had a significant negative correlation with these profitability metrics This suggests that banks with greater Scale-efficiency tend to achieve higher ROA and ROE, supporting the market power hypothesis, which posits that commercial banks with substantial market share can generate more profits However, the model results do not affirm the Efficiency Structure (ES) hypothesis during this period As illustrated in Table 5.4, the relationship between X-efficiency and profitability remains consistent with the 2007-2015 period, showing statistical significance Conversely, Scale-inefficiency negatively impacts ROA and ROE, indicating that larger banks positively influence profitability The contrasting results are insufficient to definitively accept or reject the ES hypothesis for this timeframe.

Table 5.5: Necessary condition estimation for ES hypothesis in stage 1

Dependent variable =HERF Dependent variable =MS

for variables that are significant at 0.1, 0.05, and 0.01, respectively.

In Stage 1, we examine the X-efficiency and S-efficiency hypotheses, focusing on the necessary conditions for the ES hypothesis According to Table 5.5, the necessary condition is satisfied if the coefficients for X-inefficiency and Scale-inefficiency are significantly negative with respect to the two dependent variables The results indicate that X-inefficiency is significantly and negatively correlated with both the Herfindahl-Hirschman index and market share, while Scale-inefficiency shows a significant positive correlation This outcome suggests that the necessary condition for the ES hypothesis is not met, leading to the conclusion that there is insufficient evidence to accept or reject the ES hypothesis during this period.

Table 5.6: The quiet life hypothesis-estimation

1 st stage 2 nd stage 1 st stage 2 nd stage

The study utilizes various economic indicators, including the intercept (INT), Herfindahl-Hirschman index (HERF), market share (MS), X-inefficiency (XINEFF), and scale inefficiency (SINEFF), to analyze market dynamics It also considers average income per person (AIP) and incorporates a time trend variable (TT) for temporal analysis Additionally, the loan to asset ratio (LA) and equity to asset ratio (KA) are examined to assess financial stability Statistical significance is denoted by symbols *, **, and ***, indicating levels of significance at 0.1, 0.05, and 0.01, respectively.

To address the endogeneity in testing the quiet-life hypothesis, we employed two-stage least squares, resulting in the findings displayed in Table 5.6 The analysis indicates that the quiet-life hypothesis is supported when the market share index and/or the Herfindahl–Hirschman index show a significant positive relationship with X-inefficiency and Scale-inefficiency However, in the period from 2001 to 2006, the market share index was only statistically significant in relation to X-inefficiency, which does not provide enough evidence to support the quiet-life hypothesis Similarly, from 2007 to 2015, the Herfindahl–Hirschman index was significant at the 0.05 level but only positively impacted Scale-inefficiency Consequently, we conclude that there is insufficient evidence to accept the quiet-life hypothesis across both stages of analysis.

The assessment of the banking system's restructuring reveals that prior to the changes, Vietnamese commercial banks were heavily influenced by government price controls, a lack of information transparency, and an inability to leverage market power for profitability, supporting the theory of market power Additionally, during this pre-restructuring phase, banks with greater scale efficiency achieved higher profits, aligning with the X-Efficiency and S-efficiency hypotheses However, the study found no evidence of the ES theory during this period The findings indicate that both before and after restructuring, the market power hypothesis remained valid, as banks with significant market shares and scale efficiencies continued to generate high profits In the initial stage, state-owned banks benefited from government privileges, including low interest rates and preferential mechanisms, which enhanced their market power and profitability In the second stage, despite the privatization of state-owned banks, government ownership remained substantial, perpetuating monopoly and market power Consequently, the restructuring process did not significantly alter the banking system's structure, consistent with the results of Research Question 1.

This chapter reports the empirical results of testing hypotheses associated with two research questions Sections 5.2 presented and analysed the empirical test results for hypotheses related to RQ1

The restructuring methods, including the privatization of state-owned commercial banks, state intervention, and mergers and acquisitions, had minimal impact on the operational structure of these banks Following initial government-led restructurings, a market-driven consolidation process emerged, primarily led by local banks These banks were motivated to enhance their efficiency due to competitive pressures from new market entrants.

The Vietnam banking system demonstrated the Market Power hypothesis during both the pre-restructuring and restructuring periods, where commercial banks with high market shares and scale efficiency gained a distinct advantage in generating profits throughout both stages.

Chapter 6 concludes this thesis by summarizing the research questions, hypotheses, and empirical results It highlights the contributions and implications of the findings while acknowledging the limitations of the study The chapter also outlines potential directions for future research.

CONCLUSION

INTRODUCTION

This concluding chapter summarizes the main findings and contributions of the thesis, along with its implications, limitations, and future research directions Section 6.2 revisits the four research questions, their corresponding hypotheses, and the results of the tests conducted Section 6.3 highlights the key contributions of the thesis, while Section 6.4 discusses their implications In Section 6.5, the limitations of the work are acknowledged, and Section 6.6 wraps up the chapter with recommendations for future research endeavors.

REVIEW OF RESEARCH QUESTIONS, HYPOTHESES AND

6.2.1 RQ1: How restructuring measures, which were introduced as the government intervention, merger and acquisition of the commercial banks and privatization of the state-owned commercial banks, affect the performance of the commercial banks in the studied period?

The thesis investigates the effects of government measures such as intervention, mergers and acquisitions (M&As), and the privatization of state-owned commercial banks The findings indicate that M&As enhance both the input and output of these banks; however, post-restructuring, there is a decline in resource management efficiency due to the integration of weaker banks, which introduces credit and liquidity risks Consequently, government-supported banks, primarily state-owned, face elevated marginal costs and inefficiencies in generating revenue This situation reflects the reality in Vietnam, where banks benefiting from government support incur higher expenses to fulfill obligations to the government and the community.

Privatized banks have minimal impact on the performance of commercial banks, as the central bank retains significant ownership in these institutions This dominance ensures that there are no substantial changes to the structure or operations of the banking system.

6.2.2 RQ2: What are the effects of reform on Vietnam's commercial bank structure and performance?

From the test results, it can be concluded that the effects of reform on Vietnam's commercial bank structure and performance as follows:

During the research period, RMP theory remained consistent across both stages of Vietnam's banking system, with no significant negative correlation between concentration and efficiency among the four state-owned banks This indicates that banks with higher market shares were more profitable and not dependent on market concentration during the restructuring phase In the first stage, the big four state-owned commercial banks leveraged their government support and access to state funds to achieve significant market shares and profits By the second stage, despite privatization efforts, state-owned banks still held majority shares, perpetuating a monopoly Unlike the Efficiency Structure (ES) theory, which suggests that efficient firms can reduce production costs and capture market share, this was not evidenced in Vietnam's banking sector This contrasts with findings in China, where research by Ye, Xu, and Fang (2012) indicated that neither the Structure-Conduct-Performance (SCP) nor ES hypotheses were upheld during 1998–2007, with only the X-Efficiency Hypothesis being relevant The differing outcomes stem from the distinct approaches to banking restructuring in each country; Vietnam focused on mergers, acquisitions, and privatization, while China emphasized enhanced banking administration and investment in technology and risk management to boost efficiency.

CONTRIBUTIONS

This thesis highlights the initial stages of Vietnam's banking restructuring process, revealing its ineffectiveness due to the government's majority ownership, which limited changes in the operational structure of commercial banks Environmental factors significantly impacted banks during this period, necessitating the development of targeted solutions This research aligns with previous studies by Hsiao et al (2010) and Thoraneenitiyan (2009), which noted a decline in bank efficiency during restructuring due to transition costs and external variables, such as financial crises and a weak domestic economy.

The bank restructuring in Vietnam aimed to identify the most suitable model for the competitive structure of its banking sector and assess the effects of bank size and ownership Findings reveal that while the banking system has shifted towards increased profitability and efficiency, it has also become more sensitive and vulnerable, particularly in the wake of the global recession This heightened vulnerability poses challenges for domestic banks as they strive for greater integration into the international financial system Comparatively, Vietnam's banking system exhibits more instability than China's during their respective transition periods, largely due to Vietnam's smaller economic scale This aligns with the economies of scale theory, indicating that smaller economies face greater pressures and instability during transitions and integration processes.

This article introduces a theoretical framework that extends the market-power, quiet life, and efficient-structure hypotheses to examine the impact of banking restructuring on the performance of the Vietnamese banking system and other small open transition economies By addressing this gap, the study provides valuable insights into the relationship between banking reforms and economic performance in emerging markets.

IMPLICATIONS

The restructuring of state-owned commercial banks has not significantly altered their operational structures To improve the banking system, the government should minimize market intervention, promote the privatization of these banks, and implement integrative policies that foster competition and transparency.

The restructuring of Vietnam's banking system aims to facilitate integration into international financial markets, yet it faces significant challenges Emerging economies like Vietnam's are under pressure from intense competition with foreign banks, which can lead to instability in banking performance, particularly during global economic crises To navigate this transition effectively, the government must develop a comprehensive strategy that prioritizes system stability throughout the integration process Additionally, careful consideration of the timing for integration is crucial, as entering international markets during a financial crisis could exacerbate existing vulnerabilities.

LIMITATIONS

This thesis acknowledges several limitations, primarily related to data characteristics The main data sources, Orbis Bank Focus and the State Bank of Vietnam, contained missing figures that complicated variable statistics Additionally, the ongoing banking restructuring restricted the research to the initial phase, covering the period from 2007 to 2015 While various research methods were available, none proved universally superior or suitable for all scenarios; therefore, a combination of methods was employed to effectively address the research issue.

FUTURE RESEARCH DIRECTIONS

The results highlight the extensive restructuring of Vietnam's banking system, aligning with findings from previous studies on emerging markets and the integration process into international financial markets The key question remains whether Vietnamese commercial banks can maintain their market share and competitive advantage amidst increasing foreign involvement in the financial sector If they fail to adapt, they risk becoming followers in a rapidly evolving economy, facing challenges such as limited competitiveness, outdated technology, and management practices, ultimately losing market share to foreign banks and major global financial institutions.

The author aims to explore the reasons behind why certain banks successfully navigate the restructuring process and withstand the effects of the global financial crisis, allowing them to continue their growth, while others struggle to do so This investigation will serve as the foundation for future research topics.

Vo, X V., & Nguyen, H H (2018) Bank restructuring and bank efficiency—The case of Vietnam Cogent Economics & Finance, 6(1), 1520423 https://doi.org/10.1080/23322039.2018.1520423

Vo, X V., & Nguyen, H H (2015) Does Bank Restructure Lead to Higher Efficiency? An Investigation of the Vietnamese Banking System, Singapore Economics Review, Namyang University, Singapore

Vo, X V., & Nguyen, H H (2015) The impact of restructuring on structure and performance – Evidence from Vietnam, Infiniti Conferrence, University of

Alexander, W E., Davis, J M., Ebrill, L P., & Lindgren, C J (Eds.)

(1997) Systemic bank restructuring and macroeconomic policy International Monetary Fund

Angelini, P., & Cetorelli, N (2003) The effects of regulatory reform on competition in the banking industry Journal of Money, Credit, and Banking, 35(5), 663-684

Ariff, M., & Can, L (2009) IMF bank-restructuring efficiency outcomes: Evidence from East Asia Journal of Financial Services Research, 35(2), 167-187

Athanasoglou&Brissimis (2004).The effect of M&A on bank efficiency in Greece

Avkiran, N K., & Rowlands, T (2008) How to better identify the true managerial performance: state of the art using DEA Omega, 36(2), 317-324

Baer, W., & Nazmi, N (2000).Privatization and restructuring of banks in Brazil The Quarterly Review of Economics and Finance, 40(1), 3-24

Banker, R D., Chang, H., & Lee, S Y (2010) Differential impact of Korean banking system reforms on bank productivity Journal of Banking &

Bauer, P W., Berger, A N., Ferrier, G D., & Humphrey, D B (1998) Consistency conditions for regulatory analysis of financial institutions: a comparison of frontier efficiency methods Journal of Economics and business, 50(2), 85-114

Beitel, Schiereck & Wahrenburg (2004) Explaining M&A success in European banks European Financial Management, 10(1), 109-139 h

Berg, S A., Fứrsund, F R., & Jansen, E S (1992).Malmquist indices of productivity growth during the deregulation of Norwegian banking, 1980-89 The

Berger, A N (1998) The efficiency effects of bank mergers and acquisition:

A preliminary look at the 1990s data In Bank Mergers & Acquisitions (pp 79-111) Springer, Boston, MA

Berger, A N., & Hannan, T H (1997) Using efficiency measures to distinguish among alternative explanations of the structure-performance relationship in banking Managerial Finance, 23(1), 6-31

Berger, A N., & Hannan, T H (1998) The efficiency cost of market power in the banking industry: A test of the “quiet life” and related hypotheses Review of

Berger, A N., & Humphrey, D B (1992).Measurement and efficiency issues in commercial banking In Output measurement in the service sectors (pp

Berger, A N., & Mester, L J (1997) Inside the black box: What explains differences in the efficiencies of financial institutions? Journal of banking & finance, 21(7), 895-947

Berger, A N., & Mester, L J (1997) Efficiency and productivity change in the US commercial banking industry: A comparison of the 1980s and 1990s (No

Berger, A N., Clarke, G R., Cull, R., Klapper, L., & Udell, G F (2005) Corporate governance and bank performance: A joint analysis of the static, selection, and dynamic effects of domestic, foreign, and state ownership Journal of

Berger, A N., Demsetz, R S., & Strahan, P E (1999) The consolidation of the financial services industry: Causes, consequences, and implications for the future Journal of Banking & Finance, 23(2), 135-194

Berger, A N., Hasan, I., & Zhou, M (2009) Bank ownership and efficiency in China: What will happen in the world’s largest nation? Journal of Banking &

Berger, A N., Hunter, W C., & Timme, S G (1993) The efficiency of financial institutions: A review and preview of research past, present and future Journal of Banking & Finance, 17(2), 221-249

Berger, De Young & Udell (2001).Efficiency barriers to the consolidation of the European financial services industry European Financial Management, 7(1), 117-130

Berger, R (1994) Power comparison of exact unconditional tests for comparing two binomial proportions Institute of Statistics Mimeo Series, (2266)

Bonin, J P., Hasan, I., &Wachtel, P (2005) Privatization matters: Bank efficiency in transition countries Journal of Banking & Finance, 29(8), 2155-2178

Borish, M S., Long, M F., & Noel, M (1995) Restructuring banks and enterprises: recent lessons from transition countries The World Bank

Boubakri, N., Cosset, J C., Fischer, K., & Guedhami, O (2005).Privatization and bank performance in developing countries Journal of Banking &

Brissimis, S N., Delis, M D., & Papanikolaou, N I (2008).Exploring the nexus between banking sector reform and performance: Evidence from newly acceded EU countries Journal of Banking & Finance, 32(12), 2674-2683 h

Charnes, A., Cooper, W W., & Rhodes, E (1978).Measuring the efficiency of decision making units European journal of operational research, 2(6), 429-444

Chortareas, G E., Girardone, C., &Ventouri, A (2013) Financial freedom and bank efficiency: Evidence from the European Union Journal of Banking &

Chortareas, G., Garza-Garcia, J G., & Girardone, C (2007) Performance- Structure and Market Power versus Efficiency in Latin American Banking Department of Accounting, Finance and Management

Clarke, G R., Cull, R., & Shirley, M M (2005) Bank privatization in developing countries: A summary of lessons and findings Journal of Banking &

Coccorese, P., & Pellecchia, A (2010).Testing the ‘quiet life’hypothesis in the Italian banking industry Economic Notes, 39(3), 173-202

Cook, W D., Tone, K., & Zhu, J (2014) Data envelopment analysis: Prior to choosing a model Omega, 44, 1-4

Cowling, K., & Waterson, M (1976) Price-cost margins and market structure Economica, 43(171), 267-274

Cybo-Ottone& Murgia (2000).Mergers and shareholder wealth in European banking Journal of Banking & Finance, 24(6), 831-859

Daniel, M J (1997) Fiscal aspects of bank restructuring International Monetary Fund

Das, A., & Ghosh, S (2006) Financial deregulation and efficiency: An empirical analysis of Indian banks during the post reform period Review of

Demirgỹỗ-Kunt, A., Huizinga, H., & Claessens, S (1998) How does foreign entry affect the domestic banking market? Journal of Banking & Finance, 25(5),

Demsetz, H (1973) Industry structure, market rivalry, and public policy The Journal of Law and Economics, 16(1), 1-9

Di Patti, E B., & Hardy, D C (2005) Financial sector liberalization, bank privatization, and efficiency: Evidence from Pakistan Journal of Banking &

Eckel, C., Eckel, D., & Singal, V (1997) Privatization and efficiency: Industry effects of the sale of British Airways Journal of Financial

Elyasiani, E., & Mehdian, S (1995) The comparative efficiency performance of small and large US commercial banks in the pre-and post- deregulation eras Applied Economics, 27(11), 1069-1079

Fane, G., & McLeod, R H (2002) Banking collapse and restructuring in Indonesia, 1997-2001 Cato J., 22, 277

Fethi, M D., Shaban, M., & Weyman-Jones, T (2012) Turkish Banking Recapitalization and the Financial Crisis: An Efficiency and Productivity Analysis Emerging Markets Finance and Trade, 48(sup5), 76-90

Fu, X M., & Heffernan, S (2009) The effects of reform on China’s bank structure and performance Journal of Banking & Finance, 33(1), 39-52

Fukuyama, H., & Weber, W L (2002).Estimating output allocative efficiency and productivity change: Application to Japanese banks European

Goldberg, L G., & Rai, A (1996).The structure-performance relationship for European banking Journal of Banking & Finance, 20(4), 745-771

Graf, P (1999) Policy responses to the banking crisis in Mexico BIS Policy

Harvie, C., Van Hoa, T., & Van Hoa, T (1997) Vietnam’s reforms and economic growth Springer

Hasan, I., & Marton, K (2003) Development and efficiency of the banking sector in a transitional economy: Hungarian experience Journal of Banking &

Havidz, S A H., &Setiawan, C (2015).Bank efficiency and non-performing financing (NPF) in the Indonesian Islamic banks Asian Journal of Economic

Hawkins, J., & Turner, P (1999) Bank restructuring in practice: an overview BIS Policy Paper, 6

Hicks, J R (1935) Annual survey of economic theory: the theory of monopoly Econometrica: Journal of the Econometric Society, 1-20

Hsiao, H C., Chang, H., Cianci, A M., & Huang, L H (2010) First financial restructuring and operating efficiency: evidence from Taiwanese commercial banks Journal of Banking & Finance, 34(7), 1461-1471

Isik, I., & Hassan, M K (2003) Financial disruption and bank productivity: the 1994 experience of Turkish banks The Quarterly Review of Economics and

Jeon, Y., & Miller, S M (2004).The effect of the Asian financial crisis on the performance of Korean nationwide banks Applied Financial Economics, 14(5), 351-360 h

Koetter, M., & Vins, O (2008) The quiet life hypothesis in banking:

Evidence from German savings banks (No 190) Working paper series: finance & accounting

Koetter, M., Kolari, J W., & Spierdijk, L (2012) Enjoying the quiet life under deregulation? Evidence from adjusted Lerner indices for US banks Review of

Koetter, M., Kolari, J., & Spierdijk, L (2008, February).Efficient competition?Testing the quiet life of US banks with adjusted Lerner indices

In Proceedings 44th Bank Structure and Competition Conference, Federal Reserve Bank of Chicago (pp 234-252)

Koutsomanoli-Filippaki, A., Margaritis, D., & Staikouras, C (2009).Efficiency and productivity growth in the banking industry of Central and Eastern Europe Journal of Banking & Finance, 33(3), 557-567

Kraft, E., & Tırtıroğlu, D (1998) Bank efficiency in Croatia: A stochastic- frontier analysis Journal of comparative economics, 26(2), 282-300

Kraft, E., Hofler, R., & Payne, J (2006) Privatization, foreign bank entry and bank efficiency in Croatia: a Fourier-flexible function stochastic cost frontier analysis Applied Economics, 38(17), 2075-2088

Krishnasamy, G., HanuumRidzwa, A., & Perumal, V (2004) Malaysian post merger banks’ productivity: application of Malmquist productivity index Managerial Finance, 30(4), 63-74

Kumbhakar, S., & Sarkar, S (2003) Deregulation, ownership, and productivity growth in the banking industry: evidence from India Journal of

Lambson, V E (1987) Is the concentration-profit correlation partly an artifact of lumpy technology? The American economic review, 77(4), 731-733 h

Landier, A., & Ueda, K (2009) The economics of bank restructuring:

Understanding the options (No 2009-2012) International Monetary Fund

Lee, C C., & Hsieh, M F (2013).The impact of bank capital on profitability and risk in Asian banking Journal of international money and finance, 32, 251-281

Leibenstein, H (1966) Allocative efficiency vs." X-efficiency" The

Lovell, C K., & Schmidt, P (1988).A comparison of alternative approaches to the measurement of productive efficiency In Applications of modern production theory: Efficiency and productivity (pp 3-32) Springer, Dordrecht

Maseno, O P (2014) The effect of Financial restructuring on the financial performance of commercial banks in Kenya Unpublished Master of Science in

Finance Thesis: University of Nairobi

Maudos, J., Pastor, J M., Perez, F., & Quesada, J (2002).Cost and profit efficiency in European banks Journal of international financial markets, institutions and money, 12(1), 33-58

McComb, R P., Gruben, W C., & Welch, J H (1994) Privatization and performance in the Mexican financial services industry The Quarterly Review of

Meeusen, W., & van Den Broeck, J (1977) Efficiency estimation from Cobb-Douglas production functions with composed error International economic review, 435-444

Modigliani, F., & Miller, M H (1958).The cost of capital, corporation finance and the theory of investment The American economic review, 261-297

Molyneux, P., & Forbes, W (1995).Market structure and performance in European banking Applied Economics, 27(2), 155-159 h

Musara, M., & Fatoki, O (2010) Has technological innovations resulted in increased efficiency and cost savings for banks' customers? African Journal of

Nakane, M I., & Weintraub, D B (2005) Bank privatization and productivity: Evidence for Brazil Journal of Banking & Finance, 29(8), 2259-

Oleka, D C., & Mgbodile, C C (2014) Recapitalization Reform and Banks’ Performance–Empirical Evidence from Nigeria Research journal of Finance and

Paradi, J C., & Zhu, H (2013).A survey on bank branch efficiency and performance research with data envelopment analysis Omega, 41(1), 61-79

Park, K H., & Weber, W L (2006).A note on efficiency and productivity growth in the Korean banking industry, 1992–2002 Journal of Banking &

Pazarbasioglu, C., & Dziobek, M C H (1997) Lessons from systemic bank restructuring: a survey of 24 countries (No 97-161) International Monetary Fund

Peltzman, S (1977).The gains and losses from industrial concentration The

Journal of Law and Economics, 20(2), 229-263

Philippon, T., & Schnabl, P (2013).Efficient recapitalization The Journal of

Pinprayong, B., &Siengtai, S (2012) Restructuring for organizational efficiency in the banking sector in Thailand: a case study of Siam Commercial Bank Far East Journal of Psychology and Business, 8(2), 29-42

Schulz, H (2006) Foreign Banks in Mexico: New Conquistadors or Agents of Change? Wharton Financial Institutions Center Working Paper, 06-11 h

Silberston, A (1972) Economies of scale in theory and practice The

Staub, Souza & Tabak (2010) Evolution of bank efficiency in Brazil: A DEA approach European journal of operational research, 202(1), 204-213

Tan, Y (2013) Essays on the analysis of performance and competitive condition in the Chinese banking industry (Doctoral dissertation, University of Portsmouth)

Thoraneenitiyan, N., & Avkiran, N K (2009) Measuring the impact of restructuring and country-specific factors on the efficiency of post-crisis East Asian banking systems: Integrating DEA with SFA Socio-Economic Planning

Tofallis, C (2001) Combining two approaches to efficiency assessment Journal of the Operational Research Society, 52(11), 1225-1231

Ngan, T H., Huan, N H., & Thao, T P (2015) Impact of restructuring on efficiency of vietnam’s commercial banks Journal of Economic Development, (JED, Vol 22 (2)), 27-47

Unite, A A., & Sullivan, M J (2003) The effect of foreign entry and ownership structure on the Philippine domestic banking market Journal of Banking

Vander Vennet, R., & Gropp, R (2003) Cross-border mergers in European banking and bank efficiency In Foreign direct investment in the real and financial sector of industrial countries(pp 295-321) Springer, Berlin, Heidelberg

Walker, M M (1998) Leveraged recapitalizations, operating efficiency, and stockholder wealth Financial Review, 33(3), 99-114 h

Waxman, M (1998) A legal framework for systemic bank restructuring Available at SSRN 172984

Weill, L (2003) Banking efficiency in transition economies Economics of transition, 11(3), 569-592

Wheelock, D C., & Wilson, P W (2000) Why do banks disappear? The determinants of US bank failures and acquisitions Review of Economics and

Williams & Nguyen (2005) Financial liberalisation, crisis, and restructuring:

A comparative study of bank performance and bank governance in South East Asia Journal of Banking & Finance, 29(8), 2119-2154

Wruck, K H (1990) Financial distress, reorganization, and organizational efficiency Journal of financial economics, 27(2), 419-444

Xiaogang, C., Skully, M., & Brown, K (2005) Banking efficiency in China: Application of DEA to pre-and post-deregulation eras: 1993–2000 China Economic

Ye, Q., Xu, Z., & Fang, D (2012).Market structure, performance, and efficiency of the Chinese banking sector Economic Change and Restructuring, 45(4), 337-358

Yudistira (2004) Efficiency in Islamic banking: An empirical analysis of 18 banks Islamic Economic Studies, 12(1), 2004

Zaim, O (1995) The effect of financial liberalization on the efficiency of Turkish commercial banks Applied Financial Economics, 5(4), 257-264 h

APENDIX A - List of banks in the study

1 An Binh Commercial Joint Stock Bank (ABBANK)

3 Eastern Asia Commercial Joint Stock Bank (DongABank)

4 Viet nam Export – Import Commercial Joint Stock Bank (EIB)

7 LienViet Post Commercial Joint Stock Bank

8 Maritime Commercial Joint Stock Bank (Maritimebank)

9 OCEAN Commercial Joint Stock Bank (Oceanbank)

10 Military Commercial Joint Stock Bank (MBB)

12 Vietnam International Commercial Bank (VIB Bank)

13 Vietnam Prosperity commercial joint-stock bank (VPBank)

14 Vietnam Technological and Commercial Bank (Techcombank)

15 Saigon-Hanoi Commercial Joint Stock Bank (SHB)

16 SaiGonThuong Tin Commercial Joint-stock Bank (Sacombank)

17 Southern Commercial Joint Stock Bank (Southernbank)

18 outheast Asia Commercial Joint Stock Bank (SeAbank)

20 Saigon Bank For Industry And Trade (Saigonbank)

21 Industrial and Commercial Bank of Vietnam (Vietinbank)

22 Bank For Investment And Development Of Vietnam (BIDV)

23 Bank for Foreign Trade of Vietnam (Vietcombank)

24 Mekong Delta Housing Development Bank (MHB)

25 Vietnam Bank for Agriculture and Rural Development (Agribank)

26 Hanoi Building Joint-stock Commercial Bank (Habubank) h

APPENDIX B - Model for testing the relationship of bank restructuring and efficiency lambda 0337037 0445473 0.76 0.449 -.0536074 1210148 sigma_v 4784325 0427328 11.20 0.000 4015991 5699655 sigma_u 016125 01979 0.81 0.415 0014549 1787194

_cons -8.254775 2.454585 -3.36 0.001 -13.06567 -3.443877 Usigma realdomesticcreditgrowth 0483719 0058619 8.25 0.000 0368827 059861 npl 2231217 0194998 11.44 0.000 1849028 2613406 changeintermsoftrade -11.18154 1.516109 -7.38 0.000 -14.15306 -8.210019 fiscalsurplusgdp -38.77375 realinterestrate 2134821 0455895 4.68 0.000 1241283 3028358 growth 69.15366 cop 1.254408 3350119 3.74 0.000 5977968 1.911019 si 3.009368 2732049 11.02 0.000 2.473896 3.54484 ma 1.723525 4161118 4.14 0.000 9079606 2.539089 Frontier totalloan Coef Std Err z P>|z| [95% Conf Interval]

Prob > chi2 = 0.0000 lambda 2046146 0837937 2.4e+07 0.000 2046146 2046146 sigma_v 3.59e-07 0000284 0.01 0.990 2.03e-74 6.37e+60 sigma_u 735286 0837938 8.77 0.000 5881021 9193056

_cons -.6149914 2279215 -2.70 0.007 -1.061709 -.1682734 Usigma realdomesticcreditgrowth -.0286525 4.43e-06 -6466.13 0.000 -.0286612 -.0286438 npl -.0094052 0000261 -360.44 0.000 -.0094564 -.0093541 changeintermsoftrade 7952452 fiscalsurplusgdp -35.05525 realinterestrate -.1144767 0000246 -4658.17 0.000 -.1145248 -.1144285 growth 59.47478 cop 367844 si 5949205 0001595 3730.32 0.000 5946079 5952331 ma -.5843642 Frontier noninterestrevenue Coef Std Err z P>|z| [95% Conf Interval]

Prob > chi2 = 0.0000 h lambda 0280804 0839903 0.33 0.738 -.1365376 1926984 sigma_v 5729239 0669365 8.56 0.000 4556679 7203531 sigma_u 0160879 0494018 0.33 0.745 0000391 6.611725

_cons -8.259371 6.14147 -1.34 0.179 -20.29643 3.777689 Usigma realdomesticcreditgrowth -.0290335 0084242 -3.45 0.001 -.0455447 -.0125224 npl 0230249 0289081 0.80 0.426 -.033634 0796839 changeintermsoftrade -4.01676 1.809068 -2.22 0.026 -7.562467 -.4710531 fiscalsurplusgdp -38.26057 5.900356 -6.48 0.000 -49.82505 -26.69609 realinterestrate -.1118735 0333228 -3.36 0.001 -.1771851 -.046562 growth 58.75496 cop -.0364539 3549319 -0.10 0.918 -.7321077 6591998 si 1.112031 5519707 2.01 0.044 0301879 2.193873 ma 1.567048 502845 3.12 0.002 5814895 2.552606 Frontier interestrevenue Coef Std Err z P>|z| [95% Conf Interval] Log likelihood = -72.4772 Wald chi2(8) = 205.80 lambda 028584 2867232 0.10 0.921 -.5333831 5905511 sigma_v 1.134145 2903795 3.91 0.000 6866444 1.873289 sigma_u 0324184 0458364 0.71 0.479 002029 5179724

_cons -6.858058 2.827802 -2.43 0.015 -12.40045 -1.315667 Usigma realdomesticcreditgrowth 0651225 0292208 2.23 0.026 0078508 1223943 npl 2398265 0756418 3.17 0.002 0915712 3880818 changeintermsoftrade -9.533401 4.16181 -2.29 0.022 -17.6904 -1.376404 fiscalsurplusgdp -52.83027 realinterestrate 0088172 0770705 0.11 0.909 -.1422382 1598725 growth 54.59113 cop 1.126487 7416781 1.52 0.129 -.3271752 2.580149 si 1.999474 6644424 3.01 0.003 6971909 3.301757 ma 1.848791 992321 1.86 0.062 -.0961222 3.793705 Frontier totaldeposit Coef Std Err z P>|z| [95% Conf Interval]

Prob > chi2 = 0.0000 h lambda 3573690 0475197 7.5e+07 0.000 3573690 3573690 sigma_v 1.23e-07 8.10e-06 0.02 0.988 7.23e-64 2.08e+49 sigma_u 4381101 0475197 9.22 0.000 3542073 5418874

_cons -1.65057 2169305 -7.61 0.000 -2.075746 -1.225394 Usigma realdomesticcreditgrowth -.0271108 1.38e-06 -2.0e+04 0.000 -.0271135 -.0271081 npl -.0219695 7.50e-06 -2927.53 0.000 -.0219842 -.0219548 changeintermsoftrade 1.089017 fiscalsurplusgdp -21.81918 realinterestrate -.0886737 8.74e-06 -1.0e+04 0.000 -.0886908 -.0886566 growth 1028433 cop 2266742 0000719 3153.46 0.000 2265333 2268151 si -.0936247 0000341 -2745.51 0.000 -.0936915 -.0935579 ma 6413692 Frontier noninterestexpense Coef Std Err z P>|z| [95% Conf Interval]

Prob > chi2 = 0.0000 lambda 0280593 1100208 0.26 0.799 -.1875775 2436961 sigma_v 8520562 1011405 8.42 0.000 6751949 1.075245 sigma_u 0239081 0479746 0.50 0.618 0004683 1.220637

_cons -7.467075 4.013247 -1.86 0.063 -15.33289 3987455 Usigma realdomesticcreditgrowth -.0022067 0150924 -0.15 0.884 -.0317872 0273738 npl 0986949 0354996 2.78 0.005 029117 1682727 changeintermsoftrade -7.094377 2.606738 -2.72 0.006 -12.20349 -1.985265 fiscalsurplusgdp -37.49864 9.377595 -4.00 0.000 -55.87839 -19.1189 realinterestrate -.1098786 0485585 -2.26 0.024 -.2050515 -.0147057 growth 64.42692 9.910549 6.50 0.000 45.0026 83.85124 cop 2624511 5303812 0.49 0.621 -.777077 1.301979 si 2.133488 5917036 3.61 0.000 9737703 3.293206 ma 1.776933 7466254 2.38 0.017 3135739 3.240292 Frontier interestexpense Coef Std Err z P>|z| [95% Conf Interval] Log likelihood = -103.6718 Wald chi2(9) = 565.03

APPENDIX C - Model for testing on structure and performance

Model 1: Market structure and performance hypothesis and relative market power hypotheses testing ROA

(V_b-V_B is not positive definite) Prob>chi2 = 0.0000 = 46.05 chi2(7) = (b-B)'[(V_b-V_B)^(-1)](b-B) 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 year 000615 0007105 -.0000955 ka 0673252 056183 0111422 0041206 la 0024168 0032737 -.0008569 0024245 xineff -.0093205 -.0042595 -.0050611 0072389 sineff -.0150314 -.0354458 0204144 0096053 ms 0216262 0015352 0200911 0091966 herf -.0355345 -.0176601 -.0178743 fe re Difference S.E.

(V_b-V_B is not positive definite) Prob>chi2 = 0.0371 = 14.92 chi2(7) = (b-B)'[(V_b-V_B)^(-1)](b-B) Test: Ho: difference in coefficients not systematic

The analysis reveals that the coefficient B is inconsistent under the alternative hypothesis (Ha) but efficient under the null hypothesis (Ho), as derived from the xtreg model In contrast, the coefficient b is consistent under both Ho and Ha The year variable shows a slight negative trend with coefficients of -.0009097, -.0007803, and -.0001294 The ka variable indicates a positive relationship with coefficients of 0568706 and 0730965, while la presents smaller positive values of 0085752 and 0048988 The xineff variable has coefficients of 0255272 and 0310498, suggesting a positive impact, whereas sineff shows a negative trend with coefficients of -.0426981 and -.0518549 The ms variable presents minor positive values of 0100557 and 0072394 Lastly, herf shows a slight negative relationship with coefficients of -.0166797 and -.012229, indicating variability in the results, as reflected in the differences and standard errors calculated.

Prob>chi2 = 0.0000 = 47.97 chi2(7) = (b-B)'[(V_b-V_B)^(-1)](b-B) 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 year -.1672568 -.1335933 -.0336635 0412352 ka 7053506 2137022 4916484 1.399954 la -.6472157 -.9285473 2813316 1.034664 xineff -1.022181 -.5276609 -.4945196 2.917422 sineff 2.437319 -.414674 2.851993 3.820092 ms -1.739477 -2.685584 9461062 2.958796 herf -10.03555 -6.68437 -3.351177 2.530796 fe1 re1 Difference S.E.

Prob>chi2 = 0.9170 = 2.63 chi2(7) = (b-B)'[(V_b-V_B)^(-1)](b-B) Test: Ho: difference in coefficients not systematic

The analysis reveals that under the null hypothesis (Ho), the estimates are consistent, while they appear inconsistent under the alternative hypothesis (Ha) Specifically, the xtreg results show a negative relationship with year (-0.01444) and ka (-0.1102292), while la exhibits a positive correlation (0.0347888) The efficiency metrics indicate that xineff is significantly positive (0.4848769), whereas sineff shows a notable negative value (-0.7267823) Additionally, the coefficients for ms (0.3192645) and herf (-0.4357955) further illustrate these trends, highlighting the differences between fixed effects (fe1) and random effects (re1) estimations.

Model 2: Quiet-life Hypothesis Testing

Prob>chi2 = 0.8662 = 1.87 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) 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 ka 0299525 0098569 0200956 0185797 la 0062571 -.0009269 0071839 0139638 year -.0194881 -.0195577 0000696 0003958 xineff 0020883 -.0035416 0056298 0366878 sineff 0532849 0040353 0492496 0478852 fe2 re2 Difference S.E.

Prob>chi2 = 0.0004 = 22.59 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) 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 ka 0236815 0197612 0039203 0100004 la 0242945 0081251 0161694 0040591 year -.0054464 -.0053177 -.0001286 0000911 xineff -.0124501 0026064 -.0150565 0168368 sineff 0409535 0073805 033573 017407 fe2 re2 Difference S.E. (b) (B) (b-B) sqrt(diag(V_b-V_B)) Coefficients

(V_b-V_B is not positive definite) Prob>chi2 = 0.9839 = 0.68 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) 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 ka -.0761577 -.0982309 0220731 0236871 la 0409194 056582 -.0156626 013656 year -.0063514 -.0067778 0004264 0003087 xineff 3296949 286849 0428459 0343075 sineff -.3680696 -.3518779 -.0161918 0468638 fe3 re3 Difference S.E.

Prob>chi2 = 0.0517 = 10.98 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) 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 ka -.0217912 -.0023062 -.0194851 0104009 la 029326 0218213 0075047 0038528 year -.0007924 -.0007675 -.0000249 0001658 xineff 1246002 1494795 -.0248793 0135342 sineff -.1925137 -.2261162 0336025 017534 re3 fe3 Difference S.E.

(V_b-V_B is not positive definite) Prob>chi2 = 0.0096 = 15.18 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) 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 ka -.2987633 -.2612067 -.0375566 0106108 la -.0733179 -.0662357 -.0070822 0159645 year -.0008095 0001146 -.0009241 ms 1813088 0969298 084379 0581104 herf 108219 1503845 -.0421655 fe4 re4 Difference S.E.

(V_b-V_B is not positive definite) Prob>chi2 = 0.0000 = 63.69 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) 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 ka -.153569 -.1536272 0000581 la 0093939 -.0016987 0110926 0012996 year 0050685 0058003 -.0007318 ms 4765336 1815321 2950015 1258883 herf -.0087884 0787496 -.0875379 fe4 re4 Difference S.E.

Random Effects And Fixed Effects Execution

(V_b-V_B is not positive definite) Prob>chi2 = 0.0000 = 43.28 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) Test: Ho: difference in coefficients not systematic

In the analysis, the coefficients indicate that B is inconsistent under the alternative hypothesis (Ha) but efficient under the null hypothesis (Ho), while b is consistent under both Ho and Ha The results from the xtreg model show a significant difference in coefficients, with a chi-squared value of 29.07 and a p-value of 0.0000, suggesting that the difference in coefficients is systematic Specifically, the estimates for variables such as year and market share (ms) reveal variations, with B showing a higher inconsistency The calculated standard errors indicate that the variance of the differences is not positive definite, which may affect the interpretation of the results Overall, the findings underscore the importance of considering the efficiency of estimators in regression analysis.

F test that all u_i=0: F(21, 81) = 5.23 Prob > F = 0.0000 rho 78482236 (fraction of variance due to u_i) sigma_e 00384871 sigma_u 00735025

_cons -1.224516 1.504093 -0.81 0.418 -4.217189 1.768157 year 000615 0007471 0.82 0.413 -.0008715 0021014 ka 0673252 0072516 9.28 0.000 0528968 0817536 la 0024168 0052491 0.46 0.646 -.0080273 0128609 xineff -.0093205 014762 -0.63 0.530 -.0386922 0200512 sineff -.0150314 0191366 -0.79 0.434 -.0531071 0230444 ms 0216262 0133863 1.62 0.110 -.0050084 0482609 herf -.0355345 0378237 -0.94 0.350 -.1107918 0397229 roa Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = -0.3288 Prob > F = 0.0000 F(7,81) = 26.78 overall = 0.3853 max = 8 between = 0.0687 avg = 5.0 within = 0.6983 min = 1

Group variable: stt Number of groups = 22

Fixed-effects (within) regression Number of obs = 110 rho 23720536 (fraction of variance due to u_i) sigma_e 00666696 sigma_u 0037178

_cons 1.568912 92984 1.69 0.092 -.2535408 3.391365 year -.0007803 0004595 -1.70 0.089 -.0016809 0001203 ka 0730965 0134136 5.45 0.000 0468064 0993866 la 0048988 0050094 0.98 0.328 -.0049195 014717 xineff 0310498 014427 2.15 0.031 0027733 0593263 sineff -.0518549 0212077 -2.45 0.014 -.0934211 -.0102886 ms 0072394 0192255 0.38 0.707 -.030442 0449207 herf -.012229 0750633 -0.16 0.871 -.1593504 1348924 roa Coef Std Err z P>|z| [95% Conf Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 Wald chi2(7) = 50.24 overall = 0.2405 max = 9 between = 0.4111 avg = 8.4 within = 0.1625 min = 5

Group variable: stt Number of groups = 26

Random-effects GLS regression Number of obs = 218

F test that all u_i=0: F(21, 81) = 1.11 Prob > F = 0.3581 rho 18926736 (fraction of variance due to u_i) sigma_e 89475587 sigma_u 43231827

_cons 337.5712 349.6748 0.97 0.337 -358.1719 1033.314 year -.1672568 1736831 -0.96 0.338 -.5128317 1783181 ka 7053506 1.685866 0.42 0.677 -2.648993 4.059694 la -.6472157 1.220331 -0.53 0.597 -3.075291 1.78086 xineff -1.022181 3.431896 -0.30 0.767 -7.850578 5.806217 sineff 2.437319 4.448907 0.55 0.585 -6.41461 11.28925 ms -1.739477 3.112086 -0.56 0.578 -7.931553 4.452598 herf -10.03555 8.793337 -1.14 0.257 -27.53153 7.460436 roe Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = -0.0077 Prob > F = 0.7637 F(7,81) = 0.59 overall = 0.1065 max = 8 between = 0.2360 avg = 5.0 within = 0.0484 min = 1

Group variable: stt Number of groups = 22

Fixed-effects (within) regression Number of obs = 110

F test that all u_i=0: F(25, 185) = 4.05 Prob > F = 0.0000 rho 33440827 (fraction of variance due to u_i) sigma_e 07017813 sigma_u 0497435

_cons 29.1386 10.03883 2.90 0.004 9.333287 48.94392 year -.01444 0049637 -2.91 0.004 -.0242328 -.0046472 ka -.1102292 1530638 -0.72 0.472 -.4122042 1917458 la 0347888 0611042 0.57 0.570 -.0857618 1553394 xineff 4848769 2318145 2.09 0.038 0275371 9422167 sineff -.7267823 292131 -2.49 0.014 -1.303119 -.1504458 ms 3192645 5938354 0.54 0.591 -.8522954 1.490825 herf -.4357955 81159 -0.54 0.592 -2.036957 1.165366 roe Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = -0.1069 Prob > F = 0.0000 F(7,185) = 5.53 overall = 0.1139 max = 9 between = 0.0185 avg = 8.4 within = 0.1730 min = 5

Group variable: stt Number of groups = 26

Fixed-effects (within) regression Number of obs = 218 h

F test that all u_i=0: F(21, 83) = 0.18 Prob > F = 1.0000 rho 1508443 (fraction of variance due to u_i) sigma_e 01155314 sigma_u 00486935

_cons 39.24072 1.200129 32.70 0.000 36.85372 41.62773 ka 0299525 0211564 1.42 0.161 -.0121266 0720317 la 0062571 0156739 0.40 0.691 -.0249177 0374319 year -.0194881 0005976 -32.61 0.000 -.0206766 -.0182995 xineff 0020883 0423044 0.05 0.961 -.0820535 08623 sineff 0532849 0547609 0.97 0.333 -.0556323 162202 herf Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = -0.0870 Prob > F = 0.0000 F(5,83) = 262.34 overall = 0.9407 max = 8 between = 0.9718 avg = 5.0 within = 0.9405 min = 1

Group variable: stt Number of groups = 22

Fixed-effects (within) regression Number of obs = 110 rho 0 (fraction of variance due to u_i) sigma_e 00639293 sigma_u 0

_cons 10.79384 3872726 27.87 0.000 10.0348 11.55288 ka 0197612 0095533 2.07 0.039 001037 0384854 la 0081251 0032423 2.51 0.012 0017704 0144799 year -.0053177 0001929 -27.56 0.000 -.0056958 -.0049396 xineff 0026064 0096972 0.27 0.788 -.0163999 0216126 sineff 0073805 015346 0.48 0.631 -.0226971 0374581 herf Coef Std Err z P>|z| [95% Conf Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 Wald chi2(5) = 937.95 overall = 0.8156 max = 9 between = 0.7652 avg = 8.4 within = 0.8212 min = 5

Group variable: stt Number of groups = 26

Random-effects GLS regression Number of obs = 218

F test that all u_i=0: F(21, 83) = 35.54 Prob > F = 0.0000 rho 8740695 (fraction of variance due to u_i) sigma_e 03264392 sigma_u 08600226

_cons 12.75725 3.391017 3.76 0.000 6.012655 19.50185 ka -.0761577 0597783 -1.27 0.206 -.1950543 0427388 la 0409194 0442874 0.92 0.358 -.0471663 1290052 year -.0063514 0016884 -3.76 0.000 -.0097096 -.0029932 xineff 3296949 119533 2.76 0.007 0919486 5674412 sineff -.3680696 1547293 -2.38 0.020 -.67582 -.0603193 ms Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = 0.0867 Prob > F = 0.0000 F(5,83) = 6.93 overall = 0.1018 max = 8 between = 0.1043 avg = 5.0 within = 0.2944 min = 1

Group variable: stt Number of groups = 22

Fixed-effects (within) regression Number of obs = 110

F test that all u_i=0: F(25, 187) = 142.96 Prob > F = 0.0000 rho 97404394 (fraction of variance due to u_i) sigma_e 00873717 sigma_u 05352302

_cons 1.563113 584294 2.68 0.008 4104584 2.715768 ka -.0023062 0189017 -0.12 0.903 -.039594 0349817 la 0218213 0071 3.07 0.002 0078148 0358278 year -.0007675 0002916 -2.63 0.009 -.0013428 -.0001923 xineff 1494795 0265545 5.63 0.000 0970946 2018644 sineff -.2261162 031715 -7.13 0.000 -.2886814 -.1635511 ms Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = -0.3483 Prob > F = 0.0000 F(5,187) = 15.50 overall = 0.0466 max = 9 between = 0.1346 avg = 8.4 within = 0.2930 min = 5

Group variable: stt Number of groups = 26

Fixed-effects (within) regression Number of obs = 218 rho 78830775 (fraction of variance due to u_i) sigma_e 03511076 sigma_u 06775411

_cons -.1098049 13.69999 -0.01 0.994 -26.9613 26.74169 ka -.2612067 0503499 -5.19 0.000 -.3598907 -.1625226 la -.0662357 0436361 -1.52 0.129 -.1517609 0192895 year 0001146 0068044 0.02 0.987 -.0132219 0134511 ms 0969298 0996703 0.97 0.331 -.0984203 29228 herf 1503845 3436103 0.44 0.662 -.5230794 8238483 xineff Coef Std Err z P>|z| [95% Conf Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 Wald chi2(5) = 33.25 overall = 0.0000 max = 8 between = 0.0439 avg = 5.0 within = 0.3408 min = 1

Group variable: stt Number of groups = 22

Random-effects GLS regression Number of obs = 110

F test that all u_i=0: F(25, 187) = 32.18 Prob > F = 0.0000 rho 85744547 (fraction of variance due to u_i) sigma_e 02632508 sigma_u 06456282

_cons -10.08815 3.68309 -2.74 0.007 -17.35389 -2.822402 ka -.153569 0542807 -2.83 0.005 -.2606502 -.0464879 la 0093939 022793 0.41 0.681 -.0355706 0543584 year 0050685 0018201 2.78 0.006 001478 008659 ms 4765336 1927319 2.47 0.014 0963254 8567419 herf -.0087884 299389 -0.03 0.977 -.5994023 5818256 xineff Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = -0.4711 Prob > F = 0.0000 F(5,187) = 16.82 overall = 0.0027 max = 9 between = 0.0057 avg = 8.4 within = 0.3103 min = 5

Group variable: stt Number of groups = 26

Fixed-effects (within) regression Number of obs = 218 rho 6815349 (fraction of variance due to u_i) sigma_e 02708452 sigma_u 0396218

_cons -13.66064 11.29293 -1.21 0.226 -35.79437 8.473094 ka -.2147845 0406126 -5.29 0.000 -.2943837 -.1351853 la 0357204 0344124 1.04 0.299 -.0317267 1031675 year 0068037 0056089 1.21 0.225 -.0041895 017797 ms -.0944605 0755141 -1.25 0.211 -.2424654 0535444 herf 3358774 2826845 1.19 0.235 -.2181741 8899289 sineff Coef Std Err z P>|z| [95% Conf Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 Wald chi2(5) = 33.08 overall = 0.0126 max = 8 between = 0.1618 avg = 5.0 within = 0.3911 min = 1

Group variable: stt Number of groups = 22

Random-effects GLS regression Number of obs = 110

F test that all u_i=0: F(25, 189) = 15.06 Prob > F = 0.0000 rho 88608878 (fraction of variance due to u_i) sigma_e 02083797 sigma_u 05811802

_cons -6.528026 2.897542 -2.25 0.025 -12.2437 -.8123481 ka 0611579 0418474 1.46 0.146 -.0213901 1437058 la 0263284 0178633 1.47 0.142 -.0089086 0615653 year 0032427 0014319 2.26 0.025 000418 0060673 ms -.7630294 1522052 -5.01 0.000 -1.063269 -.4627901 herf 4788155 2356289 2.03 0.044 0140151 9436159 sineff Coef Std Err t P>|t| [95% Conf Interval] corr(u_i, Xb) = -0.8615 Prob > F = 0.0000 F(5,189) = 8.28 overall = 0.0889 max = 9 between = 0.1861 avg = 8.5 within = 0.1798 min = 5

Group variable: stt Number of groups = 26

Fixed-effects (within) regression Number of obs = 220 h

Mean VIF 6.90 ka 1.25 0.798853 la 1.36 0.734213 ms 1.38 0.723367 xineff 2.17 0.460818 sineff 2.35 0.424641 year 19.84 0.050412 herf 19.95 0.050134 Variable VIF 1/VIF

Mean VIF 2.90 ka 1.41 0.708057 la 1.73 0.578814 xineff 1.93 0.517091 sineff 2.04 0.490137 ms 2.08 0.480379 herf 5.43 0.184105 year 5.69 0.175607 Variable VIF 1/VIF

Mean VIF 6.90 ka 1.25 0.798853 la 1.36 0.734213 ms 1.38 0.723367 xineff 2.17 0.460818 sineff 2.35 0.424641 year 19.84 0.050412 herf 19.95 0.050134 Variable VIF 1/VIF

Mean VIF 2.90 ka 1.41 0.708057 la 1.73 0.578814 xineff 1.93 0.517091 sineff 2.04 0.490137 ms 2.08 0.480379 herf 5.43 0.184105 year 5.69 0.175607 Variable VIF 1/VIF

Mean VIF 1.56 year 1.02 0.975942 ka 1.07 0.938072 la 1.21 0.826495 xineff 2.16 0.462354 sineff 2.35 0.426331 Variable VIF 1/VIF

Mean VIF 1.45 ka 1.08 0.926481 year 1.24 0.809223 la 1.35 0.740180 sineff 1.72 0.580131 xineff 1.86 0.538957 Variable VIF 1/VIF

Mean VIF 1.56 year 1.02 0.975942 ka 1.07 0.938072 la 1.21 0.826495 xineff 2.16 0.462354 sineff 2.35 0.426331 Variable VIF 1/VIF

Mean VIF 1.45 ka 1.08 0.926481 year 1.24 0.809223 la 1.35 0.740180 sineff 1.72 0.580131 xineff 1.86 0.538957 Variable VIF 1/VIF

Mean VIF 8.69 la 1.13 0.882075 ka 1.20 0.832231 ms 1.38 0.726256 year 19.79 0.050523 herf 19.94 0.050156 Variable VIF 1/VIF

Mean VIF 3.09 ka 1.26 0.791246 la 1.59 0.629968 ms 1.75 0.570295 herf 5.41 0.184936 year 5.45 0.183323 Variable VIF 1/VIF

Mean VIF 8.69 la 1.13 0.882075 ka 1.20 0.832231 ms 1.38 0.726256 year 19.79 0.050523 herf 19.94 0.050156 Variable VIF 1/VIF

Mean VIF 3.10 ka 1.27 0.787582 la 1.58 0.631229 ms 1.73 0.576566 herf 5.44 0.183905 year 5.47 0.182717 Variable VIF 1/VIF h

Instruments: herf year la ka L.ms

_cons 5414847 22.58186 0.02 0.981 -44.34976 45.43273 ka 2683197 09798 2.74 0.008 073542 4630975 la 0816033 0414903 1.97 0.052 -.0008766 1640833 year -.0002923 0112162 -0.03 0.979 -.0225894 0220048 herf 0256058 5581079 0.05 0.964 -1.083876 1.135088 ms 0302671 0748334 0.40 0.687 -.1184967 179031 sineff Coef Std Err t P>|t| [95% Conf Interval]

Total 332529195 91 003654167 Root MSE = 05713 Adj R-squared = 0.1069 Residual 280650685 86 00326338 R-squared = 0.1560 Model 05187851 5 010375702 Prob > F = 0.0115 F( 5, 86) = 3.16 Source SS df MS Number of obs = 92 Instrumental variables (2SLS) regression

ivregress 2sls sineff (ms=L.ms) herf year la ka, small

Instruments: herf year la ka L.ms

_cons 16.07092 27.31464 0.59 0.558 -38.22879 70.37062 ka 2035502 118515 1.72 0.089 -.0320498 4391501 la -.0507959 050186 -1.01 0.314 -.1505624 0489705 year -.0079467 013567 -0.59 0.560 -.0349169 0190235 herf -.2747091 6750781 -0.41 0.685 -1.61672 1.067302 ms 0264193 0905172 0.29 0.771 -.153523 2063616 xineff Coef Std Err t P>|t| [95% Conf Interval]

Total 433090249 91 004759234 Root MSE = 0691 Adj R-squared = -0.0032 Residual 41061785 86 004774626 R-squared = 0.0519 Model 022472399 5 00449448 Prob > F = 0.4664 F( 5, 86) = 0.93 Source SS df MS Number of obs = 92 Instrumental variables (2SLS) regression

ivregress 2sls xineff (ms=L.ms) herf year la ka, small h

Instruments: herf year la ka L.ms

_cons -7.129513 4.138207 -1.72 0.087 -15.2928 1.033774 ka 3290458 0640629 5.14 0.000 2026713 4554204 la -.0235908 0194028 -1.22 0.226 -.0618661 0146844 year 0035294 0020425 1.73 0.086 -.0004999 0075586 herf 2724843 3672989 0.74 0.459 -.4520726 9970412 ms 4537173 0624751 7.26 0.000 3304751 5769596 sineff Coef Std Err t P>|t| [95% Conf Interval]

Total 26871563 193 001392309 Root MSE = 03217 Adj R-squared = 0.2566 Residual 194595607 188 001035083 R-squared = 0.2758 Model 074120024 5 014824005 Prob > F = 0.0000 F( 5, 188) = 15.06 Source SS df MS Number of obs = 194 Instrumental variables (2SLS) regression

ivregress 2sls sineff (ms=L.ms) herf year la ka, small

Instruments: herf year la ka L.ms

_cons -24.51434 7.221897 -3.39 0.001 -38.7617 -10.26698 ka 2165374 1166372 1.86 0.065 -.0135644 4466392 la -.1409412 033906 -4.16 0.000 -.207831 -.0740514 year 0122365 0035644 3.43 0.001 0052047 0192683 herf 6634921 6410707 1.03 0.302 -.6012122 1.928196 ms 18762 1094473 1.71 0.088 -.0282977 4035377 xineff Coef Std Err t P>|t| [95% Conf Interval]

Total 68435133 191 003582991 Root MSE = 05567 Adj R-squared = 0.1350 Residual 576435514 186 003099116 R-squared = 0.1577 Model 107915816 5 021583163 Prob > F = 0.0000 F( 5, 186) = 7.12 Source SS df MS Number of obs = 192Instrumental variables (2SLS) regression h