Credit risk management case study of BIDV

The information they provided, especially frommanagers/vice managers, allowed me to get deeperunderstanding about credit risk management in BIDV andderiving the findings of this study..

MINISTRY OF EDUCATION AND TRAINING

UNIVERSITY OF ECNOMICS HOCHIMINH CITY

BÙI NGUYÊN NGỌC

CREDIT RISK MANAGEMENT:

CASE STUDY OF BIDV

MASTES’S THESIS

In Banking Ology code: 60.31.12 Supervisor: Dr Nguyễn Văn Phúc

Ho Chi Minh City – 2010

 -Page i

ACKNOWLEDGMENT

I owe a debt of gratitude to many people who helped mecomplete this thesis I would like to acknowledge the help of all

First of all I would like to express my deepest acknowledgement

to my supervisor, Doctor Phóc Nguyễn Văn, for his valuable adviceand recommendations

Then, I would like to thank my superiors and colleagues whoagreed to be interviewed and/or completed the surveyquestionnaires The information they provided, especially frommanagers/vice managers, allowed me to get deeperunderstanding about credit risk management in BIDV andderiving the findings of this study

Finally, I want to express my sincere thanks to every member of

my family for their encouragement and support during the time Idevoted to this dissertation

Page i

Page ii

ABSTRACT

Credit risk is one of the most popular risks in banks due totheir intermediary functions: lending and borrowing An excessivelevel of credit risk may destroy not only banks’ profitability butalso the stability of global banking system Therefore, it isnecessary for banks to develop an effective credit riskmanagement strategy not only to protect themselves but also toprevent banking crises

In case of BIDV, BIDV is one of four State Banks establishedwhen Viet Nam banking system is at a very early stage ofdevelopment For a long time, BIDV was controlled in allocatingloans by government Therefore, credit risk management has beenthe main challenge facing the board of BIDV managers With thebest try of this board, since 2008, BIDV has controlled creditrisk that comply with international standard (non-performing-loan ratio was less than 3%) This is the main reason thatdrove this study to describe credit risk management in BIDV, toknow why non-performing loan ratio in BIDV has been sharplyreduced from 38.3% in 2004 to 2.82% in 2009

Both secondary data and primary data are needed for this study.Collected data is analyzed by Statistical Package for SocialStudies version 16.0 (SPSS) Cronbach alpha is used to measurecoefficient of reliability and t-test technique is used to test thehypotheses about the four factors influence reduction of non-performing-loan ratio in BIDV These techniques and tools help

Page ii

collected data transform into information that will answer theresearcher’s questions

Page 3

LIST OF FIGURES

Figure 1.1: Structure of chapter 1 2

Figure 1.2: Field of research problem 4

Figure 1.3: Method of secondary data 7

Figure 1.4: Population and sampling 8

Figure 1.5: Quota sampling method 9

Figure 1.6: Structure of the study 12

Figure 3.1: BIDV Organization Chart 40

Figure 3.2: BIDV’s non-performing loans 45

Figure 3.3: BIDV’s loan structure by collateral 46

Figure 3.4: BIDV’s loan structure by economic sector 47

Figure 4.1: Respondents’ position 57

Figure 4.2: Respondents’ working years in BIDV 58

Page 4

LIST OF TABLES

Table 2.1: Level of specific provision 20

Table 2.2: Example of a loan rating system and bond rating mapping .23

Table 2.3: Strategies for reducing and scoping with portfolio credit risk 26

Table 3.1: BIDV’s key performance indicators 41

Table 3.2: BIDV’s credit indicators 43

Table 3.3: Loan classification in BIDV 49

Table 3.4: Summarize four factors influencing NPL ratio in BIDV 52

Table 4.1: Four variables with different aspects 58

Table 4.2: Level of agreement in survey questionnaire 59

Table 4.3: The overall score of Cronbach’s alpha 60

Table 4.4: The t-test result 61

Table 4.5: Summary of hypotheses testing results 64

LIST OF APPENDICES Appendix A 74

Appendix B 75

Appendix C 79

TABLE OF CONTENTS

Acknowledgment iAbstract iiList

of figures iiiList

of tables ivList

of appendices ivTable of contents v

Chapter 1: Introduction 1

1.1 Introduction 1

1.2 Rationale of the study 2

1.3 Statement of the problem and the scope of the study3

1.4 Research questions and objectives4

1.5 Methodology 5

1.5.1 Research design5

1.5.2 Data collection 6

2.4.4 Non-performing loan21

2.5 Credit risk measurement21

2.5.1 Traditional approaches21

2.5.2 Modern approaches24

2.6 External factors that affect the level of credit risk27

2.6.1 Financial deregulation28

2.6.2 Supervision and re-regulation28

2.6.3 Competition.29

2.6.4 The recent financial crisis30

2.7 Internal factors that affect the level of credit risk30

2.7.1 Credit information.30

2.7.2 Technology32

2.7.3 Credit staffs33

2.7.4 Loan policy34

2.8 Summary35

Chapter 3: Case study of BIDV 37

3.1 Introduction37

3.2 Overview of BIDV37

3.2.2 Organization structure37

3.2.3 BIDV business performance41

3.3 Lending business 433.3.1 Overview43

3.3.2 Non-performing loans and loan loss provision44

3.3.3 Loan structure45

3.4 Internal factors that influence non-performing-loan ratio in BIDV47

3.4.1 Credit information.47

3.4.2 Technology48

3.4.3 Credit staff50

3.4.4 Loan policy51

Page vii

3.4.5 Suggesting hypotheses52

3.5 Summary55

Chapter 4: Data analysis and findings 56

4.1 Introduction56

4.2 Data collection results56

4.3 Data analysis57

4.3.1 Descriptive statistic57

4.3.2 Measures of reliability58

4.3.3 Statistical hypotheses testing (t-test)60

4.4 Comparison and discussion of findings62

4.4.1 Credit information.62

4.4.2 Technology62

4.4.3 Credit staffs63

4.4.4 Loan policy63

4.5 Result of hypotheses testing64

4.6 Summary64

Chapter 5: Recommendation and Conclusion 66

5.1 Introduction66

Page vii

5.2 Reviewing research questions66

5.3 Recommendation for BIDV66

5.3.1 Credit information66

5.3.2 Technology67

5.3.3 Credit staffs67

5.3.4 Loan policy68

5.4 Recommendation for other banks68

5.5 Limitation of the research69

5.6 Summarizing and concluding the dissertation69

References 70

Section 1.1 provides a general introduction to the chapter andsection 1.2 examines the research background where the researchproblem is identified Section 1.3 defines the statement of theproblem and scope of the study.

Section 1.4 which includes two subsections 1.4.1 and 1.4.2defines the research questions and research objectives.Subsection 1.4.1 addresses the research questions that will berespectively answered in chapters of the study Subsection 1.4.2presents research objectives that the study covers in theprocess of solving the research problem defined

Section 1.5 discusses the aspects of research methodology such

as selecting from alternative types of research, research designand research techniques Section 1.6 points out the significanceand scope of the study, and finally section 1.7 describes overallstructure of the thesis

Chapter 1:

Section 1: IntroductionSection 2: Rationale of the study

Section 3: Statement of the problem and

scope of the study

Section 4: Research questions and objectives

Section 5: Methodology

Section 6: Significance of the studySection 7: Structure of the study

Figure 1.1: Structure of chapter 1

1.2Rationale of the study:

In today’s world, in order to meet customers’ requirements, there

is a need for banks to diversify their business including otheractivities such as payments, leasing, and investments besidesthe two traditional functions of lending and borrowing However,lending still plays an important role in banks because banks’revenue primarily comes from lending revenue which contributesover a half of bank total operating (about 70% in case of BIDV).The traditional way that banks make their profit is to take risk inexchange for an acceptable return to not only cover the cost offunding but also maintain their profitability Thus, the mainbusiness of banks is not, as everyone might assume, takingdeposits and making loans but minimizing the risk in collecting

interest and principle from the loans which is known as managingcredit risk (Burton & Lombra 2006).

Credit risk is usually associated with banks because of theirintermediary function which is channeling funds from people whohave fund surplus to those who have fund deficit for theirinvestment opportunities (Mishkin & Eakins 2006) Historically,financial crises are usually derived from the failure of banks tomanage credit risk from poor quality loans or high probability ofcustomers’ default (Yarbrough & Yarbrough 2006).

In case of BIDV, BIDV is one of four State Banks establishedwhen Viet Nam banking system is at a very early stage ofdevelopment For a long time, BIDV was controlled in allocatingloans by government Therefore, credit risk management has beenthe main challenge facing the board of BIDV managers With thebest try of this board, since 2008, BIDV has controlled creditrisk that comply with international standard (non-performing-loan ratio was less than 3%) This is the main reason that drovethis study to describe BIDV credit risk management, to know whynon-performing loan ratio in BIDV has been rapidly reduced from38.3% in 2004 to 2.82% in 2009

1.3Statement of the problem and scope of the study

This study conducts with particular emphasis on why performing-loan ratio in BIDV has been rapidly reduced from38.3% in 2004 to 2.82% in 2009

non-Stemming from the reason mentioned in the rationale part, thisresearch will focus two main parts: First is credit risk managementbackground; second is case study of BIDV credit riskmanagement The first part provides some background knowledgeabout credit risk, credit risk measurements, management and thefactors that influence credit risk The second part analyses BIDV’scase in reducing non- performing-loan ratio This part presents

three main subparts including overview of BIDV; analysis creditactivities, application of credit risk management theory to itscredit activity practice; consideration of four factors includingcredit information, technology, credit staff and loan policy duringthe period 2004-2009, and suggesting hypotheses These analysisand consideration help the researcher realize that credit riskmanagement is one of significant achievements of BIDV because

at the end of 2009, BIDV has controlled credit risk underinternational standard (non-performing-

The four factors will be presented throughout this study Firstly,this study reviews literature related to the four factors and creditrisk management theory Secondly, by analyzing BIDV’s creditbusiness practice, this paper shows how the four main factorsinfluence BIDV on reducing non-performing-loan ratio Finally,findings from the survey by questionnaires confirm the aboverelationships.

Figure 1.2: Fields of the research problem1.4 Research Questions and Research Objectives:

1.4.1 Research Questions

Research questions involve the research translation of “problem”into the need for inquiry (Zikmund, 1997, p.88) The researchproblem defined above leads to the following research questions:

What are factors that influence non-performing-loan ratio in BIDV?

How has BIDV applied credit risk management theory to

practice?

1.4.2 Research Objectives

This study is conducted with the purpose of:

To know the main factors leading to BIDV success in reducing non-performing loan ratio,

To consider whether BIDV applies theory to manage its credit risk

or not

1.4.3 Research hypotheses

Aiming to confirm the influence of four factors including creditinformation, technology, credit staffs and loan policy on reducingnon-performing-loan ratio in BIDV, the researcher assumeshypotheses as follows:

H1: Credit information variable influences non-performing-loan

ratio in BIDV H2: Technology variable influences

non-performing-loan ratio in BIDV

H3: Credit staffs variable influences non-performing-loan

ratio in BIDV H4: Loan policy variable influences

non-performing-loan ratio in BIDV

of agreement about factors that influence non-performing- loanratio,< ) and non-numerical data (respondents’ background,position and their suggestion to help BIDV continuously reducenon-performing-loan ratio,< ) are needed to answer the researchquestions

According to G.Zikmund (1997), there are four basic designtechniques: survey, experiment, secondary data and observation.This research utilizes both survey and secondary data methods.Based on the objectives of the research, survey method helpsthe researcher collect primary data in order to indentify of fourfactors influencing non-performing-loan ratio in BIDV whilesecondary data methods is necessary for the researcher tounderstand credit risk background and describe BIDV’s situation

in managing credit risk or reducing non-performing-loan ratio

Since this study applies perception survey which investigatesthe feeling of respondents about the research problem, thefindings are influenced by subjective judgment of respondents.However, the researcher also utilizes secondary technique for thepurpose of exploring evidences to confirm the research problem.1.5.2 Data collection

This section will describe the way in which data including bothprimary and secondary data from a variety source of informationwas collected Secondary data was collected from availablesources such as books, previous researches, BIDV’s annualreports, financial journals and magazines while primary data wasobtained through surveys and interviews conducted by the author

1.5.2.1 Secondary data

There are many advantages by using secondary data inconducting a research First, it is economical in the way thatcollecting available data is almost always less expensive thancollecting firsthand data through a study In addition, collectingsecondary data help researcher save a huge amount of timespending analyzing and interpreting the data collected Second,

in some cases, secondary data is the only source that researchercan collect from the previous periods Finally, unlike primary data,secondary data is generally permanent and available in a formthat is easily checked and collected by others (Zikmund, 1997).There are many types of secondary data such as documentarysecondary data, multiple source secondary data and surveybased secondary data (Saunders, Lewis

& Thornhill 2007) However, this study focuses only on thedocumentary secondary data source which includes BIDVinternal materials (like BIDV regulations and annual reports

which collected from the internet and intranet website of BIDV)and other written materials (such as previous researches, books,journals, newspapers and magazines) These kinds of secondarydata are important raw data sources for this study.

For this study, the information from written materials like previous researches, books, financial magazines, journals were used to build up literature review while BIDV s annual reports and regulations which collected

Secondary data

Previous researches, books,

journals, newspapers Annual reports, regulations (BIDV internal data)

from the official internet and intranet websites of BIDV were used to provide a clear picture about credit risk management of BIDV.

Figure 1.3: Method of secondary data collection

Based on the above advantages of secondary data, the

researcher decided to use secondary data as one of the sources of information in order to conduct this study.

1.5.2.2 Primary data

Beside secondary data, the researcher uses primary data

in order to get the feelings of respondents about the problem of the study The purpose of this study is to know the

factors leading to BIDV success in credit risk management.Therefore, the target population in this research is all BIDV creditstaff whose daily work relevant to lending business

It would be impracticable for this study to collect all dataavailable from the entire population of all BIDV credit staffsbecause of the limited time and financial sources Thus, thisresearch will be conducted with a sample size of 100 BIDVcredit staffs including 20 managers/vice managers, 30 creditdepartment leaders and 50 credit officers

20 managers

30 credit dept leaders

Population (BIDV credit

staffs)

50 credit officers

Sampling

Figure 1.4: Population, sample and

sampling methods Source: Adapted from G.Zikmund (1997)

The quota sampling technique is used in this study because of itsadvantages in term of time, finance and convenience Threesteps of the technique are described as below:

First, the whole population of BIDV credit staffs is dividedinto three significant classes: managers/vice managers,credit department leaders and credit officers Thisclassification is based on the researcher judgment that thehigher position the credit staffs are in, the more reasonablefeeling they have

Secondly, each class is determined the desired proportion.Managers/vice managers group occupies 20% of thesample, credit leaders group occupies 30% of the sampleand 50% of the sample is credit officers group Thedetermination is based on the researcher judgmentmentioned in the first step Since this is a perceptionsurvey, the findings are influenced by subjective judgment

of respondents Therefore, a half of selected sample are

credit staffs with high positions Thanks to the advantage ofbeing a member of BIDV, it is quite easy for the researcher

to communicate with managers/vice managers and creditdepartment leaders about the research problem

50staffs Credit officers

In this study, number of credit officers occupies only 50% ofsample while it is 90% in Chau’s (2009) This is the mainreason leads the researcher retest the four hypothesesconducted by Chau (2009)

Finally, quota sample (100 respondents) is fixed The sample

of this study is about 100 respondents (over 5 times ofobserved variables) including 20 managers/vice managers,

30 credit department leaders and 50 credit officers Thisnumber was decided after considering some previousresearches For example, see Tho & Trang (2008, p.35) orTrong & Ngoc (2008, p31)

To obtain the desired sample size, a total of 150 administered questionnaires were distributed to the respondents

self-by the researcher Of these, 100 questionnaires were returnedmaking effective response rate 67%

Figure 1.5 Quota sampling method

Source: Adapted from Adapted from

G.Zikmund (1997)This study utilizes two techniques to obtain primary data:interview (telephone interview and face to face interview) and selfadministered questionnaire Interview method is used to gatherresponse of manager group while self administered

questionnaire is used to collect response of credit departmentleader group and credit officer groups.

Before conducting survey, the researcher carries out interview and pretest in order to increase quality of datacollection

depth-Depth-interview

Beside reviewing different aspects of four factors includingcredit information, technology, credit staff and loan policy viabooks, previous researches, the researcher will also carry out andepth interview in order to draw other practical aspects of thefour factors such as: credit information selecting andsystemizing, credit staff competence and technology matching,frequency of facility maintenance, important role of board ofdirectors The objects of this interview are one manager and threecredit department leaders

Pretest

A list of questions used for getting information from respondents(Appendix A) was created In order to improve the response rate,the researcher implemented a pilot test to refine the questionsand make sure that respondents have no problems inunderstanding or answering those questions

This pilot test was conducted through a group of 10 peopleincluding one professor and four classmates from UEH, onemanager and 4 BIDV credit officers The feedback from thesepeople who have practical experience and academic knowledgehelped the researcher to improve the questions in order to getthe highest response rate from respondents

1.5.3 Data analysis

The process of analysis begins after the data have been collected(G.Zikmund, 1997, p.507) Data collected must be analyzed inorder to create meaningful findings for the study Data analysisplays an important role in analyzing the data If the collected data

is not properly analyzed, the result may be invalid SPSSsoftware version 16.0 was used for data analysis because of itsmany powerful statistical features The main objectives of thedata analysis are to test the quality of the data collected and thehypotheses studied (Sekaran, 2003)

Firstly, collected data must be recorded by using numericalcodes By doing so, the researcher can input the dataquickly into the system using the numeric keypad on thekeyboard with very few errors.

Second, once the collected data is input and coded, theresearcher can then enter them into the computer manually

Finally, when the data has been already recorded andentered, the researcher can proceed to the data analysisphase

Data analysis technique

The study uses descriptive statistics to summarize thebackground information of respondents’ in the surveyquestionnaires The detail about the frequency and percentage ofrespondents’ working years, positions and backgrounds will beshowed in this section

Reliability measures were used to test the meaning of thedifferent variable combinations The four hypotheses in thisstudy are the assumptions about the effects of four variables

to the effectiveness of credit risk management strategy.Therefore, the survey questionnaire consists of 19 questionsrelated to these four variables However, as it was difficult totest these hypotheses based on separate individual aspects, theresearcher decided to combine different aspects of each variableinto one Therefore, it was necessary for the researcher to test themeaning of this combination process Cronbach’s alpha is acommonly used number to test the reliability of the combination

of different individual variables The value of Cronbach’s alphavaries between 1 (perfect internal reliability) and 0 (no internalreliability) According to Bryman and Bell (2003), the value of 0.80

is an acceptable level of internal reliability However, manywriters accept a slightly lower figure like Vogt (2007) argued that

an alpha of 0.70 or higher is often considered satisfactory formost studies

An application of hypothesis testing is used to quantifyrespondents’ perception of research problem on a five-pointscale, where 1 indicates strongly disagree and 5 indicatesstrongly agree The scale is assumed to be an interval scale T-test technique is used to estimate confidence intervals for themean

Chapter 1: IntroductionChapter 2: Literature ReviewChapter 3: Case study of BIDVChapter 4: Data analysis and findingsChapter 5: Recommendations and conclusions

The later chapter, chapter four: “Data analysis and findings”will discuss about analysis technique in detail

1.6Significance of the study

This study helps readers realize the crucial importance of creditrisk management and know the main factors that influence thereduction of non-performing-loan ratio in BIDV

1.7Structure of the study

Figure 1.6: Structure of the studyThis main purpose of the study is to understand the impact ofcredit information, technology, loan policy and credit staffs on

reduction of non-performing-loan ratio in BIDV Beside this

chapter one currently discussed, the study also consists of the

following five chapters:

Chapter 2: Literature Review: This chapter will provide general

theories related to credit risk management Furthermore, theresearcher’s insights on these theories will also be discussed

Chapter 3: Case study of BIDV: This chapter provides an

overview of bank for investment and development of Vietnam(BIDV) and BIDV’s credit risk management is the main part of thischapter

Chapter 4: Data analysis and findings: analyzing the collected

data in order to get results to test the hypotheses and answer theresearch questions in chapter one

Chapter 5: Recommendation and conclusion: based on

these analysis and findings from chapter five, some suggestions

or recommendations about the credit risk management strategiesthat BIDV can adopt to manage credit risk will be given